Split (1)

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

identifier stringlengths 11 32	pdf_url stringlengths 17 4.62k ⌀	lang stringclasses 120 values	error stringclasses 1 value	title stringlengths 2 500 ⌀	source_name stringlengths 1 435 ⌀	publication_year float64 1.9k 2.02k	license stringclasses 3 values	word_count int64 0 1.64M	text stringlengths 1 9.75M
https://openalex.org/W2745467839	https://www.nature.com/articles/s41598-017-10893-y.pdf	English	null	Real-time monitoring system for evaluating the acid-producing activity of oral squamous cell carcinoma cells at different environmental pH	Scientific reports	2,017	cc-by	6,174	Real-time monitoring system for evaluating the acid-producing activity of oral squamous cell carcinoma cells at different environmental pH Received: 8 March 2017 Accepted: 16 August 2017 Published: xx xx xxxx Received: 8 March 2017 Accepted: 16 August 2017 Published: xx xx xxxx Hiromitsu Morishima1,2, Jumpei Washio1, Jun Kitamura1,2, Yuta Shinohara1,3, Tetsu Takahashi2 & Nobuhiro Takahashi1 Hiromitsu Morishima1,2, Jumpei Washio1, Jun Kitamura1,2, Yuta Shinohara1,3, Tetsu Takahashi2 & Nobuhiro Takahashi1 This study aimed to establish a real-time monitoring system for evaluating the acid-producing activity of cells and the effects of microenvironmental pH on their metabolism. Oral squamous cell carcinoma (HSC-2, HSC-3) and normal (HaCaT) cells were used. Their acid-producing activity from glucose, glutamine, and glutamate was monitored at various pH values using a pH stat system. Their production of lactic acid and ammonia was also measured. The acid-producing activity was monitored successfully. Both the cancer and normal cells produced acids from glucose, glutamine, and glutamate. All of the cells decreased their acid-producing activity as the environmental pH fell, but in glucose-derived acid- producing activity the cancer cells were more acid-tolerant than HaCaT cells. In the cancer cells, the proportion of lactic acid among all acids produced from glucose at the acidic environment tended to be higher than that in HaCaT cells. All of the cells produced ammonia from glutamine, while only HaCaT cells produced ammonia from glutamate. We established a real-time monitoring system for evaluating the acid-producing activity of cells. Our results suggest that the cancer cells possess acid-tolerant glucose metabolism with a tendency of metabolic shift to lactic acid production at acidic pH and they metabolise glutamate without ammonia production. Many cancer-related genes, such as Myc and p53, have been reported1, 2, and it has become clear that the meta- bolic activity of cancer cells is regulated by these oncogenes3. The supply of energy and cell constituents is crucial for the infinite proliferation of cancer cells, and therefore, elucidating their metabolic systems might provide essential information about cancer cells.f It is well known that cancer cells exhibit a characteristic metabolic phenomenon called the Warburg effect4; i.e., they produce lactic acid from glucose even in the presence of abundant oxygen. Furthermore, it has been reported that cancer cells display enhanced glutamine metabolism, so-called glutaminolysis5, 6. We have con- firmed that oral squamous cell carcinoma (OSCC) cells also demonstrate similar metabolic activity7. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Received: 8 March 2017 Accepted: 16 August 2017 Published: xx xx xxxx Results id Acid production from glucose, glutamine, and glutamate. Both normal cells and cancer cells pro- duced acids from glucose, glutamine, and glutamate. The acid-producing activity of the cells was successfully monitored in real time using a pH stat system. The amount of acid produced by the cells increased with time (Figs 1A–3A), and among the three metabolic substrates glucose induced the highest acid production rate. In the experiments involving glucose, the HSC-2 cells exhibited significantly greater acid production than the other cells (Fig. 1A). Effects of environmental pH on acid-producing activity. For each substrate, the acid production activity seen at pH 7.5 was defined as 100% and compared with those seen at other pH. In all cell types (regard- less of the substrate used), acid production decreased as the environmental pH fell (Figs 1B–3B). A signifi- cant difference (p < 0.01) in acid production was observed between pH 8.0 or 7.5 and 6.5 in all cell types. The glucose-derived acid-producing activity of the cancer cells was more acid-tolerant than that of the HaCaT cells significantly (Fig. 1B), while the glutamine-derived acid-producing activity was similar among all the cell type (Fig. 2B) and the glutamate-derived acid-producing activity of the cancer cells seemed to be more acid-sensitive than that of the HaCaT cells (Fig. 3B). Lactic acid production from glucose. Both normal cells and cancer cells produced lactic acid from glu- cose, and the lactic acid production of the cells decreased as the environmental pH fell (Fig. 1C). The proportion of lactic acid among all acids (calculated from the amounts of NaOH titrated by the pH stat system) was about 35–55% (Fig. 1D). In the cancer cells the proportion of lactic acid among all acids tended to be higher than that of the HaCaT cells at acidic pH (pH 6.5). Ammonia production from glutamine and glutamate. All of the cells produced ammonia from glu- tamine (Fig. 2C), while the HaCaT cells produced ammonia from glutamate (Fig. 3C). Ammonia production from glutamine was reduced at acidic pH in all cell types. The effect of 2DG on the acid-producing activity of the HSC-2 cells. The addition of 2 mM 2DG at 10 min after glucose-induced acid production inhibited the acid-producing activity from glucose by the HSC-2 cells, while the addition of saline showed no effect (Fig. 4A). The inhibition% was 35.9 ± 12.8 (Fig. 4B). www.nature.com/scientificreports/ kinase 1, an enzyme responsible for the inhibition of pyruvate dehydrogenase12. Moreover, it has been suggested that cancer metastasis was increased in mice by acidic pH13, and in human head and neck cancer tissue a high concentration of lactic acid was found to increase the risk of metastasis14. It was also reported that a low pH microenvironment affected the permeability of a weakly alkaline drug and was associated with resistance to anti- cancer drugs15. g However, it remains unclear how environmental pH directly affects metabolic activity, probably because the biological activity of cancer cells has mainly been evaluated based on their proliferation potency in previous studies, and thus, no method for the real-time monitoring of metabolic activity at a fixed pH has been developed. Therefore, we attempted to establish a method for monitoring the acid-producing activity of cells in real time and to evaluate the direct effects of microenvironmental pH on the metabolic processes of cancer cells in comparison with normal cells.if Furthermore, to confirm that this monitoring system can also be applied to evaluate the effect of the anti- cancer agent, we attempted to measure the acid-producing activity from glucose in the presence and absence of 2-deoxy-D-glucose (2DG), one of anticancer agents which is known as a metabolic inhibitor16. Real-time monitoring system for evaluating the acid-producing activity of oral squamous cell carcinoma cells at different environmental pH These obser- vations suggest that the pH of the microenvironment around cancer cells tends to change in response to the levels of acidic and alkaline metabolic products, such as lactic acid and ammonia. Extracellular acidosis is also reported to be a feature of cancer cells8, 9, and so the microenvironmental pH of cancer cells is considered to be different from that of normal cells.h The relationships between environmental factors and cancer cells have been examined by many researchers, and it is becoming clear that environmental factors, such as acidic pH and low oxygen levels, are involved in the expression of genes, such as those encoding glucose transporter 1 and hexokinase II, through hypoxia-inducible factor-110, 11, and the inhibition of the tricarboxylic acid (TCA) cycle via the expression of pyruvate dehydrogenase 1Division of Oral Ecology and Biochemistry, Department of Oral Biology, Tohoku University Graduate School of Dentistry, Sendai, Japan. 2Division of Oral and Maxillofacial Surgery, Tohoku University Graduate School of Dentistry, Sendai, Japan. 3Division of Advanced Prosthetic Dentistry, Tohoku University Graduate School of Dentistry, Sendai, Japan. Hiromitsu Morishima and Jumpei Washio contributed equally to this work. Correspondence and requests for materials should be addressed to N.T. (email: nobu-t@dent.tohoku.ac.jp) 1 SCiENTifiC ReportS \| 7: 10092 \| DOI:10.1038/s41598-017-10893-y www.nature.com/scientificreports/ www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 1. (A) Glucose-derived acid-production of each cell type at pH 7.5 (n = 5). The amount of NaOH indicates the amount of NaOH added during acid production from glucose after the glucose addition. Error bars represent standard deviations. #Significant difference between each cells, p < 0.05 An ANOVA and Tukey’s test. (B) Relative glucose-derived acid-producing activity at various pH values (the activity at pH 7.5 was defined as 100%) (n = 5). Error bars represent standard deviations. Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. Significant difference between each pH condition, p < 0.01 An ANOVA and Tukey’s test. #Significant difference between each cells at the same pH condition, p < 0.05 An ANOVA and Tukey’s test. (C) Lactic acid production from glucose at various pH values (n = 5). The amount of lactic acid indicates the amount of lactic acid produced during acid production from glucose after the glucose addition. Error bars represent standard deviations. Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. *Significant difference between each pH condition, p < 0.01 An ANOVA and Tukey’s test. #Significant difference between each cells at the same pH condition, p < 0.05 An ANOVA and Tukey’s test ##Significant difference between each cells at the same pH condition p<0 01 An ANOVA and Figure 1. (A) Glucose-derived acid-production of each cell type at pH 7.5 (n = 5). The amount of NaOH t Figure 1. (A) Glucose-derived acid-production of each cell type at pH 7.5 (n = 5). The amount of NaOH indicates the amount of NaOH added during acid production from glucose after the glucose addition. Error bars represent standard deviations. #Significant difference between each cells, p < 0.05 An ANOVA and Tukey’s test. (B) Relative glucose-derived acid-producing activity at various pH values (the activity at pH 7.5 was defined as 100%) (n = 5). Error bars represent standard deviations. Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. *Significant difference between each pH condition, p < 0.01 An ANOVA and Tukey’s test. #Significant difference between each cells at the same pH condition, p < 0.05 An ANOVA and Tukey’s test. (C) Lactic acid production from glucose at various pH values (n = 5). Discussion In previous studies, the metabolic activity of cells has mainly been evaluated based on the expression of meta- bolic enzymes12, cell proliferation capacity17, and lactic acid production during the proliferation process18, etc. However, it is difficult to estimate real-time metabolic activity based on the gene expression of metabolic enzymes or lactic acid accumulation during cell proliferation. In the present study, we successfully monitored the acid-producing activity of cancer and normal cells in the presence of glucose, glutamine, or glutamate at a fixed pH by continuously quantifying their acid production using a pH stat system (Figs 1A–3A). In general, glucose is metabolised via glycolysis, the TCA cycle, the electron transfer system, and the pentose phosphate pathway, and the products created during these processes are organic acids, such as lactic acid, CO2, and H2O (Fig. 5A). CO2 can dissolve in water, which results in its conversion to carbonic acid. Therefore, we can estimate the glucose-derived acid-producing activity of cells by monitoring the levels of these acids. All of the cells produced acids from glucose, and the amount of acid produced increased linearly with time (Fig. 1A), indicating that the cells metabolised glucose smoothly and that the acid-producing activity of cells can be estimated in real-time by monitoring their acid production. y y g It is generally accepted that glucose is metabolised to pyruvate via glycolysis, and pyruvate can be converted into lactic acid under anoxic conditions or further metabolised in the TCA cycle and electron transfer system with CO2 production under normoxic conditions. In addition, cancer cells are known to produce mainly lactic acid from glucose even under normoxic conditions (Warburg effect). Unexpectedly, the proportion of lactic acid among all acids, which was calculated from the amounts of NaOH titrated by the pH stat system, was 35–55%, and there was no significant difference in this parameter between the OSCC cells and the HaCaT cells (Fig. 1D). These findings suggest that the Warburg effect can occur in both cancer cells and normal cells. In recent years, the Warburg effect has been reexamined19, 20. Although it is well accepted that cancer cells possess a series of oncogene-directed metabolic reprogramming, which results in aerobic lactate production (Warburg effect)21, 22, SCiENTifiC ReportS \| 7: 10092 \| DOI:10.1038/s41598-017-10893-y 2 www.nature.com/scientificreports/ The amount of lactic acid indicates the amount of lactic acid produced during acid production from glucose after the glucose addition. Error bars represent standard deviations. Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. *Significant difference between each pH condition, p < 0.01 An ANOVA and Tukey’s test. #Significant difference between each cells at the same pH condition, p < 0.05 An ANOVA and Tukey’s test. ##Significant difference between each cells at the same pH condition, p < 0.01 An ANOVA and Tukey’s test. (D) The proportion of lactic acid among all acids at various pH values in the presence of glucose (n = 5). Error bars represent standard deviations. Figure 1. (A) Glucose-derived acid-production of each cell type at pH 7.5 (n = 5). The amount of NaOH indicates the amount of NaOH added during acid production from glucose after the glucose addition. Error bars represent standard deviations. #Significant difference between each cells, p < 0.05 An ANOVA and Tukey’s test. (B) Relative glucose-derived acid-producing activity at various pH values (the activity at pH 7.5 was defined as 100%) (n = 5). Error bars represent standard deviations. Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. *Significant difference between each pH condition, p < 0.01 An ANOVA and Tukey’s test. #Significant difference between each cells at the same pH condition, p < 0.05 An ANOVA and Tukey’s test. (C) Lactic acid production from glucose at various pH values (n = 5). The amount of lactic acid indicates the amount of lactic acid produced during acid production from glucose after the glucose addition. Error bars represent standard deviations. Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. *Significant difference between each pH condition, p < 0.01 An ANOVA and Tukey’s test. #Significant difference between each cells at the same pH condition, p < 0.05 An ANOVA and Tukey’s test. ##Significant difference between each cells at the same pH condition, p < 0.01 An ANOVA and Tukey’s test. (D) The proportion of lactic acid among all acids at various pH values in the presence of glucose (n = 5). Error bars represent standard deviations. SCiENTifiC ReportS \| 7: 10092 \| DOI:10.1038/s41598-017-10893-y 3 www.nature.com/scientificreports/ Figure 2. (A) Glutamine-derived acid-producing activity at pH 7.5 (n = 5). Error bars represent standard deviations. www.nature.com/scientificreports/ (B) Relative glutamine-derived acid-producing activity at various pH values (the activity at pH 7.5 was defined as 100%) (n = 5). Error bars represent standard deviations. Significant difference between each pH condition, p < 0.01 An ANOVA and Tukey’s test. *Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. (C) Ammonia production from glutamine at various pH values (n = 5). Error bars represent standard deviations. Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. Figure 2. (A) Glutamine-derived acid-producing activity at pH 7.5 (n = 5). Error bars represent standard deviations. (B) Relative glutamine-derived acid-producing activity at various pH values (the activity at pH 7.5 was defined as 100%) (n = 5). Error bars represent standard deviations. Significant difference between each pH condition, p < 0.01 An ANOVA and Tukey’s test. Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. (C) Ammonia production from glutamine at various pH values (n = 5). Error bars represent standard deviations. Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. it is also known that proliferating cells enhance glycolysis for rapid ATP production and macromolecular biosyn- thesis and subsequently produce lactate aerobically23, 24. In the present study, the HaCaT cells were collected at proliferating stage, so the cells might be able to express such metabolic properties. it is also known that proliferating cells enhance glycolysis for rapid ATP production and macromolecular biosyn- thesis and subsequently produce lactate aerobically23, 24. In the present study, the HaCaT cells were collected at proliferating stage, so the cells might be able to express such metabolic properties. p g g g p p p At the acidic environment such as pH 6.5, the proportion of lactic acid in the OSCC cells tended to increase (Fig. 1D), suggesting that these cells shift the metabolic flow of glucose to lactic acid production in response to pH changes. The acid-producing activity of the HSC-2 cells was significantly greater than those of the other cell types (Fig. 1A). However, acid-producing activity can vary due to differences in cell lines and culture conditions, and therefore, it is difficult to generalise this phenomenon. www.nature.com/scientificreports/ fi g p On the other hand, during the metabolism of glutamine or glutamate, ammonia was produced as well as acid (Figs 2C and 3C), and therefore, some of the acid produced would have been neutralised by the ammonia, which would have resulted in the underestimation of acid-producing activity. However, it was still possible to monitor the cells’ acid-producing activity (Figs 2A and 3A) in the presence of glutamine or glutamate, indicating that the cells were producing greater amounts of acid than ammonia from glutamine and glutamate, as expected in the SCiENTifiC ReportS \| 7: 10092 \| DOI:10.1038/s41598-017-10893-y 4 www.nature.com/scientificreports/ Figure 3. (A) Glutamate-derived acid-producing activity at pH 7.5 (n = 5). Error bars represent standard deviations. #Significant difference between each cells, p < 0.05 An ANOVA and Tukey’s test. (B) Relative glutamate-derived acid-producing activity at various pH values (the activity at pH 7.5 was defined as 100%) (n = 5). Error bars represent standard deviations. Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. Significant difference between each pH condition, p < 0.01 An ANOVA and Tukey’s test. (C) Ammonia production from glutamate at various pH values (n = 5). ND: not detected. Error bars represent standard deviations. Figure 3. (A) Glutamate-derived acid-producing activity at pH 7.5 (n = 5). Error bars represent standard deviations. #Significant difference between each cells, p < 0.05 An ANOVA and Tukey’s test. (B) Relative glutamate-derived acid-producing activity at various pH values (the activity at pH 7.5 was defined as 100%) (n = 5). Error bars represent standard deviations. Significant difference between each pH condition, p < 0.05 An ANOVA and Tukey’s test. *Significant difference between each pH condition, p < 0.01 An ANOVA and Tukey’s test. (C) Ammonia production from glutamate at various pH values (n = 5). ND: not detected. Error bars represent standard deviations. glutaminolysis where glutamine/glutamate can be converted to lactic acid via the last half part of the TCA cycle and malic enzymes in OSCC cells7 and other cancer cells25 (Fig. 5B). Acid production from glutamate increased linearly with time (Fig. 3A), while acid production from glutamine was rapid in the first 1–2 min, stopped for next few minutes, and then started again (Fig. 2A). www.nature.com/scientificreports/ Glutamine is con- glutaminolysis where glutamine/glutamate can be converted to lactic acid via the last half part of the TCA cycle and malic enzymes in OSCC cells7 and other cancer cells25 (Fig. 5B). Acid production from glutamate increased linearly with time (Fig. 3A), while acid production from glutamine was rapid in the first 1–2 min, stopped for next few minutes, and then started again (Fig. 2A). Glutamine is con- idered to be metabolised to glutamate by glutaminase (Fig. 5B). This two-step reaction might result in pulsation owever, further study is needed to elucidate the underlying metabolic control mechanisms.h y y g The HaCaT cells produced ammonia from both glutamine and glutamate, while the OSCC cells produced ammonia mainly from glutamine (Figs 2C and 3C). There was no difference in glutamine-derived metabolic activity between the OSCC cells and HaCaT cells, which suggested that the pathways involved in glutamate deg- radation differ between the cancer and normal cells. Glutamine is usually deaminated to glutamate, producing ammonia, by glutaminase, as stated above. Glutamate is further deaminated to α-ketoglutarate with the produc- tion of ammonia by glutamate dehydrogenase (Fig. 5B). In the normal cells, ammonia production was observed in the presence of glutamine or glutamate (Figs 2C and 3C), supporting the existence of these metabolic pathways. SCiENTifiC ReportS \| 7: 10092 \| DOI:10.1038/s41598-017-10893-y 5 www.nature.com/scientificreports/ Figure 4. (A) An example of real-time monitoring of acid-producing activ 2DG addition at 10 min. (B) Relative acid-producing activity after the 2DG control was defined as 100%) (n = 3). Error bars represent standard deviati An ANOVA and Tukey’s test. Figure 4. (A) An example of real-time monitoring of acid-producing activity from glucose and the effect of the 2DG addition at 10 min. (B) Relative acid-producing activity after the 2DG addition (the activity observed in control was defined as 100%) (n = 3). Error bars represent standard deviations. Significant difference, p < 0.05 An ANOVA and Tukey’s test. Figure 4. (A) An example of real-time monitoring of acid-producing activity from glucose and the effect of the 2DG addition at 10 min. (B) Relative acid-producing activity after the 2DG addition (the activity observed in control was defined as 100%) (n = 3). Error bars represent standard deviations. *Significant difference, p < 0.05 An ANOVA and Tukey’s test. However, the absence of ammonia production from glutamate in the OSCC cells suggests that cancer-specific metabolic pathways also exist. www.nature.com/scientificreports/ Recently, it was demonstrated that in human pancreatic ductal adenocarcinoma cells glutamate is converted to α-ketoglutarate by aspartate aminotransferase, producing aspartate from oxalo- acetate. Thus, glutamate could be used to maintain the pool of amino acids26. OSCC cells might possess a similar metabolic pathway involving aminotransferases which transfer glutamate-derived amino group to α-keto acids to form amino acids without ammonia production (Fig. 5B). Extracellular acidosis is reported to be a feature of cancer tissue8. Hence, it is assumed that cancer cells have a greater capacity to adapt to low pH environments than normal cells. However, in the present study metabolic activity decreased as the environmental pH fell in all cell types regardless of the substrate present (Figs 1B–3B). The glucose-derived acid-producing activity of the cancer cells was more acid-tolerant than that of the HaCaT cells significantly (Fig. 1B), while the glutamine-derived metabolic activity was similar among all the cell type SCiENTifiC ReportS \| 7: 10092 \| DOI:10.1038/s41598-017-10893-y 6 www.nature.com/scientificreports/ Figure 5. Proposed metabolic pathways in OSCC and normal cells. (A) Glucose metabolic pathways (dark lines). Main end-products are highlighted. Key metabolic enzymes are indicated in italic. (B) Glutamine and glutamate metabolic pathways (dark lines). Main end-products are highlighted. Key metabolic enzymes are indicated in italic. Figure 5. Proposed metabolic pathways in OSCC and normal cells. (A) Glucose metabolic pathways (dark lines). Main end-products are highlighted. Key metabolic enzymes are indicated in italic. (B) Glutamine and glutamate metabolic pathways (dark lines). Main end-products are highlighted. Key metabolic enzymes are indicated in italic. (Fig. 2B) and the glutamate-derived metabolic activity of the cancer cells seemed to be more acid-sensitive than that of the HaCaT cells (Fig. 3B). These observations suggest that differences in the acid sensitivity of metabolic pathways might exist between cancer cells and normal cells. Further study is needed to clarify these issues.f In the present study, the real-time monitoring method was applied to evaluate the inhibitory effect of 2DG, one of the anticancer reagents on the cancer cells. The inhibition of 2DG was clearly shown and quantified by this method (Fig. 5A,B), indicating that this method is suitable to evaluate the inhibitory effect of anticancer agents based on the metabolic activity. In conclusion, the method established in the present study made it possible to monitor the acid-producing activity of cells in real time at a fixed pH. www.nature.com/scientificreports/ There are various existing methods/machines that enable to monitor cellular metabolic activity; however, one of the popular bioanalyzers measures pH decline during metabolism, so the reaction pH cannot be maintained, while measurement of cellular oxygen consumption during metabolism is not suitable to evaluate aerobic lactate production (Warburg effect) due to limit of oxygen utilization. In addition, the method using a pH-stat is inexpensive and easily set up under various conditions. This method might provide new topics for research into cancer cell metabolism and aid the evaluation of anticancer drugs. In most previous studies of anticancer drugs targeting metabolic enzymes in cancer cells, the drugs were mainly evaluated based on their ability to inhibit cell growth27. Our method can evaluate such inhibitory effects directly and could contribute to the establishment of more effective anticancer drug treatments. Materials and Methods Cell culture. Two lines of OSCC cells were used in this study. HSC-3 (RCB1975) cells, which are highly aggressive OSCC cells derived from the tongue, and HSC-2 cells (RCB1945), which are weakly aggressive OSCC cells derived from the mouth, were purchased from Riken Cell Bank (Tsukuba, Japan) and cultured as recom- mended. These cells were cultured in Eagle’s minimal essential medium (Wako Pure Chemical industires Ltd., Tokyo, Japan) supplemented with 2 mmol/L L-alanyl-L-glutamine solution (Wako Pure Chemical Industries Ltd., Tokyo, Japan), 10% heat-inactivated fetal bovine serum, 100 µg/mL streptomycin, and 100 U/mL penicillin at 37 °C with humidified 5% CO2 (CO2 incubator; model MCO-18AC, Sanyo, Tokyo, Japan). HaCaT cells (a human immortalised keratinocyte line derived from normal skin) were also purchased from Cosmo Bio (Tokyo, Japan) and used as normal cells. The HaCaT cells were cultured in Dulbecco’s modified Eagle’s medium (10-013-CVR, Corning, NY, USA) supplemented with 10% heat-inactivated fetal bovine serum, 100 mg/mL streptomycin, and 100 U/mL penicillin at 37 °C with humidified 5% CO2 (CO2 incubator; model MCO-18AC, Sanyo, Tokyo, Japan). All f h ll b l d 3 4 d i i l i h i h ll d 80 90% pi y y p All of the cells were sub-cultured every 3–4 days to maintain logarithmic growth, collected at 80–90% con- fluence, and suspended in saline at a density of 1.0 × 106 cells/ml. Before and after the experiment, the cells were confirmed to be alive using 0.4 w/v % trypan blue solution (Wako Pure Chemical Industries Ltd., Tokyo, Japan)28. SCiENTifiC ReportS \| 7: 10092 \| DOI:10.1038/s41598-017-10893-y 7 7 www.nature.com/scientificreports/ Real-time monitoring of acid-producing activity with a pH stat system. After 10 min pre-incubation at 37 °C, the cell suspension was mixed with the relevant metabolic substrate (5 mM glucose, glu- tamine, or glutamate), and acid production was monitored for 20 min using a pH stat system (AUTO pH STAT; model AUT-211S, TOA Electronics, Tokyo, Japan) and 0.01 M NaOH as a titrant. When the pH of the reaction mixture began to fall due to acid production by the cells, NaOH was automatically added to the mixture when the pH became less than predetermined pH (6.5, 7.0, 7.5, or 8.0) by 0.01 pH unit, so every 2–10 second, 0.1 µL of NaOH was added and mixed in the reaction mixture to maintain the predetermined pH, and the amount of NaOH added was monitored by an integrator. Materials and Methods The acid-producing activity of the cells was estimated from the amount of NaOH. All of the experiments were performed in normal air. Measurement of lactic acid and ammonia. At the start and end of the reaction, samples of the reaction mixture were collected and centrifuged immediately at 12,000 rpm for 2 min. The supernatant was used to meas- ure the levels of lactic acid and ammonia. The supernatant was stored at −80 °C until the measurements were obtained. After the supernatant had been pre-treated as reported previously29, the lactic acid level was measured using capillary electrophoresis-time-of-flight mass spectrometry (CE-TOFMS: G1600AX and G1969A; Agilent Technologies, Waldbronn, Germany). A fused silica capillary (H3305-2002; Human Metabolome Technologies) was used to separate out the metabolites during the CE. The applied voltage was set at −30 kV when the electro- spray ionisation was performed in negative ion mode, and the capillary voltage was set at 3.5 kV in negative ion mode. The flow rate of the heated dry nitrogen gas (300 °C) was maintained at 7 L/min. The metabolites separated out by CE were mixed with sheath liquid (H3302-1020; Human Metabolome Technologies), which provides the electric contact as well as appropriate flow and solvent conditions for ensuring the optimal ionisation of the metabolites, and continuously sent to the TOFMS system for mass analysis. y y y Ammonia was measured with an ammonia meter (AMICHECK METER, Arkray, Kyoto, Japan), as reported reviously30. Application of this real-time monitoring system to evaluate the inhibitory effect of anticancer agents. 2-Deoxy-D-glucose (2DG; Wako) was used as a anticancer agent. The reaction mixture consisted of 1 mL of cell suspensions (1.0 × 106 cells/mL) of the HSC-2 cells and 3.5 mL of saline. Firstly, this reaction mixture was set to the pH stat at pH 7.5 and pre-incubated at 37 °C with stirring for 10 min. Then, glucose (250 μL) was added to the reaction mixture at a final concentration of 10 mM, and their acid-producing activity was moni- tored for 10 min. At 10 min after the glucose addition, 2DG (250 μL) was added to the reaction mixture at a final concentrations of 2 mM and the acid-producing activity was further monitored for 20 min. Similarly, saline was added instead of the 2DG in control group. Statistical analysis. References 1. Gabay, M., Li, Y. & Felsher, D. W. MYC activation is a hallmark of cancer initiation and maintenance. Cold Spring Harb Perspect Med 4, doi:10.1101/cshperspect.a014241 (2014). p p 2. Marcel, V., Catez, F. & Diaz, J. J. p53, a translational regulator: contribution to its tumour-suppressor activity. Oncogene 5513–5523 (2015). ( ) 3. Cairns, R. A., Harris, I. S. & Mak, T. W. Regulation of cancer cell metabolism. Nat Rev Cancer. 11, 85–95 (2011). 4. Warburg, O. On the origin of cancer cells. Science. 123, 309–14 (1956).h 4. Warburg, O. On the origin of cancer cells. Science. 123, 309–14 (1956).h 5. Kovacević, Z. A. & Morris, H. P. The role of glutamine in the oxidative metabolism of malignant cells. Cancer res. 32, 326–333 (1972). ( ) 6. Cetindis, M. et al. Glutaminolysis and carcinogenesis of oral squamous cell carcinoma. Eur Arch Otorhinolaryngol. 273, 495–503 (2016). 7. Ogawa, T., Washio, J., Takahashi, T., Echigo, S. & Takahashi, N. Glucose and glutamine metabolism in oral squamous cell carcinoma insight from a quantitative metabolomic approach. Oral Surg Oral Med Oral Pathol Oral Radiol. 118, 218–225 (2014). k h ll l d l l l f h 8. Gerweck, L. E. & Seetharaman, K. Cellular pH gradient in tumor versus normal tissue: potential exploitation for the treatment o cancer. Cancer Res. 56, 1194–1198 (1996).f 9. Schornack, P. A. & Gillies, R. J. Contributions of cell metabolism and H+ diffusion to the acidic pH of tumors. Neoplasia. 5, 135–45 (2003). ( ) 10. Chen, C., Pore, N., Behrooz, A., Ismail-Beigi, F. & Maity, A. Regulation of glut1 mRNA by hypoxia-inducible factor-1. Interaction between H-ras and hypoxia. J Biol Chem. 276, 9519–9525 (2001). yp 1. Riddle, S. R. et al. Hypoxia induces hexokinase II gene expression in human lung cell line A549. Am J Physiol Lung Cell Mol Physiol 278, L407–16 (2000). ( ) 12. Kim, J. W., Tchernyshyov, I., Semenza, G. L. & Dang, C. V. HIF-1-mediated expression of pyruvate dehydrogenase kinase: a metabolic switch required for cellular adaptation to hypoxia. Cell Metab. 3, 177–185 (2006). 3. Rofstad, E. K., Mathiesen, B., Kindem, K. & Galappathi, K. Acidic extracellular pH promotes experimental metastasis of human melanoma cells in athymic nude mice. Cancer res. 66, 6699–6707 (2006). y 4. Walenta, S. et al. Correlation of high lactate levels in head and neck tumors with incidence of metastasis. Am J Pathol. 150, 409–415 (1997). 15. Materials and Methods An ANOVA was performed and then Tukey’s test was employed to determine the statis- tical significance of the differences between the cancer cells and normal cells. P-values of < 0.05 were considered to be statistically significant.hh y gi The statistical comparison of the inhibition rates by 2DG was also analyzed using Tukey’s test. The data were analyzed using the Statflex Ver. 6 (Artech Co., Ltd., Osaka). SCiENTifiC ReportS \| 7: 10092 \| DOI:10.1038/s41598-017-10893-y References 29 T k h hi N W hi J & M i G M b l i f i i l l d l b i J D R 89 1383 1388 (2010) 9. Takahashi, N., Washio, J. & Mayanagi, G. Metabolomics of supragingival plaque and oral bacteria. J Dent Res 89, 1383–1388 (2010) 29. Takahashi, N., Washio, J. & Mayanagi, G. Metabolomics of supragingival plaque and oral bacteria. J Dent Res 89, 1383–1388 (2010). 30. Seligson, D. & Hirahara, K. The measurement of ammonia in whole blood, erythrocytes, and plasma. J Lab Clin Med. 49, 962–974 (1957). Author Contributions J.W. and N.T. conceived the experiments, H.M., J.W., J.K and Y.S. conducted the experiments, H.M., J.W. and N.T. analysed the results. All authors reviewed the manuscript. J.W. and N.T. conceived the experiments, H.M., J.W., J.K and Y.S. conducted the experiments, H.M., J.W. and N.T. analysed the results. All authors reviewed the manuscript. Acknowledgements g N.T. received the Grants-in-Aid for Challenging Exploratory Research (Nos 23659866 and 25670786) from Japan ociety for the Promotion of Science. References Trédan, O., Galmarini, C. M., Patel, K. & Tannock, I. F. Drug resistance and the solid tumor microenvironment. J Natl Cancer Inst. 99, 1441–1454 (2007). 16. Maher, J., Krishan, A. & Lampidis, T. Greater cell cycle inhibition and cytotoxicity induced by 2-deoxy-D-glucose in tumor cells treated under hypoxic vs aerobic conditions. Cancer Chemother Pharmacol. 53, 116–122 (2004).h yp 7. Teppo, S. et al. The hypoxic tumor microenvironment regulates invasion of aggressive oral carcinoma cells. Exp Cell Res. 319 376–389 (2013). SCiENTifiC ReportS \| 7: 10092 \| DOI:10.1038/s41598-017-10893-y 8 www.nature.com/scientificreports/ 8. Mazzio, E. A., Boukli, N., Rivera, N. & Soliman, K. F. Pericellular pH homeostasis is a primary function of the Warburg effect inversion of metabolic systems to control lactate steady state in tumor cells. Cancer Sci. 103, 422–432 (2012). 9. Koppenol, W. H., Bounds, P. L. & Dang, C. V. Otto Warburg’s contributions to current concepts of cancer metabolism. Nat Rev Cancer. 11, 325–337 (2011).hfi 20. Liberti, M. V. & Locasale, J. W. The Warburg Effect: How does it benefit cancer cells? Trends Biochem Sci. 41, 211–218 (2016). h gfi 21. Zong, W. X., Rabinowitz, J. D. & White, E. Mitochondria and cancer. Molecullar Cell. 61, 667–676 (2016). 2. Garcia-Heredia, J. M. & Carnero, A. Decoding Warburg’s hypothesis:tumor-related mutations in the mitochondrial respiratory chain. Oncotarget 6, 41582–41599 (2015). g 23. Heiden, M. G. V., Cantley, L. C. & Thompson, C. B. Understanding the Warbrug Effect: The metabolic requirements of cell proliferation. Science. 324, 1029–1033 (2009).hh p 4. DeBerardinis, R. J., Lum, J. J., Hatzivassiliou, G. & Thompson, C. B. The biology of cancer: metabolic reprogramming fuels cel growth and proliferation. Cell Metab 7, 11–20 (2008). growth and proliferation. Cell Metab 7, 11–20 (2008) g p 5. Deberardinis, R. J. et al. Beyond aerobic glycolysis: Transformed cells can engage in glutamine metabolism that exceeds the requirement for protein and nucleotide synthesis. PNAS. 104, 19345–19350 (2007). 26. Son, J. et al. Glutamine supports pancreatic cancer growth through a KRAS-regulated metabolic pathway. Nature. 496, 101–105 (2013). ( ) 27. Zhao, Y., Butler, E. B. & Tan, M. Targeting cellular metabolism to improve cancer therapeutics. Cell Death Dis. 4, e532 (2013). 28. Martin-Piedra, M. A. et al. Cell viability and proliferation capability of long-term human dental pulp stem cell cultures. Cytotherapy. 16, 266–77 (2014). Additional Informationh Competing Interests: The authors declare that they have no competing interests. Competing Interests: The authors declare that they have no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. SCiENTifiC ReportS \| 7: 10092 \| DOI:10.1038/s41598-017-10893-y 9
https://openalex.org/W1529579233	https://europepmc.org/articles/pmc3552701?pdf=render	English	null	HDAM: a resource of human disease associated mutations from next generation sequencing studies	BMC medical genomics	2,013	cc-by	4,354	© 2013 Jia et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. HDAM: a resource of human disease associated mutations from next generation sequencing studies Meiwei Jia1, Yanli Liu1, Zhongchao Shen1, Chen Zhao1, Meixia Zhang2, Zhenghui Yi3, Chengping Wen4, Youping Deng5, Tieliu Shi1* From The 2011 International Conference on Bioinformatics and Computational Biology (BIOCOMP’11) Las Vegas, NV, USA. 18-21 July 2011 * Correspondence: tieliushi01@gmail.com 1Center for Bioinformatics and Computational Biology, Shanghai Key Laboratory of Regulatory Biology, the Institute of Biomedical Sciences and School of Life Science, East China Normal University, Shanghai 200241, China Full list of author information is available at the end of the article Abstract Background: Next generation sequencing (NGS) technologies have greatly facilitated the rapid and economical detection of pathogenic mutations in human disorders. However, mutation descriptions are hard to be compared and integrated due to various reference sequences and annotation tools adopted in different articles as well as the nomenclature of diseases/traits. Description: The Human Disease Associated Mutation (HDAM) database is dedicated to collect, standardize and re-annotate mutations for human diseases discovered by NGS studies. In the current release, HDAM contains 1,114 mutations, located in 669 genes and associated with 125 human diseases through literature mining. All mutation records have uniform and unequivocal descriptions of sequence changes according to the Human Genome Sequence Variation Society (HGVS) nomenclature recommendations. Each entry displays comprehensive information, including mutation location in genome (hg18/hg19), gene functional annotation, protein domain annotation, susceptible diseases, the first literature report of the mutation and etc. Moreover, new mutation-disease relationships predicted by Bayesian network are also presented under each mutation. Conclusion: HDAM contains hundreds rigorously curated human mutations from NGS studies and was created to provide a comprehensive view of these mutations that confer susceptibility to the common disorders. HDAM can be freely accessed at http://www.megabionet.org/HDAM. powers to detect susceptible mutations. Thus, significant outcomes have been accumulated. Systems biology stu- dies have shown its critical role in mechanism illustration and correlation prediction between mutation and dis- ease/traits [6]. However, almost all of these studies have to take great extra efforts on data integration from publi- cations for both mutation description standardization and disease category because of inconsistence between data descriptions from different resources. Inspired by the great achievement of some popular databases, such as the database of Genotypes and Phenotypes (dbGaP), which was developed to archive and distribute the results of studies that have investigated the correlation of geno- type and phenotype [7], we believe that a database focus- ing on collecting mutation-disease relation and uniform Jia et al. BMC Medical Genomics 2013, 6(Suppl 1):S16 http://www.biomedcentral.com/1755-8794/6/S1/S16 Jia et al. BMC Medical Genomics 2013, 6(Suppl 1):S16 http://www.biomedcentral.com/1755-8794/6/S1/S16 Open Access Background Mutation is a major cause of human diseases. Scientific studies of mutational mechanisms in human genes and the function of mutated genes in pathological pathway to disease are always research hotspots. With the success of next generation sequencing technologies in past several years, exome sequencing [1], target resequencing [2] and whole genome sequencing [3-5] applied in studying human disorders, such as monogenic diseases, complex inherited diseases and cancers, have already proven their * Correspondence: tieliushi01@gmail.com 1Center for Bioinformatics and Computational Biology, Shanghai Key Laboratory of Regulatory Biology, the Institute of Biomedical Sciences and School of Life Science, East China Normal University, Shanghai 200241, China Full list of author information is available at the end of the article Jia et al. BMC Medical Genomics 2013, 6(Suppl 1):S16 http://www.biomedcentral.com/1755-8794/6/S1/S16 Jia et al. BMC Medical Genomics 2013, 6(Suppl 1):S16 http://www.biomedcentral.com/1755-8794/6/S1/S16 Page 2 of 6 In the process of extracting mutation-level informa- tion, a crucial criterion has been followed: only the mutation that can be accurately located on genome is collected. The mutation-level information includes: Genomic position, base change on genome, transcript position, base change on transcript, protein position, amino acid change, gene, chromosome number. mutation description will be meaningful and useful, espe- cially with more and more data accumulated in future. mutation description will be meaningful and useful, espe- cially with more and more data accumulated in future. Nevertheless, mutations are frequently documented with only protein and/or DNA sequence without being clarified with reference sequence accession number or version, which make those mutations unavailable for re- analysis and integration. For example, mutation records in the Online Mendelian Inheritance in Man (OMIM) [8] Allelic Variant field often have such problems. Many early records are described as sequence variation at pro- tein level only based on amino acid sequences reported in publication without protein accession IDs and versions [9]. Since the currently annotated protein often differs from these early sequences, the positions of some muta- tions cannot be inferred on the current sequences. In another word, such mutation records will lose their value in use. Several problems were encountered during text mining: Firstly, not all above information can be extracted from every paper. For a mutation, if only position information on genome and base change are available, all possible cDNA level and protein level descriptions are kept. For example, mutation NC_000011.8:g.9794133G>A may be located either in gene LOC283104 or gene SBF2. Background Because of no more detail information about this mutation in ori- ginal text, both of the two genes were collected into our database. Secondly, a gene may have several cDNA refer- ence sequences. If the cDNA accession number for a mutation is not assigned, we preferred to keep all possi- ble cDNA level descriptions that indicate the same amino acids mutation as shown in original articles. Besides, inconsistence between cDNA description and protein description was also observed in some articles. Under such circumstance, we used UCSC BLAT tool [12] to check and correct the discordance. During the whole data annotation process, we utilized Mutalyzer 2.0 b-8 [13] to check all collected data and complete infor- mation of each mutation when some information is not available from articles. Mutalyzer is a tool primarily designed to check descriptions of sequence variants according to the standard human sequence variant nomenclature of HGVS. Therefore, it makes it possible for us to standardize all mutation information from dif- ferent articles. In this project, two independent reviewers extracted information from the same articles in parallel to ensure that they applied our inclusion criteria in the same manner. Then, all results were checked and cor- rected by a third reviewer. To overcome these limitations, HDAM database is dedicated to collect and annotate mutations in nuclear genes underlying human diseases from published NGS studies. The core work is to uniform mutation nomencla- ture, locate mutation at different reference genome, record association between mutation and disease, and integrate meaningful annotation from both text-mined results and other databases. Particularly, the standard syntax of mutation description is critical for its sustain- able utilization in both molecular effects and biological functions. Another major accomplishment of this article is to predict new mutation-disease association through combining OMIM Allelic Variants data. Mitochondrial genome mutations are not included in current release. Construction and content Literature mining The mutation-disease relationship records in the current release were collected through PubMed searches using the terms ‘(exome OR whole OR deep OR high-throughput OR (next AND generation) OR (massively AND parallel)) AND sequencing AND mutation’. Through the abstract scanning, only the articles which focus on studying human diseases by the next generation sequencing were selected for further review. The Bayesian network approach As a representation of biofunctional system, Protein- Protein Interaction (PPI) network has been widely adopted in exploring the common or related mechanisms of com- plex phenotypes, such as network based function module discovery [6] and pathway tracing [16]. To facilitate under- standing of the biological linkage between diseases and mutations, a Bayesian network based scoring system was built in our platform, which integrates both topological and biological features, including disease neighbor density by common module profiling score (CMP) [17], disease gene function similarity by Shared Smallest Biological Pro- cesses score (SSBP) [18]. A disease mutation preferential interaction score (DMPI), which represents the most fre- quently used PPIs that link the mutated genes in a given disease, was also featured in the Bayesian network. The DMPI is treated as an efficacious bridge that makes the level-2 neighbor in PPI network to be available in the pre- diction [19]. These three types of the features are hier- archically organized in the Bayesian network. Topological (CMP and DMPI) and biological features (SSBP) are con- ditionally independent, while CMP and DMPI are non- independent given a disease-mutation linkage. Golden Standard Positive (GSP) dataset contains the disease- mutation relations that were manually collected from pub- lished sequencing results and OMIM Allelic Variants. As less result is available for the non-disease oriented muta- tion, Golden Standard Negative (GSN) dataset was con- structed by random combining diseases with the mutations that were annotated in a sufficient distance, where the distance was calculated in the HPO category tree. The confidence score for each inferred disease and mutation linkage was defined as the posterior probability of the hierarchical Bayesian network. Totally, 229,309 new mutation-disease relations have been predicted based on HPO, with Likelihood Ratio larger than 685 and FDR as 0.03. Data integration To further help in understanding the relationship between mutation and disease, biological annotations under differ- ent levels were also integrated into HDAM. On the gene level, gene function annotation by Gene Ontology [14], OMIM gene description, sequence information was incor- porated into the database. On the protein level, the domain annotation was accomplished by HMMSCAN (http://hmmer.janelia.org/search/hmmscan), and the pro- tein structure 3D graph with mutation labels was visua- lized by Jmol. Information on the following study-level fields was extracted from related papers: PubMed ID, PubMed URL, publication title, patient ID, patient ethnicity, the technique and the platform used in the study, disease information, validation experiments. Particularly, the disease name was checked by searching OMIM, Wikipe- dia (http://www.wikipedia.org/), Human phenotype Ontology (HPO) [10], Orphanet [11] and International Classification of Diseases (ICD, http://apps.who.int/clas- sifications/apps/icd/icd10online/). If available, the OMIM IDs, HPO IDs and/or ICD numbers were also assigned to corresponding phenotype. OMIM represents the most complete and up-to-date repository of all known disease genes and the disorders. The “omim.txt” file was downloaded from OMIM website, consisting of 21,728 OMIM records. We parsed omim file Jia et al. BMC Medical Genomics 2013, 6(Suppl 1):S16 http://www.biomedcentral.com/1755-8794/6/S1/S16 Page 3 of 6 C®T substitution happened at position 2944 on cDNA sequence NM_020975.4; NP_009008.2:p.Arg42* indicates that Arginine at position 42 on protein sequence NP_009008.2 is mutated to a stop codon. All mutation symbols used in our database follow the HGVS nomen- clature recommendations faithfully [20]. and extracted gene records containing Allelic Variants (AV) field, which records the relations between the muta- tions or polymorphisms of the genes and the phenotypes. The disease-gene association mediated by mutations can be more proper for exploring pathological mechanism of mutation induced diseases. Totally, 16,574 mutations, 18,510 mutation-disease pairs and 4,549 gene-disease pairs were extracted from 2,604 gene records. A HDAM entry comprises a mutation located gene, chromosomal location, protein domain, associated disease, the reference to the literature report of this mutation and some details in the literature (eg. The technique, platform, sample ethnicity and description of this mutation). Besides, annotation integrated from other sources, such as OMIM and GO, is also available. Moreover, predicted new mutation-disease pairs with Bayesian Network are pre- sented in each mutation entry. Utility and discussion Utility To provide a systemic gateway for mutation-disease rela- tionship exploration, we implemented several search fea- tures on the website, including mutation description syntaxes (e.g. ‘p.Gly1349Asp’), reported gene name (e.g. ‘CFTR’), chromosomal region (e.g. ‘chr1:100000-200000’), disease (string search or multiple option search), or OMIM disease IDs. To browse the results, the reference genome version (hg18/NCBI36 and hg19/NCBI37) must be selected before search. After submission, a result list is displayed, including mutation ID, mutation description on genome level, mutation description on cDNA level, muta- tion description on protein level, mutation type, mutation effect, associated diseases and PubMed IDs (Figure 1.1). To see more details about a mutation, the user can click the “view” button and a new page will appear to display all information about this mutation, both from manual col- lection and integration. The mutation page can be separated into three sections: 1) the basic annotation, 2) the reference and 3) the predic- tion (Figure 1.2). The basic annotation includes: 1) the position both on hg18 and hg19; 2) the standard descrip- tion on genome level, cDNA level and protein level; 3) the reported genes and the gene function annotation (GO); 4) the protein domain containing mutation and the protein structure 3D graph labeled with mutation (Figure 1.3). The reference information reveals the ethnicity report of the mutation, sequencing sample, methods, platform and description of mutation in articles. In the prediction results, new mutation-disease associations with scores and preferential proteins mediating such predicted linkages are shown to users. For example (Figure 1.4), the mutation NM_000492.3:c.350G>A (id: M301052)in gene CFTR is associated with Cystic fibrosis as reported with a signifi- cant preferred PPI with RNF5(PMID:21228398), which may contribute to new disease-mutation prediction in level-2 PPI neighbors. The preferred PPI network can be Record description Each mutation recorded in HDAM was given a unique identifier. Each mutation has three level symbols, includ- ing genome, cDNA and protein. The symbol is con- structed with sequence accession number in GenBank, sequence type symbol, mutation position and sequence change. For example, NM_020975.4:c.2944C>T refers a Jia et al. BMC Medical Genomics 2013, 6(Suppl 1):S16 http://www.biomedcentral.com/1755-8794/6/S1/S16 Jia et al. BMC Medical Genomics 2013, 6(Suppl 1):S16 http://www.biomedcentral.com/1755-8794/6/S1/S16 Page 4 of 6 Figure 1 Searching for information of a mutation in gene EGFR via the simple gene search interface. The panel 1 on the upper left corner is the search interface of HDAM and result table. The panel 2 on the right in the middle of the figure illustrates the content of a mutation page. Panel 3 on the top right corner shows the protein domain containing mutation and the protein structure 3D graph labeled with mutation via Jmol. Panel 4 on the bottom right conner shows the preferred path in PPI network (a) and the preferred PPI network (b). Figure 1 Searching for information of a mutation in gene EGFR via the simple gene search interface. The panel 1 on the upper left corner is the search interface of HDAM and result table. The panel 2 on the right in the middle of the figure illustrates the content of a mutation page. Panel 3 on the top right corner shows the protein domain containing mutation and the protein structure 3D graph labeled with mutation via Jmol. Panel 4 on the bottom right conner shows the preferred path in PPI network (a) and the preferred PPI network (b). shown in GUESS applet in the link page “SHOW THE PREFERRED PPI PATHWAYS”(Figure 1.4.a), while, a smaller network with only pathway nodes contained in the link page “SHOW THE PREFERRED PPI NETWORK” (Figure 1.4.b).All neighbor genes will be listed in disease gene Clique tab if all the genes associated with the given disease are significantly enriched in the level-1 neighbors of the queried mutation TGM3 or Transglutaminase 3 belongs to transglutami- nase family which are enzymes that catalyze the cross- linking of proteins by epsilon-gamma glutamyl lysine isopeptide bonds. In this protein family, another member TGM2 is related with autism spectrum disorders [22], indicating that there is a possibility that TGM3 has underlying effect on mechanism of autism. Record description Therefore, new mutated genes discovered by NGS technology should be pointed out, in need of further validation experiments. The quality and quantity of data are the most essential and crucial characteristics for a database. Therefore, the mutation collection will be a long-term task with great efforts. Considering that mutation infor- mation is widely utilized in the biomedical researches, periodically updating genomic annotation of mutations (e.g. position) is necessary for several reasons: 1) novel mutations are published; 2) more annotation data and references of existent data need to be added into Discussion HDAM currently contains 1,114 mutations, 669 genes and 125 human diseases (Figure 2). There are 444 genes in HDAM have no allelic variant records in OMIM. Actually, these genes discovered by NGS technology were confirmed with mutation events in patients, which have potential biological function and clinical applica- tions. For example, a de novo missense mutation is iden- tified in the gene TGM3 in an autistic proband [21]. Page 5 of 6 Jia et al. BMC Medical Genomics 2013, 6(Suppl 1):S16 http://www.biomedcentral.com/1755-8794/6/S1/S16 database; 3) new version of human reference genome is bl h d ) b d ll d f susceptibility to the common disorders. These data will d l bl d d d d f l Figure 2 Mutation and gene distribution on circos diagrams. The outer circle represents the chromosomes of the reference genome hg19 with chromosome band labeled. The blue rectangles in light blue ring are the mutated genes in OMIM. The green triangles in light green ring are genes in HDAM. The red dots in light red ring are the mutations recorded in HDAM. The text in middle represents statistics of mutation types in HDAM. Figure 2 Mutation and gene distribution on circos diagrams. The outer circle represents the chromosomes of the reference genome hg19 with chromosome band labeled. The blue rectangles in light blue ring are the mutated genes in OMIM. The green triangles in light green ring are genes in HDAM. The red dots in light red ring are the mutations recorded in HDAM. The text in middle represents statistics of mutation types in HDAM. Figure 2 Mutation and gene distribution on circos diagrams. The outer circle represents the chromosomes of the reference genome hg19 with chromosome band labeled. The blue rectangles in light blue ring are the mutated genes in OMIM. The green triangles in light green ring are genes in HDAM. The red dots in light red ring are the mutations recorded in HDAM. The text in middle represents statistics of mutation types in HDAM. database; 3) new version of human reference genome is published; 4) bugs and comments are collected from feedbacks. To meet the demand, HDAM will be continu- ously improved and updated. susceptibility to the common disorders. These data will provide an available and standardized resource for large- scale studies. Discussion However, due to the unformatted mutation descriptions in the publications, it is a time-consuming and labor intensive task for the mutation examination and disease category. HDAM extends the effect of standar- dized database, and makes mutation descriptions more standard in publications. Large-scale mutation disease/ traits studies provide more systematic and functional view Conclusion HDAM contains hundreds of rigorously curated human mutations from NGS studies and was created to provide a comprehensive view of these mutations that confer Jia et al. BMC Medical Genomics 2013, 6(Suppl 1):S16 http://www.biomedcentral.com/1755-8794/6/S1/S16 Page 6 of 6 3. Roach JC, Glusman G, Smit AF, Huff CD, Hubley R, Shannon PT, Rowen L, Pant KP, Goodman N, Bamshad M, et al: Analysis of genetic inheritance in a family quartet by whole-genome sequencing. Science 2010, 328:636-639. by integrating amount of mutations, disease/traits, and also various types of omics data. As a core of genome- wide annotation, genomic position is considered as the most important criterion in data correction. Nonetheless, disease category is hard to be selected, updated and com- pared with others, which is more complicated than muta- tion correction. Thus, standardization in the related field is urgently in demand. HDAM tries to contain HPO, OMIM and ICD definitions simultaneously, and builds a mutation-directed mapping table for further evaluation. The new predicted mutation-disease relations based on integrating OMIM allelic variants data and protein-protein network also provide possible new underlying causes for the related diseases. 4. Wu J, Xiao J, Zhang R, Yu J: DNA sequencing leads to genomics progress in China. Sci China Life Sci 2011, 54:290-292. 5. Jiang T, Yang L, Jiang H, Tian G, Zhang X: High-performance single-chip exon capture allows accurate whole exome sequencing using the Illumina Genome Analyzer. Sci China Life Sci 2011, 54:945-952. 6. Goh KI, Cusick ME, Valle D, Childs B, Vidal M, Barabasi AL: The human disease network. Proc Natl Acad Sci USA 2007, 104:8685-8690. 7. Mailman MD, Feolo M, Jin Y, Kimura M, Tryka K, Bagoutdinov R, Hao L, Kiang A, Paschall J, Phan L, et al: The NCBI dbGaP database of genotypes and phenotypes. Nat Genet 2007, 39:1181-1186. 5. Jiang T, Yang L, Jiang H, Tian G, Zhang X: High-performance single-chip exon capture allows accurate whole exome sequencing using the Illumina Genome Analyzer. Sci China Life Sci 2011, 54:945-952. y 6. Goh KI, Cusick ME, Valle D, Childs B, Vidal M, Barabasi AL: The human disease network. Proc Natl Acad Sci USA 2007, 104:8685-8690. 7. Mailman MD, Feolo M, Jin Y, Kimura M, Tryka K, Bagoutdinov R, Hao L, Kiang A, Paschall J, Phan L, et al: The NCBI dbGaP database of genotypes and phenotypes. Nat Genet 2007, 39:1181-1186. 8. Author details 1 1Center for Bioinformatics and Computational Biology, Shanghai Key Laboratory of Regulatory Biology, the Institute of Biomedical Sciences and School of Life Science, East China Normal University, Shanghai 200241, China. 2Department of Ophthalmology, West China Hospital, Sichuan University, Chengdu, Sichuan 610041, China. 3Schizophrenia Program, Shanghai Mental Health Center, Shanghai Jiao Tong University School of Medicine, Shanghai 200030, China. 4TCM Clinical Basis Institute, Zhejiang University of Chinese Medicine, Hangzhou, Zhejiang, 310053, China. 5Rush University Cancer Center, Department of Internal Medicine, Rush University Medical Center, Chicago, IL 60612, USA. 20. den Dunnen JT, Antonarakis SE: Mutation nomenclature extensions and suggestions to describe complex mutations: a discussion. Hum Mutat 2000, 15:7-12. 21. O’Roak BJ, Deriziotis P, Lee C, Vives L, Schwartz JJ, Girirajan S, Karakoc E, Mackenzie AP, Ng SB, Baker C, et al: Exome sequencing in sporadic autism spectrum disorders identifies severe de novo mutations. Nat Genet 2011, 43:585-589. 22. Rosenspire A, Yoo W, Menard S, Torres AR: Autism spectrum disorders are associated with an elevated autoantibody response to tissue transglutaminase-2. Autism Res 2011, 4:242-249. Conclusion Hamosh A, Scott AF, Amberger JS, Bocchini CA, McKusick VA: Online Mendelian Inheritance in Man (OMIM), a knowledgebase of human genes and genetic disorders. Nucleic Acids Res 2005, 33:D514-517. 9. 9. Yandell M, Moore B, Salas F, Mungall C, MacBride A, White C, Reese MG: Genome-wide analysis of human disease alleles reveals that their locations are correlated in paralogous proteins. PLoS Comput Biol 2008, 4: e1000218. Availability and requirements 10. Robinson PN, Kohler S, Bauer S, Seelow D, Horn D, Mundlos S: The Human 10. Robinson PN, Kohler S, Bauer S, Seelow D, Horn D, Mundlos S: The Human Phenotype Ontology: a tool for annotating and analyzing human hereditary disease. Am J Hum Genet 2008, 83:610-615. The Human Disease Associated Mutation (HDAM) database is publicly available at http://www.megabionet. org/HDAM. Phenotype Ontology: a tool for annotating and analyzing human hereditary disease. Am J Hum Genet 2008, 83:610-615. 11. Ayme S, Schmidtke J: Networking for rare diseases: a necessity for Europe. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2007, 50:1477-1483. 12. Kent WJ: BLAT–the BLAST-like alignment tool. Genome Res 2002, 12:656-664. 12. Kent WJ: BLAT–the BLAST-like alignment tool. Genome Res 2002, 12:656-664. Published: 23 January 2013 doi:10.1186/1755-8794-6-S1-S16 Cite this article as: Jia et al.: HDAM: a resource of human disease associated mutations from next generation sequencing studies. BMC Medical Genomics 2013 6(Suppl 1):S16. Acknowledgements This work was supported by the National 973 Key Basic Research Program (Grant Nos. 2010CB945401 and 2012CB910400), the National Natural Science Foundation of China (Grant No. 31171264, 31071162, 31000590 and 81171272) and the Science and Technology Commission of Shanghai Municipality (11DZ2260300). 15. Jmol: an open-source Java viewer for chemical structures in 3D. [http:// www.jmol.org/]. 16. Yu L, Gao L, Li K: A method based on local density and random walks for complexes detection in protein interaction networks. J Bioinform Comput Biol 2010, 8(Suppl 1):47-62. This article has been published as part of BMC Medical Genomics Volume 6 Supplement 1, 2013: Proceedings of the 2011 International Conference on Bioinformatics and Computational Biology (BIOCOMP’11). The full contents of the supplement are available online at http://www.biomedcentral.com/ bmcmedgenomics/supplements/6/S1. Publication of this supplement has 17. George RA, Liu JY, Feng LL, Bryson-Richardson RJ, Fatkin D, Wouters MA: Analysis of protein sequence and interaction data for candidate disease gene prediction. Nucleic Acids Res 2006, 34:e130. 18. Rhodes DR, Tomlins SA, Varambally S, Mahavisno V, Barrette T, Kalyana- Sundaram S, Ghosh D, Pandey A, Chinnaiyan AM: Probabilistic model of the human protein-protein interaction network. Nat Biotechnol 2005, 23:951-959. been supported by the International Society of Intelligent Biological Medicine. 19. Chua HN, Sung WK, Wong L: Exploiting indirect neighbours and topological weight to predict protein function from protein-protein interactions. Bioinformatics 2006, 22:1623-1630. Authors’ contributions TS and MJ conceived and designed the study. MJ, YL, ZS and CW collected the data, MJ and CZ performed the data analysis and built the system. TS, MJ and YD wrote the manuscript, TS finalize the manuscript. 13. Wildeman M, van Ophuizen E, den Dunnen JT, Taschner PE: Improving sequence variant descriptions in mutation databases and literature using the Mutalyzer sequence variation nomenclature checker. Hum Mutat 2008, 29:6-13. References 1. Ng SB, Buckingham KJ, Lee C, Bigham AW, Tabor HK, Dent KM, Huff CD, Shannon PT, Jabs EW, Nickerson DA, et al: Exome sequencing identifies the cause of a mendelian disorder. Nat Genet 2010, 42:30-35. 2. Goriely A, Hansen RM, Taylor IB, Olesen IA, Jacobsen GK, McGowan SJ, Pfeifer SP, McVean GA, Rajpert-De Meyts E, Wilkie AO: Activating mutations in FGFR3 and HRAS reveal a shared genetic origin for congenital disorders and testicular tumors. Nat Genet 2009, 41:1247-1252. Competing interests 14. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, et al: Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 2000, 25:25-29. The authors declare that they have no competing interests. 1. Ng SB, Buckingham KJ, Lee C, Bigham AW, Tabor HK, Dent KM, Huff CD, Shannon PT, Jabs EW, Nickerson DA, et al: Exome sequencing identifies the cause of a mendelian disorder. Nat Genet 2010, 42:30-35.
https://openalex.org/W2999091083	https://ieeexplore.ieee.org/ielx7/6221020/9056600/08954690.pdf	English	null	Objective Analysis of Neck Muscle Boundaries for Cervical Dystonia Using Ultrasound Imaging and Deep Learning	IEEE journal of biomedical and health informatics	2,020	cc-by	12,929	IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, VOL. 24, NO. 4, APRIL 2020 IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, VOL. 24, NO. 4, APRIL 2020 1016 I. INTRODUCTION Kobylecki are with the Department of Neurology, Salford Royal NHS Foundation Trust, Manchester Academic Health Sciences Centre, The University of Manch- ester, Oxford Road Manchester, M13 9PT Manchester, U.K. (e-mail: monty.silverdale@manchester.ac.uk; christopher.kobylecki@ manchester.ac.uk). Salford, U.K. (e mail: abdul.siddique@srft.nhs.uk). M. Silverdale and C. Kobylecki are with the Department of Neurology, Salford Royal NHS Foundation Trust, Manchester Academic Health Sciences Centre, The University of Manch- ester, Oxford Road Manchester, M13 9PT Manchester, U.K. (e-mail: monty.silverdale@manchester.ac.uk; christopher.kobylecki@ manchester.ac.uk). Objective Analysis of Neck Muscle Boundaries for Cervical Dystonia Using Ultrasound Imaging and Deep Learning Ian Loram , Member, IEEE, Abdul Siddique , María B. Sánchez , Pete Harding, Monty Silverdale , Christopher Kobylecki , and Ryan Cunningham , Member, IEEE Ian Loram , Member, IEEE, Abdul Siddique , María B. Sánchez , Pete Harding, Monty Silverdale , Christopher Kobylecki , and Ryan Cunningham , Member, IEEE Abstract—Objective: To provide objective visualization and pattern analysis of neck muscle boundaries to inform and monitor treatment of cervical dystonia. Methods: We recorded transverse cervical ultrasound (US) images and whole-body motion analysis of sixty-one standing participants (35 cervical dystonia, 26 age matched controls). We manually annotated 3,272 US images sampling posture and the functional range of pitch, yaw, and roll head movements. Using previously validated methods, we used 60-fold cross validation to train, validate and test a deep neural network (U-net) to classify pixels to 13 categories (ﬁve paired neck muscles, skin, ligamentum nuchae, vertebra). For all participants for their normal standing posture, we segmented US images and classiﬁed condition (Dystonia/Control), sex and age (higher/lower) from segment boundaries. We performed an explanatory, visualization analysis of dystonia muscle-boundaries. Results: For all segments, agreement with manual labels was Dice Coefﬁcient (64 ± 21%) and Hausdorff Distance (5.7 ± 4 mm). For deep muscle layers, boundaries predicted central injection sites with average precision 94 ± 3%. Using leave-one-out cross-validation, a support-vector- machine classiﬁed condition, sex, and age from predicted muscle boundaries at accuracy 70.5%, 67.2%, 52.4% respectively, exceeding classiﬁcation by manual labels. From muscle boundaries, Dystonia clustered optimally into three sub-groups. These sub-groups are visualized and explained by three eigen-patterns which correlate signiﬁcantly with truncal and head posture. Conclusion: Using US, neck muscle shape alone discriminates dystonia from healthy controls. Signiﬁcance: Using deep learning, US imaging allows online, automated visualization, and diagnostic analysis of cervical dystonia and segmentation of individual muscles for targeted injection. Index Terms—Deep learning, ultrasound imaging, cervical dystonia, segmentation, muscle boundaries, diagnosis. I. INTRODUCTION C C ERVICAL Dystonia (CD), also called spasmodic torti- collis, is a painful condition in which the neck muscles contract involuntarily, causing the head to twist, turn, and pull into an abnormal posture. This neurological movement disorder affects an estimated 18,000 adults in the UK [1]. The reported mean duration from symptom onset to diagnosis is 44 months, withconsultationssoughtfromameanof3.5differenthealthcare providers before reaching a diagnosis and receiving effective therapy [2]. For CD, the diagnosis is based on expert clinical assessment since laboratory testing and imaging of the brain or spine is typically unrevealing [2]. Manuscript received April 30, 2019; revised October 31, 2019 and December 23, 2019; accepted December 30, 2019. Date of publica- tion January 9, 2020; date of current version April 6, 2020. This work was funded by The Dystonia Society Grant A clinical tool for real-time analysis and visualization of cervical muscles for cervical dystonia, by Department of Life Sciences, Manchester Metropolitan University (MMU) for R Cunningham and by Salford Royal Foundation Trust (SRFT) Clinical Research Fellowship for A Siddique. (Corresponding authors: Ian Loram; Ryan Cunningham.) Treatment of CD is symptomatic and the established protocol is injecting the neck muscles with botulinum neurotoxin (BoNT) [3]. Clinical experience shows the main causes for treatment failure are suboptimal neck muscle selection or BoNT dosing, indicating the importance of appropriate targeting of overactive muscles [4], [5]. Furthermore, monitoring the effectiveness of treatment is confounded by use of differing rating scales and assessment methods [6]. There is a clinical need to diagnose CD more promptly, to improve analysis and identiﬁcation of dystonic muscles, to improve delivery of injection and dose to speciﬁc muscles, to provide objective recording of injection sites for retention within medical records and to track longitudinally the effect of injections on individual muscles [7]. I. Loram is with Cognitive Motor Function Research Group, Research Centre for Musculoskeletal Science & Sports Medicine, MMU, M1 5GD Manchester, U.K. (e-mail: i.loram@mmu.ac.uk). I. Loram is with Cognitive Motor Function Research Group, Research Centre for Musculoskeletal Science & Sports Medicine, MMU, M1 5GD Manchester, U.K. (e-mail: i.loram@mmu.ac.uk). ( ) M. B. Sánchez is with the Department of Health Professions, MMU (e-mail: m.sanchez.puccini@mmu.ac.uk). R. Cunningham is with the Centre for Advanced Computa- tional Science, Department of Health Professions, MMU (e-mail: ryan.cunningham@mmu.ac.uk). A. Siddique is with the Department of Neurology, SRFT, M6 8HD Salford, U.K. (e-mail: abdul.siddique@srft.nhs.uk). M. Silverdale and C. ( ) M. B. Sánchez is with the Department of Health Professions, MMU (e-mail: m.sanchez.puccini@mmu.ac.uk). R. Cunningham is with the Centre for Advanced Computa- tional Science, Department of Health Professions, MMU (e-mail: ryan.cunningham@mmu.ac.uk). This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see http://creativecommons.org/licenses/by/4.0/ A. Current Clinical Methods H4, Cervical Dystonia can be reduced to signiﬁcant eigen-patterns of muscle shape (Figs. 6–8). H5, Eigen-patterns of neck muscle shape are associated with features of whole body posture (Fig. 8). The objective of this study is to provide an automated, ob- jective visualization of neck muscle boundaries and to analyze whether these boundaries have diagnostic value discriminating patterns of cervical dystonia from healthy controls. If successful, thesemethodsdemonstrateproofofconceptforaclinicaltoolfor objective online diagnosis, injection guidance and monitoring, with minimal requirement for operator expertise and minimal burden on clinical time. Ultrasound (US) offers non-invasive visualisation of muscle structures with easy contralateral comparison, is readily avail- able, and improves the precision of injections [12], [13]. How- ever, use of US requires training, is dependent upon operator expertise, and remains subjective [12]. A. Current Clinical Methods For the whole dataset, images were acquired at level 3.8 ± 0.6 (mean ± SD). Fig. 1. Posture (linked US-MRI) datase. (A) Representative axial ultrasound image targeted at vertebral level C4. The probe plane was marked with four cod liver oil capsules. (B), (C), (D): Axial, coronal and sagittal MR images of the same participant showing the ultrasound image plane marked by four cod liver oil capsules (blue circles). Here, the ultrasound image plane lies between cervical vertebrae C3 and C4. Note the challenge of extracting muscle boundaries from the ultrasound image. For the whole dataset, images were acquired at level 3.8 ± 0.6 (mean ± SD). needle placement, based on the clinician’s knowledge of func- tional anatomy of the neck muscles, directed by head position, shoulder elevation and assessment of muscle tone and hyper- trophy on palpation [8]. Within the clinic, this method may be efﬁcacious for superﬁcial neck muscles but is compromised for deeper muscles - typically the deeper neck muscles (e.g., spinalis cervicis, multiﬁdus) are difﬁcult to assess clinically and inject. There is increasing awareness of the clinical relevance of deep cervical muscles in the pathogenesis and potential therapy of CD [9], [10], but the tools to assess and treat these muscles are currently not fully developed. Compared with intramuscular electromyographic (iEMG) mapping of cervical muscle activity, the sensitivity of clinical examination has been reported as 59% and the speciﬁcity 75% [8]. The positive predictive value of shoulder elevation and muscle hypertrophy is reportedly only 70% and head position does not provide added value, because individuals with solitary dystonic head postures do not have muscle dystonia following simple patterns [8]. Without iEMG mapping, 41% of dystonic muscles would not be recognized and 25% of inactive muscles would be judged dystonic [8]. However, iEMG is time consuming, requires substantial expertise, is inva- sive and cannot be performed in individuals on anticoagulants. Other methods including measurement of electrical impedance have been proposed to be sensitive to muscle changes in CD, but are not as yet validated [11]. needle placement, based on the clinician’s knowledge of func- tional anatomy of the neck muscles, directed by head position, shoulder elevation and assessment of muscle tone and hyper- trophy on palpation [8]. A. Current Clinical Methods P. Harding is with the Elements Technology Platforms Ltd, S1 2NS Shefﬁeld, U.K. (e-mail: pete@elementstechnology.co.uk). P. Harding is with the Elements Technology Platforms Ltd, S1 2NS Shefﬁeld, U.K. (e-mail: pete@elementstechnology.co.uk). The most common method of identifying and injecting the muscles involved in CD is clinical examination and manual reative Commons Attribution 4.0 License. For more information, see http://creativecommons.org/licenses/by/4.0/ nsed under a Creative Commons Attribution 4.0 License. For more information, see http://creativecommons.org/lic his work is licensed under a Creative Commons Attribution 4.0 License. For more information, see http://creativec LORAM et al.: OBJECTIVE ANALYSIS OF NECK MUSCLE BOUNDARIES FOR CERVICAL DYSTONIA USING US IMAGING 1017 Fig. 1. Posture (linked US-MRI) datase. (A) Representative axial ultrasound image targeted at vertebral level C4. The probe plane was marked with four cod liver oil capsules. (B), (C), (D): Axial, coronal and sagittal MR images of the same participant showing the ultrasound image plane marked by four cod liver oil capsules (blue circles). Here, the ultrasound image plane lies between cervical vertebrae C3 and C4. Note the challenge of extracting muscle boundaries from the ultrasound image. For the whole dataset, images were acquired at level 3.8 ± 0.6 (mean ± SD). Fig. 2. Hypothesis pipeline: Using Deep Learning Semantic Seg- mentation, Boundary Extraction, support vector machine (SVM) Clas- siﬁcation, Clustering, Pattern Analysis, Visualisation and whole body motion analysis, we sequentially test ﬁve hypotheses concerning trans- verse ultrasound images of the human neck. H1, Segmentation is ac- curate enough to guide injection of deep neck muscles (Figs. 4, 5). H2, Cervical Dystonia can be classiﬁed from age matched controls using muscle shape alone (Table III). H3, Muscle shape clusters into subtypes of Cervical Dystonia (Fig. 5). H4, Cervical Dystonia can be reduced to signiﬁcant eigen-patterns of muscle shape (Figs. 6–8). H5, Eigen-patterns of neck muscle shape are associated with features of whole body posture (Fig. 8). Fig. 1. Posture (linked US-MRI) datase. (A) Representative axial ultrasound image targeted at vertebral level C4. The probe plane was marked with four cod liver oil capsules. (B), (C), (D): Axial, coronal and sagittal MR images of the same participant showing the ultrasound image plane marked by four cod liver oil capsules (blue circles). Here, the ultrasound image plane lies between cervical vertebrae C3 and C4. Note the challenge of extracting muscle boundaries from the ultrasound image. A. Current Clinical Methods Within the clinic, this method may be efﬁcacious for superﬁcial neck muscles but is compromised for deeper muscles - typically the deeper neck muscles (e.g., spinalis cervicis, multiﬁdus) are difﬁcult to assess clinically and inject. There is increasing awareness of the clinical relevance of deep cervical muscles in the pathogenesis and potential therapy of CD [9], [10], but the tools to assess and treat these muscles are currently not fully developed. Compared with intramuscular electromyographic (iEMG) mapping of cervical muscle activity, the sensitivity of clinical examination has been reported as 59% and the speciﬁcity 75% [8]. The positive predictive value of shoulder elevation and muscle hypertrophy is reportedly only 70% and head position does not provide added value, because individuals with solitary dystonic head postures do not have muscle dystonia following simple patterns [8]. Without iEMG mapping, 41% of dystonic muscles would not be recognized and 25% of inactive muscles would be judged dystonic [8]. However, iEMG is time consuming, requires substantial expertise, is inva- sive and cannot be performed in individuals on anticoagulants. Other methods including measurement of electrical impedance have been proposed to be sensitive to muscle changes in CD, but are not as yet validated [11]. Fig. 2. Hypothesis pipeline: Using Deep Learning Semantic Seg- mentation, Boundary Extraction, support vector machine (SVM) Clas- siﬁcation, Clustering, Pattern Analysis, Visualisation and whole body motion analysis, we sequentially test ﬁve hypotheses concerning trans- verse ultrasound images of the human neck. H1, Segmentation is ac- curate enough to guide injection of deep neck muscles (Figs. 4, 5). H2, Cervical Dystonia can be classiﬁed from age matched controls using muscle shape alone (Table III). H3, Muscle shape clusters into subtypes of Cervical Dystonia (Fig. 5). H4, Cervical Dystonia can be reduced to signiﬁcant eigen-patterns of muscle shape (Figs. 6–8). H5, Eigen-patterns of neck muscle shape are associated with features of whole body posture (Fig. 8). Fig. 2. Hypothesis pipeline: Using Deep Learning Semantic Seg- mentation, Boundary Extraction, support vector machine (SVM) Clas- siﬁcation, Clustering, Pattern Analysis, Visualisation and whole body motion analysis, we sequentially test ﬁve hypotheses concerning trans- verse ultrasound images of the human neck. H1, Segmentation is ac- curate enough to guide injection of deep neck muscles (Figs. 4, 5). H2, Cervical Dystonia can be classiﬁed from age matched controls using muscle shape alone (Table III). H3, Muscle shape clusters into subtypes of Cervical Dystonia (Fig. 5). B. Contribution of This Study 10 frames per second) on a modest PC or laptop. Hyperparameters and augmentation were ﬁxed, chosen based on the experience and well-established literature. Hyperparameters: Dropout = 0.25,L^2 = 0.0005, Adam(α = 0.00005, β_1 = 0.9, β_2 = 0.999), BatchSize = 1,Epochs = 40. Data Augmentation: Local Contrast Normalisation = 31 × 31, Rotation = ± 8°, Transx = ± 128, Transy = ± 64 Fig. 3. U-Net Model Architecture. This ﬁgure details the best per- forming model, according to Table I Supplementary Material. The model consists of 2D convolutional and pooling layers in the encoder part of the network (blocks to the left), and 2D up-sampling, concatenation and convolutional layers in the decoder part of the network (blocks to the right), where concatenation layers concatenate up-sampled layers along the feature channels with compatible layers in the encoder network enabling ﬂow of information and gradients in forward and backward passes, respectively. This neural network has over 51,000,000 trainable parameters, and over 21,000,000 functional outputs, and operates in real-time (approx. 10 frames per second) on a modest PC or laptop. Hyperparameters and augmentation were ﬁxed, chosen based on the experience and well-established literature. Hyperparameters: Dropout = 0.25,L^2 = 0.0005, Adam(α = 0.00005, β_1 = 0.9, β_2 = 0.999), BatchSize = 1,Epochs = 40. Data Augmentation: Local Contrast Normalisation = 31 × 31, Rotation = ± 8°, Transx = ± 128, Transy = ± 64 Our primary interest is to establish whether from a single axial image, neck muscle shape allows differentiation of cervical dys- tonia from healthy controls. Our secondary interest is whether or not segmentation is accurate enough to guide injections to deep neck muscles. Cervical dystonia is characterized by the sustained contrac- tion of speciﬁc muscles. Our general hypothesis is that those contracting muscles cause an identiﬁable pattern of neck muscle shape and an associated pattern of whole body posture. Dystonic muscles are speciﬁc to an individual, but common combinations are observed. Each dystonic combination should produce a pattern of neck muscle shape or posture away from the normal distribution, along a dimension which is distinct from other combinations. axial US images of the posterior neck targeted at level C4 were recorded, each with renewed probe placement (Fig. 1A). Four cod liver oil capsules were taped (using Transpore medical tape) to the neck, two either side of the neck approximately in the image plane of the probe. B. Contribution of This Study The probe was removed, leaving the capsules in place, and an MRI scan (0.3T open MRI scanner, G-Scan, Esaote, Italy) was obtained with participants lying supine on the scanning bed and their neck positioned central within a cervical imaging coil. Axial scans (Spin T1-weighted HF, matrix 512 × 512) were performed in a range from the upper jaw line to the clavicle, orthogonal to the spine, in 19 equidistant sections (Fig. 1B-D). Manual labelling of muscle boundaries in US images is challenging. The purpose of collecting linked MRI-US data is to train experts in labelling US images using methods reported previously [24]. We test in sequence ﬁve speciﬁc hypotheses (Fig. 2): (i) Segmentation will identify injection points within deep neck muscles accurately. (ii) Cervical dystonia can be classiﬁed from age matched controls using neck muscle shape alone. (iii) Dysto- nia clusters into natural sub-groups using neck muscle shape (iv) Dystonic muscle shapes can be expressed as signiﬁcant eigen- patterns, (v) Dystonic muscle eigen-patterns are associated with patterns of whole body posture. II. METHODS B. Contribution of This Study The use of deep learning to extract information from limited quality images (c.f. Fig. 1A) is progressing rapidly. Application IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, VOL. 24, NO. 4, APRIL 2020 1018 Fig. 3. U-Net Model Architecture. This ﬁgure details the best per- forming model, according to Table I Supplementary Material. The model consists of 2D convolutional and pooling layers in the encoder part of the network (blocks to the left), and 2D up-sampling, concatenation and convolutional layers in the decoder part of the network (blocks to the right), where concatenation layers concatenate up-sampled layers along the feature channels with compatible layers in the encoder network enabling ﬂow of information and gradients in forward and backward passes, respectively. This neural network has over 51,000,000 trainable parameters, and over 21,000,000 functional outputs, and operates in real-time (approx. 10 frames per second) on a modest PC or laptop. Hyperparameters and augmentation were ﬁxed, chosen based on the experience and well-established literature. Hyperparameters: Dropout = 0.25,L^2 = 0.0005, Adam(α = 0.00005, β_1 = 0.9, β_2 = 0.999), BatchSize = 1,Epochs = 40. Data Augmentation: Local Contrast Normalisation = 31 × 31, Rotation = ± 8°, Transx = ± 128, Transy = ± 64 to US is under-developed and application to skeletal muscle is rare. This study builds upon previous work by our group realizing the scientiﬁc and clinical value of in-vivo skeletal muscle analysis [14]–[19], applying deep learning to skeletal muscle US [20]–[22] and speciﬁcally developing methods for analysis of the neck muscles [23], [24]. Recently we contributed a dataset, a methodology for labelling training images suitable for participants with involuntary head movement, and a bench- mark deep learning method for segmenting the neck muscles [24], [25]. Here we apply that methodology to cervical dystonia. Fig. 3. U-Net Model Architecture. This ﬁgure details the best per- forming model, according to Table I Supplementary Material. The model consists of 2D convolutional and pooling layers in the encoder part of the network (blocks to the left), and 2D up-sampling, concatenation and convolutional layers in the decoder part of the network (blocks to the right), where concatenation layers concatenate up-sampled layers along the feature channels with compatible layers in the encoder network enabling ﬂow of information and gradients in forward and backward passes, respectively. This neural network has over 51,000,000 trainable parameters, and over 21,000,000 functional outputs, and operates in real-time (approx. A. Data Collection Two thousand US images (∼30 per participant) from the Head Motion (linked US-Vicon) dataset, sampling uniformly the range of pitch, yaw and roll head rotations [24] were labelled manually for the same 13 segments. Two thousand US images (∼30 per participant) from the Head Motion (linked US-Vicon) dataset, sampling uniformly the range of pitch, yaw and roll head rotations [24] were labelled manually for the same 13 segments. Fig. 4. Boundary Extraction and Injection Point Analysis. This illustrates the output and visualisation provided by the neural network approach. This participant had cervical dystonia (note the asymmetry). Top row: Left: ultrasound image. Middle: Manually deﬁned labels. Right: neural network predicted classiﬁcation of pixels. 2nd row: Left: boundaries (green) estimated around each segment. Middle: Injection point (red star) deﬁned as pixel most distant from predicted segment boundary. Colour spectrum blue to yellow shows decreasing distance from predicted boundary and hence increasing target margin around injection point. Boundary of right multiﬁdus label (green). 3rd row: Left and middle: as 2nd row middle, but predicted pixels shown restricted to conﬁdence greater than 60 and greater than 80% respectively. Right: conﬁdence of the neural network (yellow = high conﬁdence) in clas- sifying pixels. Right: Average precision of pixels within target region for right multiﬁdus. Lines red, yellow, magenta, green show pixels re- stricted to conﬁdence greater than 20, 40, 60, 80% respectively. The clear visualisation of predicted segment (top right) will be appreciated by clinicians. The visualisation of conﬁdence of classifying pixels (2nd row right) gives the user feedback regarding optimal probe location and orientation. Fig. 4. Boundary Extraction and Injection Point Analysis. This illustrates the output and visualisation provided by the neural network approach. This participant had cervical dystonia (note the asymmetry). Top row: Left: ultrasound image. Middle: Manually deﬁned labels. Right: neural network predicted classiﬁcation of pixels. 2nd row: Left: boundaries (green) estimated around each segment. Middle: Injection point (red star) deﬁned as pixel most distant from predicted segment boundary. Colour spectrum blue to yellow shows decreasing distance from predicted boundary and hence increasing target margin around injection point. Boundary of right multiﬁdus label (green). 3rd row: Left and middle: as 2nd row middle, but predicted pixels shown restricted to conﬁdence greater than 60 and greater than 80% respectively. Right: conﬁdence of the neural network (yellow = high conﬁdence) in clas- sifying pixels. C. Machine Learning We divided data into independent training, validation and test datasets. We report a ‘Testing Mode’, and an ‘Analysis Mode’ of division. ‘Testing Mode’ is used for testing Segmentation, Boundary extraction and SVM classiﬁcation of dystonia, sex and age (Fig. 2). A. Data Collection Right: Average precision of pixels within target region for right multiﬁdus. Lines red, yellow, magenta, green show pixels re- stricted to conﬁdence greater than 20, 40, 60, 80% respectively. The clear visualisation of predicted segment (top right) will be appreciated by clinicians. The visualisation of conﬁdence of classifying pixels (2nd row right) gives the user feedback regarding optimal probe location and orientation. Image-labels from this Cervical Dystonia Project (CDP) were supplemented by our previous posture dataset (25 linked US- MRI neck image-labels) [23] and our previous Head Motion dataset (1100 linked US-Vicon neck image-labels) [24]. These supplementary image labels were acquired using a different US probe and machine (7.5 MHz T shaped probe, taped to the neck, Aloka) at the same neck location and for the same posture and head motion tasks. A. Data Collection Using a probe (7.5 MHz, SonixTouch, Ultrasonix, USA) held transversely to the posterior neck, B mode US images (depth 5cm), were recorded from 61 adults: 35 cervical dystonia (mean age 61 ± 10, 15 male) and 26 age matched controls, (mean age 59 ± 14 years, 18 male) while standing and while performing head rotation tasks deﬁning their range of pitch, yaw and head rotation. Power and contrast were adjusted per participant using visual feedback. We disabled image enhancement processes to reduce internal frame averaging. Images were interpolated (bi- linear) to a size, and resolution common to our previous datasets (491 × 525 pixel, 10 pixels per mm) [24]. These experiments, performed in the Faculty of Science and Engineering, Manch- ester Metropolitan University (MMU), received ethical approval from the NHS Health Research Authority (REC: 15/NW/0016, IRAS:169803)andfromMMUScienceandEngineeringFaculty Ethics Committee. The study was conducted in accordance with the Declaration of Helsinki guidelines. All values are reported as mean ± SD unless stated otherwise. Head Motion (Linked US-Vicon) Dataset: In a separate ses- sion, forty seven retroreﬂective markers were attached to the body to allow motion analysis of eighteen body segments (head, neck, thorax, pelvis, thighs, shanks, feet, clavicles, upper arms, forearms, hands). Participants stood in the middle of the calibrated volume and were instructed to perform pitch (ﬂexion/extension), yaw (right/left) and roll (right side/left side) head rotations, turning their heads as far as possible in both directions. Each trial was repeated starting in the opposite direction. Body motion was recorded by a 9 camera Vicon MX motion capture system. For each trial, the US probe was held to the posterior neck targeted at level C4, to allow free movement of the head and image of 5 bilateral layers of muscles. Images were saved digitally at 10 Hz with start time synchronized to the Vicon recording. The purpose of collecting linked Vicon-US data is to inves- tigate the relationship between neck muscle boundaries and posture/movement. Whole body kinematic data provides an additional modality of explanation and validation of the infor- mation content of neck muscle boundaries in US images Posture (Linked US-MRI) Dataset: These posture images are the subject of this paper. Participants stood upright, observing a monitor at 1m distance, just below eye level. A. Data Collection The visualisation of conﬁdence of classifying pixels (2nd row right) gives the user feedback regarding optimal probe location and orientation. Fig. 5. Precision of Injection points. Testing Mode results: (A) Average Precision for all pixels of conﬁdence more than 80% in the target region for a variety of sizes of target region. (B) Average Precision for varying levels of conﬁdence. Precision is number accurate as a percentage of pixels predicted to be within muscles. Average pre- cision is precision averaged with respect to recall, sorted in descending conﬁdence. Two thousand US images (∼30 per participant) from the Head Motion (linked US-Vicon) dataset, sampling uniformly the range of pitch, yaw and roll head rotations [24] were labelled manually for the same 13 segments. Image-labels from this Cervical Dystonia Project (CDP) were supplemented by our previous posture dataset (25 linked US- MRI neck image-labels) [23] and our previous Head Motion dataset (1100 linked US-Vicon neck image-labels) [24]. These supplementary image labels were acquired using a different US probe and machine (7.5 MHz T shaped probe, taped to the neck, Aloka) at the same neck location and for the same posture and head motion tasks. C. Machine Learning We divided data into independent training, validation and test datasets. We report a ‘Testing Mode’, and an ‘Analysis Mode’ of division. ‘Testing Mode’ is used for testing Segmentation, B d i d SVM l iﬁ i f d i Fig. 5. Precision of Injection points. Testing Mode results: (A) Average Precision for all pixels of conﬁdence more than 80% in the target region for a variety of sizes of target region. (B) Average Precision for varying levels of conﬁdence. Precision is number accurate as a percentage of pixels predicted to be within muscles. Average pre- cision is precision averaged with respect to recall, sorted in descending conﬁdence. Fig. 5. Precision of Injection points. Testing Mode results: (A) Average Precision for all pixels of conﬁdence more than 80% in the target region for a variety of sizes of target region. (B) Average Precision for varying levels of conﬁdence. Precision is number accurate as a percentage of pixels predicted to be within muscles. Average pre- cision is precision averaged with respect to recall, sorted in descending conﬁdence. A. Data Collection Three or more LORAM et al.: OBJECTIVE ANALYSIS OF NECK MUSCLE BOUNDARIES FOR CERVICAL DYSTONIA USING US IMAGING LORAM et al.: OBJECTIVE ANALYSIS OF NECK MUSCLE BOUNDARIES FOR C Fig. 4. Boundary Extraction and Injection Point Analysis. This illustrates the output and visualisation provided by the neural network approach. This participant had cervical dystonia (note the asymmetry). Top row: Left: ultrasound image. Middle: Manually deﬁned labels. Right: neural network predicted classiﬁcation of pixels. 2nd row: Left: boundaries (green) estimated around each segment. Middle: Injection point (red star) deﬁned as pixel most distant from predicted segment boundary. Colour spectrum blue to yellow shows decreasing distance from predicted boundary and hence increasing target margin around injection point. Boundary of right multiﬁdus label (green). 3rd row: Left and middle: as 2nd row middle, but predicted pixels shown restricted to conﬁdence greater than 60 and greater than 80% respectively. Right: conﬁdence of the neural network (yellow = high conﬁdence) in clas- sifying pixels. Right: Average precision of pixels within target region for right multiﬁdus. Lines red, yellow, magenta, green show pixels re- stricted to conﬁdence greater than 20, 40, 60, 80% respectively. The clear visualisation of predicted segment (top right) will be appreciated by clinicians. The visualisation of conﬁdence of classifying pixels (2nd row right) gives the user feedback regarding optimal probe location and orientation. 1019 Fig. 4. Boundary Extraction and Injection Point Analysis. This illustrates the output and visualisation provided by the neural network approach. This participant had cervical dystonia (note the asymmetry). Top row: Left: ultrasound image. Middle: Manually deﬁned labels. Right: neural network predicted classiﬁcation of pixels. 2nd row: Left: boundaries (green) estimated around each segment. Middle: Injection point (red star) deﬁned as pixel most distant from predicted segment boundary. Colour spectrum blue to yellow shows decreasing distance from predicted boundary and hence increasing target margin around injection point. Boundary of right multiﬁdus label (green). 3rd row: Left and middle: as 2nd row middle, but predicted pixels shown restricted to conﬁdence greater than 60 and greater than 80% respectively. Right: conﬁdence of the neural network (yellow = high conﬁdence) in clas- sifying pixels. Right: Average precision of pixels within target region for right multiﬁdus. Lines red, yellow, magenta, green show pixels re- stricted to conﬁdence greater than 20, 40, 60, 80% respectively. The clear visualisation of predicted segment (top right) will be appreciated by clinicians. B. Image Labelling 2), we selected the mode (‘Testing’ v ‘Analysis’) with the highest SVM classiﬁcation of dystonia and sex. Augmentation: Each US image and corresponding label was ﬂipped about the vertical line of symmetry, to double each training, validation and test set and to remove asymmetry bias from each process of training, validation and testing. Implementation: Following previous work [24], and using software written within this group, we conducted extensive evaluation of 99 trained neural networks (c.f. SM for detail). The best encode-decoder neural-network (U-Net) was trained (Fig. 3). Training error between labels and network predic- tion was computed using a class-weighted cross-entropy cost function −m i=0 β(k)y(i) (k) log ˆy(i) where m is the number of pixels in a single image, i is the index of a pixel in a single image, k is the class associated with the pixel i, y(i) (k) is the label category (0 or 1), ˆy(i) is the SoftMax response, and β(k), which up-weights (β(k) ≥1) a given class using, β(k)= max(C) Ck , where Ck is the total count of pixels of class k, and max(C) is the total count of pixels of the class with the maximum total pixel count. Fig. 6. Classifying Dystonia, sex and age from segment bound- aries. For 36 deep neural networks of varying architecture and hyper parameters dating from the very start of this investigation tested on the Posture dataset (61 participants), panels show Jaccard Index, and SVM classiﬁcation accuracy for Dystonia, sex and age v mean predicted con- ﬁdence. Points of conﬁdence <75%, are early 10 fold cross validation networks using only the Posture dataset (∼200 image-labels). Points of conﬁdence >75% include addition of Head Motion dataset (∼3000 image-labels) to neural network trained in Analysis Mode, except one point which is the Testing Mode U-net. Red horizontal lines show SVM classiﬁcation using the manually annotated boundaries used to train the neural networks. Key Points: Neural networks become better at extracting information than the human annotated ground truth used to train them. Ultrasound neck images contain the information required to classify Dystonia. Sex information is contained less well and age is not revealed within these US images. Jaccard Index increases with conﬁ- dence. Classiﬁcation of condition, sex or age requires US information content in addition to accurate segmentation. B. Image Labelling Using published methods two annotators were trained to a common standard using MRI images linked to US images [24]; their agreement is shown Supplementary Material (SM) (SM-Table IV). US images (192 total, ∼3 per participant) from the Posture (linked US-MRI) dataset were labelled manually by annotating the boundaries around ten muscles, vertebra, ligamentum nuchae and skin. As described previously [23], [24], MRI images showing the same cod liver oil capsule marked plane were annotated and registered to the US images to guide annotation of the US images. Testing Mode: 60-fold leave-one-out (LOO) cross validation. The 61 CDP participants (2192 image-labels) were assigned into 60 folds (one per fold, except one fold contained two partic- ipants). These 60 folds provided 30 groups each containing a ‘test’ (one participant), ‘validation’ (one participant) and ‘train’ (58 participants) dataset. Each ‘train’ dataset was supplemented by the 1100 Head Motion image-labels [24]. Analysis Mode: All 3100 (2000 + 1100) Head Mo- tion image-labels were assigned to the ‘train’ dataset. All 217 (192 + 25) Posture image-labels were assigned to the IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, VOL. 24, NO. 4, APRIL 2020 1020 Fig. 6. Classifying Dystonia, sex and age from segment bound- aries. For 36 deep neural networks of varying architecture and hyper parameters dating from the very start of this investigation tested on the Posture dataset (61 participants), panels show Jaccard Index, and SVM classiﬁcation accuracy for Dystonia, sex and age v mean predicted con- ﬁdence. Points of conﬁdence <75%, are early 10 fold cross validation networks using only the Posture dataset (∼200 image-labels). Points of conﬁdence >75% include addition of Head Motion dataset (∼3000 image-labels) to neural network trained in Analysis Mode, except one point which is the Testing Mode U-net. Red horizontal lines show SVM classiﬁcation using the manually annotated boundaries used to train the neural networks. Key Points: Neural networks become better at extracting information than the human annotated ground truth used to train them. Ultrasound neck images contain the information required to classify Dystonia. Sex information is contained less well and age is not revealed within these US images. Jaccard Index increases with conﬁ- dence. Classiﬁcation of condition, sex or age requires US information content in addition to accurate segmentation. or testing. To select the mode with best descriptive power for Pattern Analysis and Visualisation (Fig. B. Image Labelling Network training consisted of online learning, interrupted every quarter pass (550 learning iterations) through the training set, to record cross entropy test results from the validation and test (test) sets. If the cross-entropy loss for either test set was lower than any previous recorded loss for that test set, the network was saved to long term storage. Each selected network was tested by the other set, and vice versa for both networks. This process yielded held-out test results for all images in both test sets. Training terminated after 35 epochs. The Posture dataset and predicted output was used for post neural network analysis. Please refer to Fig. 2 which deﬁnes the ﬂow of hypothesis, methods and results. ‘validation’ and ‘test’ datasets and participants were assigned alternately to the ‘validation’ and ‘test’ datasets. The ‘Testing Mode’ maintains strict independence between training, validation and test dataset since participants do not overlap folds. To maximize the training set in each fold, we performed LOO cross-validation. Since, some participants pro- duce poorer quality images (e.g., deep fat layer, indistinct mus- cles), validation (selection of the training iteration to use for testing) will be sub-optimal. Typically, validation scores over ﬁt prematurely before the neural network fully encodes the generalizable content of the data. ‘Testing Mode’ represents the harshest possible testing regime. D. Boundary Extraction THE PENULTIMATE COLUMN ALONE REPORTS ALL SEGMENTS USING ‘ANALYSIS MODE’. THE FINAL COLUMN REPORTS AGREEMENT BETWEEN THE TWO ANNOTATORS WHO CONTRIBUTED THE LABELS FOR THIS DATASET TABLE I BOUNDARY ACCURACY BETWEEN BOUNDARIES EXTRACTED FROM NEURAL NETWORK OUTPUT AND MANUAL ANNOTATION. JACCARD INDEX (JI) AND DICE COEFFICIENT (DC) SHOW PERCENTAGE INTERSECTION OVER UNION. HAUSDORFF DISTANCE (HD, IN mm) SHOWS THE GREATEST DISTANCE, AND MODIFIED HAUSDORFF DISTANCE (MHD, IN mm) SHOWS THE MEAN DIFFERENCE BETWEEN PREDICTED AND MANUALLY ANNOTATED BOUNDARIES. THIS TABLE REPORTS MEAN ± S.D. FOR ALL IMAGES FROM THE POSTURE DATASET (N = 384 = 192 + 192 REFLECTIONS) USING ‘TESTING MODE’. THE PENULTIMATE COLUMN ALONE REPORTS ALL SEGMENTS USING ‘ANALYSIS MODE’. THE FINAL COLUMN REPORTS AGREEMENT BETWEEN THE TWO ANNOTATORS WHO CONTRIBUTED THE LABELS FOR THIS DATASET TABLE I BOUNDARY ACCURACY BETWEEN BOUNDARIES EXTRACTED FROM NEURAL NETWORK OUTPUT AND MANUAL ANNOTATION. JACCARD INDEX (JI) AND DICE COEFFICIENT (DC) SHOW PERCENTAGE INTERSECTION OVER UNION. HAUSDORFF DISTANCE (HD, IN mm) SHOWS THE GREATEST DISTANCE, AND MODIFIED HAUSDORFF DISTANCE (MHD, IN mm) SHOWS THE MEAN DIFFERENCE BETWEEN PREDICTED AND MANUALLY ANNOTATED BOUNDARIES. THIS TABLE REPORTS MEAN ± S.D. FOR ALL IMAGES FROM THE POSTURE DATASET (N = 384 = 192 + 192 REFLECTIONS) USING ‘TESTING MODE’. THE PENULTIMATE COLUMN ALONE REPORTS ALL SEGMENTS USING ‘ANALYSIS MODE’. THE FINAL COLUMN REPORTS AGREEMENT BETWEEN THE TWO ANNOTATORS WHO CONTRIBUTED THE LABELS FOR THIS DATASET S MODE’. THE FINAL COLUMN REPORTS AGREEMENT BETWEEN THE TWO D THE LABELS FOR THIS DATASET Fig. 7. Clustering Dystonia into sub-groups Dystonia participants were clusterred into sub-groups using segment boundaries and k- means algorithm. (A) Clustering success metric v numer of clusters. (B) Division of Dystonia into optimal number of sub-groups (Dystonia 1-3, n = 9, 17, 9 respectively) using Analysis Mode. For each group we show: Top Row. Group averaged ultrasound images, Bottom Row. Mean segment bounadaries. Left side of all images represents the left anatomical side of the participants. Dystonia groups 1-3 appear left side compressed, right side compressed and both sides compressed respectively. corresponding pixels in the manually labeled image, we com- pute average precision, for varying target region. This analysis was iterated using pixels only of predicted conﬁdence (Soft- Max scores) greater than 0, 0.2, 0.4, 0.6 and 0.8 respectively (Figs. 4, 5). F. Clustering of Boundaries Using k-means, we tested the extent to which the 61 × 2600 matrix of segment boundaries clustered into groups. We used Matlab function kmeans with correlation as the distance metric. We used the CalinskiHarabasz value to evaluate separation into 2 to 10 groups. (Fig. 7). E. SVM Classiﬁcation Using Boundaries Boundaries from reﬂected images were discarded. For all 61 participants we computed the mean segment boundaries. This generated a matrix of 61 rows by 2600 columns. With Matlab functions ﬁtcsvm, crossval and kfoldLoss, we used a support vector machine, with 61-fold LOO cross validation to test ability of the boundaries to classify clinical condition (Dystonia v Control), Sex (male v female), and age (higher, lower) where age as divided into two groups around the median value (Table III). D. Boundary Extraction The Posture dataset was used for boundary extraction and in- jection point analysis (Figs. 4, 5). To the classiﬁed pixels (Fig. 4) we applied an 8 × 8 pixel median ﬁlter, ﬁlled holes, smoothed the boundaries and extracted boundaries using MATLAB functions (medﬁlt2, imﬁll, imclose and bwtraceboundary respectively. All boundaries were extracted clockwise, starting from a key point, deﬁned as the most medial pixel for muscles, and interpolated to 100 evenly spaced points (Fig. 3). For one image, the pattern of 13 segments is described by a row vector or 2,600 numbers (100 horizontal, then 100 vertical coordinates for each segment). Accuracy of extracted boundaries was assessed using Jaccard Index (JI), Dice Coefﬁcient (DC), Hausdorff Distance (HD) and modiﬁed Hausdorff Distance (MHD) (Table I) [25]. ‘Analysis Mode’ maximizes the training set (all Head Motion data), and also maximizes the number of participants (all Posture data) in each validation and test set. We propose motion of the head, (and to a lesser extent repositioning the probe to a new location,pressureandorientation),producesmoreindependence between images than changing participant. Moving the head changes the depth, muscle shape, scale, texture and dropout of each image. We propose, ‘Analysis Mode’ maintains inde- pendence between training (Head Motion) and validation/test (Posture) sets, and allows learning to extract more fully the content of the data. For each segment, the central predicted injection point was deﬁned as the pixel of maximum distance (dmax) from any boundary point (Fig. 3, middle row). We iteratively increased the margin around this injection point by distance t = 0, 1, 2… dmax mm. The pixels enclosed by this boundary at distance dmax – t from the predicted segment boundary provided a series of target injection regions (Fig. 4). By comparison with Manual annotators were blind to the condition and sex of the image and these labels played no part in training, validation LORAM et al.: OBJECTIVE ANALYSIS OF NECK MUSCLE BOUNDARIES FOR CERVICAL DYSTONIA USING US IMAGING 1021 TABLE I BOUNDARY ACCURACY BETWEEN BOUNDARIES EXTRACTED FROM NEURAL NETWORK OUTPUT AND MANUAL ANNOTATION. JACCARD INDEX (JI) AND DICE COEFFICIENT (DC) SHOW PERCENTAGE INTERSECTION OVER UNION. HAUSDORFF DISTANCE (HD, IN mm) SHOWS THE GREATEST DISTANCE, AND MODIFIED HAUSDORFF DISTANCE (MHD, IN mm) SHOWS THE MEAN DIFFERENCE BETWEEN PREDICTED AND MANUALLY ANNOTATED BOUNDARIES. THIS TABLE REPORTS MEAN ± S.D. FOR ALL IMAGES FROM THE POSTURE DATASET (N = 384 = 192 + 192 REFLECTIONS) USING ‘TESTING MODE’. B. SVM Classiﬁcation Using Segment Boundaries Using Testing Mode results, and predicted neck segment boundaries alone as input, a support vector machine with LOO cross validation, classiﬁed Dystonia from age matched controls with accuracy 70.4%, which was higher than sex (67.2%) or classiﬁcation age (52.4%) (Table III). This classiﬁcation was higher than classiﬁcation from manual annotated boundaries at 54%, 57% and 49% for condition, sex and age respectively. Classiﬁcation of Dystonia and sex from boundaries was higher using Analysis Mode (77.0%, 68.9% respectively), than us- ing Testing Mode (Table III). Results since the start of our investigation have been consistent: ‘Dystonia’ manifests more clearly in neck US images than sex; age cannot be classiﬁed from these images; and neural networks out-perform the human annotators (Fig. 6). III. RESULTS Supplementary Material, presents comparative analysis of 96 trained neural networks from ten different models extending our previous work [24] and justifying the best model selected for this paper. The video in SM demonstrates live neural network output of the selected model. or more to select injection points (Fig. 4), accuracy indicated by average precision is improved (Fig. 5). Setting minimum predic- tion conﬁdence to 80%, accuracy of injection points for the deep muscles (multiﬁdus, spinalis cervicis, spinalis capitis, splenius capitis) is indicated by average precision 93.5 ± 3% (Table II). These results support our hypothesis (i) that segmentation of deep muscles will identify injection points within the designated muscle. Figure 2 deﬁnes the ﬂow of hypotheses and results presented below. We report ﬁve main ﬁndings: (i) accuracy of extracted boundaries and of injection points within neck muscles, (ii) classiﬁcation of condition, sex and age from muscle boundaries, (iii) the optimal clustering of dystonia into sub-groups, (iv) reduction to eigen-patterns of muscle shape associated with cervical dystonia and (v) the association of neck muscle eigen- patterns with whole body posture. Body Posture For each participant, we computed their median multi- segment posture (51 angular components from 17 joints) from all their trials in the Head Motion (linked US-Vicon) dataset. To identify joint angles associated with neck boundary eigen- functions, we calculated the structure matrix showing corre- lation (n = 61) of all joint angles with each eigen func- tion and limited the lists to those signiﬁcant at p < 0.05 (Table IV). G. Pattern Analysis and Visualisation Fig. 7. Clustering Dystonia into sub-groups Dystonia participants were clusterred into sub-groups using segment boundaries and k- means algorithm. (A) Clustering success metric v numer of clusters. (B) Division of Dystonia into optimal number of sub-groups (Dystonia 1-3, n = 9, 17, 9 respectively) using Analysis Mode. For each group we show: Top Row. Group averaged ultrasound images, Bottom Row. Mean segment bounadaries. Left side of all images represents the left anatomical side of the participants. Dystonia groups 1-3 appear left side compressed, right side compressed and both sides compressed respectively. Fig. 7. Clustering Dystonia into sub-groups Dystonia participants were clusterred into sub-groups using segment boundaries and k- means algorithm. (A) Clustering success metric v numer of clusters. (B) Division of Dystonia into optimal number of sub-groups (Dystonia 1-3, n = 9, 17, 9 respectively) using Analysis Mode. For each group we show: Top Row. Group averaged ultrasound images, Bottom Row. Mean segment bounadaries. Left side of all images represents the left anatomical side of the participants. Dystonia groups 1-3 appear left side compressed, right side compressed and both sides compressed respectively. In a series of steps, we reduced the boundaries to the sta- tistically signiﬁcant eigen-patterns which discriminate dystonia sub-groups and healthy controls (Figs. 8, 9). Using all images from the Posture dataset, we reduced 2600 columns to 100 principal components. Each component rep- resents a pattern of variation from the mean shape. Using un- reﬂectedcasesonly,wecomputedthemeanprincipalcomponent scores for the dystonia and control participants (n = 61). We selected the principal components which reconstruct the group membership (Dystonia 1-3, Control). To select, we computed a univariate ANOVA for each principal component. Then, us- ing MATLAB functions sequentialfs (10-fold cross validation, 50 monte-Carlo repetitions, forward entry starting with signif- icant univariate components), and classify (‘diaglinear’, naive Bayes), we selected the combination of principal components which predicts group membership. To reduce the model to statistically signiﬁcant discriminant eigen-functions, maximizing separation of the groups, we per- formed one-way Multivariate Analysis of Variance (n = 61), using MATLAB function manova1 (Figs. 8, 9); IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, VOL. 24, NO. 4, APRIL 2020 1022 TABLE II MUSCLE INJECTION ACCURACY. SHOWS CLASSIFICATION ACCURACY FOR PIXELS OF CONFIDENCE GREATER THAN 80% WITHIN TARGET REGION. VALUES ARE MEAN FOR ALL PARTICIPANTS IN TEST SET (N = 61). DEEP MUSCLES SHOWS MEAN ± S.D. FOR MULTIFIDUS TO SPLENIUS. TABLE III H. Correlation of Eigen-Functions With Whole CLASSIFICATION OF DYSTONIA, SEX AND AGE FROM MUSCLE BOUNDARIES. A SUPPORT VECTOR MACHINE, WITH LOO 60 FOLD CROSS VALIDATION PREDICTED CONDITION (DYSTONIA V CONTROL), SEX (MALE VS. FEMALE) AND AGE (ABOVE MEDIAN VS. BELOW MEDIAN) FROM POSTURE DATASET OF 61 PARTICIPANTS G. Pattern Analysis and Visualisation INJECTION POINT IS THE PIXEL FURTHEST FROM ALL BOUNDARY POINTS. TARGET RADIUS: DECREASE IN DISTANCE FROM EDGE TO TARGET REGION INCREASE TARGET AREA. DISTANCE FROM EDGE: DISTANCE OF TARGET BOUNDARY FROM SEGMENT BOUNDARY. N POS, N NEG: NUMBER OF PIXELS IN IMAGE. ACC: PERCENTAGE OF PIXELS CLASSIFIED CORRECTLY. TP RATE, FP RATE, FN RATE, TN RATE: TRUE POSITIVE, FALSE POSITIVE, FALSE NEGATIVE, AND TRUE NEGATIVE RATE RESPECTIVELY. AP: AVERAGE PRECISION (PRECISION AVERAGED WITH RESPECT TO RECALL SORTED IN DECREASING CONFIDENCE) TABLE II MUSCLE INJECTION ACCURACY. SHOWS CLASSIFICATION ACCURACY FOR PIXELS OF CONFIDENCE GREATER THAN 80% WITHIN TARGET REGION. VALUES ARE MEAN FOR ALL PARTICIPANTS IN TEST SET (N = 61). DEEP MUSCLES SHOWS MEAN ± S.D. FOR MULTIFIDUS TO SPLENIUS. INJECTION POINT IS THE PIXEL FURTHEST FROM ALL BOUNDARY POINTS. TARGET RADIUS: DECREASE IN DISTANCE FROM EDGE TO TARGET REGION INCREASE TARGET AREA. DISTANCE FROM EDGE: DISTANCE OF TARGET BOUNDARY FROM SEGMENT BOUNDARY. N POS, N NEG: NUMBER OF PIXELS IN IMAGE. ACC: PERCENTAGE OF PIXELS CLASSIFIED CORRECTLY. TP RATE, FP RATE, FN RATE, TN RATE: TRUE POSITIVE, FALSE POSITIVE, FALSE NEGATIVE, AND TRUE NEGATIVE RATE RESPECTIVELY. AP: AVERAGE PRECISION (PRECISION AVERAGED WITH RESPECT TO RECALL SORTED IN DECREASING CONFIDENCE) H. Correlation of Eigen-Functions With Whole Body Posture TABLE IV TABLE IV POSTURE ASSOCIATED WITH NECK EIGEN-PATTERNS (STRUCTURE MATRIX). SHOWS JOINT ANGLES ASSOCIATED WITH THREE DISCRIMINANT EIGEN-FUNCTIONS. DF1 ASSOCIATES WITH THE WHOLE BODY LEANING RIGHTWARDS, THE HEAD TILTING DOWN AND TO THE LEFT, THE LEFT SHOULDER RAISED THE RIGHT WRIST COCKED. DF2 ASSOCIATES WITH THE WHOLE BODY LEARNING RIGHTWARDS. DF3 ASSOCIATES WITH RIGHT ARM OUT, THE HEAD TILTED TO THE RIGHT, THE LEFT KNEE BUCKLED, AND THE RIGHT FOOT ROLLED OUT Fig. 8. Reconstruction of Dystonia sub-groups using eigen- patterns of neck muscle shape. We have reduced description of dysto- nia sub-groups and controls to three discriminant eigen-functions (DF1, DF2, DF3, Wilk’s lambda, p-values shown panel A). Each eigen-function represents a weighted combination of principal components of US mus- cle boundaries from all participants (35 dystonia, 26 age matched con- trols). N.B. “Muscle” shape refers to all 13 segments (muscle, vertebra, ligament, skin). (A) Distance (mean linkage) between group centres using Mahalanobis distance (i.e., units of within group variance, so 16 = 4 S.D. of within group variation). (B) Axes shown ﬁrst two canonical discriminant function scores (c1, c2 for DF1, DF2 respectively, n = 61). Dystonia sub-groups differ from healthy controls in directions which are distinct from each other. Each eigen-function represents a pattern (Fig. 9). Fig. 9. Reconstruction of Dystonia using patterns of neck muscle shape. Sub-types of dystonia can be reconstructed from three signif- icant eigen-patterns of neck muscle shape shown in order of signiﬁ- cance. Green is sample mean principal component. Yellow and Blue show respectively +0.5 and +1 standard deviation of the eigen function. Image left shows anatomical left. DF1 shows the right muscles of the neck compressed and all muscles compressed to the skin. DF2 shows enlargement, fattening of left splenius. DF3 shows a squashing of all muscles to the skin, more compression on left side, but more symmetri- cal than DF1. (Dystonia 3), (Fig. 7). A large gap could represent a thicker fat layer or could represent altered pitch of the head. The right/left asymmetry could result from a tilted head (roll), a turned head (yaw), an elevated shoulder, a laterally shifted neck or a combination. D. Can Dystonia be Reduced to Statistically Signiﬁcant Eigen-Patterns of Neck Muscle Boundaries? Four groups can be discriminated by a maximum of 3 eigen- functions. Having reduced the segment boundaries to 100 prin- cipal components, the feature selection procedure selected 15 components to predict group membership (Dystonia 1-3, Con- trol) robustly and reconstructed group membership correctly at 85.2% using LOO classiﬁcation. Discriminant function analysis of the 61 case × 15 component matrix revealed three signiﬁcant eigen-functions (DF1, DF2, DF2, p = 1.1 × 10-11, p = 7.3 × 10-7, p = 0.0004, Fig. 8). Fig. 8A shows that with a separation of 16 units of mahalanobis distance (i.e., 4 S.D of within group variance), Dystonia3 (‘deep’) differs substantially from Dystonia 1-2 and Control. At 13.5 units, Dys- tonia 1 ‘left’ differs substantially from Dystonia2 and Control whereas Dystonia2 ‘right’ differs from Control by only 8 units (Fig. 8A). Thus we expect ‘deep’ to represent the largest pattern of difference from controls. Fig. 9. Reconstruction of Dystonia using patterns of neck muscle shape. Sub-types of dystonia can be reconstructed from three signif- icant eigen-patterns of neck muscle shape shown in order of signiﬁ- cance. Green is sample mean principal component. Yellow and Blue show respectively +0.5 and +1 standard deviation of the eigen function. Image left shows anatomical left. DF1 shows the right muscles of the neck compressed and all muscles compressed to the skin. DF2 shows enlargement, fattening of left splenius. DF3 shows a squashing of all muscles to the skin, more compression on left side, but more symmetri- cal than DF1. The three discriminant functions (DF1-3) represent the sig- niﬁcant dimensions separating the groups (Fig. 8B) and those functions also represent patterns of altered neck muscle shape distinguishing the groups (Fig. 9). A. Accuracy of Extracted Boundaries and Injection Points For the Posture dataset, accuracy of all segment boundaries using metrics JI, DC, HD and MHD was equal using ‘Testing Mode‘, or using ‘Analysis Mode’ and both marginally higher than inter-annotator agreement (Table I). These values were typical for muscles deep to the surface (Multiﬁdus, Spinalis Cervicis, Spinalis Capitis, Splenius Capitis), (Table I). A meaningful assessment of accuracy is provided by the ques- tion “would an injection into the predicted segment target the proposed muscle accurately?” Predicted classiﬁcation of pixels is more conﬁdent towards the center of the muscles rather than at the boundary (Fig. 4). By using SoftMax conﬁdence at 80% LORAM et al.: OBJECTIVE ANALYSIS OF NECK MUSCLE BOUNDARIES FOR CERVICAL DYSTONIA USING US IMAGING 1023 Fig. 8. Reconstruction of Dystonia sub-groups using eigen- patterns of neck muscle shape. We have reduced description of dysto- nia sub-groups and controls to three discriminant eigen-functions (DF1, DF2, DF3, Wilk’s lambda, p-values shown panel A). Each eigen-function represents a weighted combination of principal components of US mus- cle boundaries from all participants (35 dystonia, 26 age matched con- trols). N.B. “Muscle” shape refers to all 13 segments (muscle, vertebra, ligament, skin). (A) Distance (mean linkage) between group centres using Mahalanobis distance (i.e., units of within group variance, so 16 = 4 S.D. of within group variation). (B) Axes shown ﬁrst two canonical discriminant function scores (c1, c2 for DF1, DF2 respectively, n = 61). Dystonia sub-groups differ from healthy controls in directions which are distinct from each other. Each eigen-function represents a pattern (Fig. 9). B. Rationale for Methods of Analysis This focus of this paper is primarily scientiﬁc. In other words, do transverse ultrasound images of the neck, obtained from an ordinary ultrasound machine, contain the information necessary to inform understanding and diagnosis of dystonia and to aid delivery and monitoring of treatment by botulinum toxin injections? The development of methods for annotating images, training neural networks and evaluating deep learning architectures to segment muscle boundaries in US images of the posterior neck is fully discussed in our preceding work [24]. The technical challenge of segmenting muscles is already solved [24] although as shown in supplementary material (SM) this paper demonstrates considerable improvement since our previously published work [24]. Here, we apply the deep learning methods developed in our lab and we test a series of hypothesis con- cerning the value of neck muscle boundaries for understanding dystonia (Fig. 2). By comparison with classiﬁcation of sex and This whole-body motion analysis, provides validation that dystonic patterns of neck muscles identiﬁed from US images, have functional correlates in the standing posture. E. Correspondence Between Neck Muscle Patterns and Whole-Body Posture The Dystonia sub-groups, clustered from neck muscle bound- aries alone are associated with patterns of whole body standing posture. The median standing postures of these groups (Dystonia 1-3, Control) show differences in whole body truncal alignment and head turn which are described in the legend. Univariate ANOVA of joint angles conﬁrms the most signiﬁcant difference between groups lies in whole body frontal lean to the right (Right AnkleAngle_y, p = 0.0005), left head tilt (AtlantoOccipitalAn- gle_y, p = 0.003) and head pitch (AtlantoOccipitalAngle_x, p = 0.01). ii) Cervical dystonia is associated with visible, explainable patterns of neck muscle shape (Fig. 7, Fig. 9). This sample showed optimally three dystonia sub-groups, resulting in three signiﬁcant eigen-patterns of neck muscle shape (Figs. 7, 9). iii) Each pattern showed characteristic changes in muscle depth, midline asymmetry-curvature and left-right mus- cle imbalance (Fig. 7, 9). The ﬁrst (DF1) is associated with a postural pattern of head pitch, head turn, shoulder elevation and truncal tilt (Fig. 10, Table IV). The second (DF2),associatedmoststronglywithtruncaltilt.Thethird (DF3) associated most strongly with head tilt. The discriminant neck muscle eigen-patterns associate with postural joint rotations. The associations with posture give validation and explanation to the neck muscle eigen-patterns. As shown in Table IV, DF1 is associated with whole body lean rightwards, head looking downwards (pitch/extension), left shoulder elevation, head turning leftwards (yaw) and right wrist curled (extension). DF2 is associated with whole body lean leftward and has little association with head-neck rotations. DF3 is associated with right upper arm elevation (abduction), head tilt rightwards (roll), left knee inward and right foot roll (supination). iv) Segmentation is accurate enough to guide injections to speciﬁc muscles (Table II). v) Supervised deep learning of US muscle images, can en- code information with a veracity exceeding the manual annotation of its human supervisors (Table III). A. Contribution of This Study: The Main Results This study reports the ﬁrst application of deep learning to the segmentation, analysis and visualization of axial neck US images to participants with cervical dystonia. From a sample of 35 participants with cervical dystonia, and 26 age matched controls, we classiﬁed image pixels, extracted neck muscle boundaries, and tested ability to classify Dystonia, sex or age from muscle boundaries. We further clustered dystonia partic- ipants into sub-groups (Dystonia 1-3) identiﬁed the signiﬁcant eigen-patterns, reconstructing dystonia and related those eigen- patterns to posture. The second dimension (DF2) provides an axis proceed- ing from all Dystonia groups (negative) to controls (positive) (Fig. 8B). As a pattern, DF2 shows asymmetric enlargement of the left muscles and compression of the right muscles, but no general superﬁcial movement or sideways displacement of the midline (Fig. 9). DF2 is associated with reduced area of right splenius capitis (r = −0.34, p = 0.008). DF3 shows superﬁcial shift of all structures, leftward displacement of the midline (Fig. 9) and is associated with signiﬁcant reduction of left muscles and enlargement of right muscles. The most salient ﬁndings are: - i) Cervical dystonia can be discriminated from age matched healthycontrols,usinganaxialUSimageoftheneckmus- cles. Leave-one-out classiﬁcation of Dystonia v Control using SVM was correct at 70% (Table III) C. Does Dystonia Cluster Naturally Into Sub-Groups? Using neck segment boundaries alone, the 35 Dystonia par- ticipants clustered optimally into three groups (Fig. 7A), using both Analysis Mode and Testing Mode. Given higher SVM clas- siﬁcation, Analysis Mode was selected for descriptive clustering and pattern analysis. The mean boundaries of these groups show ‘right’ asymmetry (Dystonia 2), ‘left’ asymmetry (Dystonia1) and ‘deep’ segments with a large gap between skin and muscle The ﬁrst dimension (DF1) provides an axis separating Dystonia3-‘deep’ from Dystonia 2-‘right’, with Dystonia1 and Controls in the middle (Fig. 8B). DF1 as a patterns shows super- ﬁcial movement of all structures, and rightwards displacement of the midline and a relative depth-wise compression of the right IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, VOL. 24, NO. 4, APRIL 2020 1024 muscles (Fig. 9B). Correlation of muscle areas (calculated as percentage area of all segments) with DF1 at (p < 0.05) shows reduced area of right Splenius (r = −0.47, p = 0.00014), right spinalis capitis (r = −0.47, 0.00015), right trapezius (r = −0.43, p = 0.00046) and increased area of left splenius (r = 0.41, p = 0.001), left spinalis capitis (r = 0.3, p = 0.002) and left spinalis cervicis (r = 0.28, p = 0.027). images of the neck provide a better basis for diagnosing cervical dystonia than body posture. F. Comparison of Diagnosis by US With Diagnosis by Posture The green lines show the axis of frontal rotation and typically points forwards. The blue lines show the axis of axial rotation and point along the segments long axis which is typically vertical. For the head, red, green blue axes indicate pitch, roll and yaw. For the trunk, red green and blue axes indicate forward lean, rightwards lean and right turn. Dystonia2 show neck deviated to their right relative to head, trunk deviated to their left relative to neck and back to the midline. Dystonia1 shows neck deviated to their left relative to head, trunk also deviated to their left. Dystonia 3 shows neck deviated to their right relative to head and trunk deviated further to their right. g. 11. Examples of Dystonia sub-groups. Each block con e three example cases closest to the group center using discrim nction scores from Fig. 8. Each case shows the US image (above redicted labels (below). Fig. 10. Median posture Dystonia sub-groups facing the reader. Shows joint angles, median from all linked US-Vicon trials for each par- ticipant, averaged across all participants in group. This head referenced presentation shows the kinematic chain reconstructed from the head segment which is presented vertical and forward looking for each group. The red lines show the axis of extension and typically points to the right of the participant. The green lines show the axis of frontal rotation and typically points forwards. The blue lines show the axis of axial rotation and point along the segments long axis which is typically vertical. For the head, red, green blue axes indicate pitch, roll and yaw. For the trunk, red green and blue axes indicate forward lean, rightwards lean and right turn. Dystonia2 show neck deviated to their right relative to head, trunk deviated to their left relative to neck and back to the midline. Dystonia1 shows neck deviated to their left relative to head, trunk also deviated to their left. Dystonia 3 shows neck deviated to their right relative to head and trunk deviated further to their right. age, our results demonstrate that US images of the neck muscles contain the information necessary to visualize, understand and potentially diagnose cervical dystonia (Table III, Figs. 7–11). Having tested segmentation and classiﬁcation of Dystonia using 60-fold cross validation (‘Testing Mode’) we sought to maximize the accuracy of analysis and visualization of dystonia. F. Comparison of Diagnosis by US With Diagnosis by Posture Our switch to use of ‘Analysis Mode’, for description and ex- planation of the predicted boundaries is justiﬁed by the purpose of the analysis. The purpose of cross-validation is to answer the question “how well do you expect your system to perform out in the real world on unseen data?” We have answered that with 60-Fold cross-validation. The purpose of our hypotheses is to see in optimal circumstances how dystonia manifests in ultrasound and how dystonia manifests in cross-sectional shapes of muscles. For that we used clinical labels to select (using SVM) the best training Mode. The clinical labels were not used in the process of annotation, training, validating and testing the neural networks.Inpractice,analysisandvisualizationofTestingMode gave similar results to those presented, but the quality of the reconstruction of groups was lower (∼72% rather than 85%, meaning the description was less accurate or complete. F. Comparison of Diagnosis by US With Diagnosis by Posture Classiﬁcation of clinical condition (dystonia v control) by neck muscle boundaries was superior to classiﬁcation by stand- ing posture (Table III). Using the SVM, with LOO cross val- idation (n = 61), we provide a comparative classiﬁcation of condition, sex and age using whole body motion data. From the 51 components of joint rotation (17 joints × 3 degrees of rotation), the SVM classiﬁed condition with lower accuracy than the US based classiﬁcation, sex with higher accuracy and could not predict age (Table III). These results conﬁrm the US LORAM et al.: OBJECTIVE ANALYSIS OF NECK MUSCLE BOUNDARIES FOR CERVICAL DYSTONIA USING US IMAGING 1025 Fig. 10. Median posture Dystonia sub-groups facing the reader. Shows joint angles, median from all linked US-Vicon trials for each par- ticipant, averaged across all participants in group. This head referenced presentation shows the kinematic chain reconstructed from the head segment which is presented vertical and forward looking for each group. The red lines show the axis of extension and typically points to the right of the participant. The green lines show the axis of frontal rotation and typically points forwards. The blue lines show the axis of axial rotation and point along the segments long axis which is typically vertical. For the head, red, green blue axes indicate pitch, roll and yaw. For the trunk, red green and blue axes indicate forward lean, rightwards lean and right turn. Dystonia2 show neck deviated to their right relative to head, trunk deviated to their left relative to neck and back to the midline. Dystonia1 shows neck deviated to their left relative to head, trunk also deviated to their left. Dystonia 3 shows neck deviated to their right relative to head and trunk deviated further to their right. 11. Examples of Dystonia sub-groups. Each block con hree example cases closest to the group center using discrim on scores from Fig. 8. Each case shows the US image (above cted labels (below). Fig. 10. Median posture Dystonia sub-groups facing the reader. Shows joint angles, median from all linked US-Vicon trials for each par- ticipant, averaged across all participants in group. This head referenced presentation shows the kinematic chain reconstructed from the head segment which is presented vertical and forward looking for each group. The red lines show the axis of extension and typically points to the right of the participant. IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, VOL. 24, NO. 4, APRIL 2020 1026 the source of visual-vestibular head referenced sensory frames, location of sensory integration and motor planning) and the mass distribution of the body (trunk, upper and lower limbs). Abnormal action of neck muscles causes local changes in head and shoulder position and to maintain vertically aligned balance, these local changes require compensatory changes in whole body posture of the trunk and limbs. Neck muscles provide sophisticated proprioceptive sensation and have a primary role in integration of head-referenced with ground referenced coor- dinate frames which is also subject to interference by abnormal neck muscle activity. Altered body posture and sensory feed- back is a consequence of abnormal neck muscle action. Thus, analysis of the neck muscles provides direct insight into cervical dystonia. Prior to this study, it was an open question whether infor- mation contained within images of the neck muscles was of any value for diagnosis and understanding of cervical dystonia. This study afﬁrms the US information content with respect to objective clinical labels (control, dystonia) and with respect to motion analysis. Manual annotations provide only an approximation to the true muscle boundaries. With training, neural networks should learn image features that correlate consistently with the labels. In principle, machine learning should discard random error in human labels, and converge to the on-average correct answer within and between labelers. The exciting result reported here (Table III, Fig. 6) is that these neural networks out performed their supervisors: we urge the reader to study SM Fig. 6 which shows very nice examples. These results conﬁrm that neural networks encode information in the data consistent with supervisory labels, but which reduces the random error and non-generalizable compo- nent of the labels. These results (Table III) validate altogether, the method of labelling, the method of segmentation and the information content of ultrasound muscle images regarding dystonia. US imaging analysis can quantify dystonic muscle attributes (Figs. 8, 9, 10). US does not require participants with movement disorders to remain still and avoids limitations of MRI. US is relatively low cost and available in clinics. The conﬁdence measure provided by this neural network analysis (Fig. 4) gives inexperienced operators feedback to improve the quality of their US probe location and image. This analysis within clinic could facilitate communication between patient and clinician and would inform patients about their neck muscles and their speciﬁc dystonia. IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, VOL. 24, NO. 4, APRIL 2020 The objective recording of images and analysis provides a potential tool for guiding and recording the location of injections, for monitoring change and improvement with treatment, and thus is expected to improve the patient experience. In addition, our ﬁndings reinforce the potential critical role in CD of deep neck muscles, which have previously not been amenable to assessment or treatment. This study demonstrates proof of concept of the feasibility of US imaging analysis of the neck muscles for understanding and diagnosing cervical dystonia. Figure 11 shows examples of dystonia sub-group categorization on the basis of eigen function scores and illustrate the immediate diagnosis that could be possible in the clinic. This proof of concept motivates further development of US technology. If deployed widely in clinics, there is potential to collect large quantities of data from the estimated 18,000 adults with this condition in the UK [1]. Combined with further exploration of neural network methods, there is potential for this tool to become very robust, and for a new domain (automated ultrasound muscle analysis) to be established. Evaluation of the effect of therapeutic interventions e.g., BoNT on the patterns of change on US would also be critical todeterminetheutilityofthistooltomonitorchangesindystonia severity, and to evaluate its utility as a potential biomarker. D. Relationship to Previous Work The application of machine learning and speciﬁcally deep learning to analysis of ultrasound images of muscle is rare [20], [22], [26]. While under-developed, the domain of muscle diagnosis is valuable since unlike visual observation, manual palpation or surface electromyography, ultrasound can see mus- cles deep within the body. Segmentation is the foundation of muscle-speciﬁc analysis and recent methods providing segmentation of the neck muscles include computer vision [23] and deep learning approaches [24]. As shown in Supplementary Material, this study applies the most recent deep learning methods for this application. Following [24], this study uses direct manual annotation of US images to provide training labels. This approach allows us to develop training datasets for participants with a movement disorder who cannot remain still in an MRI machine [24]. Using metrics of JI and HD (Table I), the accuracy of segmentation achieved is consistent with existing benchmarks [23], [24]. The metric MHD (Table I) shows boundaries are typically accurate 1.9 ± 1.8 mm, and for deep muscles this accuracy allows for injection at average precision more than 90% for target sizes of several millimeter (SM Fig. A) and a margin from the muscle boundary of 3.7 ± 1 mm (Table II). These ﬁndings have important clinical implications, as freehand injections of botulinum toxin have been shown to have potentially suboptimal accuracy [27]. C. Scientiﬁc and Clinical Value of Results Cervical dystonia is a neurological disorder of sensorimotor integration characterized by abnormal postures of the head and neck. Abnormal involuntary dystonic activation of neck muscles is a primary symptom, but is a direct cause of pain, abnor- mal whole-body posture, and constraints on movement. Neck muscles traverse the primary link between the head (which is Fig. 11. Examples of Dystonia sub-groups. Each block contains the three example cases closest to the group center using discriminant function scores from Fig. 8. Each case shows the US image (above) and predicted labels (below). REFERENCES [1] G. Defazio, J. Jankovic, J. L. Giel, and S. Papapetropoulos, “Descrip- tive epidemiology of cervical dystonia,” Tremor Other Hyperkinet. Mov. (N. Y)., 2013, doi: 10.7916/D80C4TGJ. [21] R. J. Cunningham, P. J. Harding, and I. D. Loram, “The application of deep convolutional neural networks to ultrasound for modelling of dynamic states within human skeletal muscle,” 2017, arXiv preprint arXiv:1706.09450. [2] E. Tiderington et al., “How long does it take to diagnose cervical dysto- nia?,” J. Neurol. Sci., vol. 335, no. 1–2, pp. 72–74, 2013. [22] R. J. Cunningham, P. Harding, and I. D. Loram, “Deep residual networks for quantiﬁcation of muscle ﬁber orientation and curvature from ultra- sound images,” in Communications in Computer and Information Science, vol. 723, Cham: Springer, 2017, pp. 63–73. [3] A. Albanese et al., “Practical guidance for CD management involving treatment of botulinum toxin: A consensus statement,” J. Neurol., vol. 262, no. 10, pp. 2201–2213, 2015. [4] G. Reichel, “Cervical dystonia: A new phenomenological classiﬁcation for botulinum toxin therapy,” Basal Ganglia, vol. 1, no. 1, pp. 5–12, 2011. [23] R. J. Cunningham, P. J. Harding, and I. D. Loram, “Real-time ultrasound segmentation, analysis and visualisation of deep cervical muscle struc- ture,” IEEE Trans. Med. Imag., vol. 36, no. 2, pp. 653–665, Feb. 2017. [5] H. A. Jinnah, E. Goodmann, A. R. Rosen, M. Evatt, A. Freeman, and S. Factor, “Botulinum toxin treatment failures in cervical dystonia: Causes, management, and outcomes,” J. Neurol., vol. 263, no. 6, pp. 1188–1194, Jun. 2016. [24] R. J. Cunningham, M. B. Sánchez, and I. D. Loram, “Ultrasound segmen- tation of cervical muscle during head motion: A dataset and a benchmark using deconvolutional neural networks,” engrXiv, Feb. 26, 2019, doi: 10.31224/osf.io/fsa3c. [6] W. H. Jost, H. Hefter, A. Stenner, and G. Reichel, “Rating scales for cervical dystonia: A critical evaluation of tools for outcome assessment of botulinum toxin therapy,” J. Neural Transm., vol. 120, no. 3, pp. 487–496, 2013. [25] M. P. Dubuisson and A. K. Jain, “A modiﬁed Hausdorff distance for object matching,” in Proc. 12th Int. Conf. Pattern Recognit., 1994, vol. 1, pp. 566–568. [7] M. F. Contarino, M. Smit, J. van den Dool, J. Volkmann, and M. A. J. Tijssen, “Unmet needs in the management of cervical dystonia,” Front. Neurol., vol. 7, p. 165, 2016, doi: 10.3389/fneur.2016.00165. [26] P. Burlina, S. Billings, N. Joshi, and J. V. CONCLUSION [12] A. Schramm, T. Bäumer, U. Fietzek, S. Heitmann, U. Walter, and W. H. Jost, “Relevance of sonography for botulinum toxin treatment of cervical dystonia: An expert statement,” J. Neural Transmiss., vol. 122, no. 10, pp. 1457–1463, 2015. This study provides the ﬁrst application of deep learning to US imaging of the neck muscles in cervical dystonia and provides an automated objective visualization (c.f. Video) and subsequent pattern analysis of neck muscle boundaries. These results demonstrate that muscle boundaries extracted from a single axial image of the neck muscles have the information content to discriminate cervical dystonia from healthy controls andtovisualizeandunderstandthedystonicpatternofneckmus- cles. This proof of concept demonstrates potential for a clinical tool to provide objective online diagnosis of cervical dystonia, guidance and objective logging of injection sites, and objective monitoring of the effect of treatment with minimal requirement for operator expertise and minimal burden on clinical time. This work supports a case for further evaluation of an automated US-based tool in a larger longitudinal dataset. [13] A. I. Grigoriu, M. Dinomais, O. Rémy-Néris, and S. Brochard, “Impact of injection-guiding techniques on the effectiveness of botulinum toxin for the treatment of focal spasticity and dystonia: A systematic review,” Arch. Physical Med. Rehabil., vol. 96, no. 11, pp. 2067–2078.e1, 2015. Physical Med. Rehabil., vol. 96, no. 11, pp. 2067–2078.e1, 201 [14] I. D. Loram, C. N. Maganaris, and M. Lakie, “Use of ultrasound to make noninvasive in vivo measurement of continuous changes in human muscle contractile length,” J. Appl. Physiol., vol. 100, no. 4, pp. 1311–1323, 2005. [15] I. D. Loram, B. Bate, P. Harding, R. Cunningham, and A. Loram, “Proac- tive selective inhibition targeted at the neck muscles: This proximal constraint facilitates learning and regulates global control,” IEEE Trans. Neural Syst. Rehabil. Eng., vol. 25, no. 4, pp. 357–369, Apr. 2017. [16] K. Bibbings, P. J. Harding, I. D. Loram, N. Combes, and E. F. Hodson-Tole, “Foreground detection analysis of ultrasound image sequences identiﬁes markers of motor neurone disease across diagnostically relevant skele- tal muscles,” Ultrasound Med. Biol., vol. 45, no. 5, pp. 1164–1175, Mar. 2019. [17] I. D. Loram, C. N. Maganaris, and M. Lakie, “The passive, human calf muscles in relation to standing: The non-linear decrease from short range to long range stiffness,” J. Physiol., vol. 584, no. 2, pp. 661–675, 2007. ACKNOWLEDGMENT With appreciation we acknowledge the contribution of all our participants who willingly gave their time, the support of Dr. Gregory May and Brian Bate to assist data collection and the encouragement and support of Joanne Day to develop this work to assist those with cervical dystonia. [18] P. J. Harding, I. D. Loram, N. Combes, and E. F. Hodson-Tole, “Ultrasound based detection of fasciculations in healthy and diseased muscles,” IEEE Trans. Biomed. Eng.,vol. 63, no. 3, pp. 512–518, Mar. 2016. [19] I. D. Loram, C. N. Maganaris, and M. Lakie, “Active, non-spring-like muscle movements in human postural sway: How might paradoxical changes in muscle length be produced?,” J. Physiol., vol. 564, no. 1, pp. 281–293, 2005. [20] R. Cunningham, M. Sánchez, G. May, and I. Loram, “Estimating full regional skeletal muscle ﬁbre orientation from B-Mode ultrasound im- ages using convolutional, residual, and deconvolutional neural networks,” J. Imag., vol. 4, no. 2, p. 29, Jan. 2018. E. Limitations The current work contains several limitations. First, we have a relatively small number of cases, which may not encom- pass the full and expanding spectrum of neck movements seen in cervical dystonia [10] However we contribute our data to address the shortage of publically available examples (DOI: 10.23634/MMUDR.00624643). Further validation in a larger and independent clinical cohort would be desirable. Using more data, a clinical classiﬁer would most logically be embedded within the neural network architecture. Second, this analysis is limited to an axial image at level C4. A larger number of probe locations/orientations and muscle images would be desirable. Third, this work predicts and interprets muscle shape, excluding prediction of texture and muscle activity. Further work will exploit the ultrasound information content revealing muscle function [20], [21] as well as geometry for a larger range of probe locations and orientations. We would expect segmentation LORAM et al.: OBJECTIVE ANALYSIS OF NECK MUSCLE BOUNDARIES FOR CERVICAL DYSTONIA USING US IMAGING LORAM et al.: OBJECTIVE ANALYSIS OF NECK MUSCLE BOUNDARIES FOR CERVICAL DYSTONIA USING US IMAGING 1027 accuracy, boundary analysis and classiﬁcation of clinical condi- tion, all to depend to some extent on the quality of the image. We recommend to those readers replicating and extending our work, to choose an ultrasound machine, probe and machine settings providing the best possible image quality at full depth down to the vertebra. [9] A. Schramm, D. Huber, C. Mobius, A. Munchau, Z. Kohl, and T. Baumer, “Involvement of obliquus capitis inferior muscle in dystonic head tremor,” Parkinsonism Relat. Disord., vol. 44, pp. 119–123, Nov. 2017. [9] A. Schramm, D. Huber, C. Mobius, A. Munchau, Z. Kohl, and T. Baumer, “Involvement of obliquus capitis inferior muscle in dystonic head tremor,” [10] L. Tatu and W. H. Jost, “Anatomy and cervical dystonia : ‘Dysfunc- tion follows form’,” J. Neural Transmiss., vol. 124, no. 2, pp. 237–243, Feb. 2017. [11] C. Lungu, A. W. Tarulli, D. Tarsy, P. Mongiovi, V. G. Vanderhorst, and S. B. Rutkove, “Quantifying muscle asymmetries in cervical dystonia with electrical impedance: A preliminary assessment,” Clin. Neurophysiol., vol. 122, no. 5, pp. 1027–1031, May 2011. REFERENCES Albayda, “Automated diagno- sis of myositis from muscle ultrasound: Exploring the use of machine learning and deep learning methods,” PLoS One, vol. 12, no. 8, 2017, Art. no. e0184059. p [8] J. A. Van Gerpen, J. Y. Matsumoto, J. E. Ahlskog, D. M. Maraganore, and P. G. McManis, “Utility of an EMG mapping study in treating cervical dystonia,” Muscle and Nerve, vol. 23, no. 11, pp. 1752–1756, 2000. [27] A. Schnitzler et al., “Manual needle placement: Accuracy of botulinum toxin a injections,” Muscle Nerve, vol. 46, no. 4, pp. 531–534, Oct. 2012.
https://openalex.org/W3119247560	https://www.researchsquare.com/article/rs-140360/latest.pdf	English	null	Deficiency and absence of endogenous isoprene in adults, disqualified its putative origin	Heliyon	2,021	cc-by	7,025	Deficiency and absence of endogeno adults, disqualified its putative origin Pritam Sukul (  pritam.sukul@uni-rostock.de ) University Medicine Rostock https://orcid.org/0000-0001-5114-1776 Research Article Keywords: Breath isoprene, Disease biomarker, Cholesterol biosynthesis, Mevalonate pathway, PTR-MS, Metabolic disorders, Volatile metabolomics, Breathomics Posted Date: January 22nd, 2021 DOI: https://doi.org/10.21203/rs.3.rs-140360/v2 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Version of Record: A version of this preprint was published on January 12th, 2021. See the published version at https://doi.org/10.1016/j.heliyon.2021.e05922. Page 1/21 Page 1/21 Abstract Background: Isoprene (C5H8) is a clinically important breath metabolite. Although, hundreds of studies have reported differential expressions in isoprene exhalation as breath biomarker for diverse diseases, the substance couldn’t enter to clinical practice as diagnostic marker. Moreover, many experimental/basic observations upon breath isoprene remained unrelated to the corresponding pathophysiological effects on its putative metabolic origin (i.e. mevalonate pathway). Here, we investigated the fundamental reason that hindered the rational interpretation and translation of this marker from basic to clinical science. Methods: Via high-resolution mass-spectrometry based breathomics in 1026 human subjects, we discovered adults with significant deficiency (order of magnitude lower than the normal) and complete absence of breath isoprene. We prospectively applied real-time breathomics, quantitative gene expression analysis of the mevalonate pathway enzymes, lipid-profiling and hemodynamic monitoring on those isoprene deficient subjects and controls. Additionally, the subject with absence of isoprene was followed up throughout different phases of her womanhood. Results: In contrast to convention, we witnessed that adults can live healthy without exhaling isoprene or with significant deficiency. This rare phenotype represents a recessive inheritance. Despite physio- metabolic changes during menstrual cycle (that is known to profoundly affect isoprene exhalation) and profoundly increased plasma cholesterol during pregnancy and after childbirth, isoprene remained absent. All genes of mevalonate pathway enzymes were normally expressed in all participants, without any down-regulation or compensatory up-regulation. Conclusions: Absence/deficiency of isoprene despite normal lipid profiles and no mevalonate pathway malfunction disqualifies the long-believed metabolic origin of isoprene from cholesterol biosynthesis. Thus, clinical translation of breath isoprene expressions should not be generally attributed to corresponding pathophysiological effects onto mevalonate/cholesterol pathway. Our finding has refined and optimized the clinical interpretation of isoprene as biomarker in volatile metabolomics and breathomics. Future studies will address the correct metabolic origin of isoprene to imply this important marker to routine practice. Introduction Isoprene (C5H8) is the second most abundant (concentration mostly range: ~100 - 300 ppbV in adults) endogenous volatile organic compound (VOC) in human breath. Hundreds of clinical studies has randomly reported differential-expressions of exhaled isoprene concentrations as non-invasive biomarker to detect/monitor lung diseases[1,2], myocardial infraction[3], hypercholesterolemia[4], oxidative stress[5], cancers[6,7] chronic liver disease[8] and many other health conditions[9]. Moreover, relative changes in exhaled concentrations mirror various physio-metabolic effects[10–14]. Isoprene is believed to be a by- product of cholesterol biosynthesis[15,16], where mevalonic acid is deduced to non-sterol isoprenoid molecules[17]. According to literatures, absence of non-sterol isoprene is attributed to inherited errors (e.g. Page 2/21 Page 2/21 hyper immunoglobulinemia D syndrome, autoinflammatory periodic fever syndrome and mevalonic aciduria) in mevalonate pathway[18–20]. Breath isoprene deficiency in Duchenne muscle dystrophy patients supports its storage and washout from the muscle compartments[21]. Despite repeated claims to date, isoprene could not be translated to routine clinical practice as a disease specific biomarker. Most surprisingly, observed expressions of breath isoprene in many experimental/clinical studies (including our own) remained unexplained while considering the corresponding pathophysiological effects onto cholesterol biosynthesis (mevalonate pathway) and observed changes are randomly attributed to other effects such as endothelial cell damage, aging and oxidative stress, inflammation, ventilation and so on [22–26]. This clearly summons the long-ignored investigation to address if we are addressing the correct biochemical origin while interpreting breath isoprene data as disease biomarker. As the answer to such fundamental mystery may hide within the rare cases of extremities (i.e. isoprene deficiency/absence), we started our search for such subjects, since 2013. Until 2019, we have identified and quantified hundreds of VOCs (via high-resolution mass-spectrometry) within the exhaled breaths from >1000 humans, who belong to diverse demography. Eventually, we came across a few adults with significant isoprene deficiency and even without any traceable breath concentration. Looking at their health and lifestyle, we questioned the putative endogenous origin of breath isoprene[27,4,15] i.e. hypothesized since 1984 via an in vitro rat model[28]. In order to address that question, we have conducted a cross-disciplinary experiments involving real-time breathomics in parallel to lipid profiling and down-stream genetic analysis. Our present study will refine certain longstanding belief/conception upon the in vivo source of breath isoprene. This will offer rational interpretations of isoprene exhalations, which may translate this important breath biomarker towards clinical practice in the future. Measurement protocol for index, isoprene deficient subjects and controls Breath VOC measurement were repeated three times (on three different days) in all subjects. Knowing the possible up- or down-regulation of cholesterol homeostasis under endocrine changes (e.g. female sex-hormone milieu)[29,30] we measured index’s breath compositions, lipid profile and hemodynamic parameters throughout the menstrual cycle (i.e. periods, mid-follicular, ovulation and mid- luteal phases), during her first pregnancy (monthly) and also after the successful child birth for any possible occurrence of isoprene or up- or down-regulation of any other endogenous substances. The index, her family members (except the infant, due to ethical limitation) and four unrelated healthy adult controls (selected from the measured population) with normal breath isoprene levels, voluntarily provided blood for down-stream genetic analysis. Breath VOC screening in search of human subjects with deficiency and/or absence of breath isoprene Since 2013 until 2019, we have enrolled 1026 subjects, who belong to diverse age, gender, BMI, ethnicity and health conditions within our breath test (i.e. real-time mass-spectrometry based). Demographic data are listed in Supplementary Table 1. In 2015, we spotted the first young and healthy woman (will be referred as ‘index’ within the manuscript) without any traceable breath isoprene. As her blood-relatives also turned out with significant breath isoprene deficiency, we also included her immediate family members (i.e. her parents, -sibling sister, infant daughter and her husband) into further experiments. Ethics and recruitment of human participants: In accordance with the amended Declaration of Helsinki guidelines, an ethical approval (Approval number: A 2019-0040) from the Institutional Ethics Committee (University Medicine Rostock, Germany) and signed informed consent from all subjects were obtained prior to inclusion in the study. For breath sampling in infants, the consent of both parents was obtained. Human subjects were recruited from several clinical studies that were conducted in parallel at the University Medicine Rostock. Among those initial studies, the physiological monitoring experiments involved mainly healthy subjects and metabolic, pathophysiological and/or therapeutic monitoring studies involved both healthy and sick subjects (i.e. age, gender and BMI matched) in both cross-sectional and longitudinal mode. Thus, healthy subjects and patients from different age (<01 – 100 years old), gender, BMI and ethnicity (mainly European and Asian) were recruited via printed- and/or online advertisements. Within our present observational study, we considered a stratified random recruitment of subjects (i.e. before analysing their actual breath isoprene Page 3/21 Page 3/21 data) from those afore-mentioned demographic groups to attain a population of at least 1000 subjects (as indicated by the statistical power calculation i.e. described in 'Statistical analysis' section). While looking at the very rare occurrence of the feature (isoprene deficiency/absence in adults), it is very unlikely to assume a positive selection-bias (i.e. a greater number of individuals with anticipated feature had participated in this study) and our observational study cannot address any probable negative selection-bias (i.e. a smaller number of individuals with anticipated feature had participated in this study aka. response bias) without conducting an entire population screening; i.e. a far-reaching goal in academic research. Study setup Figure 1 represents an overview of our study setup. Proton transfer reaction - time of flight - mass spectrometry (PTR-ToF-MS 8000; Ionicon Analytik GmbH, Innsbruck, Austria) based real-time analysis of Page 4/21 Page 4/21 breath VOCs along with the non-invasive hemodynamic monitoring were synchronised together. Venous blood samples were collected for lipid profiling and genetic analysis by the study physician. breath VOCs along with the non-invasive hemodynamic monitoring were synchronised together. Venous blood samples were collected for lipid profiling and genetic analysis by the study physician. Lipid profiling Total cholesterol, lipoproteins (a), high density lipoproteins (HDL), low density lipoproteins (LDL) and triglycerides were analysed via conventional clinical method. Blood sampling Antecubital venous blood samples (20 ml/measurement/subject) were collected for the analysis of conventional lipid profiles and genetic analysis. Breath sampling and analysis Subjects maintained a normal sitting posture[12] and performed oral breathing[31] via custom made Teflon-mouthpiece of 2·5 cm diameter[32]. The transfer-line of PTR-ToF-MS was connected to the sterile mouthpiece to record VOC concentrations for continuous breath-resolved analysis. After one minute of metronome controlled paced breathing (respiratory rate = 12/min) subjects performed spontaneous breathing for five more minutes. This sampling protocol significantly reduces ventilatory variations and breath samples from 3rd minute (of spontaneous rhythm) onward were considered for analysis[33]. The working principles and pre-optimized experimental conditions of PTR-ToF-MS are described already[34]. Concisely, the soft ionisation of VOCs is based on a non-dissociative proton transfer reaction that ionises VOCs with relatively higher proton-affinity than water. Protonated VOCs are then detected via a high-resolution time-of-flight mass-spectrometer according to their mass to charge ratio. Our instrument can detect breath VOCs (including isoprene) up to low pptV range. For a detailed description of the method, see the Transparent Methods in Supplementary Material. Hemodynamic monitoring We performed non-invasive measurement of cardiac output, stroke volume, pulse rate and mean arterial pressure etc. via our optimised volume clamp method[13] to track physiological effects during pregnancy. For a detailed description of the method, see the Transparent Methods in Supplementary Material. Gene expression analysis We checked the presence/absence of transcripts, which code for the enzymes of the mevalonate pathway. Here, we applied quantitative polymerase chain reaction (qPCR) upon the gene expressions of all principal enzymes of the mevalonate pathway. The regulating genes are acetyl-CoA acetyltransferase 2 (ACAT2), HMG-CoA synthase 1 -cytosolic (HMGCS1), HMG-CoA reductase (HMGCR), mevalonate kinase (MVK), phosphomevalonate kinase (PMVK), di-phosphomevalonate decarboxylase (MVD), isopentenyl- diphosphate delta isomerase 1 (IDI1), farnesyl-diphosphate synthase 1 (FDPS1), geranylgeranyl diphosphate synthase 1 (GGPS1), farnesyl-diphosphate farnesyltransferase 1 (FDFT1), lanosterol Page 5/21 synthase (LSS), 7-dehydrocloesterol reductase (DHCR7). The working principles and optimised process of our qPCR is described before[35]. Primers used for gene expression analysis are listed in Supplementary Table 2. Relative gene expression is normalized to GAPDH (i.e. housekeeping gene) and calculated using the 2-ΔΔCt formula. Measurements were repeated thrice in all subjects. For a detailed description of the method, see the Transparent Methods in Supplementary Material. Alveolar isoprene exhalation in different groups Figure 2 represents the exhaled alveolar isoprene concentrations from the entire study population of 1026 human subjects (aged between 01 – 90 years), measured from 2013 to 2019. Among them there were 7 pre-school children (3 males & 4 females; age: 0 – 05 years), 75 pre-adolescents (35 males & 40 females; age: 06 – 12 years), 193 teenagers (110 males & 83 females; age: 13 – 19 years), 538 adults (281 males & 257 females; age: 20 – 59 years) and 213 seniors (46 males & 167 females; age: 60 – 90 years). Pre-school kids had exhaled low (near inspiratory) isoprene concentrations (mean±SD) of 08·20±2·74 ppbV. Concentrations in pre-adolescents ranged at 87·39±60·25 ppbV and in teenagers ranged at 112·75±70·16 ppbV. In seniors, exhaled isoprene concentrations ranged at 117·08±42.78 ppbV. Among the adults, exhaled isoprene concentrations in 532 adults were at the normal range of 137·52±48.2 ppbV. Only the index woman (German ethnicity) did not exhale any detectable/quantifiable isoprene (with respect to the minimal inspiratory room-air concentrations) in her breath. Absolute isoprene concentrations (repeated measurements mean±SD) in her isoprene deficient blood-relatives i.e. sibling sister, mother, father and infant daughter (<1 year old) were 19·87±4, 27·55±5, 39·25±6 and 05±2 ppbV, respectively. The same in her husband was 111·46±15 ppbV. In 2017 and 2018, we spotted two other young and healthy women (27 & 24 years old) with deficient breath isoprene concentrations of 11·45±3 ppbV (Spanish ethnicity) and 6·41±2 ppbV (Columbian ethnicity), respectively. As they were in short-term study visit and soon moved to their home countries, we couldn’t conduct any follow-up or genetic analysis. Surprisingly, neither the index nor the isoprene deficient subjects were suffering from any acute or chronic disease or any abnormality in cholesterol and lipid profile. Exhaled isoprene concentrations in four controls (02 males & 02 females, who were included in genetic analysis) were 114·52±6·2, 144·51±12·8, 176·70±11·2 and 227·96±10·4 ppbV, respectively. Exhaled isoprene concentrations from the above-mentioned age cohorts are presented in Fig. 3. Statistically significant differences are presented with respect to the group of adults (n=532) with normal isoprene levels. Detailed results of statistical comparisons are listed in Supplementary Table 3. Differences between inspiratory and expiratory concentrations of breath isoprene from index, index family and unrelated healthy controls are presented in Fig. 4 (a). No significant differences were found in the expiratory isoprene concentrations of index and her daughter with respect to inspiratory values. Statistical analysis Sample size was computed while designing this study. We applied analysis of variance (ANOVA) test by considering a minimum detectable difference (as reported clinical studies) in mean isoprene intensities of 40 cps, an expected standard deviation of 150 (i.e. varying between 100 – 200 according to age, gander, BMI health conditions etc.), 6 groups, a desired test-power of 0·99 and the alpha value of 0·005. Thus, the resulting sample size was 1020 (to represent relevance to the general population). In this study we have included 1026 subjects to detect less than 5% differences in exhaled isoprene concentrations up to low pptV levels. Repeated measurements (thrice) of breath VOCs were performed in all subjects on different days. Mean values (as data are normally distributed) of expiratory and inspiratory isoprene concentrations were calculated over a minute of breath-resolved data in each measurement/subject. Statistically significant differences in expiratory isoprene concentrations between different age groups (pre-school children, pre-adolescents, teenagers, adults and seniors) were judged via repeated measurement ANOVA on ranks (Friedman repeated measures analysis of variance on ranks, Shapiro-Wilk test for normal distribution and post hoc Student–Newman–Keuls method for pairwise multiple comparisons between all groups; p-value ≤0·005) in SigmaPlot (version 14) software. The adults are divided into three distinct cohorts depending on the normal levels (n=532), absence (n=01) and deficiency (n=5) of exhaled isoprene. All groups were compared to each other. and in Fig. 3 we presented those referring to the group of adults with normal isoprene levels. Statistically significant differences in expiratory and inspiratory concentrations (obtained via repeated measurements) were compared between the index subject, her parents, sibling-sister, infant daughter, husband and unrelated controls. A repeated measurement ANOVA on ranks (as indicated above) was applied for all pair-wise multiple comparisons. Page 6/21 Lipid profile, hemodynamics and VOC exhalation during womanhood of the subject without isoprene Changes in absolute values of lipid profile, hemodynamic parameters and relative changes in exhaled alveolar concentrations of 28 VOCs took place during her menstrual cycle phases, pregnancy and after child birth (Figure 5). Pronounced physiological and metabolic effects are observed on other breath VOCs during menstrual cycle and pregnancy. Despite profound fluctuation in cardiac output and increased cholesterol production during pregnancy, her isoprene remained absent. In order to reduce confounding effects, only compounds with expiratory abundances significantly higher than the inspiratory room-air abundances (+standard deviations of room-air abundances) were selected for analysis. Among those, we selected 28 compounds in order to address a broad spectrum of chemical classes, physico-chemical properties and substance-specific physiological/metabolic effects. These substances reflect important aspects for human breathomics e.g. diverse origins (endogenous and blood borne, oral cavity, pre-exposure) and dependencies on physiological and metabolic changes and effects. Alveolar isoprene exhalation in different groups Detailed results of statistical comparisons are listed in Supplementary Table 4. Page 7/21 Page 7/21 Gene expression of mevalonate and cholesterol biosynthesis pathway enzymes All genes analysed were expressed in control samples, index and all family members, indicating normal pathway functionality in Fig. 4 (b). This is in line with the data obtained from blood lipid and cholesterol measurements. Compared to control samples the gene expression of the index subject was within normal range with enzymes responsible for the upper part of the cholesterol synthesis (ACAT2, HMGCS1, HMGCR, MVK, PMVK, MVD, IDI1, FDPS1, GGPS1) slightly increased (1·04‑fold to 1·77‑fold). In contrast, FDFT1, LSS and DHCR7 were down-regulated (0·76‑, 0·72‑ and 0·88‑fold, respectively) but also within normal range. Her mother has a gene expression profile that is rather adverse, with upper pathway enzymes rather decreased and a slight up-regulation of downstream genes. The profile of the father resembles that of the mother while her husband demonstrates similarities with the index. Of note, the group of controls are quite heterogeneous with two persons (Control-1, male; Control-4, female) exhibiting overall low gene expression patterns and the other two (also one male, one female) a rather up-regulated cholesterol synthesis pathway. Detailed results of gene expression analysis are listed in the Supplementary Table 5. Analysing the overall distribution, we could not detect a cluster or pattern separating controls and index family members. With all isoprene-deficient or ‑low family members well within the normal range and no clustering detected either within the family cohort or between family cohort and controls we conclude that aberrant cholesterol pathway gene expression is not a factor involved in absent isoprene synthesis. Discussion Page 8/21 Page 8/21 In accordance with the previous evidences[24,36], a distinct dependency of isoprene exhalation on ageing is observed within our study population. In contrast to conventional knowledge[18,19], we witnessed that adults can live healthy without exhaling isoprene or with significant isoprene deficiency. Isoprene exhalations in index family indicate a recessive inheritance of this phenotype. Here, her isoprene deficient parents and sibling sister are carriers of the character whereas, tiny presence is quantified in her infant daughter (<1 year old) as her husband has normal levels (Please see Supplementary Fig. 1). Any functional loss of one pathway member would probably result in compensatory upregulation of other enzymes. As this was not observed in gene expressions, we conclude that there are no functional aberrations in cholesterol synthesis. Very heterogenous gene expressions in controls and in persons without isoprene and with isoprene deficiency have indicated that mevalonate pathway gene expression does not influence isoprene synthesis and thereby, cholesterol biosynthesis is likely not the source of isoprene in human breath. Since the inception of human breathomics, isoprene has been the substance of potential interest especially, in the field of non-invasive physiological and metabolic monitoring. Due to its low aqueous solubility and high volatility, isoprene mirrors the pulmonary ventilation-perfusion phenomenon[13]. Its systemic transport to alveolar capillary is cardiac output (lung perfusion determinant) dependent[12,37] and pulmonary elimination is based on intra-alveolar pressure gradient driven diffusion[32]. Simple physiological changes in hemodynamics and/or ventilation are immediately reflected in isoprene exhalation[11,12,31,37]. We observed diverse changes in breath isoprene levels throughout menstrual cycles – with/without administration of monthly oral contraceptive pills[14]. As physiological effects (ventilatory and hemodynamic) and metabolic changes (endocrine interplay) are readily observed in isoprene exhalation, this compound is supposed to fluctuate significantly during pregnancy. Despite pronounced physiological increase in her plasma cholesterol levels throughout gestation[38] and elevated cardiac output, isoprene did not appear (Figure 5). Exhaled alveolar concentrations of endogenous VOCs changed mainly because of pregnancy induced hormonal changes (causing homeostatic metabolic effects on in vivo origins of substances), obvious oxidative stress (embryonic development driven), changes in broncho-pulmonary gas-(VOCs)-exchange under increased cardiac output (counter-balancing oxygen- and nutrients demand of the foetus) and altered respiratory rate, tidal volume and minute ventilation (due to uplifted diaphragm of the mother). Here, we have restrained ourselves from discussing substance specific behaviour of VOCs as this is not the principal focus of this manuscript. Discussion In volatile metabolomics and breathomics, clinical interpretations of differential-expressions/changes in endogenous VOC markers rely upon the primary effects of a pathophysiology onto the biochemical/metabolic origin of substances. Many clinical observations/results upon the effects of a disease onto cholesterol-genesis and isoprene exhalation remained unexplainable and are attributed to other phenomena such as endothelial damage[22], aging[36], inflammatory process[26], ventilatory effects[25] etc. For instance, we observed no relation (Pearson correlation based on 968 independent observations at p-Value significance of ≤0·05. Please refer to the supplementary material of the corresponding article by Trefz et al.) between plasma cholesterol levels and exhaled isoprene concentrations within large cohort of healthy and diabetic children[23]. In line with those, our present finding is clearly reporting that the long anticipated in vivo origin of breath isoprene is incorrect. Thus, we have addressed a general misconception that hinders trustworthy bio-medical interpretation of this important biomarker in our field of omics. Discussion Literatures state that dimethylallyl pyrophosphate (DMAPP) is converted to isoprene by isoprene synthase enzyme in plants[39]. The enzyme is reasonably well studied in plants[40] but poorly characterized in microbes and animals. DMAPP undergoes non-enzymatic degradation in animal liver[28] but the reaction rate is too slow to constantly source breath isoprene. Although enzymes with multiple functional domains (catalytic sites with broad binding and conversion capabilities) can convert DMAPP to isoprene, bioinformatic sequence alignment (whole sequence/functional domain based) search for the plant isoprene synthase with human genes failed to find enzyme homologs. Consequently, the deficiency/absence of isoprene in healthy adults may indicate that the conversion of DMAPP to isoprene Page 9/21 Page 9/21 might not be essential for life. Seeing its emission from bacteria, marine algae, plants, animals and us it’s hard to disregard the phylogenetic/evolutionary significance of its endogenous production[27,41]. As nature doesn’t play dices, it’s unlikely for any redundant biochemical pathway to sustain throughout the course of evolution. Although isoprene protects thermal damage of photosynthetic membranes in plants[39], the biological importance in animal or human is unknown. From a toxicological point of view, isoprene is a carcinogen by nature and prolonged exposure to inhaled isoprene has shown cancer growth in rodents[42–44]. Thus, isoprene exhalation might be a blood-breath excretory axis of hydrocarbon homeostasis in blood. Considering the muscle compartments as potential storage of isoprene[21], exercise increases its elimination[10] and thereby supports its benefit towards physiological detoxification. Mevalonate kinase (MVK) enzyme deficiency is an ultra-rare autosomal recessive (gene locus: chromosome 12q24) inborn defect of metabolism (aka. Mevalonic aciduria) with an autoinflammatory phenotype[45,46] i.e. characterised by absence of isoprene. Allogenic bone marrow transplantation from heterogenous carrier of the mutant gene has shown symptomatic improvement in mevalonic aciduria patient[20]. In contract to that our index is healthy and all genes of the mevalonate pathway enzymes are transcribed despite the same recessive phenotype (i.e. absence of isoprene) as of mevalonic aciduria patients. Thus, one may assume that one of these multi-functional enzymatic sites is malfunctioning. Hence, she does not have isoprene but she doesn't suffer any consequences of the genetic abnormality. This has summoned future down-stream genome analysis (e.g. exome sequencing) to spot this healthy mutation in order to find the actual biochemical origin of isoprene in human. Our future work will address those investigations. Conclusions We have pioneered that absence/deficiency of breath isoprene is not necessarily a pathological phenotype. The recessive pedigree and healthy manifestation of this phenotype along with no functional aberration in cholesterol- and/or mevalonate pathway genes have contradicted the conventionally Page 10/21 Page 10/21 hypothesised origin of isoprene in human breath. Our data shouts for future investigations to find the actual metabolic process/biochemical pathway of isoprene production. Otherwise, clinical interpretations of breath isoprene will be misleading and at random. Absence/suppression of true isoprene production process in our healthy index/index family have summoned further studies to track the possible presence of any compensatory mechanism in such individuals. Clinical translation of our basic findings may open up an unexplored frontier towards the limited clinical understanding of certain ultra-rare diseases as well as will trailblaze the rational interpretations of isoprene exhalation under various health conditions as reliable breath biomarkers in future. Limitations of the Study While looking at the present limitations, despite screening >1000 subjects, we were able to find only one adult without isoprene and five adults with significant deficiency of isoprene and therefore, we could not run a rational statistical correlation analysis (between breath isoprene and gene-expression) within this study. However, our finding indicates a rare prevalence of this phenotype. Furthermore, considering the observed negligible amounts of exhaled isoprene in kids below the age of five years, it is hard to draw further conclusions onto observed concentrations of index’s daughter before 2025. Declarations Author contribution statement: PS conceived the idea, designed the study and spotted the subjects with absence and deficiency of breath isoprene. JKS and WM contributed to conception and study design. PS recruited volunteers, performed experiments, analysed data. PS, JKS and WM contributed to clinical and analytical interpretation. JKS performed the blood collections. AR preformed the genetic analysis, data interpretation and supported recruitment of subjects. WM supported analytical instrumentations. All authors reviewed the manuscript and contributed in writing. Funding statement: The PTR-ToF-MS-8000 instrument used in this study was entirely funded by the European fund for regional development (EFRE). This study was financially supported by Marie-Curie 7th European Community Framework ITN Programme (FP7-PEOPLE-ITN-PIMMS project, Grant agreement no. 287382), EU Horizon-2020 grant (H2020-PCH-HEARTEN project, Grant agreement no. 643694) and Inno- INDIGO-NCDs-CAPomics Project (BMBF 01DQ16010). The funding bodies had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript. Declaration of interest statement: The authors declare no conflict of interest. Declaration of interest statement: The authors declare no conflict of interest. Page 11/21 Page 11/21 Page 11/21 Additional information: All data generated or analysed during this study are included in this published article [and its supplementary information files]. Any further information is available from the corresponding author on reasonable request. Acknowledgements We thank all the volunteers for participating in this study. Special thanks to Dr. Patricia Fuchs (RoMBAT, University Medicine Rostock, Germany) for supporting hemodynamic monitoring. References [1] L.T. McGrath, R. Patrick, P. Mallon, L. Dowey, B. Silke, W. Norwood, S. Elborn, Breath isoprene during acute respiratory exacerbation in cystic fibrosis, Eur. Respir. J. 16 (2000) 1065–1069. [1] L.T. McGrath, R. Patrick, P. Mallon, L. Dowey, B. Silke, W. Norwood, S. Elborn, Breath isoprene during acute respiratory exacerbation in cystic fibrosis, Eur. Respir. J. 16 (2000) 1065–1069. [2] J.K. Schubert, W.P.E. Müller, A. Benzing, K. Geiger, Application of a new method for analysis of exhaled gas in critically ill patients, Intensive Care Med. 24 (1998) 415–421. https://doi.org/10.1007/s001340050589. [3] S. Mendis, P.A. Sobotka, D.E. Euler, Expired Hydrocarbons in Patients with Acute Myocardial Infarction, Free Radic. Res. 23 (1995) 117–122. https://doi.org/10.3109/10715769509064026. [4] T. Karl, P. Prazeller, D. Mayr, A. Jordan, J. Rieder, R. Fall, W. Lindinger, Human breath isoprene and its relation to blood cholesterol levels: new measurements and modeling, J. Appl. Physiol. 91 (2001) 762– 770. https://doi.org/10.1152/jappl.2001.91.2.762. [5] F. Pabst, W. Miekisch, P. Fuchs, S. Kischkel, J.K. Schubert, Monitoring of oxidative and metabolic stress during cardiac surgery by means of breath biomarkers: an observational study, J. Cardiothorac. Surg. 2 (2007) 37. https://doi.org/10.1186/1749-8090-2-37. [6] D. Fuchs, H. Jamnig, P. Heininger, M. Klieber, S. Schroecksnadel, M. Fiegl, M. Hackl, H. Denz, A. Amann, Decline of exhaled isoprene in lung cancer patients correlates with immune activation, J. Breath Res. 6 (2012) 027101. https://doi.org/10.1088/1752-7155/6/2/027101. [7] D. Poli, P. Carbognani, M. Corradi, M. Goldoni, O. Acampa, B. Balbi, L. Bianchi, M. Rusca, A. Mutti, Exhaled volatile organic compounds in patients with non-small cell lung cancer: cross sectional and nested short-term follow-up study, Respir. Res. 6 (2005) 71. https://doi.org/10.1186/1465-9921-6-71. [7] D. Poli, P. Carbognani, M. Corradi, M. Goldoni, O. Acampa, B. Balbi, L. Bianchi, M. Rusca, A. Mutti, Exhaled volatile organic compounds in patients with non-small cell lung cancer: cross sectional and nested short-term follow-up study, Respir. Res. 6 (2005) 71. https://doi.org/10.1186/1465-9921-6-71. [8] N. Alkhouri, T. Singh, E. Alsabbagh, J. Guirguis, T. Chami, I. Hanouneh, D. Grove, R. Lopez, R. Dweik, Isoprene in the Exhaled Breath is a Novel Biomarker for Advanced Fibrosis in Patients with Chronic Liver Disease: A Pilot Study, Clin. Transl. Gastroenterol. 6 (2015) e112. https://doi.org/10.1038/ctg.2015.40. [8] N. Alkhouri, T. Singh, E. Alsabbagh, J. Guirguis, T. Chami, I. Hanouneh, D. Grove, R. Lopez, R. Dweik, Isoprene in the Exhaled Breath is a Novel Biomarker for Advanced Fibrosis in Patients with Chronic Liver Disease: A Pilot Study, Clin. Transl. Gastroenterol. References 6 (2015) e112. https://doi.org/10.1038/ctg.2015.40. Page 12/21 Page 12/21 [9] R. Salerno-Kennedy, K.D. Cashman, Potential applications of breath isoprene as a biomarker in modern medicine: a concise overview, Wien. Klin. Wochenschr. 117 (2005) 180–186. https://doi.org/10.1007/s00508-005-0336-9. [9] R. Salerno-Kennedy, K.D. Cashman, Potential applications of breath isoprene as a biomarker in modern medicine: a concise overview, Wien. Klin. Wochenschr. 117 (2005) 180–186. https://doi.org/10.1007/s00508-005-0336-9. [10] J. King, A. Kupferthaler, K. Unterkofler, H. Koc, S. Teschl, G. Teschl, W. Miekisch, J. Schubert, H. Hinterhuber, A. Amann, Isoprene and acetone concentration profiles during exercise on an ergometer, J. Breath Res. 3 (2009) 027006. https://doi.org/10.1088/1752-7155/3/2/027006. [11] J. King, A. Kupferthaler, B. Frauscher, H. Hackner, K. Unterkofler, G. Teschl, H. Hinterhuber, A. Amann, B. Högl, Measurement of endogenous acetone and isoprene in exhaled breath during sleep, Physiol. Meas. 33 (2012) 413–428. https://doi.org/10.1088/0967-3334/33/3/413. [12] P. Sukul, P. Trefz, S. Kamysek, J.K. Schubert, W. Miekisch, Instant effects of changing body positions on compositions of exhaled breath, J. Breath Res. 9 (2015) 047105. https://doi.org/10.1088/1752- 7155/9/4/047105. [13] P. Sukul, J.K. Schubert, P. Oertel, S. Kamysek, K. Taunk, P. Trefz, W. Miekisch, FEV manoeuvre induced changes in breath VOC compositions: an unconventional view on lung function tests, Sci. Rep. 6 (2016) 28029. https://doi.org/10.1038/srep28029. [14] P. Sukul, J.K. Schubert, P. Trefz, W. Miekisch, Natural menstrual rhythm and oral contraception diversely affect exhaled breath compositions, Sci. Rep. 8 (2018) 10838. https://doi.org/10.1038/s41598- 018-29221-z. [15] B.G. Stone, T.J. Besse, W.C. Duane, C.D. Evans, E.G. DeMaster, Effect of regulating cholesterol biosynthesis on breath isoprene excretion in men, Lipids. 28 (1993) 705–708. [16] B. Buszewski, M. Kesy, T. Ligor, A. Amann, Human exhaled air analytics: biomarkers of diseases, Biomed. Chromatogr. BMC. 21 (2007) 553–566. https://doi.org/10.1002/bmc.835. [17] N.M.F.S.A. Cerqueira, E.F. Oliveira, D.S. Gesto, D. Santos-Martins, C. Moreira, H.N. Moorthy, M.J. Ramos, P.A. Fernandes, Cholesterol Biosynthesis: A Mechanistic Overview, Biochemistry. 55 (2016) 5483– 5506. https://doi.org/10.1021/acs.biochem.6b00342. [18] G. Hoffmann, K.M. Gibson, I.K. Brandt, P.I. Bader, R.S. Wappner, L. Sweetman, Mevalonic Aciduria — An Inborn Error of Cholesterol and Nonsterol Isoprene Biosynthesis, N. Engl. J. Med. 314 (1986) 1610– 1614. https://doi.org/10.1056/NEJM198606193142504. [19] D. Haas, G.F. Hoffmann, Mevalonate Kinase Deficiency and Autoinflammatory Disorders, Http://Dx.Doi.Org/10.1056/NEJMp078083. (2009). https://doi.org/10.1056/NEJMp078083. Page 13/21 [20] B. Neven, V. Valayannopoulos, P. Quartier, S. Blanche, A.-M. Prieur, M. Debré, M.-O. Rolland, D. Rabier, L. Cuisset, M. Cavazzana-Calvo, P. de Lonlay, A. Fischer, Allogeneic Bone Marrow Transplantation in Mevalonic Aciduria, N. Engl. J. Med. 356 (2007) 2700–2703. https://doi.org/10.1056/NEJMoa070715. Page 13/21 [21] J. References Sukul, J.K. Schubert, S. Kamysek, P. Trefz, W. Miekisch, Applied upper-airway resistance instantly affects breath components: a unique insight into pulmonary medicine, J. Breath Res. 11 (2017) 047108. https://doi.org/10.1088/1752-7163/aa8d86. [32] P. Sukul, J.K. Schubert, S. Kamysek, P. Trefz, W. Miekisch, Applied upper-airway resistance instantly affects breath components: a unique insight into pulmonary medicine, J. Breath Res. 11 (2017) 047108. https://doi.org/10.1088/1752-7163/aa8d86. [33] P. Sukul, J.K. Schubert, K. Zanaty, P. Trefz, A. Sinha, S. Kamysek, W. Miekisch, Exhaled breath compositions under varying respiratory rhythms reflects ventilatory variations: translating breathomics towards respiratory medicine, Sci. Rep. 10 (2020) 14109. https://doi.org/10.1038/s41598-020-70993-0. [34] P. Trefz, M. Schmidt, P. Oertel, J. Obermeier, B. Brock, S. Kamysek, J. Dunkl, R. Zimmermann, J.K. Schubert, W. Miekisch, Continuous real time breath gas monitoring in the clinical environment by proton- transfer-reaction-time-of-flight-mass spectrometry, Anal. Chem. 85 (2013) 10321–10329. https://doi.org/10.1021/ac402298v. [34] P. Trefz, M. Schmidt, P. Oertel, J. Obermeier, B. Brock, S. Kamysek, J. Dunkl, R. Zimmermann, J.K. Schubert, W. Miekisch, Continuous real time breath gas monitoring in the clinical environment by proton- transfer-reaction-time-of-flight-mass spectrometry, Anal. Chem. 85 (2013) 10321–10329. https://doi.org/10.1021/ac402298v. [35] A. Richter, S. Sender, A. Lenz, R. Schwarz, B. Hinz, G. Knuebel, A. Sekora, H. Murua Escobar, C. Junghanss, C. Roolf, Influence of Casein kinase II inhibitor CX-4945 on BCL6-mediated apoptotic signaling in B-ALL in vitro and in vivo, BMC Cancer. 20 (2020) 184. https://doi.org/10.1186/s12885-020- 6650-9. [35] A. Richter, S. Sender, A. Lenz, R. Schwarz, B. Hinz, G. Knuebel, A. Sekora, H. Murua Escobar, C. Junghanss, C. Roolf, Influence of Casein kinase II inhibitor CX-4945 on BCL6-mediated apoptotic signaling in B-ALL in vitro and in vivo, BMC Cancer. 20 (2020) 184. https://doi.org/10.1186/s12885-020- 6650-9. [36] N. Nelson, V. Lagesson, A.R. Nosratabadi, J. Ludvigsson, C. Tagesson, Exhaled Isoprene and Acetone in Newborn Infants and in Children with Diabetes Mellitus, Pediatr. Res. 44 (1998) 363–367. https://doi.org/10.1203/00006450-199809000-00016. [37] P. Sukul, P. Trefz, J.K. Schubert, W. Miekisch, Immediate effects of breath holding maneuvers onto composition of exhaled breath, J. Breath Res. 8 (2014) 037102. https://doi.org/10.1088/1752- 7155/8/3/037102. [38] Ä. Bartels, K. O’Donoghue, Cholesterol in pregnancy: a review of knowns and unknowns, Obstet. Med. 4 (2011) 147–151. https://doi.org/10.1258/om.2011.110003. [39] T.D. Sharkey, Isoprene synthesis by plants and animals, Endeavour. 20 (1996) 74–78. https://doi.org/10.1016/0160-9327(96)10014-4. [40] T.D. Sharkey, S. Yeh, A.E. Wiberley, T.G. Falbel, D. Gong, D.E. Fernandez, Evolution of the Isoprene Biosynthetic Pathway in Kudzu, Plant Physiol. 137 (2005) 700–712. https://doi.org/10.1104/pp.104.054445. [41] P.C. Harley, R.K. Monson, M.T. References King, P. Mochalski, K. Unterkofler, G. Teschl, M. Klieber, M. Stein, A. Amann, M. Baumann, Breath isoprene: Muscle dystrophy patients support the concept of a pool of isoprene in the periphery of the human body, Biochem. Biophys. Res. Commun. 423 (2012) 526–530. https://doi.org/10.1016/j.bbrc.2012.05.159. [21] J. King, P. Mochalski, K. Unterkofler, G. Teschl, M. Klieber, M. Stein, A. Amann, M. Baumann, Breath isoprene: Muscle dystrophy patients support the concept of a pool of isoprene in the periphery of the human body, Biochem. Biophys. Res. Commun. 423 (2012) 526–530. https://doi.org/10.1016/j.bbrc.2012.05.159. [22] P. Trefz, J. Obermeier, R. Lehbrink, J.K. Schubert, W. Miekisch, D.-C. Fischer, Exhaled volatile substances in children suffering from type 1 diabetes mellitus: results from a cross-sectional study, Sci. Rep. 9 (2019) 15707. https://doi.org/10.1038/s41598-019-52165-x. [23] P. Trefz, S.C. Schmidt, P. Sukul, J.K. Schubert, W. Miekisch, D.-C. Fischer, Non-Invasive Assessment of Metabolic Adaptation in Paediatric Patients Suffering from Type 1 Diabetes Mellitus, J. Clin. Med. 8 (2019). https://doi.org/10.3390/jcm8111797. [24] I. Kushch, B. Arendacká, S. Stolc, P. Mochalski, W. Filipiak, K. Schwarz, L. Schwentner, A. Schmid, A. Dzien, M. Lechleitner, V. Witkovský, W. Miekisch, J. Schubert, K. Unterkofler, A. Amann, Breath isoprene-- aspects of normal physiology related to age, gender and cholesterol profile as determined in a proton transfer reaction mass spectrometry study, Clin. Chem. Lab. Med. 46 (2008) 1011–1018. https://doi.org/10.1515/CCLM.2008.181. [25] P. Lirk, F. Bodrogi, H. Raifer, K. Greiner, H. Ulmer, J. Rieder, Elective haemodialysis increases exhaled isoprene, Nephrol. Dial. Transplant. 18 (2003) 937–941. https://doi.org/10.1093/ndt/gfg049. [26] M.L. T, P. Robin, S. Bernard, Breath isoprene in patients with heart failure, Eur. J. Heart Fail. 3 (2001) 423–427. https://doi.org/10.1016/S1388-9842(01)00128-3. [27] T. Kuzuyama, H. Seto, Diversity of the biosynthesis of the isoprene units, Nat. Prod. Rep. 20 (2003) 171–183. https://doi.org/10.1039/B109860H. [28] E.S. Deneris, R.A. Stein, J.F. Mead, Invitro biosynthesis of isoprene from mevalonate utilizing a rat liver cytosolic fraction, Biochem. Biophys. Res. Commun. 123 (1984) 691–696. https://doi.org/10.1016/0006-291X(84)90284-5. [29] J.E. Metherall, K. Waugh, H. Li, Progesterone Inhibits Cholesterol Biosynthesis in Cultured Cells ACCUMULATION OF CHOLESTEROL PRECURSORS, J. Biol. Chem. 271 (1996) 2627–2633. https://doi.org/10.1074/jbc.271.5.2627. [30] S.L. Mumford, S. Dasharathy, A.Z. Pollack, E.F. Schisterman, Variations in lipid levels according to menstrual cycle phase: clinical implications, Clin. Lipidol. 6 (2011) 225–234. https://doi.org/10.2217/clp.11.9. [31] P. Sukul, P. Oertel, S. Kamysek, P. Trefz, Oral or nasal breathing? Real-time effects of switching sampling route onto exhaled VOC concentrations, J. Breath Res. 11 (2017) 027101. https://doi.org/10.1088/1752-7163/aa6368. Page 14/21 Page 14/21 [32] P. References Lerdau, Ecological and evolutionary aspects of isoprene emission from plants, Oecologia. 118 (1999) 109–123. https://doi.org/10.1007/s004420050709. [42] R.L. Melnick, R.C. Sills, J.H. Roycroft, B.J. Chou, H.A. Ragan, R.A. Miller, Inhalation toxicity and carcinogenicity of isoprene in rats and mice: comparisons with 1,3-butadiene, Toxicology. 113 (1996) 247–252. https://doi.org/10.1016/0300-483X(96)03453-1. Page 15/21 Page 15/21 [43] National Toxicology Program, NTP Toxicology and Carcinogenesis Studies of Isoprene (CAS No. 78- 79-5) in F344/N Rats (Inhalation Studies), Natl. Toxicol. Program Tech. Rep. Ser. 486 (1999) 1–176. [43] National Toxicology Program, NTP Toxicology and Carcinogenesis Studies of Isoprene (CAS No. 78- 9-5) in F344/N Rats (Inhalation Studies), Natl. Toxicol. Program Tech. Rep. S 79-5) in F344/N Rats (Inhalation Studies), Natl. Toxicol. Program Tech. Rep. Ser. 486 (1999) 1–176. [44] D. Anderson, Genetic and reproductive toxicity of butadiene and isoprene, Chem. Biol. Interact. 135– 136 (2001) 65–80. https://doi.org/10.1016/s0009-2797(01)00171-5. [45] R. van der Burgh, N.M. ter Haar, M.L. Boes, J. Frenkel, Mevalonate kinase deficiency, a metabolic autoinflammatory disease, Clin. Immunol. 147 (2013) 197–206. https://doi.org/10.1016/j.clim.2012.09.011. [45] R. van der Burgh, N.M. ter Haar, M.L. Boes, J. Frenkel, Mevalonate kinase deficiency, a metabolic autoinflammatory disease, Clin. Immunol. 147 (2013) 197–206. https://doi.org/10.1016/j.clim.2012.09.011. [46] S. Zhang, Natural history of mevalonate kinase deficiency: a literature review, Pediatr. Rheumatol. Online J. 14 (2016). https://doi.org/10.1186/s12969-016-0091-7. Figures Page 16/21 Figure 1 Figure 1 Page 16/21 Schematic overview of the study setup. Continuous real-time breath analysis (via PTR-ToF-MS) and non- invasive hemodynamic monitoring (via ClearSight System – volume clamp method) were performed along with real-time gene expression analysis (via quantitative-PCR) and clinical lipid-profiling from collected blood samples. Moments from breath analysis of the index subject (before and during her pregnancy) and of her kid (daughter) are presented at the center. Figure 2 Exhaled alveolar isoprene concentrations from entire study population (1026 humans), measured since 2013 till 2019. X-axis represents calendar years since 2013. Y-axis represents exhaled alveolar concentrations of isoprene in all participants. Primary Y-axis stands for alveolar abundances in counts per second (cps) and secondary Y-axis stands for the absolute concentrations in parts-per-billion by volume (ppbV) i.e. obtained via standard calibration driven quantification. Each coloured dot represents the exhaled breath isoprene value (mean over a minute) from each subject. The horizontal dotted line represents the inspiratory (i.e. room air) concentration range of isoprene. The adult without isoprene (n=01) and adults isoprene deficiency (n=05) are also indicated clearly. Other data points at the low range (i.e. below 500 cps) are from pre-school kids and pre-adolescents. Exhaled alveolar isoprene concentrations from entire study population (1026 humans), measured since 2013 till 2019. X-axis represents calendar years since 2013. Y-axis represents exhaled alveolar concentrations of isoprene in all participants. Primary Y-axis stands for alveolar abundances in counts per second (cps) and secondary Y-axis stands for the absolute concentrations in parts-per-billion by volume (ppbV) i.e. obtained via standard calibration driven quantification. Each coloured dot represents the exhaled breath isoprene value (mean over a minute) from each subject. The horizontal dotted line represents the inspiratory (i.e. room air) concentration range of isoprene. The adult without isoprene (n=01) and adults isoprene deficiency (n=05) are also indicated clearly. Other data points at the low range (i.e. below 500 cps) are from pre-school kids and pre-adolescents. Page 17/21 Page 17/21 gure 3 Figure 3 Distribution of exhaled isoprene concentrations according to age groups. X-axis represents different age groups with number of measured subjects in each. Y-axis represents exhaled alveolar concentrations of isoprene (ppbV). The adults are presented into three distinct cohorts depending on the normal levels (n=532), absence (n=01) and deficiency (n=5) of isoprene. Statistical significances of differences were tested by means of one-way repeated measurement-ANOVA on ranks (p-value ≤ 0·05). From all pairwise- Page 18/21 Page 18/21 multiple comparisons, statistically significant differences (with respect to adults with normal isoprene levels) are marked with blue-colored asterisk. Figure 4 Comparison of differences in exhaled alveolar isoprene concentrations in all subjects (a). Gene expression analysis of enzymes involved in cholesterol synthesis in all subjects (b). (a) Y-axis represents the absolute values of repeatedly measured isoprene concentrations (ppbV) in each subject’s expiratory breath and in the corresponding inspiratory room air. X-axis represents study subjects and room air. Index subject is indicated as ‘She’ and rest of the family members are assigned in relation to her. Four healthy control subjects (unrelated to her) are also listed. Statistical significances were tested by means of one- way repeated measurement-ANOVA on ranks (p-value ≤ 0·05). From all pairwise-multiple comparisons, statistically significant differences (with respect to inspiratory room air) are marked with blue-colored asterisk. (b) Gene expression was determined by qPCR and subsequent ddCT determination relative to GAPDH. The mean of all controls was calculated and used as baseline value for fold change determination of controls and index family members. The regulating genes are acetyl-CoA acetyltransferase 2 (ACAT2), HMG-CoA synthase 1 -cytosolic (HMGCS1), HMG-CoA reductase (HMGCR), mevalonate kinase (MVK), phosphomevalonate kinase (PMVK), di-phosphomevalonate decarboxylase (MVD), isopentenyl-diphosphate delta isomerase 1 (IDI1), farnesyl-diphosphate synthase 1 (FDPS1), geranylgeranyl diphosphate synthase 1 (GGPS1), farnesyl-diphosphate farnesyltransferase 1 (FDFT1), lanosterol synthase (LSS), 7-dehydrocloesterol reductase (DHCR7). Page 19/21 Figure 5 Heatmap of exhaled breath VOC compositions, lipid profile and hemodynamic values of the index subject during different phases of her womanhood. Normalized (onto maximum value) expression (i.e. Heatmap) of relative changes in exhaled VOC concentrations throughout her menstrual cycle (Period – day after period – mid follicular – ovulation phase – mid luteal – 2nd period), pregnancy (from 4th – 10th months) and the immediate follicular and luteal phases of the menstrual cycle after the successful child birth is presented as a heatmap. For VOCs, changes in colours from red to blue symbolise relative Heatmap of exhaled breath VOC compositions, lipid profile and hemodynamic values of the index subject during different phases of her womanhood. Normalized (onto maximum value) expression (i.e. Heatmap) of relative changes in exhaled VOC concentrations throughout her menstrual cycle (Period – day after period – mid follicular – ovulation phase – mid luteal – 2nd period), pregnancy (from 4th – 10th months) and the immediate follicular and luteal phases of the menstrual cycle after the successful child birth is presented as a heatmap. For VOCs, changes in colours from red to blue symbolise relative Page 20/21 Page 20/21 changes from higher to lower concentrations and vice versa (Please refer to the color scale). The sum formula of protonated isoprene [i.e. (C5H8)H+] is indicated in red color. Absolute values of lipid profile (cholesterol, HDL, LDL, lipoproteins and triglycerides) and hemodynamic parameters (cardiac output, stroke volume and pulse rate) along with corresponding healthy reference range (age, gender and BMI matched) are presented before, during and after pregnancy. In spite of pregnancy driven pronounced physiological and metabolic effects on many other VOCs, isoprene remained absent in the index. Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. SupplementarymaterialSukuletalHeliyonResearchSquare.docx Page 21/21
https://openalex.org/W3009265656	http://jurnal.atmaluhur.ac.id/index.php/sisfokom/article/download/792/610	Indonesian	null	Implementasi Simple Additive Weighting Penentuan Prioritas Penanganan Sumber Air Bersih	Jurnal Sistem Informasi dan Komputer/Jurnal Sisfokom	2,020	cc-by-sa	3,999	Jurnal SISFOKOM (Sistem Informasi dan Komputer), Volume 09, Nomor 01, PP 77 - 82 Jurnal SISFOKOM (Sistem Informasi dan Komputer), Volume 09, Nomor 01, PP 77 - 82 Implementasi Simple Additive Weighting Penentuan Prioritas Penanganan Sumber Air Bersih Evi Yulianingsih[1], Nia Oktaviani[2], Usman Ependi[3] * Program Studi Sistem Informasi, Fakultas Ilmu Komputer[1], [2], [3] Universitas Bina Darma Palembang, Indonesia evi_yulianingsih@binadarma.ac.id [1], niaoktaviani@binadarma.ac.id [2], u.ependi@binadarma.ac.id [3] * Abstract— Clean water is the need of every human being, for that the necessity of clean water is a priority for the government in providing services. The current problem is the lack of clean water that has not been evenly distributed due to government limitations. For that reason, in providing services, it must be according to priority in order to have a broad impact on the community. For this reason, in this research the development of an information system can be used to assist in determining priorities in handling clean water sources. In the development process the waterfall method is used while the determination of priorities uses simple additive weighting. The results showed that the simple additive weighting method can provide a clear value in determining the priority of clean water handling. This condition can be seen from the existing process in the information system, which starts from weighting, filling the survey value to simple additive weighting calculations. besides that, the information system can run well in accordance with the expected proven by the results of testing using black box testing to get the test results received. This condition is proven by the function of the information system that produces the calculation correctly in accordance with the provisions of the SAW. mendapatkan hasil pengujian yang diterima. Kondisi ini dibuktikan dengan fungsi sistem informasi yang menghasilkan perhitungan dengan benar sesuai dengan ketentuan SAW. Kata Kunci—Simple Additive Weighting, Waterfall, Air Bersih p-ISSN 2301-7988, e-ISSN 2581-0588 DOI : 10.32736/sisfokom.v9.i1.792, Copyright ©2020 Submitted : 8 Februari 2020, Revised : 24 Februari 2020, Accepted : 7 Maret 2020, Published : 9 Maret 2020 PENDAHULUAN Sesuai dengan Undang Undang Dasar 1945 yang terdapat pada pasal 33 ayat 2 menyatakan bahwa “bumi dan air dan kekayaan alam yang terkandung di dalamnya dikuasai oleh negara dan dipergunakan untuk sebesar-besar kemakmuran rakyat” [1]. Dengan demikian penjaminan kegunaan sumber daya alam terutama sumber daya air menjadi keharusan bagi penyelengara negara terutama pemerintah. salah satu usaha nyata pemerintah yaitu adanya instasnsi yang membidangi masalah pengairan yaitu Dinas PU Cipta Karya dan Tata Ruang. Dimana dinas ini terdapat disetiap kabupaten/kota walaupun dengan nama yang berbeda namun dengan fungsi yang sama. Salah satu dinas tersebut yaitu Dinas PU Cipta Karya dan Tata Ruang yang ada di Kabupaten Musi Banyuasin. Dinas ini memiliki peran sebagai pelaksana bidang PU dan pengelolaan sumber daya air yang ada di lingkungan Kabupaten Musi Banyuasin. Pengelolaan sumber daya air tidak hanya terkait dengan pengelolaan kebutuhan pertanian, pertambangan namun juga berkaitan dengan pengelolaan kebutuhan air bersih untuk rumah tangga yang meliputi wilayah kota dan desa. Keywords— Simple Additive Weighting, Waterfall, Clean Water Abstrak—Air bersih adalah kebutuhan setiap manusia, untuk itu kebutuhan air bersih merupakan prioritas bagi pemerintah dalam memberikan layanan. Masalah saat ini adalah kurangnya air bersih yang belum merata karena keterbatasan pemerintah. Oleh Sebab itu, dalam memberikan layanan, harus sesuai prioritas agar berdampak luas pada masyarakat. Untuk itu, dalam penelitian ini dilakukan pengembangan sistem informasi yang dapat digunakan untuk membantu dalam menentukan prioritas dalam penanganan sumber air bersih. Dalam proses pengembangan metode air terjun digunakan sementara penentuan prioritas menggunakan pembobotan aditif sederhana. Hasil penelitian menunjukkan bahwa metode pembobotan aditif sederhana dapat memberikan nilai yang jelas dalam menentukan prioritas penanganan air bersih. Kondisi ini dapat dilihat dari proses yang ada dalam sistem informasi, yang dimulai dari pembobotan, pengisian nilai survei hingga perhitungan pembobotan aditif sederhana. Selain itu, sistem informasi dapat berjalan dengan baik sesuai dengan yang diharapkan dibuktikan dengan hasil pengujian menggunakan pengujian black box untuk Air bersih merupakan salah satu kebutuhan dasar bagi masyarakat. Kebutuhan tersebut bersifat berkesinambungan sehingga pelayanan penyediaan air bersih menjadi keharusan [2]. Kondisi tersebut tentunya terjadi disetiap daerah khususnya di Kabupaten Musi Banyuasin. Untuk memenuhi kebutuhan air bersih kadang kala tidak sebanding dengan ketersediaan air bersih itu sendiri sehingga tidak terpenuhinya permintaan air bersih. Selain itu juga dapat disebabkan oleh penigkatan permukinan penduduk atau berkembangnya sebuah kawasan serta menigkatnya perekonomian masyarakat itu sendiri. II. METODOLOGI PENELITIAN Untuk melakukan penelitian sebagai upaya penyelesaian malasah maka metode yang digunakan didalam penelitian ini dapat dijelaskan seabgai berikut PENDAHULUAN 77 Jurnal SISFOKOM (Sistem Informasi dan Komputer), Volume 09, Nomor 01, PP 77 - 82 juga dapat digunakan untuk membuat sistem informasi pendukung keputusan dalam melakukan evaluasi keputusan dalam proses “evaluasi alternatif pemilihan pemenang pengadaan asset” [11]. Sesuai dengan kondisi tersebut maka timbul beberapa permasalahan di sisi pemerintah sebagai penyedia air bersih. Diantara permasalahan tersebut yaitu: pertama, banyaknya permintaan masyarakat. Kedua, terbatasnya sumber daya dalam memenuhi kebutuhan ari bersih, dan ketiga penentuan prioritas pemasangan air bersih. Permasalahan yang timbul tersebut tentunya harus diberikan solusi bagi pemerintah, terutama dalam penentuan keputusan. salah satu solusi yang dapat digunakan yaitu membuat mekanisme penentuan prioritas dalam membuat fasilitas air bersih utuk masyarakat. Proses atau mekanisme yang dapat diberikan yaitu melalui sistem informasi yang dapat membantu pemerintah dalam menentukan atau mengambil kebijakan tentang prioritas pemasangan air bersih dari semua usulan yang diberikan oleh masyarakat. Sesuai dengan kondisi yang terjadi berkaitan dengan penentuan prioritas penyediaan air bersih pada Kabupaten Musi Banyuasin maka salah satu solusi yang dapat dilakukan yaitu menggunakan sistem informasi. Untuk itu dalam penelitian akan dilakukan pengembangan sistem informasi dengan menerapkan metode simple additive weighting (SAW) sebagai metode untuk menentukan prioritas dalam membuat fasilitas air bersih. Penentuan prioritas tersebut berdasarkan kriteria yang terdiri dari jumlah penduduk, kebutuhan air, debit air, jarak pipa irigasi, beda tinggi, gesekan dan tekanan. Penentuan atau penyeleksian tersebut dapat dilakukan dengan menerapkan berbagai metode pada sistem informasi untuk mengambil keputusan atau penunjang keputusan. Salah satu metode tersebut yaitu metode simple additive weighting (SAW). Metode simple additive weighting (SAW) merupakan metode yang menekankan pada pembobotan dari semua atribut yang menjadi penilaian [3]. Simple additive weighting (SAW) memiliki konsep dengan cara melakukan penjumlahan terhadap atribut atau indikator dalam penilaian untuk menentukan rangking [4]. Metode simple additive weighting (SAW) juga melakukan proses normalisasi terhadap matrik kedalam skala tertentu untuk dilakukan perbandingan dalam penentuan rating atau perangkingan [5]. Hasil penilaian dalam bentuk rating atau perangkingan didasari oleh nilai kritertia atau pembobotan sesuai ketentuan [6]. A. Metode Penelitian Dalam melakukan peneltian implementasi simple additive weighting dalam penentuan prioritas penanganan sumber air bersih ini metode yang digunakan untuk menjelaskan hasil peneltian menggunakan teknik atau metode deskriptif. Dalam banya penelitian dijelaskan bahwa metode ini merupakan metode yang digunakan untuk menjelasakan fakta dan penomena sebuah objek yang diteliti berdasarkan pandangan peneliti [12]. Dapat juga dikatakan bahwa metode penelitian dekkriptif merupakan paparan sebuah kedaan fakta yang diteliti berdasarkan perspektif peneliti [13]. Penelitian terkait dengan metode simple additive weighting (SAW) digunakan untuk membuat sistem informasi pendukung keputusan. Diantara penelitian tersebut dilakukan untuk menentukan seleksi penerima beasiswa. Kriteria penilaian yang digunakan sebanyak lima kriteria yaitu indek prestasi komulatif, penghasilan orang tua, prestasi siswa, organisasi yang diikuti dan tanggungan orang tua [7]. Penelitian lain juga berkaitan dengan sistem pendukung keputusan penerimaan beasiswa, dimana di dalam penelitian ini yang menjadi kriteria penilaian yaitu penghasilan orang tua, semester, tanggungan, saudara kandung dan nilai [8]. Kajian lain juga berkaitan dengan simple additive weighting (SAW) dilakukan untuk menentukan jurusan pada sekolah tingkat menengah atas [9]. Selain dari penentuan beasiswa dan jurusan metode simple additive weighting (SAW) digunakan untuk menentukan siswa bermasalah pada sebuah sekolah. Di dalam kajian tersebut yang menjadi kriteria yaitu kepatuhan, nilai rapot, kehadiran dan jiwa sosial [10]. Metode simple additive weighting (SAW) B. Metode Pengembangan Sistem Informasi B. Metode Pengembangan Sistem Informasi Untuk menghasilkan sistem informasi yang sesuai dengan kebtuhan dalam pemencahan masalah penentuan prioritas atau rangking pembangaunan penyediaan ari bersih, maka dalam proses pengembangan diguakan sebuah metode pengembangan. Metode pengembangan yang digunakan yaitu metode pengembangan kasik yaitu waterfall. Metode ini telah mengalami beberapa perkembangan dalam langkah-langkah pengembangan. Tahapan metode waterfall terdapat dua jenis yaitu pertama: requirement analysis, design, implementation, testing, dan maintenance [14]. Kedua yaitu: communication, planning, modeling, construction, dan deployment [15]. Untuk itu dalam proses pengembangan sistem informasi penentuan prioritas air bersih dengan mengimplementasi simple additive weighting (SAW) digunakan proses pengembangan cara kedua. Proses pengembangan dengan cara kedua ini seperti yang diperlihatkan pada Gambar 1. p-ISSN 2301-7988, e-ISSN 2581-0588 DOI : 10.32736/sisfokom.v9.i1.792, Copyright ©2020 Submitted : 8 Februari 2020, Revised : 24 Februari 2020, Accepted : 7 Maret 2020, Published : 9 Maret 2020 78 Gambar 1. Proses Pengembangan Waterfall [16] Gambar 1. Proses Pengembangan Waterfall [16] Gambar 1. Proses Pengembangan Waterfall [16] 78 Jurnal SISFOKOM (Sistem Informasi dan Komputer), Volume 09, Nomor 01, PP 77 - 82 82 diagram digunakan untuk menggambarkan interaksi dari sistem informasi. Sesuai Gambar 1 merupakan proses pengembangan sistem informasi maka dapat dijelaskan pekerjaan atau aktivitas yang dilakukan untuk setiap tahapan pengembagan sebagai beriktu [15]: Use case diagram digunakan dalam pemodelan sistem informasi ini sebagai alat untuk menggambarkan peran pengguna yang berinteraksi dengan sistem informasi. Selain itu juga use case diagram digunakan untuk melihat komponen fungsional sistem informasi yang dapat diakses oleh pengguna [18]. Dalam pengembagnan sistema informasi penentuan prioritas sumber air bersih ini dapat diketahui bahwa pengguna dapat melakukan berbagai aktivitas yaitu: memasukkan data lokasi, informasi alternatif kriteria, informasi normalisasi, informasi hasil normalisasi dan keputusan pembuatan sumber air bersih. Kondisi tersebut seperti yang digambarkan pada use case diagram yang terlihat pada Gambar 2.  Communication, aktivitas dan pekerjaan yang dilakukan pada tahapan communication yaitu analisis kebutuhan awal sesuai dengan project initiation. Pada tahapan ini terdapat aktivitas pengumpulan data yang akan digunakan dalam proses pengembangan sistem informasi.  Planning, aktivitas dan pekerjaan yang dilakukan pada tahapan planning yaitu membuat perencanaan dalam proses pengembangan sistem infomrasi. Perancangan tersebut meliputi tiga aspek yaitu estimating, scheduling dan tracking.  Modeling, aktivitas dan pekerjaan yang dilakukan pada tahapan modeling yaitu membuat sebuah model analisis dan perancangan antarmuka dari sistem infomrasi yang akan dihasilkan. Namun dalan penelitian ini hanya akan dibuat dari hasil model anaslsis menggunakan unified modeling language (UML). Gambar 2. B. Metode Pengembangan Sistem Informasi Use Case Diagram uc Use Case Model user input data lokasi info hasil alternatif kriteria info normalisasi Info Hasil Normalisasi Keputusan «include» «include» «include» «include»  Construction, aktivitas dan pekerjaan yang dilakukan pada construction yaitu proes penterjemahan hasil modeling kedalam bentuk coding. Selain itu juga dilakukan implementasi simple additive weighting (SAW) sebagai metode dalam menentukan prioritas penyediaan air bersih.  Deployment, tahapan deployment merupakan proses terakhir yang dilakukan. Pada tahapan ini dilakukan pengiriman (delivery) sistem informasi kepada pemakai.  Deployment, tahapan deployment merupakan proses terakhir yang dilakukan. Pada tahapan ini dilakukan pengiriman (delivery) sistem informasi kepada pemakai. Gambar 2. Use Case Diagram C. Metode Pengujian Activity Diagram merupakan diagram yang dapat digunakan untuk melihat alur kerja, logika prosedur dan prosis bisnis dari sebuah sistem informasi. Activity diagram juga dapat menunjukkan alir sebuah proes dalam system informasi. Namun memiliki perbedaan yang mendasar dengan flowchart. Perbedaan tersebut terlihat dari bagaimana activity diagram dalam memperlihatkan behavior dari sistem informasi [19]. Metode pengujian merupakan metode yang digunakan untuk menguji sistem informasi yang dihasilkan. Metode pengujian yang digunakan yaitu metode black box testing. Metode ini merupakan metode yang melihat sistem informasi dari sudut pandang input dan output. Metode ini juga sering dikenal dengan metode pengujian yang melihat fungsional sebuah sistem informasi, dalam artian melihat apakah komponen sistem informasi yang dihasilkan telah berjalan sesuai dengan yang diharapkan ataukah sebaliknya [17]. Gambar 3. Activity Diagram analysis Activ ity Diagram Start Menu Utama Lokasi Normalisasi Hasil Normalisasi Hasil Alternatif Kriteria Keputusan end p-ISSN 2301-7988, e-ISSN 2581-0588 DOI : 10.32736/sisfokom.v9.i1.792, Copyright ©2020 Submitted : 8 Februari 2020, Revised : 24 Februari 2020, Accepted : 7 Maret 2020, Published : 9 Maret 2020 D. Pemodelan Sistem Informasi Pemodelan sistem informasi merupakan proses penterjemahan kebutuhan sistem informasi dari kebutuhan sistem informasi. Dimana dalam implementasi simple additive weighting dalam penentuan prioritas penanganan sumber air bersih kebutuhan yang ingin dicapai yaitu bagaimana menentukan prioritas berdasarkan kriteria yang telah ditentukan. Untuk itu di dalam pemodelan sistem informasi ini dibuat dua jenis pemodelan sebagai gambaran bagaimana sistem informasi yang akan dihasilkan. Dalam melakukan pemodelan ini digunakan dua diagram untuk menggambarkan bagaimana sistem informasi yang akan dihasilkan. Dua diagram tersebut yaitu use case diagram dan activity diagram. Use case diagram digunakan untuk menggambarkan tingkah laku atau behavior dari sistem informasi sedangkan activity Gambar 3. Activity Diagram Untuk itu dalam implementasi metode simple additive weighting (SAW) untuk penentuan prioritas penanganan sumber air bersih activity diagram seperti yang diperlihatkan pada Gambar 3. Dari Gambar 3 dapat diketahui bahwa semua 79 Jurnal SISFOKOM (Sistem Informasi dan Komputer), Volume 09, Nomor 01, PP 77 - 82 82 yaitu sangat tinggi yang berarti nilai yang digunakan yaitu angka 5. Namun dapat juga dibuat dengan nilai yang berbeda tergantung dengan prioritas yang diinginkan seperti tinggi dengan nilai 4, cukup dengan nilai 3, rendah dengan nilai 2 dan sangat rendah dengan nilai 1. komponen sistem informasi berjalan melalui menu utama. Proses yang dilakukan dalam system informasi dimulai dari penentuan lokasi dan dapat juga melalui alternative kriteria. Selanjutnya dilakukan normalisasi. Kemudian dilanjutkan dengan hasil normalisasi dan terakhir mendapatkan keputusan. Gambar 4. Bobot Kriteria Penilaian p-ISSN 2301-7988, e-ISSN 2581-0588 DOI : 10.32736/sisfokom.v9.i1.792, Copyright ©2020 Submitted : 8 Februari 2020, Revised : 24 Februari 2020, Accepted : 7 Maret 2020, Published : 9 Maret 2020 80 Gambar 9. Hasil Rangking A. Implementasi Sistem Informasi Dalam melakukan implementasi sistem informasi yang dapat digunakan untuk menentukan prioritas penyedian air bersih maka terdapat kriteria dalam melakukan penilaian. Kriteria penilaian tersebut meliputi C1 untuk jumlah penduduk, C2 untuk kebutuhan air, C3 untuk debit air, C4 untuk jarak pipanisasi, C5 untuk beda tinggi, C6 untuk gesekan dan C7 untuk tekanan. Kriteria penilaian tersebut seperti yang diperlihatkan pada Tabel 1. TABEL I KRITERIA PENILAIAN Kriteria Keterangan C1 Jumlah Penduduk C2 Kebutuhan Air C3 Debit Air C4 Jarak Pipanisasi C5 Beda Tinggi C6 Gesekan C7 Tekanan Gambar 4. Bobot Kriteria Penilaian Setelah mengetahui nilai bobot kriteria yang digunakan maka selanjutnya dapat melakukan pengisian hasil survey. Pengisian hasil survey didapat dari survey lapangan yang dilakukan pihak Dinas PU Cipta Karya dan Tata Ruang yang ada di Kabupaten Musi Banyuasin. Survey yang dilakukan dengan cara melakukan kunjungan langsung ke kecamatan- kecamatan yang ada di Kabupaten Musi Banyuasin. Selain itu juga dapat dilakukan dengan cara mewakilkan kepada pihak kecamtan untuk melakukan survey lapangan terhadap daerah atau pedesaan yang ada di setiap kecamtan. Nilai survey yang dimasukkan pada sistem informasi ini berupa nilai keadaan umum untuk setiap kecamatan. Untuk memasukkan data hasil survey pada sistem informasi dimulai dari memilih kode kecamatan dan akan otomatis tampil nama kecamatan yang dimaksud. Setelah nama kecamatan muncul maka pengguna dapat dapat memasukkan data nilai hasil survey berupa jumlah penduduk, kebutuhan air, debit mata air, jarak pipanisasi, beda tinggi, gaya gesekan dan tinggi tekanan. Proses pengisian tersebut seperti yang diperlihatkan pada Gambar 5. Untuk menentukan penilaian digunakan bilangan fuzzy yang terdiri dari lima kriteria yaitu 5 berarti sangat tinggi, 4 berarti tinggi, 3 berarti cukup, 2 berarti rendah, dan 1 berarti sangat rendah. Penggunaan nilai fuzzy tersebut untuk memudahkan perhitungan implementasi simple additive weighting (SAW) agar penentuan prioritas dapat dilakukan dengan mudah. Pada Tabel 2 dapat dilihat kriteria bilangan fuzzy yang digunakan. TABEL II KRITERIA BILANGAN Bilagan Fuzzy Nilai Sangat Tinggi 5 Tinggi 4 Cukup 3 Rendah 2 Sangat Rendah 1 Gambar 5. Hasil Survey Sumber Air Lapangan Setelah mengetahui kriteria dan kriteria bilangan yang digunakan untuk melakukan perhitungan prioritas penyediaan sumber air bersih maka selanjutnya adalah melakukan penentuan bobot kriteria pada sistem informasi yang dihasilkan. Seperti yang terlihat pada Gambar 4 merupakan form yang dapat digunakan untuk menentukan bobot kriteria tersebut. Pada Gambar 4 dapat dilihat penentuan bobot kriteria Gambar 5. , py g ted : 8 Februari 2020, Revised : 24 Februari 2020, Accepted : 7 Maret 2020, Published : 9 Maret 2020 A. Implementasi Sistem Informasi Hasil Survey Sumber Air Lapangan 80 Jurnal SISFOKOM (Sistem Informasi dan Komputer), Volume 09, Nomor 01, PP 77 - 82 Perhitungan dilakukan untuk menentukan perangkingan atau pemeringkatan prioritas dalam penanganan sumber air bersih. Proses perhitungan tersebut seperti yang diperlihatkan pada Gambar 8. Pada Gambar 8 dapat dilihat perhitungan dilakukan untuk setiap daerah yang dimasukkan dan berdasarkan kriteria penialaian. Untuk melihat hasil perhitungan setelah selesai memaukkan data hasil survey maka dapat dilakukan pada menu hasil perhitungan SAW. Menu ini akan menampilan secara keseluruhan hasil perhitungan dari forim yang ditampilkan daman bentuk atau berupa tabel hasil survey seperti yang diperlihatkan pada Gambar 6. Hasil pengisian ditampilan dikelompokan berdasarkan nama tempat dan kriteria penilaian yaitu C1 untuk jumlah penduduk, C2 untuk kebutuhan air, C3 untuk debit air, C4 untuk jarak pipanisasi, C5 untuk beda tinggi, C6 untuk gesekan dan C7 untuk tekanan. Proses terakhir dari simple additive weighting (SAW) adalah menentukan rangking dari setiap nilai yang diproses. Untuk menentukan rangkin maka dilakukan jumlah untuk setiap komponen kriteria penialaian yaitu mulai dari C1 untuk jumlah penduduk, C2 untuk kebutuhan air, C3 untuk debit air, C4 untuk jarak pipanisasi, C5 untuk beda tinggi, C6 untuk gesekan dan C7 untuk tekanan. Dari penjumlahan tersebut maka didapat nilai untuk masing-masing daerah yang dimasukkan. Semakin tinggi nilai yang didapat maka menjadi prioritas utama dalam pengembangan atau pengadaan sumber air. Hasil perhitungan dalam penentuan prioritas penanganan sumber air bersih seperti yang diperlihatkan pada Gambar 9. Gambar 6. Alternatif Setiap Kriteria Gambar 9. Hasil Rangking Gambar 6. Alternatif Setiap Kriteria Hasil masukkan seperti yang diperlihatkan pada Gambar 6 dilakukan konversi nilai berdasarkan kriteria bilangan fuzzy yaitu dari 1 sampai dengan 5. 5 berarti sangat tinggi, 4 berarti tinggi, 3 berarti cukup, 2 berarti rendah, dan 1 berarti sangat rendah. Salah satu hasil konversi bilangan tersebut seperti yang diperlihatkan pada Gambar 7. Gambar 9. Hasil Rangking IV. PENUTUP Sesuai dengan proses penelitian yang telah dilakukan yaitu implementasi simple additive weighting dalam penentuan prioritas penanganan sumber air bersih maka dapat disimpulkan bahwa untuk menentukan perangkingan dalam menentukan prioritas digunakan kriteria penilaian dan nilai bobot agar nilai pengukuran dapat dilakukan perhitungan dengan mudah. Selain itu juga simple additive weighting dapat diimplementasikan pada sistem informasi dengan sangat sederhana dan mendapatkan hasil yang jelas. Untuk melihat sistem informasi yang dihasilkan sesuai dengan keinginan juga dapat dilakukan dengan melakukan pengujian seperti black box testing. Dari hasil pengujian menunjukkan bahwa semua fungsional sistem informasi berjalan dengan baik dan menghasilkan output atau nilai yang benar. [10] H. T. Sihotang and M. S. Siboro, “Aplikasi Sistem Pendukung Keputusan Penentuan Siswa Bermasalah Menggunakan Metode Saw Pada Sekolah Smp Swasta Mulia Pratama Medan,” J. Inform. Pelita Nusant., vol. 1, no. 1, 2016. [11] R. Setiadi, C. Suhery, and R. Hidayati, “SISTEM PENDUKUNG KEPUTUSAN EVALUASI PEMILIHAN PEMENANG DALAM PELELANGAN PENGADAAN ASET JALAN DAN JEMBATAN MENGGUNAKAN METODE WEIGHTED PRODUCT (WP) BERBASIS WEB (Studi Kasus: Dinas Pekerjaan Umum Provinsi Kalimantan Barat),” Coding J. Komput. dan Apl., vol. 7, no. 03. [12] U. Ependi, A. Putra, and F. Panjaitan, “Evaluasi tingkat kebergunaan aplikasi Administrasi Penduduk menggunakan teknik System Usability Scale,” Regist. J. Ilm. Teknol. Sist. Inf., vol. 5, no. 1, pp. 63–76, 2019. [13] S. Suyanto and U. Ependi, “Pengujian Usability dengan Teknik System Usability Scale pada Test Engine Try Out Sertifikasi,” MATRIK J. Manajemen, Tek. Inform. dan Rekayasa Komput., vol. 19, no. 1, pp. 62–69, 2019. SYSTEM PENENTUAN PEGAWAI TERBAIK DENGAN METODE SAW (STUDI KASUS UNIVERSITAS BANTEN JAYA),” J. Ilm. Sains dan Teknol., vol. 2, no. 2, pp. 14–26, 2018. Komponen Pengamatan Hasil Pengujian Hasil Survey Berfungsi (Dterima) Hasil Perhitungan SAW Berfungsi (Dterima) Komponen Pengamatan Hasil Pengujian Hasil Survey Berfungsi (Dterima) Hasil Perhitungan SAW Berfungsi (Dterima) [5] A. S. Putra, D. R. Aryanti, and I. Hartati, “Metode SAW (Simple Additive Weighting) sebagai Sistem Pendukung Keputusan Guru Berprestasi (Studi Kasus: SMK Global Surya),” in Prosiding Seminar Nasional Darmajaya, 2018, vol. 1, no. 1, pp. 85–97. [6] M. Munir, F. Marisa, and D. Purnomo, “SISTEM PENUNJANG KEPUTUSAN DALAM PENGKLASIFIKASIAN MAHASISWA PENERIMA BEASISWA DENGAN METODE SIMPLE ADDITIVE WEIGHTING DI UNIVERSITAS WIDYAGAMA MALANG,” J. SPIRIT, vol. 10, no. 2, 2018. [7] W. Supriyanti, “Rancang bangun aplikasi sistem pendukung keputusan penerima beasiswa dengan metode SAW,” Creat. Inf. Technol. J., vol. 1, no. 1, pp. 67–75, 2013. [8] M. Muslihudin and D. Hartini, “Perancangan Sistem Pendukung Pengambilan Keputusan Untuk Penerimaan Beasiswa Di SMA PGRI 1 Talang Padang Dengan Model Fuzzy Multiple Attribute Menggunakan Metode Simple Additive Weighting (SAW),” J. TAM (Technology Accept. Model., vol. 4, pp. 34–40, 2017. [9] M. Elistri, J. Wahyudi, and R. Supardi, “Penerapan metode saw dalam sistem pendukung keputusan pemilihan jurusan pada Sekolah Menengah Atas Negeri 8 Seluma,” J. Media Infotama, vol. 10, no. 2, 2014. UCAPAN TERIMA KASIH [14] R. N. Hays, A. Sugiyarta, and D. E. Winungkas, “Aplikasi Inventory Terintegrasi Order System Konsumen Pada Oto Bento Perumnas Cilegon Menggunakan Metode Waterfall,” ProTekInfo (Pengembangan Ris. dan Obs. Tek. Inform., vol. 5, 2018. Terima kasih diucapkan kepada redaksi Jurnal Sisfokom yang telah meluangkan waktu dan tenada untuk melakukan review, editing dan publishing artikel ini. Terima kasih juga kami sampaikan kepada Universitas Bina Darma yang telah membiaya penelitian ini hingga terlaksana dengan baik. [15] F. Syakti, “Sistem Informasi Geografis Penderita Malaria pada Kelurahan Cereme Taba Kota Lubuklinggau,” Digit. Zo. J. Teknol. Inf. dan Komun., vol. 10, no. 2, pp. 178–188, 2019. [16] R. S. Presman, “Rekayasa Perangkat Lunak Pendekatan Praktisi Edisi 7 (Buku Satu),” Penerjemah Adi Nughroho, Georg. John Leopold Nikijuluw, theresia Herlina Rachadiani dan Ike Kurniawati Witaya. Yogyakarta Penerbit Andi, 2012. p-ISSN 2301-7988, e-ISSN 2581-0588 DOI : 10.32736/sisfokom.v9.i1.792, Copyright ©2020 Submitted : 8 Februari 2020, Revised : 24 Februari 2020, Accepted : 7 Maret 2020, Published : 9 Maret 2020 py g mitted : 8 Februari 2020, Revised : 24 Februari 2020, Accepted : 7 Maret 2020, Published : 9 Maret 2020 B. Pengujian Sistem Informasi Gambar 7. Normalisasi Pengujian sistem informasi merupakan proses validasi dan verifikasi sistem informasi. Pengujian juga dilakukan untuk melihat apakah fungsional sistem informasi telah berjalan dengan baik ataukah sebaliknya. Dalam melakukan pengujian sistem informasi digunakan teknik pengujian black box testing. Dimana pengujian ini lakukan untuk melihat fungsional dari sistem informasi yang dihasilkan. Untuk itu berikut dapat dilihat hasil pengujian sistem informasi penentuan prioritas penanganan sumber air menggunakan metode simple additive weighting (SAW). Setelah melakukan konversi hasil survey selanjutnya dilakukan perhitungan sesuai dengan ketentuan simple additive weighting (SAW). Gambar 8. Perhitungan Normalisasi p-ISSN 2301-7988, e-ISSN 2581-0588 DOI : 10.32736/sisfokom.v9.i1.792, Copyright ©2020 Submitted : 8 Februari 2020, Revised : 24 Februari 2020, Accepted : 7 Maret 2020, Published : 9 Maret 2020 81 Gambar 8. Perhitungan Normalisasi g p g j fungsional dari sistem informasi yang dihasilkan. Untuk itu berikut dapat dilihat hasil pengujian sistem informasi penentuan prioritas penanganan sumber air menggunakan metode simple additive weighting (SAW). TABEL III HASIL PENGUJIAN Komponen Pengamatan Hasil Pengujian Bobot Kriteria Berfungsi (Dterima) TABEL III HASIL PENGUJIAN Komponen Pengamatan Hasil Pengujian Bobot Kriteria Berfungsi (Dterima) 81 Jurnal SISFOKOM (Sistem Informasi dan Komputer), Volume 09, Nomor 01, PP 77 - 82 SYSTEM PENENTUAN PEGAWAI TERBAIK DENGAN METODE SAW (STUDI KASUS UNIVERSITAS BANTEN JAYA),” J. Ilm. Sains dan Teknol., vol. 2, no. 2, pp. 14–26, 2018. REFERENSI [1] E. Ruslina, “Makna Pasal 33 Undang-Undang Dasar 1945 dalam Pembangunan Hukum Ekonomi Indonesia,” J. Konstitusi, vol. 9, no. 1, p. 109449, 2012. [17] F. Fatoni and D. Irawan, “Implementasi Metode Extreme Programming dalam Pengembangan Sistem Informasi Izin Produk Makanan,” J. Sisfokom (Sistem Inf. dan Komputer), vol. 8, no. 2, pp. 159–164, 2019. [2] D. Kusumawardani, “Valuasi Ekonomi Air Bersih di Kota Surabaya,” J. Ekon. dan Bisnis Airlangga (JEBA)\| J. Econ. Bus. Airlangga, vol. 21, no. 3, 2011. [18] A. Penira, A. Zahara, M. Ramadhani, and M. L. Amin, “ANALISA DAN PERANCANGAN SISTEM E-CLAIM PADA PT ASURANSI JIWA SYARIAH BUMIPUTERA CABANG MEDAN,” J. Tek. Inform. Kaputama, vol. 4, no. 1, 2020. [3] R. Marbun and I. Gunawan, “Sistem Pendukung Keputusan Perankingan Data Konsumen Penumpang Kereta Api Dengan Menggunakan Metode Simple Additive Weighting Di PT. KAI,” Juripol, vol. 2, no. 1, pp. 15–25, 2018. [19] S. S. Harahap, “Sistem Pendukung Keputusan Penyetujuan Kartu Kredit Pada PT. Bank Rakyat Indonesia Dengan Menggunakan Metode Simple Additive Weighting (SAW),” J. Sist. Inf. dan Ilmu Komput. Prima (JUSIKOM PRIMA), vol. 2, no. 2, pp. 1–7, 2019. [4] I. Gunawan, S. Sutanto, and I. Yunita, “DECISION SUPPORT 82
https://openalex.org/W2031985211	https://pure.mpg.de/pubman/item/item_2086947_4/component/file_2152189/1367-2630_16_9_095002.pdf	English	null	Germanene: a novel two-dimensional germanium allotrope akin to graphene and silicene	New journal of physics	2,014	cc-by	5,956	Germanene: a novel two-dimensional germanium allotrope akin to graphene and silicene Germanene: a novel two-dimensional germanium allotrope akin to graphene and silicene This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 New J. Phys. 16 095002 (http://iopscience.iop.org/1367-2630/16/9/095002) This content has been downloaded from IOPscience. Please scroll down to see the full text. View the table of contents for this issue, or go to the journal homepage for more Download details: Home Search Collections Journals About Contact us My IOPscience Germanene: a novel two-dimensional germanium allotrope akin to graphene and silicene M E Dávila1, L Xian2,3, S Cahangirov2,3, A Rubio2,3,4,5 and G Le Lay6 1 Instituto de Ciencia de Materiales de Madrid-ICMM-CSIC, C/Sor Juana Inés de la Cruz, 3 Cantoblanco, 28049-Madrid, Spain 2 Nano–Bio Spectroscopy group, Universidad del País Vasco, Barrio Sarriena s/n, 48940 Leioa, Bizkaia, Spain 3 European Theoretical Spectroscopy Facility (ETSF), CFM CSIC-UPV/EHU-MPC and DIPC, Av Tolosa 72, E-20018 San Sebastián, Spain 4 Max Planck Institute for the Structure and Dynamics of Matter, Hamburg, Germany 5 Fritz-Haber-Institut Max-Planck-Gesellschaft, Berlin, Germany 6 Aix-Marseille Université, CNRS, PIIM UMR 7345, 13397 Marseille Cedex, France E-mail: mdavila@icmm.csic.es and angel.rubio@ehu.es M E Dávila1, L Xian2,3, S Cahangirov2,3, A Rubio2,3,4,5 and G Le Lay6 1 Instituto de Ciencia de Materiales de Madrid-ICMM-CSIC, C/Sor Juana Inés de la Cruz, 3 Cantoblanco, 28049-Madrid, Spain 2 Nano–Bio Spectroscopy group, Universidad del País Vasco, Barrio Sarriena s/n, 48940 Leioa, Bizkaia, Spain 3 European Theoretical Spectroscopy Facility (ETSF), CFM CSIC-UPV/EHU-MPC and DIPC, Av Tolosa 72, E-20018 San Sebastián, Spain 4 Max Planck Institute for the Structure and Dynamics of Matter, Hamburg, Germany 5 Fritz-Haber-Institut Max-Planck-Gesellschaft, Berlin, Germany 6 Aix-Marseille Université, CNRS, PIIM UMR 7345, 13397 Marseille Cedex, France E-mail: mdavila@icmm.csic.es and angel.rubio@ehu.es Received 31 May 2014, revised 24 July 2014 Accepted for publication 31 July 2014 Published 9 September 2014 New Journal of Physics 16 (2014) 095002 doi:10.1088/1367-2630/16/9/095002 IP Address: 131.169.116.119 This content was downloaded on 11/05/2015 at 15:48 Please note that terms and conditions apply. Germanene: a novel two-dimensional germanium allotrope akin to graphene and silicene M E Dávila1, L Xian2,3, S Cahangirov2,3, A Rubio2,3,4,5 and G Le Lay6 1 Instituto de Ciencia de Materiales de Madrid-ICMM-CSIC, C/Sor Juana Inés de la Cruz, 3 Cantoblanco, 28049-Madrid, Spain 2 Nano–Bio Spectroscopy group, Universidad del País Vasco, Barrio Sarriena s/n, 48940 Leioa, Bizkaia, Spain 3 European Theoretical Spectroscopy Facility (ETSF), CFM CSIC-UPV/EHU-MPC and DIPC, Av Tolosa 72, E-20018 San Sebastián, Spain 4 Max Planck Institute for the Structure and Dynamics of Matter, Hamburg, Germany 5 Fritz-Haber-Institut Max-Planck-Gesellschaft, Berlin, Germany 6 Aix-Marseille Université, CNRS, PIIM UMR 7345, 13397 Marseille Cedex, France E-mail: mdavila@icmm.csic.es and angel.rubio@ehu.es Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Introduction After the successful synthesis of silicene in 2012, which was followed by a surge of studies on elemental, novel two-dimensional (2D) materials beyond graphene, a daunting quest was to obtain germanene, the germanium-based analogue of graphene, already predicted to possibly exist in 2009. Although its fully hydrogenated form, germanane, was fabricated using a wet chemistry method in 2013, germanene has remained elusive. Here we show compelling experimental and theoretical evidence of its synthesis by dry epitaxial growth on a gold (111) surface. The discovery of graphene boosted research in nanoscience on 2D materials, especially on elemental ones. In 2012, silicene, graphene’s silicon cousin [1], was successfully synthesized on two metallic templates, namely a silver (111) surface [2, 3] and the zirconium diboride (0001) surface of a thin ﬁlm grown on a silicon (111) substrate [4]. One year later, silicene was also grown on an iridium (111) surface [5]. Germanene, another germanium-based cousin of graphene, along with silicene, had already been predicted to be stable as freestanding novel germanium and silicon 2D allotropes in a low buckled honeycomb geometry by Cahangirov et al in 2009 [6]. In the quest for germanene, its fully hydrogen-terminated partner, germanane (GeH), was ﬁrst fabricated from the topochemical deintercalation of the layered van der Waals solid calcium digermanide (CaGe2) [7]; next, the stability of germanane was improved by replacing the H atom termination with a methyl group one [8]. Since silicene has, up to now, only been synthesized in dry conditions under ultrahigh vacuum (UHV)—with silver (111) as the favored substrate—trying to synthesize germanene by also growing it on Ag(111) single crystals using germanium molecular beam epitaxy seems tempting. However, to the best of our knowledge, this has failed up to now, because (1) the ‘magic mismatch’ between three lattice constants of silicene and four of the Ag(111) plane is not fulﬁlled for germanene, and (2) germanium most probably prefers to form an ordered Ag2Ge surface alloy, where Ge atoms, up to a coverage of one-third of a monolayer (1/3 ML), substitute Ag ones at the silver surface. This surface alloy presents a complex ‘√3 × √3’ structure [9], which not only deviates in its geometry but also in its electronic properties [9, 10] from the simple √3 × √3 reconstruction envisaged earlier [11]. We have thus used a gold (111) substrate instead to avoid such a surface alloy formation. Abstract We have grown an atom-thin, ordered, two-dimensional multi-phase ﬁlm in situ through germanium molecular beam epitaxy using a gold (111) surface as a substrate. Its growth is similar to the formation of silicene layers on silver (111) templates. One of the phases, forming large domains, as observed in scanning tunneling microscopy, shows a clear, nearly ﬂat, honeycomb structure. Thanks to thorough synchrotron radiation core-level spectroscopy measurements and advanced density functional theory calculations we can identify it as a √3 × √3 R(30°) germanene layer in conjunction with a √7 × √7 R(19.1°) Au(111) supercell, presenting compelling evidence of the synthesis of the germanium- based cousin of graphene on gold. S Online supplementary data available from stacks.iop.org/NJP/16/095002/ mmedia Keywords: germanene, ARPES, STM, DFT Keywords: germanene, ARPES, STM, DFT Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journ and DOI. New Journal of Physics 16 (2014) 095002 1367-2630/14/095002+10$33.00 © 2014 IOP Publishing Ltd and Deutsche Physikalische Gesellschaft New J. Phys. 16 (2014) 095002 M E Dávila et al Results and discussion The methodology we adopted is very similar to the one we used for the synthesis of silicene on Ag(111) (see [2] for details); the silver sample was exchanged for a gold (111) one and the silicon source in the experimental set-up in Marseille was replaced by a germanium evaporator to deposit Ge atoms onto a clean Au(111) surface prepared in a standard fashion by Ar+ ion bombardment and annealing. The in situ cleaned Au(111) surface is characterized by its well- known 22 × √3 herringbone structure [25]. Low energy electron diffraction (LEED) and STM observations were performed at room temperature (RT) at different stages of the growth, carried out at several substrate temperatures to determine potential candidates for germanene in an overall multiphase diagram, as was already the case for silicon deposition onto Ag(111) [26]. p g y p g In this paper we focus on one of the phases obtained at ∼200 °C growth temperature at about 1 ML coverage, as estimated from the 32% attenuation of the Au 4f7/2 core level intensity. It covers extended regions, larger than 50 × 50 nm2 in size, with a honeycomb appearance and a very small corrugation of just 0.01 nm, as well as with a weak long range modulation in STM imaging, as displayed in ﬁgure 1. However, as was the case for the main silicene phase on Ag(111) (denoted 3 × 3/4 × 4 to illustrate the 3 × 3 reconstruction of silicene in a 4 × 4 coincidence cell on Ag(111)), here, again, the observed honeycomb arrangement is too large [2] to correspond directly to a germanene primitive cell. Instead, its cell size ﬁts to a √7 × √7 R(19.1°) superstructure in terms of Au (111), which is in accord with the LEED pattern of ﬁgure 1(b). This somewhat astonishing LEED pattern, with many extinctions, reﬂects the diffraction from three main co-existing phases, as shown schematically in ﬁgure 1(c): √19 × √19 R(23.4°), 5 × 5 and √7 × √7 R (19.1°) phases, the last one being of prime interest here, noted with reference to the Au(111) 1 × 1 basis vectors. Introduction Indeed, for silicene synthesis we deposited silicon on silver (111) surfaces because the inverse system, silver grown on Si(111) surfaces, is well-known to form atomically abrupt interfaces, without intermixing [12]. Our choice of an Au(111) substrate is based on the same strategy. It turns out that among the four noble metals on elemental semiconductor systems studied, namely, Au, Ag/Ge, Si(111) [12], the most similar in several aspects, especially in the growth mode—the Stranski–Krastanov (or layer-plus-islands) mode characterized by the formation of a √3 × √3 R30° superstructure (or wetting layer) associated with the formation of Au trimers on Ge(111) [13] or Ag ones on Si(111)—appeared to be Si/Ag(111) [14] and Ge/Au(111) [15], a trend conﬁrmed in a recent study of Au/Ge(111) [16]. This strategy has paid off. As we show, we have succeeded in growing a 2D germanium sheet with a honeycomb appearance in scanning tunneling microscopy (STM) imaging on top of the Au(111) surface. Its formation shows great similarity to the growth of silicene sheets on the Ag(111) surface [2, 3]. We identify it as a germanene sheet after detailed synchrotron radiation spectroscopy measurements of the deposit (Ge 3d) and substrate (Au 4f) shallow core-levels, and thanks to advanced density functional theory calculations of the geometry and stability of 2 New J. Phys. 16 (2014) 095002 New J. Phys. 16 (2014) 095002 M E Dávila et al the system, using for the exchange-correlation potentials the general gradient approximation (see the supplementary material, available at stacks.iop.org). Independently, and nearly in parallel, another group has concluded the formation of a germanene layer on a platinum (111) template, but with this appearing strongly distorted, based only on STM observations of a hexagonal arrangement and density functional theory calculations in the basic local density approximation [17]. the system, using for the exchange-correlation potentials the general gradient approximation (see the supplementary material, available at stacks.iop.org). Independently, and nearly in parallel, another group has concluded the formation of a germanene layer on a platinum (111) template, but with this appearing strongly distorted, based only on STM observations of a hexagonal arrangement and density functional theory calculations in the basic local density approximation [17]. Introduction One can anticipate this new discovery having a major impact because of the expected very high mobilities of the carriers [18]; the potential optical applications [19]; the predicted robust 2D topological insulator character, nearly up to room temperature, resulting from the large effective spin–orbit coupling [20, 21], opening the way to the quantum spin Hall effect [22]; the possibility of very high Tc superconductivity [23, 24]; and, last but not least, the practicability of direct integration in the current electronics industry. Results and discussion The extra spots observed at a small distance from the integer order ones suggest the presence of a distortion with a long repetition length, most probably reﬂecting the modulation seen in the STM images, which can be related to underlying remnants of the native herringbone structure of the pristine Au(111) surface. At this stage, we stress that in most circumstances the growth of silicene on Ag(111) also takes place in several phases, the most frequent LEED pattern corresponding in this case to a superposition of diffraction patterns, with many extinctions stemming essentially from √13 × √13 R(13.9°) and 4 × 4 phases (labeled with reference to Ag(111)(1 × 1), but also, typically, a √19 × √19 R(23.4°) phase [26]. Hence, 3 New J. Phys. 16 (2014) 095002 M E Dávila et al Figure 1. (a) 16.2 nm × 16.2 nm STM image of the modulated honeycomb √7 × √7 superstructure with a close-up in the bottom left corner (sample bias: −1.12 V, 1.58 nA; the √7 × √7 unit cell is drawn in black); (b) associated LEED pattern taken at 59 V; (c) schematic illustration of one sixth of the pattern, ﬁlled dots: hidden (0,0) spot and integer order spots, open circles: spots corresponding to the √7 × √7 superstructure (in red), the √19 × √19 one (in green) and the 5 × 5 (in blue). Figure 1. (a) 16.2 nm × 16.2 nm STM image of the modulated honeycomb √7 × √7 superstructure with a close-up in the bottom left corner (sample bias: −1.12 V, 1.58 nA; the √7 × √7 unit cell is drawn in black); (b) associated LEED pattern taken at 59 V; (c) schematic illustration of one sixth of the pattern, ﬁlled dots: hidden (0,0) spot and integer order spots, open circles: spots corresponding to the √7 × √7 superstructure (in red), the √19 × √19 one (in green) and the 5 × 5 (in blue). somehow, the astonishing LEED pattern we get for the present Ge/Au(111) system should not be such a surprise. √ √ In line with the silicene phases on the Ag(111) surface, the √7 × √7 R(19.1°) superstructure in terms of Au(111) we focus on here could possibly be associated with a 2 × 2 or a √3 × √3 R(30°) germanene phase on top of the Au(111) surface (symmetry, in each case, imposing a two-domain structure). Results and discussion This view is supported by the synchrotron radiation core-level spectroscopy measurements displayed in ﬁgure 2 (for details on the data acquisition and ﬁtting procedure and parameters, see the supplementary material). g p p pp y The clean Au(111) 4f core-levels are ﬁtted with bulk (B) and surface-shifted (S) components of similar intensities in the highly surface-sensitive conditions of the measurements (typically, the kinetic energy of the measured Au 4f 7/2 line at ∼46 eV corresponds to the minimum of the escape length of around 0.5 nm). After germanium deposition at ∼200 °C and the growth of the two-dimensional phases, as displayed in ﬁgure 1, the total Au 4f intensity at normal emission is reduced by ∼32%; still, a fraction of the Au(111) surface (about 25%) remains uncovered since the surface component is not totally quenched. On the high binding energy side a new component, denoted I and representing ∼15% of the total Au 4f intensity, has developed. The relative intensity ratio I/B at normal emission (0.21) increases to 0.29 at 50° off- normal emission, while the S/B ratio barely changes. This indicates that the gold atoms contributing to this I component are at the very top surface. The corresponding Ge 3d core- levels are ﬁtted with a very narrow, asymmetric single component at 50° off-normal emission, 4 M E Dávila et al New J. Phys. 16 (2014) 095002 Figure 2. Synchrotron radiation Au 4 f (a) and Ge 3d (b) core-level spectroscopy measurements at normal (NE) and 50° off-normal emission, taken at hν = 135 eV for the 2D phases of Ge grown on Au(111) at ∼200 °C; B, S and I are bulk, surface and interface components, respectively. Figure 2. Synchrotron radiation Au 4 f (a) and Ge 3d (b) core-level spectroscopy measurements at normal (NE) and 50° off-normal emission, taken at hν = 135 eV for the 2D phases of Ge grown on Au(111) at ∼200 °C; B, S and I are bulk, surface and interface components, respectively. in extremely surface sensitive conditions (escape depth estimated at ∼0.4 nm), essentially signaling a unique environment of the germanium atoms at the very top surface and their metallic character; we assign the small broad additional component (just 9% of the total intensity) at normal emission to defect sites. Results and discussion The essentially unique Ge species indicate that no formation of a surface alloy occurs, at variance with the Ag2Ge one formed in the case of Ge deposited on Ag(111) surfaces [9, 11] or the one initially formed upon Ge deposition onto the Au(110) surface [27]. In this respect, we note that the formation of a surface alloy is surface dependent; typically, it takes place upon Ge deposition onto the Ag(111) surface, but not onto the Ag(110) and Ag(100) ones. Based on these core-level results and the honeycomb appearance of the STM images of ﬁgure 1, we can assume that the germanium 2D overlayer grown on top of the Au(111) surface is composed of germanene sheets arranged in either a √19× √19 R(23.4°) supercell (with reference to Au(111) 1 × 1), a 5 × 5 one or a √7 × √7 R(19.1°) one. As mentioned above, this last supercell could correspond either to a 2 × 2 germanene reconstructed epitaxial sheet (projected in-plane Ge- Ge distance: dGe-Ge= 0.221 nm) or to a √3 × √3 R(30°) one (dGe-Ge= 0.255 nm), since the corresponding value for freestanding germanene is dGe-Ge = 0.238 nm [6], while the √19 × √19 R (23.4°) and 5 × 5 ones could correspond respectively to a 3 × 3 (projected in-plane Ge-Ge distance: dGe-Ge = 0.242 nm) and a √13× √13 R(13.9°) one (dGe-Ge = 0.231 nm). √ √ √ √ In the following we address the question of the epitaxial structures for the √7 × √7 R (19.1°) supercell, since this is the one observed in STM imaging as a honeycomb arrangement. 5 New J. Phys. 16 (2014) 095002 New J. Phys. 16 (2014) 095002 M E Dávila et al To this end we have undertaken to determine the minimum energy conﬁguration within this supercell through density functional theory calculations, even searching for a surface alloy (although very unlikely from the core-level measurements) and also allowing for possible substitution of few Au atoms within the germanene sheet; for details of the calculations, see the supplementary material. The lattice mismatch between the cell sizes of 2 × 2 freestanding germanene and those of the √7 × √7 Au (111) surface appears to be small (8.12 Å versus 7.78 Å). However, the 4.2% compression, along with the strong Ge–Au interaction, distorts the germanene lattices and induces considerable buckling in the structure. Results and discussion The atomic structures and simulated STM images of the two lowest energy structures for 2 × 2 germanene on top of the √7 × √7 Au supercell are shown in ﬁgures 3(a) and (b). The average height variations in structures 1 and 2 are 0.150 nm and 0.142 nm, respectively, much larger than what is measured in experiments. Therefore, these two structures are not the structures observed in experiments. √ √ On the other hand, a very ﬂat structure is obtained when √3 × √3 germanene is placed on top of the √7 × √7 Au surface. It corresponds to a coverage ratio θGe = 6/7 = 0.857, typically lower than θSi = 18/16 = 1.125 for the prototypical 3 × 3/4 × 4 silicene on Ag(111), which can be related to the larger size of Ge atoms compared to Si ones, since gold and silver have nearly the same lattice parameter, and also probably points to different interactions with the substrates. The lowest energy structure in this conﬁguration is shown in ﬁgure 3(c). The height variations of different Ge atoms are less than 0.05 nm. Compared with the previous two structures, this one has a lower absorption energy, which is deﬁned as: = − − ( ) E E E N E , (1) abs Ge Au Au Ge Ge / (111) (111) (1) where EGe/Au(111), EAu(111), and EGe represent the total energies of the germanene-covered Au slab, the pure Au slab and the isolated Ge atom, respectively, and NGe represents the number of Ge atoms in the supercell. As seen in table 1, the absorption energy of this structure is even lower than the bulk cohesive energy of diamond Ge, indicating that it is more energetically favorable to form such a layer structure than to form Ge clusters on the Au(111) surface. This is in contrast to the Ge overlayer formed on top of graphene [28]. In that case, the absorption or binding energy of the Ge overlayer is higher than the cohesive energy of bulk Ge, so, as the size of the Ge cluster becomes larger, 3D growth becomes more energetically favorable. The simulated STM image for structure 3 is shown in ﬁgure 3(c). As highlighted with blue circles, there is a darker region for each supercell. Results and discussion With the consideration of the tip effects, an image similar to those observed in experiments may be obtained by Gaussian smearing with width σ = 0.6 Å, as shown in ﬁgure 3(d). where EGe/Au(111), EAu(111), and EGe represent the total energies of the germanene-covered Au slab, the pure Au slab and the isolated Ge atom, respectively, and NGe represents the number of Ge atoms in the supercell. As seen in table 1, the absorption energy of this structure is even lower than the bulk cohesive energy of diamond Ge, indicating that it is more energetically favorable to form such a layer structure than to form Ge clusters on the Au(111) surface. This is in contrast to the Ge overlayer formed on top of graphene [28]. In that case, the absorption or binding energy of the Ge overlayer is higher than the cohesive energy of bulk Ge, so, as the size of the Ge cluster becomes larger, 3D growth becomes more energetically favorable. The simulated STM image for structure 3 is shown in ﬁgure 3(c). As highlighted with blue circles, there is a darker region for each supercell. With the consideration of the tip effects, an image similar to those observed in experiments may be obtained by Gaussian smearing with width σ = 0.6 Å, as shown in ﬁgure 3(d). The Au 4f surface core level shift calculated for a pure seven-layer Au slab is −0.35 eV, which agrees well with the experimental value. The Ge 3d core level shift between different Ge atoms in structure 3 has three components, as shown in table 1, with the energy of the main component set to 0.00 eV. The intensity ratio between these three components is 1:4:1. The difference between each small component and the main component is less than 0.09 eV, which is also consistent with our experimental results. The calculated Au 4f core level shift for the Au atoms below Ge atoms in structure 3 has two components: one is shifted by 0.37 eV and the other is shifted by 0.15 eV, with relative intensity ratio 4:3. These two components, the signals of which are attenuated by the germanene layer above, can be related to the component I, the interface component, located at the left of the bulk peak in experiments. 6 New J. Phys. 16 (2014) 095002 M E Dávila et al Figure 3. Results and discussion (a)–(c) Atomic structures (side and top views) and simulated STM images of three different models of germanene on the √7 × √7 Au(111) surface. Structures 1 and 2 have 2 × 2 periodicity, while structure 3 has √3 × √3 periodicity with respect to germanene. The protruding Ge atoms are highlighted in dark red. The supercells in the STM images are highlighted with yellow lines. (d) Comparison between the experimental image (left panel, close-up from ﬁgure 1(a)) and the simulated STM image for structure 3 after smearing (right panel). Figure 3. (a)–(c) Atomic structures (side and top views) and simulated STM images of three different models of germanene on the √7 × √7 Au(111) surface. Structures 1 and 2 have 2 × 2 periodicity, while structure 3 has √3 × √3 periodicity with respect to germanene. The protruding Ge atoms are highlighted in dark red. The supercells in the STM images are highlighted with yellow lines. (d) Comparison between the experimental image (left panel, close-up from ﬁgure 1(a)) and the simulated STM image for structure 3 after smearing (right panel). Although our calculations on structure 3 agree very well with our experiments we would also like to explore the possibility of forming a Ge–Au surface alloy on the Au(111) surface. For example, switching the position of one Ge atom and one Au atom in structure 3, the total energy of the system can be even lower. The lowest energy structure from switching one pair of atoms is shown in ﬁgure 4(a). The total energy of this structure (‘structure 4’) in the √7 × √7 Au supercell is lower than that of structure 3 by 0.43 eV. A larger scale calculation in a 2√7 × 2√7 Au supercell shows that the total energy decreases almost linearly with the increasing number of switching pairs, from one pair up to three pairs. The surface of such a structure is also very ﬂat, with height variations of less than 0.3 Å. The simulated STM image is shown in the bottom panel of ﬁgure 4(a). Calculations indicate that the Ge 3d core level spectra 7 New J. Phys. 16 (2014) 095002 M E Dávila et al 4 5 6 Figure 4. Atomic structure (side and top views) and simulated STM images of three different germanene structures having substitutional Au atoms. New J. Phys. 16 (2014) 095002 New J. Phys. 16 (2014) 095002 M E Dávila et al When there are only ﬁve Ge atoms on the Au(111) surface within the √7 × √7 Au supercell, the relaxation of all the surface layers will result in a structure similar to that shown in ﬁgure 4(b), i.e., one Au atom is pulled out of the Au surface and forms a honeycomb lattice with the other Ge atoms, leaving a vacancy in the subsurface (the locations of the vacancy are highlighted with red circles). Following the deﬁnition in equation (1), the absorption energy per Ge of this structure is even lower than that of structure 3, by 0.049 eV. However, with a vacancy in the subsurface, this structure is not stable. Other Au atoms will diffuse from the bare surface or from the bulk to ﬁll the vacancy to form a more stable structure. The new structure consists of a honeycomb germanene lattice with some Ge atoms substituted by Au atoms and the original Au surface. The lowest energy structures with one and two Ge atoms substituted by Au atoms are shown in ﬁgures 4(b) and (c), with the vacancies in 4(b) ﬁlled with Au atoms. However, it is not possible to compare the absorption energy of Ge atoms in structure 3 with that in structure 5, since equation (1) does not apply to the latter. Instead we introduce a new deﬁnition for absorption energy of Ge atoms in systems having Au atom substitutions like structure 5: = − − − + ( ) E E E N E N E / , (2) abs Ge Au Au Au Au bulk Au Ge Ge / (111) (111) (2) where NAu is the number of substituting Au atoms and Ebulk Au is the cohesive energy per atom of bulk gold. After introducing Au atoms in the honeycomb lattice, the Ge–Au layer is still very ﬂat, with height variations of less than 0.5 Å. The simulated STM images are shown in the lower panels of ﬁgures 4(b) and (c). For structure 5, there is a darker region in every supercell. For structure 6, besides a darker region, there is a brighter spot due to the Au atom in each supercell. With proper smearing, images similar to those observed in experiments may be acquired. New J. Phys. 16 (2014) 095002 To conclude regarding these theoretical results, our density functional theory calculations show that the atomic model composed of a √3 × √3 reconstructed germanene sheet on top of a √7 × √7 Au(111) surface matches the STM observations and the core-level measurements very well, although the germanene layer may possess some gold atom substitutions. Conclusions To summarize, a two-dimensional germanium layer, forming several phases, has been grown in situ by dry deposition of germanium onto the Au(111) surface, similarly to the formation of silicene on Ag(111). One of these phases displays a clear honeycomb structure with a very weak corrugation in STM imaging. Detailed core-level spectroscopy measurements along with advanced density functional theory calculations allow us to identify this phase as a √3 × √3 reconstructed germanene layer on top of a √7 × √7 Au(111) surface. Through this we provide compelling evidence for the growth of nearly ﬂat germanene, a synthetic graphene-like germanium allotrope that does not exist in nature. Results and discussion The Au atoms in the surface layer are highlighted in orange and the Ge atoms in the subsurface layer are highlighted in light blue. 4 5 6 Figure 4. Atomic structure (side and top views) and simulated STM images of three different germanene structures having substitutional Au atoms. The Au atoms in the surface layer are highlighted in orange and the Ge atoms in the subsurface layer are highlighted in light blue. should have three components (see table 1). In particular, the two surface components are separated by about 0.17 eV, and the component corresponding to the subsurface Ge atom shifts to 0.34 eV higher than the majority of surface components. The disagreement between the calculation and the experimental results for the Ge 3d core-level indicates that structure 4 is not the one observed in experiments. Table 1. Absorption energies and core level shifts for different germanene structures on the Au (111) surface. The cohesive energy of bulk germanium in its cubic diamond-type structure is calculated to be −3.727 eV/atom. The main components of the Ge 3d core levels are set to 0.00 eV. Energy per Ge atom (eV/atom) Buckling (Å) Ge 3d core level shifts (eV) Au 4 f core level shifts (eV) Structure 1 −3.641 1.50 Structure 2 −3.648 1.42 Structure 3 −3.744 0.47 −0.09, 0.00, 0.06 0.15, 0.37 Structure 4 −3.815 0.28 0.00, 0.17, 0.34 0.24, 0.36, 0.43 Structure 5 −3.847 0.23 −0.14, 0.00, 0.11 −0.11, 0.27, 0.33 Structure 6 −3.868 0.41 ⩽0.03 −0.12, 0.17, 0.22,0.29 Table 1. Absorption energies and core level shifts for different germanene structures on the Au (111) surface. The cohesive energy of bulk germanium in its cubic diamond-type structure is calculated to be −3.727 eV/atom. The main components of the Ge 3d core levels are set to 0.00 eV. should have three components (see table 1). In particular, the two surface components are separated by about 0.17 eV, and the component corresponding to the subsurface Ge atom shifts to 0.34 eV higher than the majority of surface components. The disagreement between the calculation and the experimental results for the Ge 3d core-level indicates that structure 4 is not the one observed in experiments. 8 Acknowledgments Drs F Cheynis, D Chiappe, A Ranguis and A Resta are warmly acknowledged for their contributions to the measurements. We are grateful to Dr M Radović and Dr R S Dhaka at Swiss Light Source and I Vobornik at the Italian synchrotron radiation source Elettra, whose 9 New J. Phys. 16 (2014) 095002 M E Dávila et al outstanding efforts made these experiments possible. Funding from the 2D-NANOLATTICES project within the 7th Framework Programme for Research of the European Commission, under FET-Open grant number 270749 is greatly appreciated. We acknowledge support from the European Research Council Advanced Grant DYNamo (ERC-2010-AdG-267374), Grupos Consolidados UPV/EHU del Gobierno Vasco (IT-578-13) and European Commission project CRONOS (grant number 280879-2). References [1] Brumﬁel G 2013 Nature 495 153 [2] Vogt P, De Padova P, Quaresima C, Avila J, Frantzeskakis E, Asensio M C, Resta A, Ealet B and Le Lay G 2012 Phys. Rev. Lett. 108 15550 [3] Lin C L, Arafune R, Kawahara K, Tsukahara N, Minamitami E, Kim Y, Takagi N and Kawai M 2012 Appl. Phys. Express 5 045802 [4] Fleurence A, Friedlein R, Osaki T, Kawai H, Wang Y and Yamada-Takamura Y 2012 Phys. Rev. Lett. 108 245501 [5] Meng L et al 2013 Nanoletters 13 685 [6] Cahangirov S, Topsakal M, Akturk E, Sahin H and Ciraci S 2009 Phys. Rev. Lett. 102 236804 [7] Bianco E, Butler S, Jiang S, Restrepo O D, Windl W and Goldberger J E 2013 ACS Nano 7 4414 [8] Jiang S, Butler S, Bianco E, Restrepo O D, Windl W and Goldberger J E 2014 Nature Comm. 5 3383 [9] W W S h il H M O i ki J R d Uh b R I G 2014 Ph R B 89 125410 [6] Cahangirov S, Topsakal M, Akturk E, Sahin H and Ciraci S 2009 Phys. Rev. Lett. 102 236804 [7] Bianco E, Butler S, Jiang S, Restrepo O D, Windl W and Goldberger J E 2013 ACS Nano 7 4414 [8] Ji S B l S Bi E R O D Wi dl W d G ldb J E 2014 N C 5 3383 [7] Bianco E, Butler S, Jiang S, Restrepo O D, Windl W and Goldberger J E 2013 ACS Nano 7 4414 [8] Jiang S, Butler S, Bianco E, Restrepo O D, Windl W and Goldberger J E 2014 Nature [9] Wang W, Sohail H M, Osiecki J R and Uhrberg R I G 2014 Phys. Rev. B 89 125410 [10] Golias E, Xenogiannopoulou E, Tsoutsou D, Tsipas P, Giamini S A and Dimoulas A 2013 Phys. Rev. B 88 075403 Rev. B 50 4470 [16] Hajjar S et al 2011 Phys. Rev. B 84 125325 Li L, Lu S-Z, Pan J, Qin Z, Wang Y-Q, Wang Y, Cao G-Y, Du S and Gao H-J 2014 Adv. Mater. 26 4820 [ ] , , , Q , g Q, g , , v [18] Shao Z-G, Ye X-S, Yang L and Wang C-L 2013 J. Appl. Phys. 114 093712 [18] Shao Z-G, Ye X-S, Yang L and Wang C-L 2013 J. Appl. Phys. 114 093712 [18] Shao Z-G, Ye X-S, Yang L and Wang C-L 2013 J. Appl. Phys. 114 093712 [19] Matthes L, Pulci O and Bechstedt F 2013 J. Phys.: Condens. Matter 25 395305 [19] Matthes L, Pulci O and Bechstedt F 2013 J. Phys.: Condens. Matter 25 395305 [20] Liu C-C, Feng W and Yao Y 2011 Phys. Rev. Lett. 107 076802 20] Liu C-C, Feng W and Yao Y 2011 Phys. Rev. L [20] Liu C-C, Feng W and Yao Y 2011 Phys. Rev. Lett. 107 076802 [21] Ezawa M 2012 Phys. Rev. Lett. 109 055502 [22] Liu C-C, Jiang H and Yao Y 2011 Phys. Rev. B 84 195430 [23] Chen L, Feng B and Wu K 2013 Appl. Phys. Lett. 102 081602 [24] Baskaran G 2013 arXiv:1309.2242 [25] Hanke F and Björk J 2013 Phys. Rev. B 87 235422 [26] Arafune R, Lin C-L, Kawahara K, Tsukahara N I, Minamitani E, Kim Y, Takagi N and Kawai M 2013 Surface Sci. 608 297 075403 [11] Oughaddou H et al 2000 Phys. Rev. B 62 16653 [12] Le Lay G 1983 Surface Sci. 132 169 [13] Le Lay G, Manneville M and Métois J J 1982 Surface Sci. 123 117 [13] Le Lay G, Manneville M and Métois J J 1982 Surface Sci. 123 117 [14] Katayama M, Williams R S, Kato M, Nomura E and Aono M 1991 Phys. Rev. Lett. 66 2762 [15] Göthelid M, Hammar M, Björkqvist M, Karlsson U O, Flodström S A, Wigren C and LeLay G 1994 Phys. Rev. B 50 4470 Surface Sci. 608 297 Chiappe D, Grazianetti C, Tallarida G, Fanciulli M and Molle A 2012 Adv. Mater. 24 5088 [27] Tchalala M R et al 2013 Appl. Phys. Lett. 102 083107 [28] Cai Y-, Chuu C-P, Wei C M and Chou M Y 2013 Phys. Rev. B 88 245408 10
https://openalex.org/W1996808507	http://www.scirp.org/journal/PaperDownload.aspx?paperID=47908	English	null	The Misleading Use of “Enthalpy” in an Energy Conversion Analysis	Natural science	2,014	cc-by	4,964	Heinz Herwig Institute for Thermo-Fluid Dynamics, Hamburg University of Technology, Hamburg, Germany Email: h.herwig@tuhh.de Received 26 May 2014; revised 30 June 2014; accepted 9 July 2014 Copyright © 2014 by author and Scientific Research Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC http://creativecommons.org/licenses/by/4.0/ py g y g This work is licensed under the Creative Commons Attribution International License (CC BY). htt // ti /li /b /4 0/ Keywords Enthalpy Misinterpretation, Energy Conversion Analysis Enthalpy Misinterpretation, Energy Conversion Analysis Natural Science, 2014, 6, 878-885 Published Online July 2014 in SciRes. http://www.scirp.org/journal/ns http://dx.doi.org/10.4236/ns.2014.611085 Natural Science, 2014, 6, 878-885 Published Online July 2014 in SciRes. http://www.scirp.org/journal/ns http://dx.doi.org/10.4236/ns.2014.611085 Natural Science, 2014, 6, 878-885 Published Online July 2014 in SciRes. http://www.scirp.org/journal/ns http://dx.doi.org/10.4236/ns.2014.611085 How to cite this paper: Herwig, H. (2014) The Misleading Use of “Enthalpy” in an Energy Conversion Analysis. Natural Sci- ence, 6, 878-885. http://dx.doi.org/10.4236/ns.2014.611085 Abstract The frequently used thermodynamic state quantity enthalpy H turns out to be very problematic when applied in an energy conversion analysis. Due to the fact that H combines two terms, the in- ternal energy U and the product pV, the interpretation of what H means in physical terms is often obscure and leads to various misinterpretations. Several examples are given and interpreted twice, i.e. with and without referring to enthalpy. Keywords 1. Introduction In basically all thermodynamic text books the enthalpy H is introduced and used in an energy conversion analy- sis, predominantly in open thermodynamic systems. For example, Moran and Shapiro ([1] page 81, 128) intro- duce H in the following way: 𝑈𝑝 In many thermodynamic analyses the sum of the internal energy 𝑈𝑈 and the product of pressure 𝑝𝑝 and vol- ume V appears. Because the sum U pV + occurs so frequently in subsequent discussions, it is convenient to give the combination a name, enthalpy, and a distinct symbol, H. By definition . H U pV = + Since , , U p and V are all properties, this combination is also a property… The appearance of the sum u pv + in the control volume energy equation is the principal reason for introducing enthalpy previously. It is brought in solely as a convenience. While Cengel ([2] page 52) under the headline “Enthalpy—A combination Property” writes: In the analysis of certain types of processes, particularly in power generation and refrigeration (...), we fre- quently encounter the combination of properties U pV + . For the sake of simplicity and convenience, this combination is defined as a new property, enthalpy, and given the symbol H. How to cite this paper: Herwig, H. (2014) The Misleading Use of “Enthalpy” in an Energy Conversion Analysis. Natural Sci- ence, 6, 878-885. http://dx.doi.org/10.4236/ns.2014.611085 Once a new physical quantity is introduced, one seeks for its physical interpretation. With the enthalpy often it is just the “combination property” explanation as in the two examples given above. Sometimes, however, H is named “heat content” or “total heat” (see [2] for details) without giving a really convincing justification for these names as a general description of what H stands for. 2. Enthalpy: Motivation, Definition, and Interpretation In the examples given in the introduction the control volume energy equation was mentioned as the principal reason for introducing enthalpy. Before this is explained in more details some general considerations about con- trol volume balance equations and the quantities appearing in these equations are necessary. 2.1. Control Volume Balance Equations With n state quantities and m process quantities involved, a general control volume balance equation for a steady thermodynamic system between two cross Sections 1 and 2, for example, reads ( ) 2 1 12 change of state quantities action of process quantities n n m n m a a b − = ∑ ∑    (1) ( ) 2 1 12 change of state quantities action of process quantities n n m n m a a b − = ∑ ∑    (1) (1) change of state quantities action of process quantities   Only when balance equations are developed according to this general form a clear physical interpretation can be given. 2.1. Control Volume Balance Equations The state of a thermodynamic system is characterized by several physical quantities like mass, energy and en- tropy which can be subject to a balance with respect to a control volume. In this kind of balance the change of a certain quantity in the system is exclusively due to a transport of this quantity across the control volume boun- daries when the quantity under consideration is a conserved quantity (like the thermodynamic total energy). For non-conserved quantities (like entropy) changes may also be due to creation and/or destruction of the quantity. q py g y In these equations two types of quantities appear which are very different in nature: • state quantities, characterizing a (instantaneous and local) state of the system at a certain time and location within the system. Let i and j be different times or locations then a general state quantity here will be named ia or ja and changes are denoted by ( ). j i a a − • state quantities, characterizing a (instantaneous and local) state of the system at a certain time and location within the system. Let i and j be different times or locations then a general state quantity here will be named ia or ja and changes are denoted by ( ). j i a a − • process quantities, characterizing a process by which the state quantities of the system are changed between their states i and j. A general process quantity here will be named ijb with the double index ij indicating that the process occurs between i and j. • process quantities, characterizing a process by which the state quantities of the system are changed between their states i and j. A general process quantity here will be named ijb with the double index ij indicating that the process occurs between i and j. p j Mathematically, both kinds of thermodynamic quantities are very different: state quantities have a total dif- ferential (and thus are path independent), process quantities do not have it. 2.3. The Enthalpy-Form of the Energy Balance In the examples given below it will be shown, however, that combining state and process quantities by using h gives rise to various misinterpretations. 2.2. The Energy Balance (1. Law of Thermodynamics) In an open system there are two work terms which are 12S W : shaft work, transferred by the shaft of a turbine or a pump 12F W : flow work, transferred at the inlet and outlet by the action of pressure forces i i p A moving their point of action with the velocity ic , so that with i i i i Ac V mv = =   • energy transfer across the control volume boundaries between 1 and 2 in form of work, i.e. due to local forces with moving points of action. In an open system there are two work terms which are 12S W : shaft work, transferred by the shaft of a turbine or a pump 12F W : flow work, transferred at the inlet and outlet by the action of pressure forces i i p A moving their point of action with the velocity ic , so that with i i i i Ac V mv = =   S 12F W : flow work, transferred at the inlet and outlet by the action of pressure forces i i p A moving the f action with the velocity ic , so that with i i i i Ac V mv = =   ( ) 12 2 2 1 1 F W m p v p v = −   (4) (4) The quantities involved in the energy conversion in an open system between 1 and 2 are sketched in Figure 2. 2.3. The Enthalpy-Form of the Energy Balance 2.3. The Enthalpy-Form of the Energy Balance When Equation (4) is formally introduced in Equation (2) the energy balance can be written as ( ) ( ) 2 1 2 2 2 1 2 2 2 2 1 1 1 1 12 12 2 2 S h h c c m p v u gz p v u gz Q W −     + + + − − + + = −         (5) (5) Here the specific enthalpy h appears by combining the state quantity u and the process quantity pv. Note that now 12S W is the only work term left since 12F W “disappeared” in h. This form of the energy equation is fre- quently used for open systems, rewritten as Here the specific enthalpy h appears by combining the state quantity u and the process quantity pv. Note that now 12S W is the only work term left since 12F W “disappeared” in h. This form of the energy equation is fre- quently used for open systems, rewritten as ( ) ( ) ( ) 2 2 2 1 2 1 2 1 12 12 2 S c c h h g z z q w − − + + − = − (6) (6) with the specific heat transfer 12 12 q Q m =   and the specific shaft work 12 12 S S w W m =  . In this equation ( ) 2 1 h h − is often interpreted as an “energy change” of the fluid within an energy conversion process. For example in Warhaft ([3] page 92) it reads with respect to pumps and turbines: They are both devices that change the energy of the fluid by means of work interactions. with the specific heat transfer 12 12 q Q m =   and the specific shaft work 12 12 S S w W m =  . In this equation ( ) 2 1 h h − is often interpreted as an “energy change” of the fluid within an energy conversion process. For example in Warhaft ([3] page 92) it reads with respect to pumps and turbines: They are both devices that change the energy of the fluid by means of work interactions. 2.2. The Energy Balance (1. Law of Thermodynamics) The well-known “1. Law of Thermodynamics” applied to a finite open control volume exemplified for a steady energy conversion is illustrated in Figure 1. The state quantity is the total thermodynamic energy, here intro- duced as specific energy e which is either E m or E m . It has three parts which are the internal energy ue u = , the kinetic energy 2 2 ke c = and the gravitational potential energy pe gz = . Process quantities are the work done (W) and the heat transferred ( Q) with respect to the control volume. According to its general form (1) the energy balance equation reads with 1 for the inlet and 2 for the exit cross section ( ) ( ) 2 2 2 1 1 1 12 12 action of heat &work on change of u k p u k p e e m e e e e e e Q W   + + − + + = −         (2) (2) The process quantities on the r.h.s. of (2) are The process quantities on the r.h.s. of (2) are • energy transfer across the control volume boundaries between 1 and 2 in form of heat, i.e. due to local tem- perature gradients, which lead to a local wall heat flux w q. Then 12 Q with A as transfer area is • energy transfer across the control volume boundaries between 1 and 2 in form of heat, i.e. due to local tem- perature gradients, which lead to a local wall heat flux w q. Then 12 Q with A as transfer area is 12 d w A Q q A = ∫   (3) 12 d w A Q q A = ∫   (3) 12 d w A Q q A = ∫   (3) H. Herwig Figure 1. State and process quantities involved in a control volume balance for an open system. Figure 1. State and process quantities involved in a control volume balance for an open system. • energy transfer across the control volume boundaries between 1 and 2 in form of work, i.e. due to local forces with moving points of action. 2.4. The Non-Enthalpy-Form of the Energy Balance The “1. Law of Thermodynamics” balances the total thermodynamic energy which is the sum of a mechanical and a thermal part. Without introducing the enthalpy h it reads in a form corresponding to that of Equation (6) above ( ) ( ) ( ) ( ) 2 2 2 1 2 1 2 1 12 12 12 2 S F c c u u g z z q w w − − + + − = − + (7) (7) H. Herwig Figure 2. Control volume balance for a steady energy conversion in an open system. Figure 2. Control volume balance for a steady energy conversion in an open system. Note that now both work terms ( 12S w and 12F w ) appear on the r.h.s. of Equation (7). For further analyzing energy conversion systems one can deduce the partial energy equations for the me- chanical and the thermal part of the total thermodynamic energy. This is done by multiplying the momentum equation with the velocity in order to get the mechanical partial energy equation. This equation in fluid mechan- ics is known as Bernoulli equation, see for example Herwig [4]. The thermal partial energy equation then ap- pears when the mechanical equation is subtracted from Equation (7)—guaranteeing that the total thermodynamic energy is the sum of its two parts (mechanical and thermal). In terms of Equation (7) the two partial equations are: ( ) mechanical: ( ) ( ) ( ) 12 12 12 1 2 2 1 2 2 2 ϕ − + − = − + − F S w w z z g c c (8) thermal: ( ) 12 12 1 2 ϕ + = − q u u (9) (8) thermal: ( ) 12 12 1 2 ϕ + = − q u u (9) Now the specific dissipation m ϕ = Φ  appears which is not explicitly present in Equation (7). The physical background is the following: Now the specific dissipation m ϕ = Φ  appears which is not explicitly present in Equation (7). The physical background is the following: Dissipation of mechanical energy converts mechanical into internal energy. This energy devaluation process happens, when flow work is not converted into shaft work but increases the internal energy of the fluid. Ther- modynamically this corresponds to an entropy generation in this irreversible process. 2.4. The Non-Enthalpy-Form of the Energy Balance Since, however, flow work ( ) 12F w and internal energy increase ( ) 2 1 u u − both appear in the total energy equation, the dissipation term 12 ϕ does not explicitly show up in the total energy Equation (6) or (7). Only in Equation (8) the specific dissi- pation ϕ appears in a way which is appropriate for the determination of this quantity.  tioned before the specific dissipation is accompanied by an entropy generation irr S according to As mentioned before the specific dissipation is accompanied by an entropy generation irr S acco irr irr TS m Ts ϕ = =   (10) (10) with irr s as specific entropy generation and T as thermodynamic temperature. Knowing ϕ or irr s is essen- tial when losses in a flow field should be determined. For a detailed discussion of this aspect of an energy con- version see Herwig and Schmandt [5]. with irr s as specific entropy generation and T as thermodynamic temperature. Knowing ϕ or irr s is essen- tial when losses in a flow field should be determined. For a detailed discussion of this aspect of an energy con- version see Herwig and Schmandt [5]. As far as the enthalpy is concerned, h does not only combine a state and a process quantity, but also one that appears in the thermal energy equation (u in Equation (9)) and one that is part of the mechanical energy equation (pv as part of 12F w in Equation (8)). That is why the enthalpy-form (6) of the energy balance cannot be further detailed by determining the mechanical and the thermal parts of it separately. Then, however, the dissipation term ϕ cannot be determined within the energy considerations based on the enthalpy concept. 3. Energy Conversion Examples Since the introduction of h is motivated by energy conversion considerations four such examples will be analysed based on Equation (6) using h as the crucial “energy quantity” and compared to an analysis avoiding the enthalpy, based on Equations (7)-(9). The interpretation of the results will be named “enthalpy interpretation” and “non- enthalpy interpretation”, respectively. 3.1. Energy Conversion in a Turbine (Dissipation Neglected) • potential energy because 1 2 z z = From Equation (7) we get From Equation (7) we get ( ) 12 12 0 S F w w = − + (13) ( ) 12 12 0 S F w w = − + (13) r: ( ) 12 12 1 2 S F w w v p p = − = − (14) (13) ( ) or: ( ) 12 12 1 2 S F w w v p p = − = − (14) (14) The non-enthalpy interpretation is: The non-enthalpy interpretation is: The non-enthalpy interpretation is: Fl k i t d i t h ft k f th t bi ith t h i th f th fl id The non-enthalpy interpretation is: Flow work is converted into shaft work of the turbine without changing the energy of the fluid. The non-enthalpy interpretation is: Flow work is converted into shaft work of the turbine without changing the energy of the fluid Flow work is converted into shaft work of the turbine without changing the energy of the fluid. 3.2. Energy Conversion in a Turbine (Dissipation Accounted for) 3.1. Energy Conversion in a Turbine (Dissipation Neglected) Figure 3 shows a turbine operated with water which may be part of a pumped storage hydro power station. As- suming incompressible fluid ( const v = ), an adiabatic process ( 12 0 q = ), equal cross sections ( 1 2 A A = ), and equal elevation head ( 1 2 z z = ), Equation (6) reduces to (note: 1 2 c c = ) 881 H. Herwig Figure 3. Control volume around the turbine in a pumped storage hydro power station. Figure 3. Control volume around the turbine in a pumped storage hydro power station. 2 1 12S h h w − = − (11) or: 12 1 2 S w h h = − (12) (11) (12) The enthalpy interpretation is: The enthalpy interpretation is: Energ of the fl id is con erted into shaft ork of the t rbine py interpretation is: he fluid is converted into shaft work of the turbine. Energy of the fluid is converted into shaft work of the turbine. Energy of the fluid is converted into shaft work of the turbine. This, however, contradicts the fact that the energy e of the fluid remains unchanged, because there is no change in This, however, contradicts the fact that the energy e of the fluid remains unchanged, because there is no change in • internal energy, because u for an incompressible fluid is a function of T only and the temperature remains (basically) the same. • internal energy, because u for an incompressible fluid is a function of T only and the temperature remains (basically) the same. • internal energy, because u for an incompressible fluid is a function of T only and the temperature remains (basically) the same. • kinetic energy because 1 2 c c = holds • kinetic energy because 1 2 c c = holds rgy because 1 2 c c = holds ergy because 1 2 z z = holds. • potential energy because 1 2 z z = holds. 3.3. Heating the Fluid Ahead of the Turbine (Dissipation Neglected) The example from Section 3.1 is resumed but now with a heat transfer of strength 01 q ahead of the turbine as sketched in Figure 4. From Equation (5) we get for this case From Equation (5) we get for this case (19) 2 0 01 12S h h q w − = − (19) or: 12 0 2 01 S w h h q = − + (20) 2 0 01 12S h h q w − = − (19) or: 12 0 2 01 S w h h q = − + (20) (20) or: 12 0 2 01 S w h h q = − + The enthalpy interpretation is: From comparing Equation (20) with Equation (12) one might conclude that 01 q increases the shaft work though it is not clear how h behaves under the heat transfer. From Equation (9) we get (with indices adopted) From Equation (9) we get (with indices adopted) 2 0 01 u u q − = (21) (21) 2 0 01 u u q − = so that Equation (7) now becomes: so that Equation (7) now becomes: ( ) 12 02 0 S F w w = − + (22) ( ) 12 02 0 S F w w = − + (22) l ( ) 12 02 0 S F w w = − + (22) (22) or with 02 12 F F w w = since ( ) u u T = only or with 02 12 F F w w = since ( ) u u T = only ( ) 12 12 1 2 S F w w v p p = − = − (23) (23) The non-enthalpy interpretation is: There is no influence of 01 q on the shaft work, the internal energy is increased by 01 q leaving the flow work unchanged. 3.4. Throttling of an Incompressible Fluid 3.2. Energy Conversion in a Turbine (Dissipation Accounted for) In real turbines dissipation occurs, often reducing the available work by the order of 10%. Based on Equation (6) one gets In real turbines dissipation occurs, often reducing the available work by the order of 10%. Based on Equation (6) one gets 2 1 12S h h w − = − (15) (15) 2 1 12S h h w − = − (15) 2 1 12S h h w − = − e same result as before, c.f. Equation (11). which is the same result as before, c.f. Equation (11). which is the same result as before, c.f. Equation (11). py p no recognizable effect of dissipation on the shaft work of a turbine. There is no recognizable effect of dissipation on the shaft work of a turbine. F E ti (9) t There is no recognizable effect of dissipation on the shaft work of a turbine. From Equation (9) we get (16) 2 1 12 u u ϕ − = (16) es: ( ) 12 12 12 S F w w ϕ = − + (17) or: ( ) 12 12 12 1 2 12 S F w w v p p ϕ ϕ = − = − − (18) 2 1 12 u u ϕ − = so that Equation (7) now becomes: ( ) 12 12 12 S F w w ϕ = − + ( ) 12 12 12 S F w w ϕ = − + or: ( ) 12 12 12 1 2 12 S F w w v p p ϕ ϕ = − = − − (18) H. Herwig 3.4. Throttling of an Incompressible Fluid Figure 5 shows a throttle which may be part of a heat pump or refrigeration cycle and in which an incompressi- ble liquid flows through the throttle. Equation (6) with 1 2 c c = , 1 2 z z = , 12 0 q = and 0 12 = S w reduces to (24) 2 1 0 h h − = (24) 2 1 0 h h − = The enthalpy interpretation is: The energy of the fluid obviously remains unchanged, “though” dissipation leads to a reduction of the pres- sure level from 1p to 2p which is the desired effect. 1 2 Equations (8) and (9), however, show that flow work by the irreversible dissipation process (c.f. Equation (10)) increases the internal energy of the fluid, so that (with 1 2 c c = and 1 2 z z = ) also the total thermodynamic energy of the fluid is increased. gy The non-enthalpy interpretation is: gy The non-enthalpy interpretation is: The total thermodynamic energy of the fluid is increased by the action of flow work, which leads to a pressure reduction. The non-enthalpy interpretation is: The non-enthalpy interpretation is: There is a recognizable effect of dissipation on the shaft work of a turbine. Shaft work is reduced by dissipa- tion (which increases the internal energy of the fluid). 3.3. Heating the Fluid Ahead of the Turbine (Dissipation Neglected) 4. Conclusion: The Misleading Aspects of Enthalpy This, however, is suppressed by introducing h, so that ϕ is not accessible and thus cannot be determined. • The explicit appearance of the dissipation term ϕ only occurs when the partial energies are balanced. This, however, is suppressed by introducing h, so that ϕ is not accessible and thus cannot be determined. With all these points in mind the final recommendation is: try to avoid enthalpy in an energy conversion analysis. With all these points in mind the final recommendation is: try to avoid enthalpy in an energy conversion analysis. 4. Conclusion: The Misleading Aspects of Enthalpy 4. Conclusion: The Misleading Aspects of Enthalpy Enthalpy, according to its definition H U pV = + is a combination of state quantities and thus itself is a state quantity. When, however, H or its specific versions h is used in control volume balance equations for an energy con- version analysis only U continues to be a state quantity while pV comes from a process quantity (flow work). Furthermore, the introduction of h in the total energy equation prevents its interpretation as the sum of the partial energy equations for the mechanical and the thermal energy. Then, however, the dissipation term does not ap- pear explicitly (and thus cannot be determined). 883 H. Herwig Figure 4. Heating the fluid ahead of the turbine. Figure 5. Control volume around a throttle. Figure 4. Heating the fluid ahead of the turbine. Figure 4. Heating the fluid ahead of the turbine. Figure 5. Control volume around a throttle. Figure 5. Control volume around a throttle. The examples given in Section 3 show several shortcomings of the enthalpy interpretation. Main aspects in this respect are: The examples given in Section 3 show several shortcomings of the enthalpy interpretation. Main aspects in this respect are: this respect are: • The energy of the fluid and its changes in conversion processes is misinterpreted. Flow work may (as in a throttle) or may not (as in a turbine) change the energy of the fluid. Taking h as an “energy” leads to wrong conclusions. p • The energy of the fluid and its changes in conversion processes is misinterpreted. Flow work may (as in a throttle) or may not (as in a turbine) change the energy of the fluid. Taking h as an “energy” leads to wrong conclusions. • The introduction of h combines one term of the thermal energy equation (u in (9)) and one term of the me- chanical energy equation (pv in (8)). Then, however, the detailed interpretation in terms of mechanical versus thermal energy is no longer possible. • The introduction of h combines one term of the thermal energy equation (u in (9)) and one term of the me- chanical energy equation (pv in (8)). Then, however, the detailed interpretation in terms of mechanical versus thermal energy is no longer possible. gy g p • The explicit appearance of the dissipation term ϕ only occurs when the partial energies are balanced. [5] Herwig, H. and Schmandt, B. (2014) How to Determine Losses in a Flow Field: A Paradigm Shift towards the Second Law Analysis. Entropy, 16, 2959-2989. http://dx.doi.org/10.3390/e16062959 References [1] Moran, M.J. and Shapiro, H.N. (1996) Fundamentals of Engineering Thermodynamics. Wiley and Sons, New York. [3] Warhaft, Z. (1997) An Introduction to Thermal-Fluid Engineering. Cambridge University Press, Cambridge. [5] Herwig, H. and Schmandt, B. (2014) How to Determine Losses in a Flow Field: A Paradigm Shift towards the Second Law Analysis. Entropy, 16, 2959-2989. http://dx.doi.org/10.3390/e16062959 [5] Herwig, H. and Schmandt, B. (2014) How to Determine Losses in a Flow Field: A Paradigm Shift towards the Second Law Analysis. Entropy, 16, 2959-2989. http://dx.doi.org/10.3390/e16062959 H. Herwig Scientific Research Publishing (SCIRP) is one of the largest Open Access journal publishers. It is currently publishing more than 200 open access, online, peer-reviewed journals covering a wide range of academic disciplines. SCIRP serves the worldwide academic communities and contributes to the progress and application of science with its publication. Other selected journals from SCIRP are listed as below. Submit your manuscript to us via either submit@scirp.org or Online Submission Portal. Other selected journals from SCIRP are listed as below. Submit your manuscript to us via eith submit@scirp.org or Online Submission Portal. Other selected journals from SCIRP are listed as below. Submit your manuscript to us via either submit@scirp.org or Online Submission Portal.
https://openalex.org/W2120443491	https://europepmc.org/articles/pmc3479071?pdf=render	English	null	Effect of etching time on morphological, optical, and electronic properties of silicon nanowires	Nanoscale research letters	2,012	cc-by	3,246	NANO EXPRESS Open Access © 2012 Nafie et al.; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background cleaning, silicon samples were immersed into the etching solution containing 0.05 M AgNO3, 40% HF, and H2O2 at room temperature for different etching times; 10, 20, 30, 40, 50, 60, 70, 80, and 90 min. After etching, samples were rinsed with deionized water to remove residual HF and immersed in a H2O-HNO3 (2 and 1 V) solution during several seconds to remove the silver film. Silicon nanowires (SiNWs) have attracted much atten- tion in the recent years due to their importance in the field of electronic devices and photovoltaic [1-4]. Hence, SiNWs could be used as an antireflection coating due to the reduction of optical loss which is an important factor to obtain efficient Si solar cells. However, when SiNWs are used as an antireflection coating, a great care should be taken to avoid degradation of the elec- tronic properties, which in turn can increase the serial resistance of the solar cell. Different methods have been employed to fabricate SiNWs, such as chemical physical deposition [5], laser ablation [6,7], thermal evaporation [8,9], and etching. In this paper, we used the silver- assisted chemical etching technique [10-15]. We fabri- cate SiNWs at different durations, ranging from 10 to 90 min. The morphology of samples was analyzed using a scanning electron microscope (SEM). We performed top and cross-section SEM images of the samples. The cross-section SEM images were used to evaluate the length of SiNWs. We measured the surface reflectivity in the 250 to 1,250 nm spectral range by a UV–vis-NIR spectrophotometer. To study the electronic properties of the formed films, we evaluate the effective diffusion lengths (L) of minority carriers in the SiNWs films. Values of L were carried out from the light-beam- induced-current (LBIC) profiles measured on metal/ SiO2/SiNWs/c-Si/metal diode. * Correspondence: Mongi.Bouaicha@crten.rnrt.tn Laboratoire de Photovoltaique, Centre de Recherches et des Technologies de l’Energie, Technopole de Borj-Cedria, BP 95, Hammam-Lif, Tunis 2050, Tunisia Effect of etching time on morphological, optical, and electronic properties of silicon nanowires Nesma Nafie, Manel Abouda Lachiheb and Mongi Bouaicha* Nafie et al. Nanoscale Research Letters 2012, 7:393 http://www.nanoscalereslett.com/content/7/1/393 Nafie et al. Nanoscale Research Letters 2012, 7:393 http://www.nanoscalereslett.com/content/7/1/393 Open Access Methods Substrates used in this study are P+ silicon wafers, boron-doped and (100) oriented, with thickness of 500 μm and resistivity of 0.01 to 0.02 Ωcm. After Abstract Owing to their interesting electronic, mechanical, optical, and transport properties, silicon nanowires (SiNWs) have attracted much attention, giving opportunities to several potential applications in nanoscale electronic, optoelectronic devices, and silicon solar cells. For photovoltaic application, a superficial film of SiNWs could be used as an efficient antireflection coating. In this work we investigate the morphological, optical, and electronic properties of SiNWs fabricated at different etching times. Characterizations of the formed SiNWs films were performed using a scanning electron microscope, ultraviolet–visible-near-infrared spectroscopy, and light-beam- induced-current technique. The latter technique was used to determine the effective diffusion length in SiNWs films. From these investigations, we deduce that the homogeneity of the SiNWs film plays a key role on the electronic properties. Keywords: Silicon nanowires, Reflectivity, Light-beam-induced-current, Diffusion length Results and discussion In Figure 1 we give the top SEM views of one sample, before (Figure 1a) and after (Figure 1b) removing the sil- ver film. One can see in Figure 1a, the silver dendrites formed during the etching process. In Figure 1b, we give a tilted SEM view of the SiNWs after removing the Ag Page 2 of 6 Nafie et al. Nanoscale Research Letters 2012, 7:393 http://www.nanoscalereslett.com/content/7/1/393 (a) (b) Figure 1 Top SEM views of one sample. (a) Before, and (b) after removing the silver film. (b) (a) igure 1 Top SEM views of one sample. (a) Before, and (b) after removing the silver film. etching by the HF solution as observed in the SEM image corresponding to 60 min. film. In Figure 2, we give cross-sectional SEM images of SiNWs films prepared at 10, 20, 30, 40, 50, 60, 70, 80, and 90 min etching times. As reported in Figure 3, the mean length of the film varies from 21 to 38 μm. From Figure 3, we notice that the etching velocity is similar for 10, 20, and 30 min. However, it is more important for 40 and 50 min, where the maximum length is reached. When the sample is etched at greater durations, the length decreases by 10 μm from its maximum value and seems to be stabilized at values around 28 μm. This was attributed to the fact that when the etching process started, the HF solution etches the silicon substrate lead- ing to an increase of the SiNWs’ length. After reaching the maximum length at 50 min, SiNWs themselves are The total reflectivity in the 250 to 1,250 nm spectral range of formed films has very low values, less than 1% in the UV domain and a maximum of 8% in the visible and near-infrared regions (Figure 4). Generally, radiations with small wavelengths (UV domain) are absorbed at small depth and this absorption depends on the surface morphology. In the case of the non- treated sample, the reflectivity in the UV region is greater than 50% as shown in the inset of Figure 4. However, values of the reflectivity of SiNWs films in the UV domain are in the 0.5% to 1.5% range, which is unusual in silicon, even with texturized morphology Figure 2 Cross-section SEM images of formed SiNWs at different durations. Figure 2 Cross-section SEM images of formed SiNWs at different durations. Results and discussion Figure 2 Cross-section SEM images of formed SiNWs at different durations. Page 3 of 6 Nafie et al. Nanoscale Research Letters 2012, 7:393 http://www.nanoscalereslett.com/content/7/1/393 Nafie et al. Nanoscale Research Letters 2012, 7:393 http://www.nanoscalereslett.com/content/7/1/393 0 10 20 30 40 50 60 70 80 90 100 20 22 24 26 28 30 32 34 36 38 40 Length of SiNWs (µm) Etching time (min) Figure 3 Variation of the length of SiNWs films vs. etching time. For the electronic characterization, we use the LBIC technique at the He-Ne wavelength. A schematic illus- tration of the LBIC technique is given in Figure 5. LBIC measurements were performed on metal-insulator- semiconductor structures formed on SiNWs films. Typ- ical LBIC profiles are given in Figure 6. We notice that we obtained the same shape of LBIC profiles for samples etched during 30, 50, 60, 70, 80, and 90 min. However, for samples etched during 10, 20, and 40 min, the LBIC profiles have approximately the same shape of that one performed on the MIS diode without SiNWs (0 min). Using the LBIC measurements (ILBIC), we determine the effective values of L. To carry out L values, we fit the LBIC theoretical expression given in Equation 1 [16] to the measured LBIC profiles. Figure 3 Variation of the length of SiNWs films vs. etching time. ILBIC / expð x LÞ x 3 2 ð1Þ ð1Þ and/or porous silicon. This was attributed to the im- portant internal surface area of SiNWs. From curves in Figure 4, we remark that in the fully used spectral range, the SiNWs film elaborated at 50 min has the minimum value of the total reflectivity. The small values of reflectivity are attributed to the multiple reflections of incident photons which are important when the length of SiNWs film is important (38 μm during 50 min). Obtained values of the effective diffusion length are plot in Figure 7. We remark that the obtained L values can be divided into two domains: red and green regions in Figure 7. The red region corresponds to the L values obtained for samples prepared during 0, 10, 20, and 40 min. The green region corresponds to samples pre- pared at 30, 50, 60, 70, 80, and 90 min. Results and discussion To understand why L values change from a sample to another, we use the cross-section SEM images of Figure 2 300 400 500 600 700 800 900 1000 1100 1200 0 1 2 3 4 5 6 7 8 9 400 600 800 1000 1200 40 60 80 100 Reflectivity (%) Wavelength (%) 0 min Total reflectivity (%) Wavelength (nm) 90 min 10 min 80 min 60 min 40 min 50 min 70 min 30 min 20 min Figure 4 Total reflectivity spectra of all SiNWs films in the 250 to 1,250 nm wavelength range. The inset is the total reflectivity of untreated silicon surface. Figure 4 Total reflectivity spectra of all SiNWs films in the 250 to 1,250 nm wavelength range. The inset is the to untreated silicon surface. Figure 4 Total reflectivity spectra of all SiNWs films in the 250 to 1,250 nm wavelength range. The inset is the total reflectivity of untreated silicon surface. Nafie et al. Nanoscale Research Letters 2012, 7:393 http://www.nanoscalereslett.com/content/7/1/393 Page 4 of 6 Si SiNWs Generation volume Laser beam Ag/Al contact Ag/Al contact ILBIC Figure 5 Schematic illustration of the LBIC technique. Laser beam Generation volume SiNWs Ag/Al contact Figure 5 Schematic illustration of the LBIC technique. contains homogeneous SiNWs, a great amount of photo-generated electrons by the laser beam can reach the top metal contact. However, when SiNWs are not homogeneous, surface recombination at small wires reduces the LBIC current value, leading to a decrease in the effective diffusion length. and the total reflectivity of films given in Figure 4. In Figure 2, we remark that durations 10, 20, and 40 min lead to inhomogeneous SiNWs films. However, the cross-section SEM images of samples prepared during 30, 50, 60, 70, 80, and 90 min show homogeneous films. In addition, taken into account that at the used wave- length in the LBIC investigations (He-Ne: 633 nm), the corresponding values of the total reflectivity (Figure 4) cannot explain, for example, why the L value of the sam- ple prepared at 90 min is greater than that one prepared at 30 min. Consequently, we attribute these variations not to the total reflectivity, but to the carriers’ trapping at surface defects. For this purpose, we consider the schemes given in Figure 8. Thus, when the MIS diode Authors’ contributions d l S NN prepared samples, MIS diodes, and performed SEM images. She also achieved the reflectivity and LBIC measurements. MAL helped on the preparation of samples, SEM investigations, carrying out the length of SiNWs, and the interpretation of LBIC profiles. Finally, MB supervised the work, did the interpretations, and wrote the text with NN and MAL. All authors read and approved the final manuscript. 9. Zhang YF, Tang YH, Lam C, Wang N, Lee CS, Bello I, Lee ST: Bulk quantity Si nanowires synthesized by SiO sublimation. J Cryst Growth 2000, 212:115. y y y 10. Peng KQ, Yang YJ, Gao SP, Zhu J: Synthesis of large-area silicon nanowires arrays via self-assembling nanoelectrochemistry. Adv Mater 2002, 14:1164. 11. Peng K, Yung Y, Gao S, Zhu J: Dendrites-assisted growth of silicon nanowires in electroless metal deposition. Adv Funct Mater 2002, 13:127. 12. Peng K, Fung H, Hu J, Wu Y, Zhu J, Yun Y, Lee ST: Metal-particle-induced, highly localized site-specific etching of Si and formation of single- crystalline Si nanowires in aqueous fluoride solution. Chem Eur 2006, 12:7942. Received: 27 April 2012 Accepted: 8 June 2012 Published: 16 July 2012 13. Srivastava SK, Kumar D, Singh PK, Kar M, Kumar V, Husain M: Excellent antireflection properties of vertical silicon nanowire arrays. Sol Energ Mater Sol Cells 2010, 94:1506. Competing interests The authors declare that they have no competing interests. 8. Yu DP, Bai ZG, Ding Y, Hang QL, Zhang HZ, Wang JJ, Zou YH, Qian W, Xiong GC, Zhou HT, Feng SQ: Nanoscale silicon wires synthesized using simple physical evaporation. Appl Phys Lett 1998, 72:3458. 13. Srivastava SK, Kumar D, Singh PK, Kar M, Kumar V, Husain M: Excellent antireflection properties of vertical silicon nanowire arrays. Sol Energ Mater Sol Cells 2010, 94:1506. Conclusions Li DY, Wu YY, Kim P, Shi L, Yang PD: Thermal conductivity of individual SiNWs. Appl Phys Lett 2003, 83:2934. 4. Cui Y, Lieber CM: Functional nanoscale electronic devices assembled using SiNWs building blocks. Science 2001, 291:851. 5. Westwater J, Gosain DP, Tomiya S, Usui S, Ruda H: Growth of silicon nanowires via gold/silane vapor–liquid–solid reaction. J Vac Sci Technol B 1997, 15:554. 6. Morales AM, Lieber CM: A laser ablation method for the synthesis of crystalline semiconductor nanowires. Science 1998, 279:208. 7. Wang N, Zhang YF, Tang YH, Lee CS, Lee ST: SiO2-enhanced synthesis of Si nanowires by laser ablation. Appl Phys Lett 1998, 73:3902. 1. Zhao XY, Wei CM, Yang L, Chou MY: Quantum confinement and electronic properties of SiNWs. Phys Rev Lett 2004, 92:236805. Conclusions Ma DDD, Lee CS, Au FCK, Tong SY, Lee ST: Small diameter SiNWs. Science 1874, 2003:299. 2. Ma DDD, Lee CS, Au FCK, Tong SY, Lee ST: Small diameter SiNWs. Science 1874, 2003:299. total reflectivity reaching a minimum less than 1% in the 250 to 400 nm and a minimum of 1.5% in the visi- ble domain. From LBIC investigations, we deduce that the homogeneity of the SiNWs film plays a key role on the electronic properties. Hence, we carried out that when the SiNWs film is inhomogeneous, surface re- combination of photo-generated carriers can decrease the effective diffusion length. 2. Ma DDD, Lee CS, Au FCK, Tong SY, Lee ST: Small diameter SiNWs. Science 1874, 2003:299. 3. Li DY, Wu YY, Kim P, Shi L, Yang PD: Thermal conductivity of individual SiNWs. Appl Phys Lett 2003, 83:2934. 4. Cui Y, Lieber CM: Functional nanoscale electronic devices assembled using SiNWs building blocks. Science 2001, 291:851. 5. Westwater J, Gosain DP, Tomiya S, Usui S, Ruda H: Growth of silicon nanowires via gold/silane vapor–liquid–solid reaction. J Vac Sci Technol B 1997, 15:554. 6. Morales AM, Lieber CM: A laser ablation method for the synthesis of crystalline semiconductor nanowires. Science 1998, 279:208. 7. Wang N, Zhang YF, Tang YH, Lee CS, Lee ST: SiO2-enhanced synthesis of Si nanowires by laser ablation. Appl Phys Lett 1998, 73:3902. 8. Yu DP, Bai ZG, Ding Y, Hang QL, Zhang HZ, Wang JJ, Zou YH, Qian W, Xiong GC, Zhou HT, Feng SQ: Nanoscale silicon wires synthesized using simple physical evaporation. Appl Phys Lett 1998, 72:3458. 9. Zhang YF, Tang YH, Lam C, Wang N, Lee CS, Bello I, Lee ST: Bulk quantity Si nanowires synthesized by SiO sublimation. J Cryst Growth 2000, 212:115. 10. Peng KQ, Yang YJ, Gao SP, Zhu J: Synthesis of large-area silicon nanowires arrays via self-assembling nanoelectrochemistry. Adv Mater 2002, 14:1164. 11. Peng K, Yung Y, Gao S, Zhu J: Dendrites-assisted growth of silicon nanowires in electroless metal deposition. Adv Funct Mater 2002, 13:127. 12. Peng K, Fung H, Hu J, Wu Y, Zhu J, Yun Y, Lee ST: Metal-particle-induced, highly localized site-specific etching of Si and formation of single- crystalline Si nanowires in aqueous fluoride solution. Chem Eur 2006, 12:7942. 13. Srivastava SK, Kumar D, Singh PK, Kar M, Kumar V, Husain M: Excellent antireflection properties of vertical silicon nanowire arrays. Sol Energ Mater Sol Cells 2010, 94:1506. 3. Conclusions In this study, we present a morphological, optical, and electronic study of SiNWs films elaborated at different durations; 10, 20, 30, 40, 50, 60, 70, 80, and 90 min. At some etching durations, a regular structure of formed SiNWs was observed. The SiNWs lengths vary from 21 to 38 μm. We notice a spectacular very low value of the 1200 1400 1600 1800 2000 2200 0,00 0,25 0,50 0,75 1,00 1,25 Normalized ILBIC Laser spot position (µm) 0min 40min 50min 90min MIS Structure Figure 6 Typical normalized LBIC profiles. Vertical dash line corresponds to the metal position on the top of the MIS structure. Normalized ILBIC igure 6 Typical normalized LBIC profiles. Vertical dash line corresponds to the metal position on the top of the MIS stru Nafie et al. Nanoscale Research Letters 2012, 7:393 http://www.nanoscalereslett.com/content/7/1/393 Page 5 of 6 0 10 20 30 40 50 60 70 80 90 100 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 L (µm) Etching time (min) Values of L for homogeneous SiNWs film Values of L for inhomogeneous SiNWs film Figure 7 Obtained values of the diffusion length vs. etching time. MIS diode with homogeneous SiNWs MIS diode with inhomogeneous SiNWs Figure 8 Schematic illustration of the MIS diode when the SiNWs film is homogeneous and not. 0 10 20 30 40 50 60 70 80 90 100 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 L (µm) Etching time (min) Values of L for homogeneous SiNWs film Values of L for inhomogeneous SiNWs film Figure 7 Obtained values of the diffusion length vs. etching time. 0 10 20 30 40 50 60 70 80 90 100 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 L (µm) Etching time (min) Values of L for homogeneous SiNWs film Values of L for inhomogeneous SiNWs film Figure 7 Obtained values of the diffusion length vs. etching time. L (µm) Figure 7 Obtained values of the diffusion length vs. etching time. MIS diode with homogeneous SiNWs MIS diode with inhomogeneous SiNWs Figure 8 Schematic illustration of the MIS diode when the SiNWs film is homogeneous and not. Figure 8 Schematic illustration of the MIS diode when the SiNWs film is homogeneous and not. 2. Competing interests Competing interests The authors declare that they have no competing interests. References 1. Zhao XY, Wei CM, Yang L, Chou MY: Quantum confinement and electronic properties of SiNWs. Phys Rev Lett 2004, 92:236805. Page 6 of 6 Page 6 of 6 Nafie et al. Nanoscale Research Letters 2012, 7:393 http://www.nanoscalereslett.com/content/7/1/393 14. Peng KQ, Xu Y, Wu Y, Yan Y, Lee ST, Zhu J: Aligned single-crystalline Si nanowires arrays for photovoltaic applications. Small 2005, 11:1062. 15. Peng KQ, Wu Y, Fang H, Zhong XY, Xu Y, Zhu J: Uniform, axial-orientation alignment of one-dimensional single-crystal silicon nanostructure arrays. Angew Chem Int Ed 2005, 44:2737. 16. Ioannou DE, Dimitriadis CA: A SEM-EBIC minority-carrier diffusion-length measurement technique. IEEE Trans Electron Dev 1982, ED-29(3):445–450. doi:10.1186/1556-276X-7-393 Cite this article as: Nafie et al.: Effect of etching time on morphological, optical, and electronic properties of silicon nanowires. Nanoscale Research Letters 2012 7:393. 14. Peng KQ, Xu Y, Wu Y, Yan Y, Lee ST, Zhu J: Aligned single-crystalline Si nanowires arrays for photovoltaic applications. Small 2005, 11:1062. 15. Peng KQ, Wu Y, Fang H, Zhong XY, Xu Y, Zhu J: Uniform, axial-orientation alignment of one-dimensional single-crystal silicon nanostructure arrays. Angew Chem Int Ed 2005, 44:2737. 16. Ioannou DE, Dimitriadis CA: A SEM-EBIC minority-carrier diffusion-length measurement technique. IEEE Trans Electron Dev 1982, ED-29(3):445–450. doi:10.1186/1556-276X-7-393 Cite this article as: Nafie et al.: Effect of etching time on morphological, optical, and electronic properties of silicon nanowires. Nanoscale Research Letters 2012 7:393. doi:10.1186/1556-276X-7-393 Cite this article as: Nafie et al.: Effect of etching time on morphological, optical, and electronic properties of silicon nanowires. Nanoscale Research Letters 2012 7:393. Submit your manuscript to a journal and beneﬁ t from: 7 Convenient online submission 7 Rigorous peer review 7 Immediate publication on acceptance 7 Open access: articles freely available online 7 High visibility within the ﬁ eld 7 Retaining the copyright to your article Submit your next manuscript at 7 springeropen.com
https://openalex.org/W2806684167	https://europepmc.org/articles/pmc5991446?pdf=render	English	null	Correction to: A new species of Xenoturbella from the western Pacific Ocean and the evolution of Xenoturbella	BMC evolutionary biology	2,018	cc-by	1,069	60729A85A Genetic information. Whole mitochondrial genome sequences (15,244 bp in holotype; 15,249 bp in paratype) and partial Histone H3 gene sequences (346 bp in holo- type; 413 bp in paratype) were determined and deposited as INSD accession numbers LC228486, LC228485, LC228579 and LC228578, respectively. (Figs. 1–3, Additional files 1–3 in Nakano et al. (2017) [1]) Etymology. Named for the locality where the speci- mens were collected. Holotype. NSMT-Xe 2, female (Figs. 1, 3, Additional files 1, 2 in Nakano et al. (2017) [1]), off Jogashima, Miura, Kanagawa, Japan, 35°06.93″ N 139°33.72″ E to 35°06.95″ N 139°33.33″ E, 380–554 m depth, December 9th, 2015. Remarks. The same species name presented in Nakano et al. (2017) [1] is nomen nudum as the publica- tion does not meet the requirement of the amendment of Article 8.5.3 of the Code [2]. The present erratum fully meets the amended provisions of the Code, and thus this work makes the specific name japonica in the combination of Xenoturbella japonica available. Paratype. NSMT-Xe 1, juvenile, sex unknown (Figs. 2, 3, Additional files 1, 3 in Nakano et al. (2017) [1]), Sanriku coast, Iwate, Japan, 39°37.86″ N 142°18.22″ E to 39°37. 00″ N 142°17.60″ E, 517–560 m depth, July 18th, 2013. Description of female. Based on holotype. Body 5.3 cm in length; pale orange with coloration getting darker Correction to: A new species of Xenoturbella from the western Pacific Ocean and the evolution of Xenoturbella toward the anterior. In live specimens, muscles hold the dorsal body wall in a W-shape (three ridges and two troughs). Body shape actively changes by contracting and elongating when alive. Ring furrow and side furrow are present. Ventral mouth present, oval-shaped, just anterior to ring furrow. Glandular network present over ventral surface, starting near anterior tip of body and ending just in front of ring furrow. Internally, body wall with epidermis, circular and longitudinal muscles, paren- chyma and gastrodermis present. Oocytes present within intestine. Statocyst situated near anterior tip of body, just inside side furrow. © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Correction After publication of Nakano et al. (2017) [1], the authors became aware of the fact that the new species-group name erected for the two specimens of a Japanese xenoturbellid species in the article is not available because Nakano et al. (2017) [1] does not meet the requirement of the amend- ment of Article 8.5.3 of the International Code of Zoo- logical Nomenclature (the Code) [2]. The authors therefore describe the two xenoturbellids as a new species again in this correction article. Methods for morphological observa- tion, DNA extraction and sequencing were as described in Nakano et al. (2017) [1]. The holotype and paratype speci- mens are deposited in the National Museum of Nature and Science, Tsukuba (NSMT), Japan. The DNA sequences obtained were deposited in the International Nucleotide Sequence Database (INSD). Description of juvenile. Based on paratype. Similar to female, but differs as follows: body 1.1 cm in length; pale orange in color; dorsal body surface in live specimen smooth, lacking longitudinal ridges and troughs, similar to that of X. bocki; gametes not observed. Ventral glan- dular network not detected externally, but observed with microCT imaging. Genus Xenoturbella Westblad, 1949 [3] Xenoturbella japonica sp. nov. http://zoobank.org/6C4EA6F8-8AC1-4511-A59B-BCB 60729A85A Genus Xenoturbella Westblad, 1949 [3] Xenoturbella japonica sp. nov. * Correspondence: h.nakano@shimoda.tsukuba.ac.jp 1Shimoda Marine Research Center, University of Tsukuba, 5-10-1, Shimoda, Shizuoka 415-0025, Japan Full list of author information is available at the end of the article Nakano et al. BMC Evolutionary Biology (2018) 18:83 https://doi.org/10.1186/s12862-018-1190-5 Nakano et al. BMC Evolutionary Biology (2018) 18:83 https://doi.org/10.1186/s12862-018-1190-5 Open Access Nakano et al. BMC Evolutionary Biology (2018) 18:83 Nakano et al. BMC Evolutionary Biology (2018) 18:83 Received: 26 December 2017 Accepted: 8 May 2018 Received: 26 December 2017 Accepted: 8 May 2018 Reference 1. Nakano, et al. A new species of Xenoturbella from the western Pacific Ocean and the evolution of Xenoturbella. BMC Evol Biol. 2017;17:245. 2. International Commission of Zoological Nomenclature. Amendment of Articles 8, 9, 10, 21 and 78 of the International Code of Zoological Nomenclature to expand and refine methods of publication. Bull Zool Nomencl. 2012;69:161–9. https://doi.org/10.21805/bzn.v.69i3a8. 3. Westblad E. Xenoturbella bocki n.g, n.sp, a peculiar, primitive turbellarian type. Ark Zool. 1949;1:3–29. Reference 1. Nakano, et al. A new species of Xenoturbella from the western Pacific Ocean and the evolution of Xenoturbella. BMC Evol Biol. 2017;17:245. 2. International Commission of Zoological Nomenclature. Amendment of Articles 8, 9, 10, 21 and 78 of the International Code of Zoological Nomenclature to expand and refine methods of publication. Bull Zool Nomencl. 2012;69:161–9. https://doi.org/10.21805/bzn.v.69i3a8. 3. Westblad E. Xenoturbella bocki n.g, n.sp, a peculiar, primitive turbellarian type. Ark Zool. 1949;1:3–29. Acknowledgements Acknowledgements The authors would like to thank Takafumi Nakano for discussions concerning this erratum. Acknowledgements The authors would like to thank Takafumi Nakano for discussions concerning this erratum. * Correspondence: h.nakano@shimoda.tsukuba.ac.jp 1Shimoda Marine Research Center, University of Tsukuba, 5-10-1, Shimoda, Shizuoka 415-0025, Japan Full list of author information is available at the end of the article Page 2 of 2 Author details 1 1Shimoda Marine Research Center, University of Tsukuba, 5-10-1, Shimoda, Shizuoka 415-0025, Japan. 2Mammalian Genetics Laboratory, National Institute of Genetics, 1111 Yata, Mishima, Shizuoka 411-8540, Japan. 3Faculty of Science, Hokkaido University, N10 W8, Kita-ku, Sapporo, Hokkaido 060-0810, Japan. 4DNA Sequencing Section, Okinawa Institute of Science and Technology Graduate University, Onna, Okinawa 904-0495, Japan. 5Marine Genomics Unit, Okinawa Institute of Science and Technology Graduate University, Onna, Okinawa 904-0495, Japan. 6Misaki Marine Biological Station, The University of Tokyo, 1024 Koajiro, Misaki, Miura, Kanagawa 238-0225, Japan. 7Present address: Sado Marine Biological Station, Faculty of Science, Niigata University, Sado, Niigata 952-2135, Japan. Received: 26 December 2017 Accepted: 8 May 2018 Received: 26 December 2017 Accepted: 8 May 2018 3. Westblad E. Xenoturbella bocki n.g, n.sp, a peculiar, primitive turbellarian type. Ark Zool. 1949;1:3–29. 2. International Commission of Zoological Nomenclature. Amendment of Articles 8, 9, 10, 21 and 78 of the International Code of Zoological Nomenclature to expand and refine methods of publication. Bull Zool Nomencl. 2012;69:161–9. https://doi.org/10.21805/bzn.v.69i3a8.
https://openalex.org/W2731010995	https://europepmc.org/articles/pmc5498578?pdf=render	English	null	Chimeras and complex cluster states in arrays of spin-torque oscillators	Scientific reports	2,017	cc-by	9,378	OPEN Received: 28 February 2017 Accepted: 22 May 2017 Published: xx xx xxxx Michael Zaks1 & Arkady Pikovsky 1,2 We consider synchronization properties of arrays of spin-torque nano-oscillators coupled via an RC load. We show that while the fully synchronized state of identical oscillators may be locally stable in some parameter range, this synchrony is not globally attracting. Instead, regimes of different levels of compositional complexity are observed. These include chimera states (a part of the array forms a cluster while other units are desynchronized), clustered chimeras (several clusters plus desynchronized oscillators), cluster state (all oscillators form several clusters), and partial synchronization (no clusters but a nonvanishing mean field). Dynamically, these states are also complex, demonstrating irregular and close to quasiperiodic modulation. Remarkably, when heterogeneity of spin-torque oscillators is taken into account, dynamical complexity even increases: close to the onset of a macroscopic mean field, the dynamics of this field is rather irregular. Synchronization in large populations of self-sustained periodic oscillators occurs in many physical, biological, engineering and social systems, see recent reviews1, 2. The basic effect, appearance of the macroscopic mean field due to pulling together the frequencies, is described by the simple solvable Kuramoto model of sine-coupled phase oscillators3. However, recently it has been realized that in more involved situations also nontrivial syn- chronization regimes can appear, such as chaotic mean fields4–6, multiplicity of synchronous states7, glassy and Griffiths states8, 9, etc. Remarkably, even for identical oscillators under global (mean field) coupling, complex synchronization regimes like partial synchrony10–12, “chimera“s13–15, heteroclinic cycles16 have been reported. For many such regimes it is still not clear, how robust they are, and whether different dynamic states can coexist. These problems are relevant for many applications in neurosciences (see, e.g. refs 17, 18), nanospintronics (see, e.g. refs 19–21), laser physics22, mechanical, electrochemical, and electronic systems23–26, etc. g p y y In this paper we report on nontrivial regimes of synchronization in an array of spin-torque oscillators (STOs). This consideration is relevant for applications in the generation of coherent field of significant amplitude by a coherent summation of outputs of many spin-torque nano-oscillators. Our results show that these oscillators, being more complex than the phase oscillators used in many studies, demonstrate also more complex properties of the collective dynamics. Spin-torque oscillator is a nanoscale spintronic device generating periodic microwave (in the frequency range of several GHz) oscillations (see refs 27, 28 for an introductory review). www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 www.nature.com/scientificreports/ www.nature.com/scientificreports/ prototype model for such a coupling has been suggested, where N STOs are connected in series and are subject to a common dc current, with a parallel resistive load. The coupling is due to the giant magnetic resistance (GMR) effect, as the resistance of an STO depends on the orientation of its magnetization, so that the redistribution of the ac current between the STO array and the load depends on the the average (over the array) value of this resistance. This setup thus corresponds to the general scheme of mean-field coupling, as discussed above. However, applica- tion of the standard Kuramoto approach here43, 47 is rather questionable, because the STOs are highly nonlinear (see refs 44–46, 48–50) and are especially sensitive close to the homoclinic gluing bifurcation45, 51, 52.h p y g g In this paper we study an array of electrically coupled STOs with an RC-load. This setup is close to that consid- ered in refs 19, 20. We will show that the STOs demonstrates a plethora of nontrivial dynamical regimes, includ- ing partial synchronization, pure and clustered “chimera” states, and clustering. Furthermore, the mean field appears to be highly irregular in a large range of parameters, making the transition to synchrony in this ensemble quite different from the usual transitions observed, e. g., in the Kuramoto model. qf g Before proceeding to the results, we make some remarks about the terminology used in this paper and its relation to the terms used in the literature. Most controversial is the definition of chimera states. Commonly, this term refers to situations where elements or their groups differ with respect to a certain characteristics or property. In a traditional approach, this characteristics is the individual frequency of oscillations, averaged over time: if fre- quencies of oscillators in a symmetric setup differ, one calls this state a chimera (see, e.g. ref. 53). Another wide- spread approach is based on spatial organization: one defines chimera as a profile with a coexistence of smooth and discontinuous in space patches (see, e.g. refs 54, 55). Results B i Basic equations. In this paper we consider an array of spin-torque oscillators, connected to the external driving current via an RC load. The external, generally time-dependent current I(t), is therefore divided between the current through the STO-array J and the current through the RC-load (capacitance C, resistance r) +  CV V r/ . Here V = JR is the voltage on the array, depending on the (generally time-dependent) resistance R of the STO elements. This relation gives the first differential equation of the system = − − + . CdV dt V r J I t( ) (1) (1) The remaining equations are those for N STO oscillators. Each STO is described by its free-layer magnetiza- tion Mi. This vector has constant unit length, and its orientation varies according to the Landau-Lifshitz-Gilbert-Slonczewski equation γ α γβ = − × + × + × × = … d dt M M H M d dt M JM M M i N ( ), 1, , , (2) i i eff i i i i 0 (2) where γ is the gyromagnetic ratio; α is the Gilbert damping constant; β contains material parameters; J is the current through the STO defined above; the effective magnetic field Heff contains an external magnetic field, an easy-axis field, and an easy-plane anisotropy field; M0 is magnetization of the fixed layer. Following44 we assume that = + − \| \| ˆ ˆ ˆ H H e H M e H M e M ( )/ eff a x k x x dz z z . The Landau-Lifshitz-Gilbert-Slonczewski equations can be rewrit- ten in terms of the spherical coordinates (φi, θi) determining orientation of Mi, these equations are coupled via the current J = V/R (see Supplementary Materials for details): θ θ φ φ θ φ =      =     .   G V R Q V R , , , , , (3) i i i i i i (3) Finally, the interaction between the dynamics of the STOs and the load voltage V is due to the dependence of the resistance R on the magnetization vectors Mi, according to the mechanism described in ref. www.nature.com/scientificreports/ Each approach, along with merits, has its obvious limita- tions: the former, being restricted to the cases of continuous-time oscillators with well-defined mean frequency, is applicable neither to the ensembles of coupled maps54, 55, nor to the spin dynamics56; the latter approach assumes spatial ordering and cannot be applied to interactions in globally coupled arrays that are insensitive to distances in physical space. Below we treat a situation where none of the above approaches is applicable: we discuss globally coupled units in the regime of strong coupling where existence of a well-defined mean frequency cannot be guar- anteed. Instead, below we adopt a definition of “chimera” state, based on the clustering property (the respective characteristics is the instantaneous position of the unit in its state space). This approach is applicable to globally coupled populations both of discrete maps and of continuous in time systems. We call a state in a population of identical units a “chimera” if it consists of a macroscopic cluster of units, the states of which coincide, and of a cloud of units, the states of which are different. Notice, that we put in this paper the term “chimera” into quotation marks, to emphasize the usage of a particular definition which may differ from other definitions existing in the literature. OPEN The generation is based on the spin-transfer torque force, with which a spin-polarized electrical current (created by sending electrons through a thick layer with fixed magnetization M0) acts on a small free precessing magnet. As the electrons with spins aligned via M0 enter the free layer, a spin transfer torque acts on its magnetization, tending to reorient it, as has been the- oretically predicted by Slonczewski29 and Berger30. As has been realized by Slonczewski29, the spin transfer torque can compensate the damping of the spin precession of the free layer, and in a constant external magnetic field a sustained oscillation (rotation of magnetization vector) takes place. After experimental observation of the gener- ation31, 32, much attention has been recently attracted to synchronization of STOs; this problem is of high practical relevance, as a way to increase the output power of otherwise rather weak individual STOs33. For two or several STOs, mainly the couplings due to spin wave interaction34–36 or due to magnetic vortex interaction37–42 have been discussed in the literature. For a large number of STOs, the most promising way of coupling the STOs to achieve synchrony, is to connect them in serial electrically via the common microwave current19–21, 43–46. In ref. 19 a 1Institute for Physics and Astronomy, University of Potsdam, Karl-Liebknecht-Str. 24/25, 14476, Potsdam-Golm, Germany. 2Research Institute for Supercomputing, Nizhny Novgorod State University, Gagarin Av. 23, 603950, Nizhny Novgorod, Russia. Correspondence and requests for materials should be addressed to A.P. (email: pikovsky@ uni-potsdam.de) Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 1 Results B i   dv dt v v X G Iv X Q Iv X X N 1 (1 ) , , , 1 , , , 1 , sin cos 1 sin cos (5) i i i i i i N i i 1 (5) The main parameters that determine the coupling are the level of magnetoresistance variations ε, and the dimensionless time constant of the load τ. The main parameter determining the dynamics of the STOs is the external current I. Below we will study the dynamics of the array in the cases of identical and non-identical STOs. The mean magnetization X will be used for the visualization of the mean field behavior. Identical oscillators. For identical spin-torque oscillators, different cluster regimes, where the states (i.e. variables (θI, φ)) of different groups of units coincide identically, are possible. The simplest of these regimes is that of full synchrony (one-cluster state), where θ θ θ φ φ φ = = … = = = … = . , (6) N N 1 2 1 2 (6) The full system of 2N + 1 equations then reduces to a three-dimensional one; for large enough values of the external current > = + + α β I I H H ( ) H a k H 2 dz the reduced system possesses a limit cycle solution, born in the supercritical Hopf bifurcation at I = IH. Mean field in this case coincides with the field of a unit. This regime has a chance to be observed if it is stable towards “evaporation” of units from the cluster. Quantitatively, this stability is measured by the “evaporation” or “split” Lyapunov exponent57, see section Methods for details. A negative evap- oration Lyapunov exponent means that the cluster is stable, a positive one indicates instability. We show the sta- bility diagram in Fig. 1. One can see the large range of the parameters of the external load τ and of the current I, where the full synchrony is stable. This result is confirmed by direct numerical calculations: when the initial states are prepared sufficiently close to each other, evolution of the array ends up in the fully synchronous periodic regime. g Direct calculations however show that the full synchrony is not globally stable. Starting the simulations of Eqs (5) from random initial conditions, we never observed the fully synchronized state after transients. Instead, different complex configurations occur. Results B i 19: Finally, the interaction between the dynamics of the STOs and the load voltage V is due to the dependence of the resistance R on the magnetization vectors Mi, according to the mechanism described in ref. 19: ∑ ρ ε θ φ = − = = R X X N (1 ), 1 sin cos , (4) i N i i 1 (4) where parameters ρ, ε depend on the resistances at parallel and antiparallel (with respect to that of the fixed layer) magnetizations. Referring for details to Supplementary Materials, we finally formulate the dynamics as a set of 2N equations for individual STOs, coupled via the mean magnetization field X through the load voltage (dimen- sionless variable v): where parameters ρ, ε depend on the resistances at parallel and antiparallel (with respect to that of the fixed layer) magnetizations. Referring for details to Supplementary Materials, we finally formulate the dynamics as a set of 2N equations for individual STOs, coupled via the mean magnetization field X through the load voltage (dimen- sionless variable v): Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 2 www.nature.com/scientificreports/ 0.006 0.008 0.01 0.012 0.014 0.016 current I -1 0 1 2 log10(τ) -0.2 -0.1 0 0.1 0.2 Figure 1. Stability of full synchrony. The evaporation Lyapunov exponent of the fully synchronous periodic state in dependence on the parameters (I, τ). This state appears in the supercritical Hopf bifurcation at IH = 0.00315 (see main text), to the left from the shown range of I. Red domains: instability, blue domains: stability. The sharp boundary around I ≈ 0.01–0.012 corresponds to the homoclinic bifurcation of the periodic orbit. 0.006 0.008 0.01 0.012 0.014 0.016 current I -1 0 1 2 log10(τ) -0.2 -0.1 0 0.1 0.2 Figure 1. Stability of full synchrony. The evaporation Lyapunov exponent of the fully synchronous periodic state in dependence on the parameters (I, τ). This state appears in the supercritical Hopf bifurcation at IH = 0.00315 (see main text), to the left from the shown range of I. Red domains: instability, blue domains: stability. The sharp boundary around I ≈ 0.01–0.012 corresponds to the homoclinic bifurcation of the periodic orbit. ∑ τ ε θ θ φ ε φ θ φ ε θ φ θ φ = − + − − =   −    =   −    = = . Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 Results B i These configurations can be classified as follows (for a practical imple- mentation of this classification see section Methods below): 1. Partially synchronous states. In this regime the states of all oscillators are different (i.e. not a single per- sistent cluster is observed), however the distribution is not uniform so that the mean field X(t) performs macroscopic oscillations. These states are of the same type, as described in refs 10–12, 58 for other models of globally coupled oscillators. Noteworthy, in the partially synchronous states described in these referenc- es, the mean frequencies of all oscillators are the same. However, the notion of partial synchronization can be also applied to systems without a well-defined frequency. i 2. Clustered states. In this case several clusters are built, and all the oscillators (except possibly for a few) belong to them. The effective dimension of the system is low. Clustered states of this type have been studied in the context of globally coupled maps and oscillators in refs 59–61.f i 2. Clustered states. In this case several clusters are built, and all the oscillators (except possibly for a few) belong to them. The effective dimension of the system is low. Clustered states of this type have been studied in the context of globally coupled maps and oscillators in refs 59–61.f g y p p 3. “Chimera” states. Here one large cluster is formed, while other units remain different and build a so- called cloud. Such states have been recently reported for globally coupled oscillators in refs 13–15, 60. Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 3 www.nature.com/scientificreports/ tificreports/ 0 0.2 0.4 0.6 0.8 1 0.006 0.007 0.008 0.009 0.01 0.011 0.012 0.013 0.014 0.015 Probability of states Current I PS Chimera Clustered Chimera Clusters Figure 2. Stacked area chart representation of the frequency of occurrence of different states in an array of identical STOs. We show probability to observe different dynamical regimes: partial synchronization (PS, red), “chimera” (green), clustered “chimera” (blue), and clustered states (grey) starting from random initial conditions in an array of 200 STOs. Vertical extensions of the corresponding colors show the probabilities to observe the corresponding states in dependence on the parameter I. Parameters of the coupling: ε = 0.3, τ = 6. 0 0.2 0.4 0.6 0.8 1 0.006 0.007 0.008 0.009 0.01 0.011 0.012 0.013 0.014 0.015 Probability of states Current I PS Chimera Clustered Chimera Clusters Figure 2. Results B i Stacked area chart representation of the frequency of occurrence of different states in an array of identical STOs. We show probability to observe different dynamical regimes: partial synchronization (PS, red), “chimera” (green), clustered “chimera” (blue), and clustered states (grey) starting from random initial conditions in an array of 200 STOs. Vertical extensions of the corresponding colors show the probabilities to observe the corresponding states in dependence on the parameter I. Parameters of the coupling: ε = 0.3, τ = 6. The dynamics here is high-dimensional, although the dimension is significantly reduced compared to the dimension 2N + 1 of the original system. + g y 4. Clustered “chimera” states. Here several clusters are formed, but also a significant number of non-clustered units in a cloud is present. This regime can be considered as a combination of regimes 2 and 3. We stress here that this classification is purely compositional one, it can be inferred from the instantaneous snapshot and does not rely on the particular dynamics of the STOs or of the mean fields; rather we discuss below how/whether dynamics is related to the compositional properties. y p p p In Fig. 2 we show the statistics of these states in a population of N = 200 spin-torque oscillators (this number is not very high because we needed here a large set of trials; below, when illustrating different states, we use ensembles with sizes up to N = 1000). Here the parameter of the external load is fixed at τ = 6. The computational procedure was as follows. For every checked value of I we localized the limit cycle corresponding to the fully synchronous state. Given a point θ φ v ( , , ) on this limit cycle, we randomly chose ca. 1800 initial conditions θi, φi, v from uniform distributions on the intervals θ θ θ < < + .0 5 i , φ φ φ < < + .0 5 i , < < + . v v v 0 1, and deter- mined the proportions of those conditions that, after a transient time of 2000 characteristic periods, ended up in one of the four states described above. One can see in Fig. 2 that there are regions where the partial synchronous states dominate, and a large region   . . Results B i Together with the time course of the mean field, we present in the right panels the Poincaré maps. Recall that a quasiperiodic attractor with two incommensurate frequencies would be seen as a smooth closed curve on the map. One can see that in all the cases the dynamics is close to a quasiperiodicity, with deviations much more noisy in “chimera” (a,b) and clustered “chimera” (c,d) regimes compared to the clustered yfi y g Above we characterized compositional properties of the complex states, by classifying them according to clus- ter structure; next we describe dynamical properties of the evolution of the mean field. For the most of the values of the current I presented in Fig. 2, nontrivial aperiodic dynamics of the mean field X(t) has been observed. We show three typical examples in Fig. 3. To make the time dynamics clear, we present it on two panels, at a long and at a relatively short time scales. Characterization of irregularity of the mean field in a system of many interacting units is not an easy task. Indeed, here the usual methods suitable for low-dimensional systems, such as calculation of Lyapunov exponents and of dimension are hardly applicable. Therefore we characterize regularity of the mean field by calculating the autocorrelation function (see section Methods for details). For small values of the current I, a very slow irregular modulation of the oscillations is observed (Fig. 3, panels (a,b)). The autocorrelation function (panel (c)) slowly decays in this case. For larger values of I the modulation is quite ordered, close to a quasipe- riodic regime with two frequencies (Fig. 3, panels (d,e)). Here the autocorrelation function (panel (f)) returns nearly to 1, as one expects for quasiperiodic processes. For even larger values of the current, the dynamics is strongly irregular (Fig. 3, panels (g,h)). Here the autocorrelation function (panel (i)) rapidly decays nearly to zero. g y g g p g p p y y y A natural question is whether the dynamical properties of the array of STOs are related to the compositional ones. We have found that this relation is very weak. For example, regimes (a,b) and (e,f) in Fig. 3, while being both compositionally partial synchronous states, demonstrate in one case quite regular dynamics and in another case rather irregular one. Results B i I 0 0063 0 0115 where partial synchrony is never observed, but instead clustered and “chimera” states are dominant. Remarkably, in this region the fully synchronized state is stable, so that possibly this stability is inher- ited by sufficiently big clusters. ited by sufficiently big clusters. Above we characterized compositional properties of the complex states, by classifying them according to clus- ter structure; next we describe dynamical properties of the evolution of the mean field. For the most of the values of the current I presented in Fig. 2, nontrivial aperiodic dynamics of the mean field X(t) has been observed. We show three typical examples in Fig. 3. To make the time dynamics clear, we present it on two panels, at a long and at a relatively short time scales. Characterization of irregularity of the mean field in a system of many interacting units is not an easy task. Indeed, here the usual methods suitable for low-dimensional systems, such as calculation of Lyapunov exponents and of dimension are hardly applicable. Therefore we characterize regularity of the mean field by calculating the autocorrelation function (see section Methods for details). For small values of the current I, a very slow irregular modulation of the oscillations is observed (Fig. 3, panels (a,b)). The autocorrelation function (panel (c)) slowly decays in this case. For larger values of I the modulation is quite ordered, close to a quasipe- riodic regime with two frequencies (Fig. 3, panels (d,e)). Here the autocorrelation function (panel (f)) returns nearly to 1, as one expects for quasiperiodic processes. For even larger values of the current, the dynamics is strongly irregular (Fig. 3, panels (g,h)). Here the autocorrelation function (panel (i)) rapidly decays nearly to zero. A natural question is whether the dynamical properties of the array of STOs are related to the compositional ones. We have found that this relation is very weak. For example, regimes (a,b) and (e,f) in Fig. 3, while being both compositionally partial synchronous states, demonstrate in one case quite regular dynamics and in another case rather irregular one. The case, where for the same value of the parameters different structures are possible (like in panels (c,d) in Fig. 3) requires special consideration. Here, for I = 0.008, according to diagram Fig. 2, clusters, clustered “chimera”, and “chimera” states can occur. In Fig. 4 we compare the dynamical regimes for these three compositional states. Results B i This has quite natural explanation: clustered state has a rather small dimension compared to situations where there is a large cloud of unclustered oscillators, thus the effective noise is smaller. Nonidentical oscillators. Above we focused on the properties of an array of identical STOs. Here we study a more realistic situation, where parameters of the STOs are different. Following ref. 19, we take parameter Hk in the effective magnetic field → Heff to be uniformly dispersed in a range Hk0 − ΔHk < Hk < Hk0 + ΔHk. According to general theory of the synchronization transition, e.g. from the exact solution of the Kuramoto model1, 2, one expects, that for a fixed coupling strength and strong enough diversity, no synchronization is observed, and this state turns into a synchronous regime with typically periodic behavior of the mean field, through a bifurcation at a certain critical value of diversity. The bifurcation is perfect in the thermodynamic limit (infinite number of oscillators) and is spoiled by finite-size fluctuations for finite ensembles. p yili Our numerical simulations only partially support this scenario, see Fig. 5. For large diversity in the array, indeed the oscillators do not synchronize and the mean field vanishes (see panels e,f). However, we do not observe, for a fixed level of coupling strength, a transition to a regular, periodic mean field like in the Kuramoto model: the mean field appears to be highly irregular close to the transition (panels c,d) and even for a small diver- sity (panels a,b). This holds for the case where identical oscillators demonstrate erratic behavior (case I = 0.012, panels b,d,f), as well as for the case where for identical oscillators a regime close to quasiperiodicity is observed (case I = 0.008, panels a,c,e). This effect is opposite to “taming spatiotemporal chaos with disorder” reported in ref. 62. Here disorder enhances collective chaos, as is especially pronounced by comparing Fig. 3(c) with Fig. 5(a,c). A more detailed exploration of the range of parameters is presented in the Supplementary Material, Figs 3 and 4. There we show the dynamics of the mean field for different levels of disorder and for different cou- pling strengths. One can see there that only for weak disorder and weak coupling above the synchronization threshold, the mean field is regular; otherwise irregular variations are observed. Results B i The case, where for the same value of the parameters different structures are possible (like in panels (c,d) in Fig. 3) requires special consideration. Here, for I = 0.008, according to diagram Fig. 2, clusters, clustered “chimera”, and “chimera” states can occur. In Fig. 4 we compare the dynamical regimes for these three compositional states. Together with the time course of the mean field, we present in the right panels the Poincaré maps. Recall that a quasiperiodic attractor with two incommensurate frequencies would be seen as a smooth closed curve on the map. One can see that in all the cases the dynamics is close to a quasiperiodicity, with deviations much more noisy in “chimera” (a,b) and clustered “chimera” (c,d) regimes compared to the clustered Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 4 www.nature.com/scientificreports/ -1 -0.5 0 0.5 1 0 1000 2000 3000 (a) 0 200 400 (b) 0 1000 2000 3000 (c) ACF -1 -0.5 0 0.5 1 0 1000 2000 3000 4000 5000 (d) 0 200 400 (e) 0 500 1000 (f) ACF -1 -0.5 0 0.5 1 0 1000 2000 3000 (g) time 0 200 400 (h) time 0 100 200 300 400 500 (i) ACF time shift Figure 3. Different nontrivial dynamical regimes of the mean field X(t). Panels (a,b,c): I = 0.006; panels (d,e,f): I = 0.008; panels (g,h,i): I = 0.012. All data are for an array of N = 1000 oscillators with parameters ε = 0.3 and τ = 6. Panels (b,e,h) show details of evolution on a shorter timescale. Panels (c,f,i) show the corresponding normalized autocorrelation functions, see section Methods for the details of their calculation. 0 1000 2000 3000 (c) ACF 0 0 (f) ACF 0 0 500 1000 0 100 200 300 400 500 (i) time shift Figure 3. Different nontrivial dynamical regimes of the mean field X(t). Panels (a,b,c): I = 0.006; panels (d,e,f): I = 0.008; panels (g,h,i): I = 0.012. All data are for an array of N = 1000 oscillators with parameters ε = 0.3 and τ = 6. Panels (b,e,h) show details of evolution on a shorter timescale. Panels (c,f,i) show the corresponding normalized autocorrelation functions, see section Methods for the details of their calculation. state (e,f). Results B i We have to stress, however, that finite-size effects may hide or destroy regularity of oscillations, therefore more detailed studies of very large arrays are needed to clarify this observation. Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 Discussion I h In this paper we studied complex collective regimes in an array of spin-torque oscillators coupled via a common RC load. Our main focus was on the compositional properties of the ensemble, and on the dynamical properties of the mean field. In almost all aspects, these properties are very much different from those in the standard mod- els of globally coupled periodic oscillators, like the Kuramoto model. We did not use for classification of the states the properties of the mean frequencies, because the STOs in some regimes demonstrate mixed mode oscillations (see Fig. 2 of Supplementary Material), for which the unique frequency cannot be defined straightforwardly. i Compositional properties can be most clearly described for the identical oscillators. Here, although stability theory predicts a possibility of a fully synchronized state, practically only complex regimes between full syn- chrony and full asynchrony are encountered. In a large range of parameters we observed a “chimera” state, where Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 5 www.nature.com/scientificreports/ www.nature.com/scientificreports/ -0.5 0 0.5 0 2000 4000 (a) X -0.02 0 0.02 -0.2 0 0.2 0.4 0.6 0.8 (b) Ymax -0.5 0 0.5 0 2000 4000 (c) X -0.02 0 0.02 -0.2 0 0.2 0.4 0.6 0.8 (d) Ymax -0.5 0 0.5 0 2000 4000 (e) X time -0.02 0 0.02 -0.2 0 0.2 0.4 0.6 0.8 (f) Ymax Xmax Figure 4. Relation between compositional and dynamical properties of the mean field. We show behaviors of the mean field X(t) for I = 0.008 and three compositionally different states. Panels (a,b): “chimera” state; panels (c,d): clustered “chimera” state; panels (e,f): clustered state. Panels (a,c,e) show time evolution of the mean field X(t). Panels (b,d,f) show Poincaré maps: values of the mean fields X and θ φ = Y cos sin i i at the moments where X reaches local maxima. We stress here that the states are classified (“chimera”, clustered “chimera”, clustered) not according to the behavior of the mean field, but through the analysis of the instantaneous states of the units in the population, as defined above and described in section Methods. -0.02 0 0.02 -0.2 0 0.2 0.4 0.6 0.8 (b) -0.02 0 0.02 -0.2 0 0.2 0.4 0.6 0.8 (d) -0.02 0 0.02 -0.2 0 0.2 0.4 0.6 0.8 (f) Xmax Ymax Ymax Ymax Figure 4. Relation between compositional and dynamical properties of the mean field. Discussion I h We show behaviors of the mean field X(t) for I = 0.008 and three compositionally different states. Panels (a,b): “chimera” state; panels (c,d): clustered “chimera” state; panels (e,f): clustered state. Panels (a,c,e) show time evolution of the mean field X(t). Panels (b,d,f) show Poincaré maps: values of the mean fields X and θ φ = Y cos sin i i at the moments where X reaches local maxima. We stress here that the states are classified (“chimera”, clustered “chimera”, clustered) not according to the behavior of the mean field, but through the analysis of the instantaneous states of the units in the population, as defined above and described in section Methods. part of the population builds a cluster and the rest remains disperse. While this regime has been recently reported for several models13–15, a more complex clustered “chimera” state, with several clusters and dispersed oscillators, appears to be a novel one. Together with these two types of “chimera”, a clustered state where all oscillators belong to several clusters, is also possible. Alternatively to appearance of clusters, the array can demonstrate a partially synchronous state where all oscillators are dispersed but remain correlated forming a relatively strong mean field. We reiterate that our definition of “chimera” is not based on the comparison of mean frequencies of oscillations, as has been suggested in some recent studies. Indeed, the frequency is a convenient observable for phase oscil- lators (or, e.g., for those obeying differential equations on a cylinder, where rotations around the cylinder can be straightforwardly counted). In our case, every individual oscillator lives on a sphere, and it cannot be guaranteed that trajectories do form in this projection a “band attractor” allowing to count oscillations and determine the mean frequency. Proper definition of a mean frequency and its analysis in different regimes may be a subject of future work.ii Dynamical properties are related to the regularity of the mean field, which is well defined for nonidentical oscillators as well. One expects that for strong diversity of the oscillator’s properties, the correlation between them will be minimal and they should sum up incoherently into a constant mean field. We confirm this property for the spin-torque oscillators as well. Close to the desynchronization threshold, usually the appearing mean field is dynamically simple (periodic in time). Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 Discussion I h For example, in the case where for identical oscillators one observes complex behavior related to a heteroclinic cycle63, by following the transition from large diversity to the small one, in ref. 64 first a transition to small periodic oscillations was detected; these oscillations undergo secondary bifurcations to more complex states only when diversity becomes sufficiently small. For the spin-torque oscilla- tors we observe a different scenario: here close to the transition to partial coherence where the mean field is small, we observe rather irregular dynamics. For smaller disorder, the dynamics becomes more regular, close to a qua- siperiodic one, for clustered states, and remains rather irregular for the partial synchronous state. This interesting observation deserves more detailed study, where especially the role of the finite size effects should be clarified. Finally we discuss how general are the results described above In this paper we focused on a particular setup y yifi Finally, we discuss how general are the results described above. In this paper we focused on a particular setup, where the common load for the array of STOs is of RC-type. In refs 20, 51 an array with an RCL-load has been Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 6 www.nature.com/scientificreports/ -1 -0.5 0 0.5 1 0 2000 4000 (a) X -1 -0.5 0 0.5 1 0 1000 2000 3000 4000 (b) -1 -0.5 0 0.5 1 0 2000 4000 (c) X -1 -0.5 0 0.5 1 0 1000 2000 3000 4000 (d) -1 -0.5 0 0.5 1 0 2000 4000 (e) X time -1 -0.5 0 0.5 1 0 1000 2000 3000 4000 (f) time Figure 5. Dynamical regimes for different diversities of parameters. Panels (a,c,e): I = 0.008, panels (b,d,f): I = 0.012. Values of the parameter ΔHk: panels (a,b): ΔHk = 0.01; panels (c,d): ΔHk = 0.02; panels (e,f): ΔHk = 0.04. For the regimes without diversity see panels (c,d,e,f) in Fig. 3 above. Figure 5. Dynamical regimes for different diversities of parameters. Panels (a,c,e): I = 0.008, panels (b,d,f): I = 0.012. Values of the parameter ΔHk: panels (a,b): ΔHk = 0.01; panels (c,d): ΔHk = 0.02; panels (e,f): ΔHk = 0.04. For the regimes without diversity see panels (c,d,e,f) in Fig. 3 above. considered. Discussion I h Preliminary calculations for this case show, that all four types of the composition are observed in this setup as well; however the frequency of their occurrence and the dynamical properties are different. Methods We used the following algorithm to distinguish dif- ferent compositional states. In a population of STOs, we first identified clusters. Two units with indices k, m are assigned to the same cluster if their states are sufficiently close: d(k, m) < dcl, where d(k, m) = \|xk − xm\| + \|yk − ym\|, dcl is a small number (we used in calculations dcl = 10−10), and the local observables xk, yk are defined in the same way as the mean fields xk = sinθkcosφk, yk = cosθksinφk. If no clusters were detected in the population, (but the macroscopic mean field was present, that has always been the case) the state was classified as a partially syn- chronous one. The “chimera” state was the state with just one cluster which however did not include all units. We defined the clustered “chimera” state as one with more than one cluster, and a sufficient (larger that 10% of all units) number of units not belonging to the clusters. Finally, a state with many clusters and possibly a small amount (less than 10% of all units) of units outside the clusters was identified as a multiclustered state. This algo- rithm was used to produce Fig. 2. Autocorrelation function and Poincaré map. The autocorrelation function of the mean field X(t) is defined as τ τ = + ˆ ˆ C X t X t ( ) ( ) ( ) where = − ˆX X t X ( ) . Practically, the autocorrelation function was calcu- lated via time averaging; the length of a time series was 16000 characteristic periods. Another way to characterize the complexity of the dynamics is a Poincaré map. In our case we have chosen a two-dimensional map, by plotting all the points at which the mean field X(t) reaches a maximum, on the plane (X, Y) where the mean field Y is defined as ∑ θ φ θ φ = = Y N cos sin 1 cos sin (9) N i i 1 (9) For the periodic motion attractor of this map reduces to a point, while for a quasiperiodic motion with two incommensurate frequencies it yields a closed curve. Poincaré maps presented in Fig. 4 are close to this structure. Methods Integration of equations. Simulation of the system (5) of ODEs governing the array of spin-torque oscil- lators is performed by virtue of the 4th-order Runge-Kutta method. In the case of identical oscillators, there is a numerical trap due to a finite precision of computer representation of real numbers: if the states of two units are closer than the accuracy of representation of numbers in double precision (16 decimal digits), these states “merge” and remain henceforth indistinguishable. This pitfall may lead to appearance of clusters which otherwise (e.g. with quadruple precision calculations) would not exist. To avoid this effect, we add a very small diversity (of the order 10−12) to parameters of the oscillators, thereby preventing exact coincidence of their dynamics. Evaporation Lyapunov exponent. To determine the stability of the cluster solution (6), one imposes a lin- ear perturbation θ1 → θ1 + δθ, θ2 − δθ, θ3 → θ3, …, θN → θN and φ1 → φ1 + δφ, φ2 → φ2 − δφ, φ3 → φ3, …, φN → φN (due to the permutation symmetry such a perturbation can be imposed to any pair of oscillators). This perturba- tion does not influence the mean field X and thus also the global variable v, it just corresponds to a small deviation of two oscillators from the cluster. The equations for δθ and δφ thus read δθ θ θ φ ε δθ φ θ φ ε δφ δφ θ θ φ ε δθ φ θ φ ε δφ = ∂ ∂   −    + ∂ ∂   −    = ∂ ∂   −    + ∂ ∂   −    . d dt G Iv X G Iv X d dt Q Iv X Q Iv X , , 1 , , 1 , , , 1 , , 1 (7) (7) Integration of these linear equations together with the full system (5) yields in the standard way the evapora- tion Lyapunov exponent λ δθ δφ δθ δφ = + + →∞ T T lim 1 2 log ( ) ( ) (0) (0) (8) ev T 2 2 2 2 (8) which determines whether the cluster is transversely unstable (λev > 0) or stable (λev < 0). Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 7 www.nature.com/scientificreports/ Distinguishing different compositional states. Methods More complex states (quasiperiodic ones with many incommensurate frequencies and chaotic ones) result in a cloud of points lying on a fractal set (for a low-dimensional chaos) or without any structure. References Physica D 238(1), 27–37 (2009). 12 Clusella P Politi A & Rosenblum M A minimal model of self consistent partial synchrony New Journal of Physics 18 093037 Physica D 238(1), 27–37 (2009). 12. Clusella, P., Politi, A. & Rosenblum, M. A minimal model of self-consistent partial synchrony. New Journal of Physics 18, 093037 Physica D 238(1), 27–37 (2009). 12. Clusella, P., Politi, A. & Rosenblum, M. A minimal model of self-consistent partial synchrony. New Journal of Physics 18, 093037 (2016) 12. Clusella, P., Politi, A. & Rosenblum, M. A minimal model of self-consistent partial synchrony. New Journal of Physics 18, 09 (2016).h ( ) 13. Sethia, G. C. & Sen, A. Chimera states: The existence criteria revisited. Phys. Rev. Lett. 112, 144101 (2014). 13. Sethia, G. C. & Sen, A. Chimera states: The existence criteria revisited. Phys. Rev. Lett. 112, 144101 (2014). 14 S h idt L S hö l b K K i h K & G í M l V C i t f h d i h i ill t di d h 14. Schmidt, L., Schönleber, K., Krischer, K. & García-Morales, V. Coexistence of synchrony and incoherence in oscillatory media u nonlinear global coupling. Chaos 24, 013102 (2014). g p g 15. Yeldesbay, A., Pikovsky, A. & Rosenblum, M. Chimeralike states in an ensemble of globally coupled oscillators. Phys. Rev. Lett. 112, 144103 (2014). 6. Hansel, D., Mato, G. & Meunier, C. Clustering and slow switching in globally coupled phase oscillators. Phys. Rev. E. 48, 3470–3477 (1993). 17. Singh, R., Menon, S. N. & Sinha, S. Complex patterns arise through spontaneous symmetry breaking in dense homogeneous networks of neural oscillators. Scientific Reports 6, 22074 (2016). fi p ( ) 18. Bolotov, M. I., Osipov, G. V. & Pikovsky, A. Marginal chimera state at cross-frequency locking of pulse-coupled neural networks. Phys. Rev. E 93, 032202 (2016). y 19. Grollier, J., Cros, V. & Fert, A. Synchronization of spin-transfer oscillators driven by stimulated microwave currents. Phys. Rev. 060409(R) (2006). ( ) ( ) 20. Pikovsky, A. Robust synchronization of spin-torque oscillators with an LCR load. Phys. Rev. E 88, 032812 (2013). 20. Pikovsky, A. Robust synchronization of spin-torque oscillators with an LCR load. Phys. Rev. E 88, 032812 (20 l l h f ll h ( ) Synchronization of spin torque nano-oscillators. Phys. Rev. B 95, 1 2. Nixon, M., Ronen, E., Friesem, A. A. & Davidson, N. Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 References p y 2. Pikovsky, A. & Rosenblum, M. Dynamics of globally coupled oscillators: Progress and perspectives. Chaos 25, 097616 (2015). 3. Kuramoto, Y. Self-entrainment of a population of coupled nonlinear oscillators. In Araki, H. (ed.) International Symposium Mathematical Problems in Theoretical Physics, 420 (Springer Lecture Notes Phys., v. 39, New York, 1975). h y p g y 4. Nakagawa, N. & Kuramoto, Y. From collective oscillations to collective chaos in a globally coupled oscillator system. Physica D 7 74–80 (1994). h 4. Nakagawa, N. & Kuramoto, Y. From collective oscillations to collective chaos in a globally coupled oscillator system. Physica D 75, 74–80 (1994). 5. Watanabe, S. & Strogatz, S. H. Constants of motion for superconducting Josephson arrays. Physica D 74, 197–253 (1994). g p g p y y 6. So, P. & Barreto, E. Generating macroscopic chaos in a network of globally coupled phase oscillators. CHAOS 21, 033127 (2011). 7 K M & Pik k A M lti li it f i l h t t i th K t d l f l d ill t Ph R g p g p y y 6. So, P. & Barreto, E. Generating macroscopic chaos in a network of globally coupled phase oscillators. CHAOS 21, 033127 (2011). 6. So, P. & Barreto, E. Generating macroscopic chaos in a network of globally coupled phase oscillators. CHAOS 21, 033127 (2011). 7. Komarov, M. & Pikovsky, A. Multiplicity of singular synchronous states in the Kuramoto model of coupled oscillators. Phys. Rev Lett. 111, 204101 (2013). 8. Iatsenko, D., McClintock, P. V. E. & Stefanovska, A. Glassy states and superrelaxation in populations of coupled phase oscillators. Nature Commun. 5, 4118 (2014). 8. Iatsenko, D., McClintock, P. V. E. & Stefanovska, A. Glassy states and superrelaxation in populations of coupled phase oscillators Nature Commun. 5, 4118 (2014). 9. Villegas, P., Moretti, P. & Muñoz, M. A. Frustrated hierarchical synchronization and emergent complexity in the human connectome network. Scientific Reports 4, 5990 (2014). fi p 10. van Vreeswijk, C. Partial synchronization in populations of pulse-coupled oscillators. Phys. Rev. E 54, 5522–5537 (1996). k k bl lf d ll h d l f l l l d 1. Pikovsky, A. & Rosenblum, M. Self-organized partially synchronous dynamics in populations of nonlinearly coupled oscillators Physica D 238(1), 27–37 (2009). 11. Pikovsky, A. & Rosenblum, M. Self-organized partially synchronous dynamics in populations of nonlinearly coupled oscillators. 1. Acebrón, J. A., Bonilla, L. L., Vicente, C. J. P., Ritort, F. & Spigler, R. The Kuramoto model: A simple paradigm for synchronization phenomena. Rev. Mod. Phys. 77, 137–175 (2005). 2. Pikovsky, A. & Rosenblum, M. Dynamics of globally coupled oscillators: Progress and perspectives. Chaos 25, 097616 (2015). 3. Kuramoto, Y. Self-entrainment of a population of coupled nonlinear oscillators. In Araki, H. (ed.) International Symposium Mathematical Problems in Theoretical Physics, 420 (Springer Lecture Notes Phys., v. 39, New York, 1975). References Clustering, coding, switching, hierarchical ordering, and control in a network of chaotic elements. Physica D 41, 137–172 (1990). 60. Daido, H. & Nakanishi, K. Diffusion-induced inhomogeneity in globally coupled oscillators: Swing-by mechanism. Phys. Rev. Lett. 96, 054101 (2006). , ( ) 61. Schmidt, L. & Krischer, K. Clustering as a prerequisite for chimera states in globally coupled systems. Phys. Rev. Lett. 114, 034101 (2015). ( 0 5). 62. Braiman, Y., Lindner, J. F. & Ditto, W. L. Taming spatiotemporal chaos with disorder. Nature 378, 465–467 (1995). 62. Braiman, Y., Lindner, J. F. & Ditto, W. L. Taming spatiotemporal chaos with disorder. Nature 378, 465–467 (1995). 63. Hansel, D., Mato, G. & Meunier, C. Clustering and slow switching in globally coupled phase oscillators. Phys. Rev. E 48, 3470–3477 (1993).h g 3. Hansel, D., Mato, G. & Meunier, C. Clustering and slow switching in globally coupled phase oscillators. Phys. Rev. E 48, 3470–3477 (1993). 4 Komarov M & Pikovsky A The Kuramoto model of coupled oscillators with a bi harmonic coupling function Physica D 289 3. Hansel, D., Mato, G. & Meunier, C. Clustering and slow switching in globally coupled phase oscillators. Phys. Rev. E 48, 3470–3477 (1993). 4. Komarov, M. & Pikovsky, A. The Kuramoto model of coupled oscillators with a bi-harmonic coupling function. Physica D 289 18–31 (2014). ( ) 4. Komarov, M. & Pikovsky, A. The Kuramoto model of coupled oscillators with a bi-harmonic coupling function. Physica D 289 18–31 (2014). Acknowledgements g Numerical part of this work conducted by A.P. was supported by the Russian Science Foundation (Project No. 17-12-01534). M.Z. was supported by DFG (grant PI 220/17-1). Author Contributions References Mancoff, F. B., Rizzo, N. D., Engel, B. N. & Tehrani, S. Phase-locking in double-point-contact spin-transfer devices. Nature 393–395 (2005).ii 36. Safin, A. R., Udalov, N. N. & Kapranov, M. V. Specific features of the phase locking of two nonidentical spin-transfer nanooscillators. Journal of Communications Technology and Electronics 59, 972–976 (2014). f gy 7. Abreu Araujo, F. & Grollier, J. Controlling the synchronization properties of two dipolarly coupled vortex based spin-torque nano oscillators by the intermediate of a third one. Journal of Applied Physics 120 (2016). 37. Abreu Araujo, F. & Grollier, J. Controlling the synchronization properties of two dipolarly coup oscillators by the intermediate of a third one. Journal of Applied Physics 120 (2016). 38. Chen, H.-H. et al. Phase locking of spin-torque nano-oscillator pairs with magnetic dipolar coupling. Phys. Rev. B 93, 224410 (2016). 9. Yogendra, K., Fan, D. & Roy, K. Coupled spin torque nano oscillators for low power neural computation. IEEE Transactions on Magnetics 51, 1–9 (2015). g 0. Abreu Araujo, F. et al. Optimizing magnetodipolar interactions for synchronizing vortex based spin-torque nano-oscillators. Phys Rev. B 92, 045419 (2015). , ( ) 41. Kendziorczyk, T., Demokritov, S. O. & Kuhn, T. Spin-wave-mediated mutual synchronization of spin-torque nano-oscillators: A micromagnetic study of multistable phase locking. Phys. Rev. B 90, 054414 (2014). 42. Demidov, V. E. et al. Synchronization of spin Hall nano-oscillators to external microwave signals. Nature Communications 5, 3179 (2014). ( ) 3. Georges, B., Grollier, J., Cros, V. & Fert, A. Impact of the electrical connection of spin transfer nano-oscillators on their synchronization: an analytical study. Appl. Phys. Lett. 92, 232504 (2008). 43. Georges, B., Grollier, J., Cros, V. & Fert, A. Impact of the electrical connection of spin transfer nano-oscillators on synchronization: an analytical study. Appl. Phys. Lett. 92, 232504 (2008).fi 44. Li, D., Zhou, Y., Zhou, C. & Hu, B. Global attractors and the difficulty of synchronizing serial spin-torque oscillators. Phys. Rev. B 82, 140407 (2010). 45. Li, D., Zhou, Y., Hu, B. & Zhou, C. Coupled perturbed heteroclinic cycles: Synchronization and dynamical behaviors of spin-torque oscillators. Phys. Rev. B 84, 104414 (2011). Å y 6. Li, D., Zhou, Y., Hu, B., Åkerman, J. & Zhou, C. Multiple synchronization attractors of serially connected spin-torque nanooscillators Phys. Rev. B 86, 014418 (2012). 47. Vegard, F., Ferran, M. & Erik, W. References Observing geometric frustration with thousands of coupled lasers. Phys. Rev Lett. 110, 184102 (2013).h 3. Martens, E. A., Thutupalli, S., Fourrière, A. & Hallatschek, O. Chimera states in mechanical oscillator networks. Proc. Natl. Acad. Sci 110, 10563–10567 (2013). ( ) 24. Zhai, Y., Kiss, I. Z. & Hudson, J. L. Control of complex dynamics with time-delayed feedback in populations of chemical oscillators: Desynchronization and clustering. Industrial & Engineering Chemistry Research 47, 3502 (2008). 25. Temirbayev, A. A., Zhanabaev, Z. Z., Tarasov, S. B., Ponomarenko, V. I. & Rosenblum, M. Experiments on oscillator ensembles with global nonlinear coupling. Phys. Rev. E 85, 015204(R) (2012). g p g y , ( ) ( ) 6. Temirbayev, A. A., Nalibayev, Y. D., Zhanabaev, Z. Z., Ponomarenko, V. I. & Rosenblum, M. Autonomous and forced dynamics o oscillator ensembles with global nonlinear coupling: An experimental study. Phys. Rev. E 87, 062917 (2013). g g y y 27. Slavin, A. & Tiberkevich, V. Nonlinear auto-oscillator theory of microwave generation by spin-polarized current. IEEE T. Magn. 45, 1875–1918 (2009). 28. Chen, T. et al. Spin-torque and spin-hall nano-oscillators. Proceedings of the IEEE 104, 1919–1945 (2016). 28. Chen, T. et al. Spin-torque and spin-hall nano-oscillators. Proceedings of the IEEE 104, 1919–1945 (2016). Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 8 www.nature.com/scientificreports/ 29. Slonczewski, J. C. Current-driven exitation of magnetic multilayers. J. Magn. Magn. Mater. 159, L1–L7 (1996). 30. Berger, L. Emission of spin waves by a magnetic multilayer traversed by a current. Phys. Rev. B 54, 9353–9358 (1996). 30. Berger, L. Emission of spin waves by a magnetic multilayer traversed by a current. Phys. Rev. B 54, 9353–9358 (1996). 31 Ki l S I t l Mi ill ti f t d i b i l i d t N t 425 380 383 (200 g p y g y y y 31. Kiselev, S. I. et al. Microwave oscillations of a nanomagnet driven by a spin-polarized current. Nature 425, 380–383 (2003). 2. Rippard, W. H., Pufall, M. R., Kaka, S., Russek, S. E. & Silva, T. J. Direct-current induced dynamics in co90fe10/ni80fe20 point contacts Phys. Rev. Lett. 92, 027201 (2004).fi y , ( ) 33. Georges, B. et al. Coupling efficiency for phase locking of a spin transfer nano-oscillator to a microwave current. Phys. Rev. Lett. 101, 017201 (2008). 34. Kaka, S. et al. Mutual phase-locking of microwave spin torque nano-oscillators. Nature 437, 389–392 (2005).f 35. References Describing synchronization and topological excitations in arrays of magnetic spin torque oscillators through the Kuramoto model. Scientific Reports 6, 32528 (2016). i 48. Tiberkevich Vasil, S., Khymyn Roman, S., Tang Hong, X. & Slavin Andrei, N. Sensitivity to external signals and synchronization properties of a non-isochronous auto-oscillator with delayed feedback. Scientific Reports 4, 3873 (2014).h 9. Slavin, A. N. & Tiberkevich, V. S. Theory of mutual phase locking of spin-torque nanosized oscillators. Phys. Rev. B 74, 10440 (2006). 50. Zhou, Y. et al. Oscillatory transient regime in the forced dynamics of a nonlinear auto oscillator. Phys. Rev. B 82, 012408 (2010). 51. Zaks, M. A. & Pikovsky, A. Frequency locking near the gluing bifurcation: Spin-torque oscillator under periodic modulati current. Physica D 335, 33–44 (2016). y 52. Turtle, J. et al. Gluing bifurcations in coupled spin torque nano-oscillators. J. Appl. Phys. 113, 114901 (2013). 53. Ashwin, P. & Burylko, O. Weak chimeras in minimal networks of coupled phase oscillators. Chaos 25, 013106 (2015).f 4. Dudkowski, D., Maistrenko, Y. & Kapitaniak, T. Different types of chimera states: An interplay between spatial and dynamical chaos Phys. Rev. E 90, 032920 (2014). 54. Dudkowski, D., Maistrenko, Y. & Kapitaniak, T. Different types of chimera states: An interplay between spatial and dynamical chaos. Phys. Rev. E 90, 032920 (2014). 55. Ghosh, S., Kumar, A., Zakharova, A. & Jalan, S. Birth and death of chimera: Interplay of delay and multiplexing. EPL 115, 60005 f Phys. Rev. E 90, 032920 (2014). 55. Ghosh, S., Kumar, A., Zakharova, A. & Jalan, S. Birth and death of chimera: Interplay of delay and multiplexing. EPL 115, 60005 y 5. Ghosh, S., Kumar, A., Zakharova, A. & Jalan, S. Birth and death of chimera: Interplay of delay and multiplexing. EPL 115, 60005 (2016). 56. Singh, R., Dasgupta, S. & Sinha, S. Chimera order in spin systems. EPL 95, 10004 (2011). 56. Singh, R., Dasgupta, S. & Sinha, S. Chimera order in spin systems. EPL 95, 10004 (2011). 57. Pikovsky, A. & Politi, A. Lyapunov Exponents. A Tool to Explore Complex Dynamics (Cambridge University Press, Cambridge, 2016) 57. Pikovsky, A. & Politi, A. Lyapunov Exponents. A Tool to Explore Complex Dynamics (Cambridge University Press, Cambr 2016). ) 58. Mohanty, P. & Politi, A. A new approach to partial synchronization in globally coupled rotators. J. Phys. A: Math. Gen. 39, L415–L421 (2006). ( ) 9. Kaneko, K. Author Contributions d f l d h M.Z. and A.P. formulated the problem, developed the methods of analysis, and performed numerical simulations A.P. wrote the main manuscript text. All authors reviewed the manuscript. Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 Additional Information Supplementary information accompanies this paper at doi:10.1038/s41598-017-04918-9h Competing Interests: The authors declare that they have no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps an institutional affiliations. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 9 www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2017 Scientific Reports \| 7: 4648 \| DOI:10.1038/s41598-017-04918-9 10
https://openalex.org/W3020529036	https://europepmc.org/articles/pmc7215756?pdf=render	English	null	Mechanical and Damping Properties of Recycled Aggregate Concrete Modified with Air-Entraining Agent and Polypropylene Fiber	Materials	2,020	cc-by	9,181	Received: 28 February 2020; Accepted: 21 April 2020; Published: 24 April 2020 Abstract: In this study, recycled aggregate concrete (RAC) modiﬁed with polypropylene ﬁber (PP) and air-entraining agent (AGA) was prepared, and the eﬀects of PP and AGA on the static (compressive strength, Young’s modulus, and splitting tensile strength) and dynamic properties (dynamic modulus of elasticity and damping ratio) of RAC were investigated. The experimental results showed that the addition of an AGA and PP had a favorable eﬀect on the damping ratio of the concrete, however, the addition of the AGA had a slightly negative eﬀect on the mechanical performance of the concrete. The AGA and PP contents required to achieve the optimum damping ratio of the concrete with the least reduction in the mechanical performance were 0.02% and 0.10%, respectively. Furthermore, the addition of AGA was more eﬀective than that of PP in improving the damping property of the concrete. Keywords: recycled aggregate concrete; interface transition zone; static properties; dynamic behavior; damping mechanism; air-entraining agent materials materials materials Materials 2020, 13, 2004; doi:10.3390/ma13082004 Article Mechanical and Damping Properties of Recycled Aggregate Concrete Modiﬁed with Air-Entraining Agent and Polypropylene Fiber Chonggang Zhou, Xingwang Pei , Wenlong Li and Yijun Liu College of Civil Engineering, Xi’an University of Architecture & Technology, Xi’an 710055, China; wisdom181@126.com (C.Z.); li1536295967@126.com (W.L.); appleliuyijun@sina.com (Y.L.) Correspondence: peixingwang@xauat.edu.cn; Tel.: +86-1872-908-1868 Chonggang Zhou, Xingwang Pei , Wenlong Li and Yijun Liu College of Civil Engineering, Xi’an University of Architecture & Technology, Xi’an 710055, China; wisdom181@126.com (C.Z.); li1536295967@126.com (W.L.); appleliuyijun@sina.com (Y.L.) Correspondence: peixingwang@xauat.edu.cn; Tel.: +86-1872-908-1868 1. Introduction Concrete is the most common construction material and has been extensively used in buildings, bridges, and dams owing to the advantages of its low cost, abundant raw materials, high strength, and excellent durability [1,2]. Aggregate accounts for approximately 60–70% of all raw materials in the concrete production [3–6]. However, concrete consumption is expected to increase with the continuous growth of the infrastructure industry [7,8]. Consequently, non-renewable aggregate resources may get exhausted [4]. Therefore, an alternative material to replace natural aggregate and to relieve the current pressure on the sustainable development of environment is urgently needed [4,9]. Recycled concrete aggregate (RCA) is manufactured from abandoned concrete blocks by a series of processes, including washing, crushing, and grading. RCA is a renewable resource and can be used to alleviate environment pollution [10,11]. Studies have indicated that the partial replacement of natural concrete aggregate (NCA) with RCA to prepare recycled aggregate concrete (RAC) has immense potential for practical applications [12]. RAC is composed of the original aggregate, an adhered old cement mortar, and two types of interface transition zones (ITZs) [4]. However, it exhibits weaker mechanical properties and lower durability than NCA due to the inferior properties of its constituents [4,12]. Dynamic characteristics govern the intrinsic properties of a material and can reﬂect the behavior of a material under dynamic loading [13]. Such characteristics facilitate the dispersion and conversion of energy during the vibration process [7,14]. Further, they can help in improving the stability and safety of the concrete structure. There are three types of damping in concrete [7]: system damping, Materials 2020, 13, 2004; doi:10.3390/ma13082004 www.mdpi.com/journal/materials www.mdpi.com/journal/materials 2 of 11 Materials 2020, 13, 2004 structural damping, and material damping. System and structural damping are external damping (viscous damping, dry damping, hysteresis damping, nonlinear damping, etc.) [15,16], and material damping is internal damping (medium damping, friction damping, etc.) [17]. The damping ratio is the dominant property that indicates the dynamic characteristics of a material [18–20]. Therefore, it has been widely used for the dynamic analysis of building structures [4,7–10]. To date, several studies have focused on the damping of RAC [11], while some have mainly focused on the eﬀect of the replacement ratio and size of RCA on the damping performance of RAC [4,7]. Studies have suggested that due to its high tensile strain, a ﬁber polymer, as an additive, can contribute to improving the performance of concrete [3,7,21,22]. 1. Introduction Furthermore, the study of the eﬀect of the type and dosage of ﬁber on the mechanical properties of RAC has shown that concrete reinforced with glass and steel ﬁbers can signiﬁcantly improve the mechanical performance of RAC when the ﬁber content is within a reasonable range [23]. Additionally, the air entraining agent (AGA) introduces many bubbles in concrete, and it may increase the damping performance of the concrete to a certain extent. Few studies have investigated the properties of RAC modiﬁed with polypropylene ﬁber (PP) and AGA. Speciﬁcally, the damping properties of pure concrete and concrete modiﬁed with AGA and PP have rarely been compared. In this study, for the ﬁrst time, RAC was incorporated with the advantage of PP as well as the high damping characteristic of AGA. The damping properties of RAC at various PP and AGA contents were compared with those of the reference concrete. Our results provide a deep understanding of the dynamic behavior of RAC modiﬁed with PP and AGA and potentially serve as an eﬀective tool for designing concrete with excellent damping property. 2. Materials and Methods 2.1. Raw Materials 2.1.1. Cementitious Materials The cementitious materials used were ordinary Portland cement (P·O 42.5) and ﬂy ash (FA). The cement was purchased from Shaanxi Qinling Cement Co., Ltd. (Yaoxian, Shaanxi, China), and the FA was purchased from Shenmu County Huatai Clean Coal Technology Development Co., Ltd. (Shenmu, Shaanxi, China). Table 1 presents the chemical composition and physical properties of cement according to Chinese Standard GB 175-2007 [24]. Table 2 shows the chemical composition of FA, which complies with the requirements of the Chinese Standard GB/T 51003-2014 [25]. Figures 1 and 2 show the scanning electron micrograph and particle size distribution of FA, respectively. Table 1. Chemical composition and physical properties of the cement. Chemical Composition (wt.%) Primary Physical Properties CaO SiO2 Al2O3 Fe2O3 SO3 MgO Na2O K2O Speciﬁc Surface Area (m2/kg) Speciﬁc Gravity (kg/m3) LOI (Loss on Ignition, %) 61.89 18.73 5.88 3.37 2.08 3.24 0.31 0.15 360 3170 3.03 Table 2. Chemical composition of the FA (wt.%). SiO2 Al2O3 Fe2O3 CaO K2O Na2O MgO SO3 ZnO TiO2 P2O5 49.72 31.09 6.04 3.81 0.41 0.17 1.53 1.49 0.08 2.58 0.42 Table 1. Chemical composition and physical properties of the cement. Table 2. Chemical composition of the FA (wt.%). 3 of 11 Materials 2020, 13, 2004 Figure 1. Scanning electron micrograph of fly ash (FA). Table 2. Chemical composition of the FA (wt.%). Fe2O3 CaO K2O Na2O MgO SO3 ZnO 6.04 3.81 0.41 0.17 1.53 1.49 0.08 2 3 4 5 Volume (%) 20 40 60 80 100 Cumulation (%) Figure 1. Scanning electron micrograph of ﬂy ash (FA). Figure 1. Scanning electron micrograph of fly ash (FA). Table 2. Chemical composition of the FA (wt.%). Fe2O3 CaO K2O Na2O MgO SO3 ZnO 6.04 3.81 0.41 0.17 1.53 1.49 0.08 95 31 36 0.4 0.43 0.47 0.51 0.57 0.67 0.79 0.95 1.13 1.38 1.65 1.97 2.23 2.45 2.63 2.78 2.97 3.2 3.46 3.76 4.07 4.36 4.56 4.64 4.58 4.43 4.14 4.03 3.89 3.71 3.49 3.24 2.93 2.53 2.07 1.56 1.08 0.54 0.01 000000000000000 0.1 1 10 100 1000 0 1 2 3 4 5 Diameter (m) Volume (%) 0 20 40 60 80 100 Cumulation (%) Figure 2. Particle size distribution of FA. Figure 2. Particle size distribution of FA. Figure 1. Scanning electron micrograph of fly ash (FA). Figure 1. Scanning electron micrograph of ﬂy ash (FA). Figure 1. Scanning electron micrograph of fly ash (FA). Table 2. Chemical composition of the FA (wt.%). 2.1.1. Cementitious Materials e2O3 CaO K2O Na2O MgO SO3 Zn 6.04 3.81 0.41 0.17 1.53 1.49 0 Figure 1. Scanning electron micrograph of fly ash (FA). Figure 1. Scanning electron micrograph of ﬂy ash (FA). Table 2. Chemical composition of the FA (wt.%). Fe2O3 CaO K2O Na2O MgO SO3 ZnO 6.04 3.81 0.41 0.17 1.53 1.49 0.0 2 3 4 5 Volume (%) 20 40 60 80 100 Cumulation (%) 95 31 36 0.4 0.43 0.47 0.51 0.57 0.67 0.79 0.95 1.13 1.38 1.65 1.97 2.23 2.45 2.63 2.78 2.97 3.2 3.46 3.76 4.07 4.36 4.56 4.64 4.58 4.43 4.14 4.03 3.89 3.71 3.49 3.24 2.93 2.53 2.07 1.56 1.08 0.54 0.01 000000000000000 0.1 1 10 100 1000 0 1 2 3 4 5 Diameter (m) Volume (%) 0 20 40 60 80 100 Cumulation (%) Figure 2. Particle size distribution of FA. Figure 2. Particle size distribution of FA. 000000000000000 Figure 2. Particle size distribution of FA. Figure 2. Particle size distribution of FA. 2 1 2 Aggregate 2.1.2. Aggregate 0000000000 0.01 0.01 0.02 0.04 0.05 0.1 0.13 0.19 0.25 0.31 0.36 0.4 0.43 0.47 0.51 0.57 0.67 0.79 0.95 1.13 1.38 1.65 1.97 2.23 2.45 2.63 2.78 2.97 3.2 3.46 3.76 4.07 4.36 4.56 4.64 4.58 4.43 4.14 4.03 3.89 3.71 3.49 3.24 2.93 2.53 2.07 1.56 1.08 0.54 0.01 000000000000000 0.1 1 10 100 1000 0 Diameter (m) Figure 2. Particle size distribution of FA. 2.1.2. Aggregate 2.1.2. Aggregate The fine aggregate used was river sand (RS) with a fineness modulus of 2.60 (medium sand); it has water absorption of 1.15% and specific gravity of 2.45. The coarse aggregates used were NCA (broken granite gravel) and RCA (laboratory waste concrete block) and with continuous grading. The maximum size of the coarse aggregates was 25 mm. Table 3 lists the properties of RCA and NCA. The properties of fine and coarse aggregates complied with the Chinese standards GB/T 14684 2011 [26] and GB/T14685 2011 [27] respectively The ﬁne aggregate used was river sand (RS) with a ﬁneness modulus of 2.60 (medium sand); it has water absorption of 1.15% and speciﬁc gravity of 2.45. The coarse aggregates used were NCA (broken granite gravel) and RCA (laboratory waste concrete block) and with continuous grading. The maximum size of the coarse aggregates was 25 mm. Table 3 lists the properties of RCA and NCA. The properties of ﬁne and coarse aggregates complied with the Chinese standards GB/T 14684-2011 [26] and GB/T14685-2011 [27], respectively. The fine aggregate used was river sand (RS) with a fineness modulus of 2.60 (medium san as water absorption of 1.15% and specific gravity of 2.45. The coarse aggregates used were N roken granite gravel) and RCA (laboratory waste concrete block) and with continuous grad he maximum size of the coarse aggregates was 25 mm. Table 3 lists the properties of RCA CA. The properties of fine and coarse aggregates complied with the Chinese standards G 4684-2011 [26] and GB/T14685-2011 [27], respectively. Table 3. Properties of coarse aggregate. ype Water Absorption (%) Crushing Value Index (%) Apparent Density (kg/m3) Bulk Den (kg/m3 CA 1.06 9.80 2561 1380 CA 5 48 15 7 2509 1253 Table 3. Properties of coarse aggregate. Type Water Absorption (%) Crushing Value Index (%) Apparent Density (kg/m3) Bulk Density (kg/m3) NCA 1.06 9.80 2561 1380 RCA 5.48 15.7 2509 1253 as river sand (RS) with a fineness mo and specific gravity of 2.45. 2 1 2 Aggregate 2.1.2. Aggregate The coar Table 3 Properties of coarse aggregate Table 3. Properties of coarse aggregate. 2.2. Mixture Preparation Nine diﬀerent mixtures of RAC modiﬁed with AGA (wt.%) and PP (V%) were prepared. Tables 5 and 6 show the mixture proportions of reference concrete and RAC, respectively. The specimens were formed based on the content. For e.g., in 50RAC-A1, 50 indicates that the replacement percentage of RCA is 50% (wt.), and A1 indicates that the content of AGA is 0.072 kg/m3. The mixing water consisted of two parts: water consumption of ordinary concrete mix design and additional water. It should be noted that the additional water did not change the eﬀective water-to-binder ratio of the mixture. Table 5. Mixture proportion of reference concrete (kg/m3). Concrete ID Cement FA RS NCA RCA Water (Additional Water) PP SP AGA R 280 70 640 1040 0 175 (0) 0 3.5 0 Table 6. Mixture proportions of reference concrete and recycled aggregate concrete (RAC). Concrete ID NCA (%) RCA (%) PP (%) AGA (%) SP(%) Water (Additional Water) (kg/m3) R 100 0 0 0 1 175 (0) 50RAC 50 50 0 0 1 175 (23.2) 100RAC 0 100 0 0 1 175 (46.4) 50RAC-A1 50 50 0 0.01 1 175 (23.2) 50RAC-A2 50 50 0 0.02 1 175 (23.2) 50RAC-A3 50 50 0 0.03 1 175 (23.2) 50RAC-P1 50 50 0.1 0 1 175 (23.2) 50RAC-P2 50 50 0.5 0 1 175 (23.2) 50RAC-P3 50 50 1.0 0 1 175 (23.2) Table 5. Mixture proportion of reference concrete (kg/m3). Table 6. Mixture proportions of reference concrete and recycled aggregate concrete (RAC). p gg g Type Water Absorption (%) Crushing Value Index (%) Apparent Density (kg/m3) Bulk Density (kg/m3) 2.1.3. Polypropylene Fiber PP with a length of 12 mm was used to improve the damping property of RAC. Table 4 shows PP with a length of 12 mm was used to improve the damping property of RAC. Table 4 shows the primary physical and mechanical properties of PP. Type Water Absorption (%) Crushing Value Index (%) Apparent Density (kg/m3) Bulk Density (kg/m3) 2.1.3. Polypropylene Fiber PP with a length of 12 mm was used to improve the damping property of RAC. Table 4 shows PP with a length of 12 mm was used to improve the damping property of RAC. Table 4 shows the primary physical and mechanical properties of PP. CA 1.06 9.80 2561 1380 CA 5.48 15.7 2509 1253 1.3. Polypropylene Fiber PP with a length of 12 mm was used to improve the damping property of RAC. Table 4 sho e primary physical and mechanical properties of PP. Table 4. Main physical and mechanical properties of polypropylene fiber (PP). Table 4. Main physical and mechanical properties of polypropylene ﬁber (PP). Diameter (µm) Density (kg/m3) Tensile Strength (MPa) Elastic Modulus (GPa) Elongation (%) 15–40 920 450 3.81 20 1.06 9.80 2561 5.48 15.7 2509 physical and mechanical properties of PP. Table 4. Main physical and mechanical properties of polypropylene ﬁber (PP). 4 of 11 Materials 2020, 13, 2004 2.1.4. Admixture Polycarboxylate superplasticizer (SP) with a water reduction ratio of 30% and solid content of 35% was used in this study. An AGA with a speciﬁc density of 1.02% was used to enhance the damping property of the concrete. 3.1. Static Properties 3.1. Static Properties 3.1.1. Compressive Strength 3.1.1. Compressive Strength Figure 5a shows the compressive strength of the reference concrete and RAC modified with PP and AGA. The compressive strength of the samples decreased by 3.37% (50RAC) and 40.04% (100 RAC) as compared to that of the reference concrete after 28 days of curing. It is evident that RCA has a marginal effect on the compressive strength of RAC with replacement percentage of 50%, while the RAC with replacement percentage of 100% is significantly affected. Similar results were reported by 3.1.1. Compressive Strength Figure 5a shows the compressive strength of the reference concrete and RAC modified with PP and AGA. The compressive strength of the samples decreased by 3.37% (50RAC) and 40.04% (100 RAC) as compared to that of the reference concrete after 28 days of curing. It is evident that RCA has a marginal effect on the compressive strength of RAC with replacement percentage of 50%, while the RAC with replacement percentage of 100% is significantly affected. Similar results were reported by Liang et al [4] This behavior can be attributed to the higher absorption and crushing value of RCA Figure 5a shows the compressive strength of the reference concrete and RAC modiﬁed with PP and AGA. The compressive strength of the samples decreased by 3.37% (50RAC) and 40.04% (100 RAC) as compared to that of the reference concrete after 28 days of curing. It is evident that RCA has a marginal eﬀect on the compressive strength of RAC with replacement percentage of 50%, while the RAC with replacement percentage of 100% is signiﬁcantly aﬀected. Similar results were reported by Liang et al. [4]. This behavior can be attributed to the higher absorption and crushing value of RCA, as shown in Table 3. p p g g y p y Liang et al. [4]. This behavior can be attributed to the higher absorption and crushing value of RCA, as shown in Table 3. Liang et al. [4]. This behavior can be attributed to the higher absorption and crushing value of RCA, as shown in Table 3. Further, a large number of micro-cracks and voids exist in the old cement mortar and ITZs of RAC, as seen in Figure 5a. Figure 4. Estimation of th Figure 4. Estimation of th 3. Results and Discussion 3. Results and Discu 3. Results and Discu 3.1. Static Properties 3. Results and Discu 3. Results and Discu 3.1. Static Properties 2.3. Test Methods where , f1, and f2 represent the damping ratio of concrete (%), frequency corresponding to an amplitude of f0 /√2 (Hz), and the resonant frequency of concrete (Hz), respectively. where ξ, f1, and f2 represent the damping ratio of concrete (%), frequency corresponding to an amplitude of f0 /√2 (Hz), and the resonant frequency of concrete (Hz), respectively. where , f1, and f2 represent the damping ratio of concrete (%), frequency corresponding to an amplitude of f0 /√2 (Hz), and the resonant frequency of concrete (Hz), respectively. where , f1, and f2 represent the damping ratio of concrete (%), frequency corresponding to an amplitude of f0 /√2 (Hz), and the resonant frequency of concrete (Hz), respectively. Figure 3. Time-magnitude signal as the acceleration response signal. Figure 3. Time-magnitude signal as the acceleration response signal. Figure 3. Time-magnitude signal as the acceleration response signal. Figure 3. Time-magnitude signal as the acceleration response signal. Figure 3. Time-magnitude signal as the acceleration response signal. Figure 3. Time-magnitude signal as the acceleration response signal. Figure 4 Estimation of the damping ratio of concrete using the half power bandwidth method [29] Figure 4. Estimation of the damping ratio of concrete using the half-power bandwidth method [29]. Figure 4. Estimation of the damping ratio of concrete using the half-power bandwidth method [29]. R lt d Di i Figure 4. Estimation of the damping ratio of concrete using the half-power bandwidth method [29] Figure 4. Estimation of the damping ratio of concrete using the half-power bandwidth method [29]. 2.3. Test Methods All specimens with a dimension of 100 mm × 100 mm × 400 mm underwent damping ratio test after 28 days of curing. The damping ratio of concrete can be expressed as follows: frequency–magnitude (m2/s) graphs, as shown in Figure 4, and then the resonant frequency (f0) was obtained [7]. The half-power bandwidth method was applied to measure the damping ratio of the concrete [30]. All specimens with a dimension of 100 mm × 100 mm × 400 mm underwent damping ratio test after 28 days of curing. The damping ratio of concrete can be expressed as follows: shown in Figure 4, and then the resonant frequency (f0) was obtained [7]. The half-power bandwidth method was applied to measure the damping ratio of the concrete [30]. All specimens with a dimension of 100 mm × 100 mm × 400 mm underwent damping ratio test after 28 days of curing. The damping ratio of concrete can be expressed as follows: ME Scope modal analysis program was used to record the data and convert it to frequency–magnitude (m2/s) graphs, as shown in Figure 4, and then the resonant frequency (f0) was obtained [7]. The half-power bandwidth method was applied to measure the damping ratio of the concrete [30]. All specimens with a dimension of 100 mm × 100 mm × 400 mm underwent damping ratio test after 28 days of curing. The damping ratio of concrete can be expressed as follows: frequency–magnitude (m2/s) graphs, as shown in Figure 4, and then the resonant frequency (f0) was obtained [7]. The half-power bandwidth method was applied to measure the damping ratio of the concrete [30]. All specimens with a dimension of 100 mm × 100 mm × 400 mm underwent damping ratio test after 28 days of curing. The damping ratio of concrete can be expressed as follows: ξ = ( f1 −f2) 2f0 (2)   0 2 1 2 f f f    (2)   0 2 1 2 f f f    (2) (2) (2) (2) where ξ, f1, and f2 represent the damping ratio of concrete (%), frequency corresponding to an amplitude of f0 /√2 (Hz), and the resonant frequency of concrete (Hz), respectively. where , f1, and f2 represent the damping ratio of concrete (%), frequency corresponding to an amplitude of f0 /√2 (Hz), and the resonant frequency of concrete (Hz), respectively. 2.3. Test Methods For testing the compressive strength, Young’s modulus, and splitting tensile strength, 150 mm × 150 mm × 150 mm cubic molds, 100 mm × 100 mm × 400 mm prism, and 150 mm × 150 mm × 300 mm prism cubic molds were used, respectively, according to GB/T 50081-2002 [28]. The test samples were demolded 24 h after the initial casting and were put in a standard curing room (20 ± 1 °C, ≥95% RH) until the testing day. The dynamic modulus of elasticity of the concrete was determined by the resonant frequent method in accordance with GB/T 50082-2009 [29]. It was calculated according to the following equation: Ed = 13.244 × 10−4 × WL3 f 2 α4 (1) (1) where Ed, W, L, f, and α represent the dynamic modulus of elasticity of RAC (MPa), weight of sample (kg), length of test sample (mm), transverse frequency of the specimen (Hz), and the side length of square cross-section sample (mm), respectively. where Ed, W, L, f, and α represent the dynamic modulus of elasticity of RAC (MPa), weight of sample (kg), length of test sample (mm), transverse frequency of the specimen (Hz), and the side length of square cross-section sample (mm), respectively. The damping ratio of concrete was evaluated by the vibration of a free-free beam. A pulse hardware/software vibration analysis system (Brüel and Kjær) was utilized to record the acceleration response signals as time magnitude and hammer force, as shown in Figure 3. The ME’Scope modal analysis program was used to record the data and convert it to frequency–magnitude (m2/s) graphs, as 5 of 11 d the 3 The rd the 3. The Materials 2020, 13, 2004 hardware/software hardware/software acceleration respon shown in Figure 4, and then the resonant frequency (f0) was obtained [7]. The half-power bandwidth method was applied to measure the damping ratio of the concrete [30]. All specimens with a dimension of 100 mm × 100 mm × 400 mm underwent damping ratio test after 28 days of curing. The damping ratio of concrete can be expressed as follows: ME Scope modal analysis program was used to record the data and convert it to frequency–magnitude (m2/s) graphs, as shown in Figure 4, and then the resonant frequency (f0) was obtained [7]. The half-power bandwidth method was applied to measure the damping ratio of the concrete [30]. 3.1. Static Properties 3.1. Static Properties 3.1.1. Compressive Strength The compressive strength of RAC slightly decreased by 1.15% (50RAC-A1) and 3.54% (50RAC-A2) when the AGA content was below 0.02%, but was sharply reduced to 10.95% (50RAC-A3) for the mixture containing 0.03% AGA as compared to that of the 50RAC sample after 28 days of curing This could be attributed to the increment of porosity due to the introduction of The compressive strength of RAC slightly decreased by 1.15% (50RAC-A1) and 3.54% (50RAC-A2) when the AGA content was below 0.02%, but was sharply reduced to 10.95% (50RAC-A3) for the mixture containing 0.03% AGA as compared to that of the 50RAC sample after 28 days of curing. This could be attributed to the increment of porosity due to the introduction of AGA in concrete. The compressive strength of RAC slightly decreased by 1.15% (50RAC-A1) and 3.54% (50RAC-A2) when the AGA content was below 0.02%, but was sharply reduced to 10.95% (50RAC-A3) for the mixture containing 0.03% AGA as compared to that of the 50RAC sample after 28 days of curing. This could be attributed to the increment of porosity due to the introduction of The compressive strength of RAC slightly decreased by 1.15% (50RAC-A1) and 3.54% (50RAC-A2) when the AGA content was below 0.02%, but was sharply reduced to 10.95% (50RAC-A3) for the mixture containing 0.03% AGA as compared to that of the 50RAC sample after 28 days of curing. This could be attributed to the increment of porosity due to the introduction of The compressive strength of RAC slightly decreased by 1.15% (50RAC-A1) and 3.54% (50RAC-A2) when the AGA content was below 0.02%, but was sharply reduced to 10.95% (50RAC-A3) for the mixture containing 0.03% AGA as compared to that of the 50RAC sample after 28 days of curing. This could be attributed to the increment of porosity due to the introduction of AGA in concrete. Th i h f h RAC d d b 18 03% (50RAC P1) 3 15% (50RAC P2) d The compressive strength of RAC slightly decreased by 1.15% (50RAC-A1) and 3.54% (50RAC-A2) when the AGA content was below 0.02%, but was sharply reduced to 10.95% (50RAC-A3) for the mixture containing 0.03% AGA as compared to that of the 50RAC sample after 28 days of curing. This could be attributed to the increment of porosity due to the introduction of AGA i t 28 days of curing. 3.1. Static Properties 3.1. Static Properties 3.1.1. Compressive Strength This could be attributed to the increment of porosity due to the introduction of AGA in concrete. The compressive strength of the RAC decreased by 18.03% (50RAC-P1), 3.15% (50RAC-P2), and 22.89% (50RAC-P3) than that of 50RAC after 28 days of curing. The PP-reinforced RAC consisted of six phases: original aggregate, natural aggregate, cement mortar (new cement mortar and old cement mortar), new ITZ (between the NCA and new cement mortar), old ITZ (between the NCA and old cement mortar), and ITZ between the fiber and cement mortar [4]. A large number of micro-cracks were observed in the ITZ between the fibers and cement mortar, as shown in Figure 6. In addition, the thickness of the fiber-cement mortar ITZ was larger than that of the old ITZ in RAC, resulting in a significant gap between the micro hardness of ITZs [1 10] This hindered the The compressive strength of the RAC decreased by 18.03% (50RAC-P1), 3.15% (50RAC-P2), and 22.89% (50RAC-P3) than that of 50RAC after 28 days of curing. The PP-reinforced RAC consisted of six phases: original aggregate, natural aggregate, cement mortar (new cement mortar and old cement mortar), new ITZ (between the NCA and new cement mortar), old ITZ (between the NCA and old cement mortar), and ITZ between the ﬁber and cement mortar [4]. A large number of micro-cracks were observed in the ITZ between the ﬁbers and cement mortar, as shown in Figure 6. In addition, the thickness of the ﬁber-cement mortar ITZ was larger than that of the old ITZ in RAC, resulting in a signiﬁcant gap between the micro-hardness of ITZs [1,10]. This hindered the mechanical performance of RAC. AGA in concrete. The compressive strength of the RAC decreased by 18.03% (50RAC-P1), 3.15% (50RAC-P2), and 22.89% (50RAC-P3) than that of 50RAC after 28 days of curing. The PP-reinforced RAC consisted of six phases: original aggregate, natural aggregate, cement mortar (new cement mortar and old cement mortar), new ITZ (between the NCA and new cement mortar), old ITZ (between the NCA and old cement mortar), and ITZ between the fiber and cement mortar [4]. A large number of micro-cracks were observed in the ITZ between the fibers and cement mortar, as shown in Figure 6. In addition, the thickness of the fiber-cement mortar ITZ was larger than that of the old ITZ in RAC, resulting in a significant gap between the micro-hardness of ITZs [1,10]. 3.1. Static Properties 3.1. Static Properties 3.1.1. Compressive Strength The accumulation of water ﬁlm at the ITZ near the surface of the aggregate and the growth of a large amount of hydration products, calcium hydroxide and ettringite (CH and AFt) on the surface of ITZ cause a porous ITZ in the RAC [4]. Furthermore, RAC has weak ITZs due to the poor adherence between the old aggregate and the new cement mortar, which causes the specimen (100RAC) to exhibit the lowest compressive strength [11,12]. Moreover, the weak properties of RCA (Table 3) and the characteristics of ITZs positively aﬀected the energy dissipation of RAC. This is discussed in Section 3.2.2. 6 of 11 ation of es ak Materials 2020, 13, 2004 properties of RCA ( RAC Thi i di ITZs due to the p the specimen (1 R 50RAC 100RAC 50RAC-A1 50RAC-A2 50RAC-A3 50RAC-P1 50RAC-P2 50RAC-P3 0 10 20 30 40 50 Compressive strength /MPa Concrete ID (a) Figure 5. Static properties of concrete after 28 days of curing: (a) compressive strength and (b) Young’s modulus Figure 5. Static properties of concrete after 28 days of curing: (a) compressive strength and (b) Young’s modulus. p p p y gy p RAC. This is discussed in Section 3.2.2. R 50RAC 100RAC 50RAC-A1 50RAC-A2 50RAC-A3 50RAC-P1 50RAC-P2 50RAC-P3 0 10 20 30 40 50 Compressive strength /MPa Concrete ID (a) Figure 5. Static properties of concrete after 28 days of curing: (a) compressive strength and (b) Young’s modulus 5 5 Concrete ID 0R 50R 50R 5 Concrete ID 5 5 Concrete ID R 50R 50R 5 Concrete ID Figure 5. Static properties of concrete after 28 days of curing: (a) compressive strength and (b) Young’s modulus Figure 5. Static properties of concrete after 28 days of curing: (a) compressive strength and (b) Young’s modulus. Figure 5. Static properties of concrete after 28 days of curing: (a) compressive strength and (b) Young’s modulus. 3.1. Static Properties 3.1. Static Properties 3.1.1. Compressive Strength This hindered the mechanical performance of RAC. echanical performance of RAC. New mortar Old ITZ Aggregate Old mortar a) New ITZ New mortar Old ITZ Aggregate Old mortar (a) New ITZ Figure 6. Interface transition zones (ITZs) in RAC: (a) ITZ between aggregate and cement mortar and (b) ITZ between ﬁber and cement mortar. New mortar New ITZ Figure 6. Interface transition zones (ITZs) in RAC: (a) ITZ between aggregate and cement mortar and (b) ITZ between ﬁber and cement mortar. 3.1.2. Young’s Modulus Figure 5b shows the Young’s modulus of the reference concrete and the RAC modiﬁed with PP and AGA. It is evident that the eﬀect of RCA on the reduction of Young’s modulus is stronger than that of the addition of AGA and PP. The Young’s modulus of the RAC decreases by 3.77% and 41.74% for RCA with replacement percentage of 50% and 100%, respectively, as compared to that of the reference concrete. Furthermore, the Young’s modulus of the RAC decreased with increasing AGA content. A slightly decreasing trend 7 of 11 d Materials 2020, 13, 2004 50RAC P2, and 5 25.07%, respectiv is observed when the AGA content is less than 0.02%. However, when the AGA content is higher than 0.02%, the Young’s modulus of the specimens (50RAC-A3) exhibit a signiﬁcant reduction. Figure 5b shows that the Young’s modulus of the mixtures 50RAC-P1, 50RAC-P2, and 50RAC-P3 containing 0.09%, 0.10%, and 0.11% PP is reduced by 15.52%, 10.15%, and 25.07%, respectively, as compared to that of 50RAC. Young s modulus of concrete [3,4]. The reduction in Young s modulus is primarily attributed to the low strength and elastic modulus of RCA. Therefore, the Young’s modulus of concrete can be significantly controlled by the properties of RCA. 3.1.3. Splitting Tensile Strength Figure 7 shows the splitting tensile strength of the reference concrete and RAC modified with Previous studies have indicated that the properties of aggregates signiﬁcantly aﬀect the Young’s modulus of concrete [3,4]. The reduction in Young’s modulus is primarily attributed to the low strength and elastic modulus of RCA. Therefore, the Young’s modulus of concrete can be signiﬁcantly controlled by the properties of RCA. igu e o e p i i g e i e e g o e e e e e o e e a A o i ie i PP and AGA. The effect of RCA on the splitting tensile strength of RCA is similar to that on the compressive strength. The splitting tensile strength of the concrete decreases by 3.97% (50RAC) and 17.78% (100RAC) as compared to that of the natural aggregate concrete (NAC) after 28 days of curing. Furthermore the splitting tensile strength of concrete decreases by 1 57% (50RAC A1) 2 42% 3.2. Dynamic Behavior The dynamic behavior, i.e., the dynamic modulus of elasticity and damping ratio of the reference concrete and the RAC modiﬁed with AGA and PP, is shown in Figure 8. 3.1.3. Splitting Tensile Strength p g (50RAC-A2), and 13.53% (50R similar conclusion as that from Figure 7 shows the splitting tensile strength of the reference concrete and RAC modiﬁed with PP and AGA. The eﬀect of RCA on the splitting tensile strength of RCA is similar to that on the compressive strength. The splitting tensile strength of the concrete decreases by 3.97% (50RAC) and 17.78% (100RAC) as compared to that of the natural aggregate concrete (NAC) after 28 days of curing. similar conclusion as that from the compressive strength test can be drawn, i.e., a higher dose of AGA adversely affects the splitting tensile property of the RACs regardless of the replacement ratio of the RCA. While a slight increment was found for the specimen containing PP, the splitting tensile strength of concrete increased by 3.85% (50RAC-P1), 4.42% (50RAC-P2), and 3.27% (50RAC-P3). Similar conclusions have been reported by earlier studies [23,31]. R 50RAC 100RAC 50RAC-A1 50RAC-A2 50RAC-A3 0 2 4 6 8 Splitting tensile strength/MPa Concrete ID (a) R 50RAC 100RAC 50RAC-P1 50RAC-P2 50RAC-P3 0 2 4 6 8 Splitting tensile strength/MPa Concrete ID (b) Figure 7. Splitting tensile strength of concrete after 28 days of curing: (a) reference concrete and RAC modified with AGA and (b) reference concrete and RAC modified with PP. Figure 7. Splitting tensile strength of concrete after 28 days of curing: (a) reference concrete and RAC modiﬁed with AGA and (b) reference concrete and RAC modiﬁed with PP. Figure 7. Splitting tensile strength of concrete after 28 days of curing: (a) reference concrete and RAC modified with AGA and (b) reference concrete and RAC modified with PP. Figure 7. Splitting tensile strength of concrete after 28 days of curing: (a) reference concrete and RAC modiﬁed with AGA and (b) reference concrete and RAC modiﬁed with PP. 3.2. Dynamic Behavior Furthermore, the splitting tensile strength of concrete decreases by 1.57% (50RAC-A1), 2.42% (50RAC-A2), and 13.53% (50RAC-A3) as compared to that of the 50RAC samples. Therefore, a similar conclusion as that from the compressive strength test can be drawn, i.e., a higher dose of AGA adversely aﬀects the splitting tensile property of the RACs regardless of the replacement ratio of the RCA. While a slight increment was found for the specimen containing PP, the splitting tensile strength of concrete increased by 3.85% (50RAC-P1), 4.42% (50RAC-P2), and 3.27% (50RAC-P3). Similar conclusions have been reported by earlier studies [23,31]. dynamic modulus o 3.2.2. Damping Ratio 3.2.2. Damping Ratio The damping ratio of concrete exhibits a sharply increasing trend with the increase in the replacement percentage of RCA, as seen in Figure 8b. The damping ratio of the samples 50RAC and 100RAC is enhanced by 7.86% and 35.00%, respectively, as compared to that of the reference concrete after 28 days of curing. Similar results have been reported by Liang et al. [4], Lin et al. [10], and Li et al. [11]. The elastic modulus of different phases significantly varies in RAC. For e.g., the elastic modulus of the aggregate is much larger than that of the cement mortar and ITZs (old ITZ The damping ratio of concrete exhibits a sharply increasing trend with the increase in the replacement percentage of RCA, as seen in Figure 8b. The damping ratio of the samples 50RAC and 100RAC is enhanced by 7.86% and 35.00%, respectively, as compared to that of the reference concrete after 28 days of curing. Similar results have been reported by Liang et al. [4], Lin et al. [10], and Li et al. [11]. The elastic modulus of diﬀerent phases signiﬁcantly varies in RAC. For e.g., the elastic modulus of the aggregate is much larger than that of the cement mortar and ITZs (old ITZ and new ITZ). Further, there is a considerable diﬀerence between the elastic modulus of various phases [4]. elastic modulus of the aggregate is much larger than that of the cement mortar and ITZs (old ITZ and new ITZ). Further, there is a considerable difference between the elastic modulus of various phases [4]. Previous reports have indicated that the indentation modulus of new cement mortar is 1.18 times that of new ITZ, while the indentation modulus of the old cement mortar is approximately 1.33 times higher than that of the old ITZ in the RAC [4], which is confirmed in Figure 9. Hence, the addition of RCA can increase the proportion of the old cement mortar and ITZs of RAC, which Previous reports have indicated that the indentation modulus of new cement mortar is 1.18 times that of new ITZ, while the indentation modulus of the old cement mortar is approximately 1.33 times higher than that of the old ITZ in the RAC [4], which is conﬁrmed in Figure 9. 3.2.1. Dynamic Modulus of Elasticity It is evident from Figure 8a that the dynamic modulus of elasticity of the RAC decreases by 9.38% (50RAC) and 56.07% (100RAC) as compared to the reference concrete. It is evident from Figure 8a that the dynamic modulus of elasticity of the RAC decreases by 9.38% (50RAC) and 56.07% (100RAC) as compared to the reference concrete. The dynamic modulus of elasticity of the samples 50RAC-A1, 50RAC-A2, and 50RAC-A3 decreases by 1.60%, 4.34%, and 13.65%, respectively, as compared to that of 50RAC specimen. It is clear seen that the dynamic modulus of elasticity decreases with the increase in the replacement percentage of RCA and the AGA content. The dynamic modulus of elasticity of the samples 50RAC-P1, 50RAC-P2, and 50RAC-P3 decreases by 22.51%, 4.62%, and 29.64%, respectively, indicating that the addition of PP can The dynamic modulus of elasticity of the samples 50RAC-A1, 50RAC-A2, and 50RAC-A3 decreases by 1.60%, 4.34%, and 13.65%, respectively, as compared to that of 50RAC specimen. It is clear seen that the dynamic modulus of elasticity decreases with the increase in the replacement percentage of RCA and the AGA content. The dynamic modulus of elasticity of the samples 50RAC-P1, 50RAC-P2, and 50RAC-P3 decreases by 22.51%, 4.62%, and 29.64%, respectively, indicating that the addition of PP can 8 of 11 ement mples Materials 2020, 13, 2004 clear seen that the percentage of RCA signiﬁcantly reduce the dynamic modulus of elasticity of RAC with the exception for concrete with 0.50% PP. y p y indicating that the addition of PP can significantly reduce the dynamic modulus of elasticity of RAC with the exception for concrete with 0.50% PP. R 50RAC 100RAC 50RAC-A1 50RAC-A2 50RAC-A3 50RAC-P1 50RAC-P2 50RAC-P3 0 10 20 30 40 Dynamic modulus of elasticity/GPa Concrete ID (a) R 50RAC 100RAC 50RAC-A1 50RAC-A2 50RAC-A3 50RAC-P1 50RAC-P2 50RAC-P3 0.0 0.5 1.0 1.5 2.0 Damping ratio/% Concrete ID (b) Figure 8. Dynamic behavior of the reference concrete and RAC modified with AGA and PP: (a) Figure 8. Dynamic behavior of the reference concrete and RAC modiﬁed with AGA and PP: (a) dynamic modulus of elasticity and (b) damping ratio. R 50RAC 100RAC 50RAC-A1 50RAC-A2 50RAC-A3 50RAC-P1 50RAC-P2 50RAC-P3 0 10 20 30 40 Dynamic modulus of elasticity/GPa Concrete ID (a) 5 5 Concrete ID 5 5 Concrete ID Figure 8. Dynamic behavior of the reference concrete and RAC modified with AGA and PP: (a) Figure 8. 3.2.1. Dynamic Modulus of Elasticity Dynamic behavior of the reference concrete and RAC modiﬁed with AGA and PP: (a) dynamic modulus of elasticity and (b) damping ratio. dynamic modulus o 3.2.2. Damping Ratio As the AGA content increases, the introduction of bubbles and voids reaches a certain amount, and the damping property of the RAC attains its optimal value. Further increase in the amount of AGA does not have a signiﬁcant eﬀect on damping ratio of RAC. On the contrary, it signiﬁcantly reduces the mechanical properties of the RAC, as discussed in Section 3.1. Therefore, for the same replacement percentage of RAC, the optimal content of AGA to simultaneously attain improved mechanical properties and dynamic characteristics is 0.02%. cushioning bag when the structure is subjected to external vibration, as shown in Figure 10. The bubbles and voids in the RAC improve the damping capacity of RAC. As the AGA content increases, the introduction of bubbles and voids reaches a certain amount, and the damping property of the RAC attains its optimal value. Further increase in the amount of AGA does not have a significant effect on damping ratio of RAC. On the contrary, it significantly reduces the mechanical properties of the RAC, as discussed in Section 3.1. Therefore, for the same replacement percentage of RAC, the optimal content of AGA to simultaneously attain improved mechanical properties and dynamic characteristics is 0.02%. Figure 10. Bubbles and voids in RAC. Figure 10. Bubbles and voids in RAC. Figure 10. Bubbles and voids in RAC. Figure 10. Bubbles and voids in RAC. The damping ratio of the samples 50RAC-P1, 50RAC-P2, and 50RAC-P3 increases by 3.18%, 0.66%, and 5.26%, respectively, as compared to that of the 50RAC specimen after 28 days of curing. This can be attributed to the significant gap between the micro-hardness of ITZs (new ITZ, old ITZ, and fiber-cement mortar ITZ), which increases the energy dissipation of RAC due to the sliding between the ITZs [18], as shown in Figure 9. The proportion of ITZs increases with the content of PP, which leads to higher energy dissipation of RAC. All these effects are conducive to improve the damping ratio of RAC. On the contrary, the bubbles and voids in the RAC adversely affect the mechanical properties of RAC, as discussed in Section 3.1. Therefore, it can be concluded that the addition of AGA is more effective than that of PP for increasing the damping ratio of RAC. dynamic modulus o 3.2.2. Damping Ratio Hence, the addition of RCA can increase the proportion of the old cement mortar and ITZs of RAC, which enhances the inhomogeneity of RAC and contribute to the relative deformation between diﬀerent phases in RAC [11,30,31]. Materials 2020, 13, 2004 9 of 11 enhances the inhomogeneity of RAC and contribute to the relative deformation between different phases in RAC [11,30,31]. New ITZ Old ITZ CH CH Figure 9. Images showing ITZs in RAC. The damping ratio of RAC increases with the increase in the AGA content. This increasing trend is more obvious when the AGA content is above 0.02%. The central function of the AGA is to introduce several air bubbles and pores into the concrete, which plays a vital role in flexible hi i b h h i bj d l ib i h i Fi 10 Th Figure 9. Images showing ITZs in RAC. The damping ratio of RAC increases with the increase in the AGA content. This increasing trend is more obvious when the AGA content is above 0.02%. The central function of the AGA is to introduce several air bubbles and pores into the concrete, which plays a vital role in ﬂexible cushioning bag eity of RAC and contribute to the relative deform New ITZ Old ITZ CH CH Figure 9. Images showing ITZs in RAC. Figure 9. Images showing ITZs in RAC. Figure 9. Images showing ITZs in RAC. Figure 9. Images showing ITZs in RAC. Figure 9. Images showing ITZs in RAC. Figure 9. Images showing ITZs in RAC. The damping ratio of RAC increases with the increase in the AGA content. This increasing trend is more obvious when the AGA content is above 0.02%. The central function of the AGA is to introduce several air bubbles and pores into the concrete, which plays a vital role in flexible The damping ratio of RAC increases with the increase in the AGA content. This increasing trend is more obvious when the AGA content is above 0.02%. The central function of the AGA is to introduce several air bubbles and pores into the concrete, which plays a vital role in ﬂexible cushioning bag 9 of 11 A is to lexible Materials 2020, 13, 2004 trend is more obvio d l when the structure is subjected to external vibration, as shown in Figure 10. The bubbles and voids in the RAC improve the damping capacity of RAC. dynamic modulus o 3.2.2. Damping Ratio The damping ratio of the samples 50RAC-P1, 50RAC-P2, and 50RAC-P3 increases by 3.18%, 0.66%, and 5.26%, respectively, as compared to that of the 50RAC specimen after 28 days of curing. This can be attributed to the signiﬁcant gap between the micro-hardness of ITZs (new ITZ, old ITZ, and ﬁber-cement mortar ITZ), which increases the energy dissipation of RAC due to the sliding between the ITZs [18], as shown in Figure 9. The proportion of ITZs increases with the content of PP, which leads to higher energy dissipation of RAC. All these eﬀects are conducive to improve the damping ratio of RAC. On the contrary, the bubbles and voids in the RAC adversely aﬀect the mechanical properties of RAC, as discussed in Section 3.1. Therefore, it can be concluded that the addition of AGA is more eﬀective than that of PP for increasing the damping ratio of RAC. 4. Conclusions 4. Conclusions We prepared RAC containing PP and AGA and evaluated their damping properties and mechanical performance at various AGA and PP contents. The important results of this study are summarized as follows: 1. The RCA had a marginal eﬀect on the mechanical properties of RAC with a replacement aggregate percentage of 50% (reduced by approximately 5%). However, the mechanical properties of the RAC with 100% replacement aggregate were signiﬁcantly aﬀected (approximately decreased by 40%) than those of the reference concrete. 2. The addition of both AGA and PP adversely aﬀected the mechanical properties of the concrete, and the eﬀect of PP addition on the reduction of the mechanical properties of RAC was stronger than that of AGA addition. The compressive strength and Young’s modulus reduced slightly when the AGA content was below 0.02%. 3. The addition of both AGA and PP favorably aﬀected the damping property of the concrete, and the addition of AGA was more eﬀective than that of PP in increasing the damping ratio of the RAC. Furthermore, the AGA and PP contents to achieve the optimum dynamic property of the concrete with the least reduction in the mechanical performance were determined to be 0.02% and 0.10%, respectively. 10 of 11 10 of 11 Materials 2020, 13, 2004 Author Contributions: Conceptualization, Y.L.; Data curation, Y.L.; Formal analysis, X.P.; Funding acquisition, X.P.; Investigation, W.L.; Project administration, X.P.; Software, C.Z.; Supervision, W.L.; Validation, Y.L.; Visualization, W.L.; Writing—original draft, C.Z.; Writing—review & editing, C.Z. All authors have read and agree to the published version of the manuscript. Author Contributions: Conceptualization, Y.L.; Data curation, Y.L.; Formal analysis, X.P.; Funding acquisition, X.P.; Investigation, W.L.; Project administration, X.P.; Software, C.Z.; Supervision, W.L.; Validation, Y.L.; Visualization, W.L.; Writing—original draft, C.Z.; Writing—review & editing, C.Z. All authors have read and agree to the published version of the manuscript. Funding: This research was funded by National Natural Science Foundation of China, grant number 51808424. Conﬂicts of Interest: The authors declare no conﬂict of interest. u d g: s esea c was u ded by Nat o a Natu a Sc e ce ou dat o o C a, g a t u be 5 808 Conﬂicts of Interest: The authors declare no conﬂict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. References 1. Lu, D.; Tang, Z.; Zhang, L.; Zhou, J.; Gong, Y.; Tian, Y.; Zhong, J. Eﬀects of combined usage of supplementary cementitious materials on the thermal properties and microstructure of high-performance concrete at high temperatures. Materials 2020, 13, 1833. [CrossRef] [PubMed] 1. Lu, D.; Tang, Z.; Zhang, L.; Zhou, J.; Gong, Y.; Tian, Y.; Zhong, J. Eﬀects of combined usage of supplementary cementitious materials on the thermal properties and microstructure of high-performance concrete at high temperatures. Materials 2020, 13, 1833. [CrossRef] [PubMed] 2. Saavedra, W.G.V.; de Gutiérrez, R.M. Performance of geopolymer concrete composed of ﬂy ash after exposure to elevated temperatures. Constr. Build. Mater. 2017, 154, 229–235. [CrossRef] 2. Saavedra, W.G.V.; de Gutiérrez, R.M. Performance of geopolymer concrete composed of ﬂy ash after exposure to elevated temperatures. Constr. Build. Mater. 2017, 154, 229–235. [CrossRef] 3. Wang, Z.; Li, L.; Zhang, Y.; Wang, W. Bond-slip model considering freeze-thaw damage eﬀect of concrete and its application. Eng. Struct. 2019, 201, 109831. [CrossRef] 3. Wang, Z.; Li, L.; Zhang, Y.; Wang, W. Bond-slip model considering freeze-thaw damage eﬀect of concrete and its application. Eng. Struct. 2019, 201, 109831. [CrossRef] 4. Liang, C.; Liu, T.; Xiao, J.; Zou, D.; Yang, Q. The damping property of recycled aggregate concrete. Constr. Build. Mater. 2016, 102, 834–842. [CrossRef] 4. Liang, C.; Liu, T.; Xiao, J.; Zou, D.; Yang, Q. The damping property of recycled aggregate concrete. Constr. Build. Mater. 2016, 102, 834–842. [CrossRef] 5. Long, G.; Yang, J.; Xie, Y. The mechanical characteristics of steam-cured high strength concrete incorporating with lightweight aggregate. Constr. Build. Mater. 2017, 136, 456–464. [CrossRef] 5. Long, G.; Yang, J.; Xie, Y. The mechanical characteristics of steam-cured high strength concrete incorporating with lightweight aggregate. Constr. Build. Mater. 2017, 136, 456–464. [CrossRef] 6. Zhang, Y.; Bicici, E.; Sezen, H.; Zheng, S. Reinforcement slip model considering corrosion eﬀects. Constr. Build. Mater. 2020, 235, 117348. [CrossRef] 6. Zhang, Y.; Bicici, E.; Sezen, H.; Zheng, S. Reinforcement slip model considering corrosion eﬀects. Constr. Build. Mater. 2020, 235, 117348. [CrossRef] 7. Tian, Y.; Shi, S.; Jia, K.; Hu, S. Mechanical and dynamic properties of high strength concrete modiﬁed with lightweight aggregates presaturated polymer emulsion. Constr. Build. Mater. 2015, 93, 1151–1156. [CrossRef] 7. Tian, Y.; Shi, S.; Jia, K.; Hu, S. Mechanical and dynamic properties of high strength concrete modiﬁed with lightweight aggregates presaturated polymer emulsion. Constr. Build. Mater. 2015, 93, 1151–1156. [CrossRef] 8. Yoon, J.Y.; Kim, J.H. References Mechanical properties of preplaced lightweight aggregates concrete. Constr. Build. Mater. lightweight aggregates presaturated polymer emulsion. Constr. Build. Mater. 2015, 93, 1151–1156. [CrossRef] 8. Yoon, J.Y.; Kim, J.H. Mechanical properties of preplaced lightweight aggregates concrete. Constr. Build. Mater. 2019, 216, 440–449. [CrossRef] 8. Yoon, J.Y.; Kim, J.H. Mechanical properties of preplaced lightweight aggregates concrete. Constr. Build. Mater. 2019, 216, 440–449. [CrossRef] 9. Li, W.; Lang, L.; Wang, D.; Wu, Y.; Li, F. Investigation on the dynamic shear modulus and damping ratio of steel slag sand mixtures. Constr. Build. Mater. 2018, 162, 170–180. [CrossRef] 10. Lin, J.; Hong, J.; Xiao, Y. Dynamic characteristics of 100% cold recycled asphalt mixture using asphalt emulsion and cement. J. Clean. Prod. 2017, 156, 337–344. [CrossRef] 11. Li, T.; Xiao, J.; Sui, T.; Liang, C.; Li, L. Eﬀect of recycled coarse aggregate to damping variation of concrete. Constr. Build. Mater. 2018, 178, 445–452. [CrossRef] 12. Wang, C.; Xiao, J.; Wang, C.; Zhang, C. Nonlinear damping and nonlinear responses of recycled ag concrete frames under earthquake loading. Eng. Struct. 2019, 201, 109575. [CrossRef] 13. Giner, V.T.; Ivorra, S.; Baeza, F.J.; Zornoza, E.; Ferrer, B. Silica fume admixture eﬀect on the dynamic pr of concrete. Constr. Build. Mater. 2011, 25, 3272–3277. [CrossRef] 14. Tang, X.; Yan, X. Acoustic energy absorption properties of ﬁbrous materials: A review. Compos. Appl. Sci. Manuf. 2017, 101, 360–380. [CrossRef] 15. Cortés, F.; Castillo, G. Comparison between the dynamical properties of polymer concrete and grey cast iron for machine tool applications. Mater. Des. 2007, 28, 1461–1466. [CrossRef] 16. Scalerandi, M.; Bentahar, M.; Mechri, C. Conditioning and elastic nonlinearity in concrete: Separation of damping and phase contributions. Constr. Build. Mater. 2018, 161, 208–220. [CrossRef] 17. Liew, K.M.; Kai, M.F.; Zhang, L.W. Mechanical and damping properties of CNT-reinforced cementitious composites. Compos. Struct. 2017, 160, 81–88. [CrossRef] 18. Bentz, D.P. Inﬂuence of internal curing using lightweight aggregates on interfacial transition zone percolation and chloride ingress in cement mortars. Cem. Concr. Compos. 2009, 31, 285–289. [CrossRef] 19. Jeon, E.B.; Ahn, S.K.; Lee, I.G.; Koh, H.I.; Park, J.; Kim, H.S. Investigation of mechanical/dynamic properties of carbon ﬁber reinforced polymer concrete for low noise railway slab. Compos. Struct. 2015, 134, 27–35. [CrossRef] 20. Wang, Z.; Li, L.; Zhang, Y.; Zheng, S. Reinforcement model considering slip eﬀect. Eng. Struct. 2019, 198, 109493. [CrossRef] 11 of 11 Materials 2020, 13, 2004 21. Al-Subaihawi, S.; Kolay, C.; Marullo, T.; Ricles, J.M.; Quiel, S.E. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). References Assessment of wind-Induced vibration mitigation in a tall building with damped outriggers using real-time hybrid simulations. Eng. Struct. 2020, 205, 110044. [CrossRef] 2. Fracois, D.L.; Horacio, C. Concrete Recycling: Research and Practice; CRC Press: Boca Raton, FL, USA, 2 ISBN 9781138724723. 23. Liu, W.H.; Li, H.; Zhu, H.M.; Xu, P.J. The Interfacial Adhesion Performance and Mechanism of a Modiﬁed Asphalt–Steel Slag Aggregate. Materials 2020, 13, 1180. [CrossRef] [PubMed] 24. China Architecture and Building. Common Portland Cement (GB175-2007); China Architecture and Building Press: Beijing, China, 2007. (In Chinese) j g 25. China Architecture and Building. Fly Ash Used for Cement and Concrete (GB/T1596-2005); China Architecture and Building Press: Beijing, China, 2005. (In Chinese) 26. China Architecture and Building. Sand for Construction (GB/T 14684-2011); China Architecture and Building Press: Beijing, China, 2011. (In Chinese) 27. China Architecture and Building. Pebble and Crushed Stone for Construction (GB/T14685-2011); China Architecture and Building Press: Beijing, China, 2011. (In Chinese) 28. China Architecture and Building. Standard for Test Method of Mechanical Properties on Ordinary Concrete (GB/T 50081-2002); China Architecture and Building Press: Beijing, China, 2002. (In Chinese) 29. China Architecture and Building. Standard for Test Method of Long-term Performance and Durability of Ordinary Concrete (GB/T 50082-2009); China Architecture and Building Press: Beijing, China, 2009. (In Chinese) 29. China Architecture and Building. Standard for Test Method of Long term Performance and Durability of Ordinary Concrete (GB/T 50082-2009); China Architecture and Building Press: Beijing, China, 2009. (In Chinese) 30. Papagiannopoulos, G.A.; Hatzigeorgiou, G.D. On the Use of the Half-Power Bandwidth Method to Estimate Damping in Building Structures. Soil Dyn. Earthq. Eng. 2011, 31, 1075–1079. [CrossRef] 31. Xuan, D.G.; Zhan, B.J.; Poon, C.S. Durability of Recycled Aggregate Concrete Prepared with Carbonated Recycled Concrete Aggregates. Cem. Concr. Compos. 2017, 84, 214–221. [CrossRef] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W1986195753	https://europepmc.org/articles/pmc2569935?pdf=render	English	null	Italian network for obesity and cardiovascular disease surveillance: A pilot project	BMC family practice	2,008	cc-by	8,840	BioMed Central BioMed Central BMC Family Practice Open Access Open Acc Research article Italian network for obesity and cardiovascular disease surveillance: A pilot project Chiara Donfrancesco1, Cinzia Lo Noce1, Ovidio Brignoli2, Gabriele Riccardi3, Paola Ciccarelli1, Francesco Dima1, Luigi Palmieri1 and Simona Giampaoli1 Address: 1Centro Nazionale di Epidemiologia, Sorveglianza e Promozione della Salute, Istituto Superiore di Sanità, Rome, Italy, 2Società Italiana di Medicina Generale, Rome, Italy and 3Università Federico II, Naples, Italy Email: Chiara Donfrancesco - chiara.donfrancesco@iss.it; Cinzia Lo Noce - cinzia.lonoce@iss.it; Ovidio Brignoli - docbri@fastwebnet.it; Gabriele Riccardi - riccardi@unina.it; Paola Ciccarelli - paola.ciccarelli@iss.it; Francesco Dima - francesco.dima@iss.it; Luigi Palmieri - luigi.palmieri@iss.it; Simona Giampaoli - simona.giampaoli@iss.it * Corresponding author Received: 4 March 2008 Accepted: 29 September 2008 Received: 4 March 2008 Accepted: 29 September 2008 This article is available from: http://www.biomedcentral.com/1471-2296/9/53 © 2008 Donfrancesco et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background comparable data and implementing prevention proce- dures at individual and population levels. g Over the last decades, a slow but steady decrease can be observed in cardiovascular disease (CVD) mortality in industrialized countries. This is partly due to improve- ments in pharmacological treatment during both acute and chronic phases, reduction in levels of established major risk factors – total cholesterol and hypertension –, better pharmacological control of risk factors, and reduc- tion in smoking habit [1,2]. An inverse trend can be observed for obesity, whose prevalence is steadily increas- ing [3]. Following the experience of other Italian studies trying to evaluate the feasibility of a surveillance network for other risk factors and diseases (alcohol, smoking, physical inac- tivity, renal diseases) [17-20], an Italian pilot project was implemented in January 2006 with the main aim of eval- uating the feasibility of a surveillance network for CVD and obesity. Secondary aims were to assess the prevalence of cardiovascular risk conditions and distribution of over- weight and obesity in men and women aged 35–74 years, to describe lifestyle habits such as diet and physical activ- ity, and to investigate the association between lifestyle, obesity and predicted cardiovascular risk in the general population. Scientific evidence accumulated over recent years has prompted the American Heart Association to identify obesity as the major and modifiable risk factor for CVD, the UK Government (Department of Health) to include obesity as one of the six key priorities in the White Paper Choosing Health: Making Healthy Choices Easier and the Italian Ministry of Health to launch the Project 'Guadagnare Salute' which foresees multidisciplinary inter- ventions in the field of diet and physical activity [4-6]. Promotion of physical activity and healthy diet are regarded as key factors to prevent and reduce overweight and obesity epidemic. In USA, the Healthy People 2010 programme identifies physical activity as one of the public health top priorities [7]. The Public Health Programme 2003–2008 of the European Union invites countries to implement initiatives that sustain and promote physical activity [8]. http://www.biomedcentral.com/1471-2296/9/53 http://www.biomedcentral.com/1471-2296/9/53 http://www.biomedcentral.com/1471-2296/9/53 Background In 2005, the Aprifel, the French agency for research and information on fresh fruits and vegetables, the Italian Ministry of Health and the National Centre for Disease Prevention and Control (CCM) organized the International Conference on Health Benefits of Mediterra- nean Style Diet – From Scientific Evidence to Health Pre- vention Actions with the aim of highlighting the latest scientific knowledge on health benefits of Mediterranean style diet and the strategic perspectives for health opera- tors [9]. The study lasted about two years and involved general practitioners (GPs) to perform screening examinations to assess cardiovascular risk, lifestyles habits (physical activ- ity, diet and smoking) and drug use among their patients. Particular focus was laid on overweight and obese per- sons. Page 2 of 11 (page number not for citation purposes) Abstract Background: Also in Mediterranean countries, which are considered a low risk population for cardiovascular disease (CVD), the increase in body mass index (BMI) has become a public health priority. To evaluate the feasibility of a CVD and obesity surveillance network, forty General Practitioners (GPs) were engaged to perform a screening to assess obesity, cardiovascular risk, lifestyle habits and medication use. Methods: A total of 1,046 women and 1,044 men aged 35–74 years were randomly selected from GPs' lists stratifying by age decade and gender. Anthropometric and blood pressure measurements were performed by GPs using standardized methodologies. BMI was computed and categorized in normal weight (BMI 18.5–24.9 kg/m2), overweight (BMI 25.0–29.9 kg/m2) and obese (BMI ≥ 30 kg/ m2). Food frequency (per day: fruits and vegetables; per week: meat, cheese, fish, pulses, chocolate, fried food, sweet, wholemeal food, rotisserie food and sugar drink) and physical activity (at work and during leisure time) were investigated through a questionnaire. CVD risk was assessed using the Italian CUORE Project risk function. Results: The percentage of missing values was very low. Prevalence of overweight was 34% in women and 50% in men; prevalence of obesity was 23% in both men and women. Level of physical activity was mostly low or very low. BMI was inversely associated with consumption of pulses, rotisserie food, chocolate, sweets and physical activity during leisure time and directly associated with consumption of meat. Mean value of total cardiovascular risk was 4% in women and 11% in men. One percent of women and 16% of men were at high cardiovascular risk (≥ 20% in 10 years). Normal weight persons were four times more likely to be at low risk than obese persons. Conclusion: This study demonstrated the feasibility of a surveillance network of GPs in Italy focusing on obesity and other CVD risk factors. It also provided information on lifestyle habits, such as diet and physical activity. Page 1 of 11 (page number not for citation purposes) (page number not for citation purposes) BMC Family Practice 2008, 9:53 Methods Screening Geographical distribution of general practitioners and patients involved in the study Figure 1 Geographical distribution of general practitioners and patients involved in the study. Geographical distribution of general practitioners and patients involved in the study Figure 1 Geographical distribution of general practitioners and patients involved in the study. could not refer to a non-smoker) and alerts GPs to missing data. ment were: total cholesterol (TC), HDL-cholesterol (HDL- C) and glycaemia. Total cardiovascular risk (TCR) was assessed using the CUORE Project risk function [21,22]. ment were: total cholesterol (TC), HDL-cholesterol (HDL- C) and glycaemia. Total cardiovascular risk (TCR) was assessed using the CUORE Project risk function [21,22]. The main aim of this pilot study was to assess the feasibil- ity of a network of GPs for CVD surveillance, prevention and health promotion. For this reason, collected informa- tion were carefully evaluated for completeness and relia- bility. A data entry software intended for GPs was developed to facilitate the process of data collection and storage during the examination of patients. At the end of the study, every GP sent collected data to the ISS for processing. Data sent by the 40 GPs referred to 1,044 men and 1,046 women, that is an average of 55 patients per GP. Written informa- tion consent was obtained from each study participant. For completeness analysis, the percentage of missing val- ues for each variable was calculated; for reliability analy- sis, anthropometric (height, weight, waist and hip circumference) and blood pressure measurements were considered: the distribution of the terminal digit of recorded blood pressure readings and the percentage of equal consecutive readings were calculated [23]. For Methods Screening g Forty GPs from the Italian Association of General Practi- tioners (SIMG) and homogeneously distributed across the country were engaged. To ensure uniformity in the appli- cation of screening procedures, a training course for GPs on standardization of measurements was organized. At least one GP per region participated in the study [Figure 1]. Each participating GP had the task to screen 56 patients aged 35–74 years, randomly selected by the National Institute of Health (ISS) from each GP's patient list in order to have 7 patients in each age decade and sex (35–44, 45–54, 55–64, 65–74 years). Selected patients were invited by letter to the ambulatory of their own GP. In Italy, a national surveillance network for overweight, obesity and other cardiovascular risk factors does not exist. Presently, latest available data on overweight and obesity come from the multi-scopes survey performed in 2005 and published in 2007 by the Italian Institute of National Statistics (ISTAT), collecting self-reported infor- mation on health status [10], and from the Epidemiologic Cardiovascular Observatory (OEC) of the Progetto CUORE, estimating prevalence of overweight, obesity and other CVD risk factors in about 10,000 Italian men and women aged 35–74 years examined between 1998 and 2002. [11]. Many initiatives are taken at the regional level, but they are mainly targeted to children and adolescents [12-16]. Therefore, there is a pressing need to establish a national surveillance system for collecting reliable and The screening protocol foresaw the completion of a ques- tionnaire containing questions on anagraphical position, such as marital status and educational level, habits and lifestyles, remote pathological anamnesis (angina pec- toris, myocardial infarction, cerebrovascular accidents, TIA, intermittent claudication), current drug treatments (antihypertensives, lipid-lowering, oral antidiabetics), family history (father, mother, sisters/brothers) of prema- ture coronary heart disease, cerebrovascular accidents and family history of diabetes. Anthropometric (weight, height, waist circumference, hip circumference) and blood pressure measurements were performed using standardized methodologies (standard electronic scale, wall height ruler, tape measure, mercury sphygmoma- nometer). Laboratory analyses considered for risk assess- Page 2 of 11 (page number not for citation purposes) Page 2 of 11 (page number not for citation purposes) BMC Family Practice 2008, 9:53 http://www.biomedcentral.com/1471-2296/9/53 Geographical distribution of general practitioners and patients involved in the study Figure 1 Geographical distribution of general practitioners and patients involved in the study. Data analysis Data on 2,090 patients included in the lists of 40 GPs were transmitted to the ISS. The TCR was assessed by GPs on 45% of those patients considered eligible for application of the CUORE Project risk function (complete information for the risk factors included in the CVD risk assessment, risk factors values within the reference limits, age range 35–69 years, free from previous cardiovascular events). For the remaining 55%, TCR was assessed within the present analysis using the risk factors values collected by GPs. Demographic characteristics (age, educational level and marital status) and the adverse conditions (moderate or low physical activity at work and during leisure time, smoking habit, obesity, hypertension, hypercholestero- lemia and pharmacological treatment) were reported as prevalence. Definitions used in the data analysis of the OEC for hypertension, hypercholesterolemia and diabetes were taken into consideration to make results comparable [11]. The distribution of overweight and obesity was cal- culated considering the following categories: normal weight as body mass index (BMI) 18.5–24.9 kg/m2, over- weight as BMI 25.0–29.9 kg/m2 and obesity as BMI ≥ 30.0 kg/m2 or waist circumference > 88 cm in women and > 102 cm in men. If not differently indicated, results and comments hereinafter refer to BMI. On the basis of the TCR value, patients were classified as low risk (TCR < 3.0%), risk to be kept under control through the adoption of a healthy lifestyle (TCR 3.0– 19.9%), high risk (TCR ≥ 20.0%). Age-adjusted ANOVA models were used to assess mean difference between nor- mal weight and obese, and the relation between BMI (or waist and hip circumferences) and TCR. After checking for homogeneity of variances, comparisons among mean values of men and women were made using the t-test. In absence of homogeneity, the Welch test was used. The independence between categorical variables was assessed using the chi-square test. Mean values were calculated for continuous major cardio- vascular risk factors and consumption of some foods. The frequency of food consumption per week was recorded for the following food items: meat, cheese, fish, pulses, choc- olate, fried food, sweet, wholemeal food, rotisserie food and sugar drink. Daily consumption was asked for fruits and vegetables. Alcohol intake was recorded as grams per day. Results associated to a p-value ≤ 0.05 were considered sta- tistically significant. Quality controls and statistical analy- ses were performed using SAS software version 8.2 (SAS Institute, Cary, NC). Quality controls f The software provided to GPs for data entry offers auto- matic checks for variables congruence (e.g. an informa- tion related to the number of daily smoked cigarettes Page 3 of 11 (page number not for citation purposes) Page 3 of 11 (page number not for citation purposes) BMC Family Practice 2008, 9:53 http://www.biomedcentral.com/1471-2296/9/53 http://www.biomedcentral.com/1471-2296/9/53 http://www.biomedcentral.com/1471-2296/9/53 The 10-year TCR was assessed by GPs through the data entry software using the individual risk score built within the longitudinal study of the Italian CUORE Project [22]. This score allows to estimate the probability of experienc- ing a first cardiovascular event (myocardial infarction, stroke) over the next 10 years knowing the level of eight risk factors for CVD (age, gender, SBP, TC, HDL-C, diabe- tes, smoking habit and use of antihypertensive medica- tion). height, the percentage of terminal zeros was calculated as well; for weight, waist and hip circumferences the termi- nal decimal and whole digits were considered. Homoge- neity of terminal digit distribution indicates a correct application of standard methodologies for risk factors measurements. height, the percentage of terminal zeros was calculated as well; for weight, waist and hip circumferences the termi- nal decimal and whole digits were considered. Homoge- neity of terminal digit distribution indicates a correct application of standard methodologies for risk factors measurements. Data analysis The relation of BMI and other CVD risk factors to con- sumption of some food items, alcohol intake and physical activity was assessed by univariate and multivariate analy- ses. For univariate analysis, age-adjusted Pearson's corre- lation coefficients (r) were calculated to assess relations between couple of variables; for multivariate analysis, multivariate linear regression models were performed considering CVD risk factors as dependent variables and consumption of food items, alcohol intake and physical activity as independent covariates (all models were also age and sex adjusted). In order to compare the effect of independent covariates, standardized coefficients of regression models were calculated (regression coefficient multiplied by the ratio of the standard deviation of the independent variable to the standard deviation of the dependent variable). Only significant coefficients were reported. Page 4 of 11 (page number not for citation purposes) Quality control As the age distribution between men and women was sim- ilar, age adjustments were not needed to compare results among sexes. The completeness of collected information was consid- ered satisfactory. Percentages of recorded information are the following: 98% for blood pressure measurements; 94% for laboratory analyses such as TC, HDL-C and gly- caemia; 98% for anthropometric measurements. Through the questionnaire, smoking habit was registered for 98% of sample, physical activity at work and during leisure time for 70% and 98% respectively, and food consump- tion frequency for 98%; information on medications use were available for 95% of sample. Page 4 of 11 (page number not for citation purposes) Page 4 of 11 (page number not for citation purposes) http://www.biomedcentral.com/1471-2296/9/53 BMC Family Practice 2008, 9:53 Table 1: Demographic characteristics and conditions at risk. The distribution of the terminal whole digit of blood pres- sure readings was found to be highly concentred: more than 50% of readings had zero as terminal digit. The pro- portion of equal readings between first and second meas- urements was 50% for SBP and 40% for DBP. Also for anthropometric measurements, the distributions of the terminal whole and decimal digits were concentrated on zero. itively correlated although not strongly with meat con- Table 1: Demographic characteristics and conditions at risk. Quality control WOMEN MEN N % N % Demographic characteristics Age classes 35–44 247 24 248 24 45–54 260 25 250 24 55–64 265 25 261 25 65–74 274 26 285 27 Marital status Single 116 11 123 12 Married or live-in-partner 770 74 867 83 Separated or divorced 50 5 37 4 Widowed 110 11 15 1 Education College or university 113 11 131 13 High school 306 30 298 29 Secondary school 335 33 404 39 Primary school 275 27 200 19 Conditions at risk Physical activity at work Very low 261 33 250 38 Low 451 57 262 40 Physical activity during leisure time Very low 562 55 440 43 Low 402 40 474 46 Body Mass Index, kg/m2 25.0 – 29.9 348 34 512 50 > = 30.0 234 23 235 23 Waist circumference, cm > 88 for women, > 102 for men 458 45 283 28 Smoking habit 183 18 269 26 Hypertension* 400 39 427 42 Antihypertensive medication 365 37 378 38 Hypercholesterolaemia 332 33 326 32 Lipid-lowering medication 117 12 156 16 Diabetes* 98 10 137 14 Diabetes specific medication 67 7 97 10 Men and women 35–74 years. Systolic blood pressure ≥ 160 mmHg or diastolic blood pressure ≥ 95 mmHg or use of antihypertensive medications. * Total cholesterol ≥ 6.2 mmol/l or use of lipid lowering medications. **Glycemia ≥ 7 mmol/l or use of specific medications for diabetes. Sample characteristics and risk conditions p Table 1 shows the main sample characteristics: more than 70% of women and 80% of men were married or cohab- itant; 41% of women and 42% of men had high school diploma or university degree (Table 1). Ninety percent of women and 78% of men declared to make low or very low physical activity at work (mostly sedentary work or involving standing/walking) (Table 1), 10% of women and 22% of men declared to make high or very high phys- ical activity at work (work involving much walking, man- ual work involving strenuous efforts). Physical activity during leisure time was low or very low (sedentary activity or moderate physical activity for no more than 4 hours a week) for both men and women (Table 1); only 5% of women and 11% of men declared to make high or very high physical activity during leisure time (enjoying sport as a hobby, doing heavy gardening, or exercising regularly for agonistic activity). Prevalence of overweight was 34% in women and 50% in men; 23% of both men and women were obese. Waist cir- cumference exceeded the threshold for obesity in 45% of women and 28% of men. Forty one percent of women and 27% of men were normal weight. Mean value of BMI fell in the overweight category for both men and women (Table 2). Page 5 of 11 (page number not for citation purposes) Dietary and physical activity habits Data on dietary habits collected through administered questionnaire showed that women ate more wholemeal food, vegetables and chocolate compared to men; men had a higher weekly consumption of alcohol, meat, fried food and sugar drinks than women (Table 3). Normal weight persons declared to have a higher average number of meals eaten away from home per week compared to obese persons. No statistically significant differences were found between normal weight and obese persons with regard to frequency of breakfast consumption away from home. Similar results were recorded using waist circum- ference as obesity indicator. Eight percent of normal weight women and 0.4% of obese women declared to make high or very high physical activity during leisure time; among men, 15% and 8% respectively. itively correlated, although not strongly, with meat con- sumption and alcohol intake (r = 0.05 and r = 0.08 respectively), while significant negative correlations were found between BMI and consumption of fish (r = -0.05), pulses (r = -0.11), wholemeal food (r = -0.05), vegetables (r = -0.04), chocolate (r = -0.07), sweets (r = -0.05), rotis- Considering food frequency, alcohol intake and physical activity singly, BMI was found to be significantly and pos- Page 5 of 11 (page number not for citation purposes) Page 5 of 11 (page number not for citation purposes) BMC Family Practice 2008, 9:53 http://www.biomedcentral.com/1471-2296/9/53 Table 2: Mean and standard deviation of blood pressure, blood test and anthropometric measurements. WOMEN MEN N mean standard deviation N mean standard deviation Systolic blood pressure, mmHg 1,016 131.9 17.4 1,027 133.5 16.0 Diastolic blood pressure, mmHg 1,016 79.4 8.9 1,027 81.1 9.2 Serum Total Cholesterol, mmol/l 978 5.5 1.0 984 5.3 1.0 Serum HDL-Cholesterol, mmol/l 978 1.5 0.4 980 1.3 0.3 Glycaemia, mmol/l 978 5.4 1.3 984 5.8 1.6 Weight, kg 1,016 67.5 12.7 1,027 80.7 13.0 Height, cm 1,016 159.2 6.9 1,027 171.2 7.2 BMI, kg/m2 1,016 26.7 5.3 1,027 27.5 4.2 Waist circumference, cm 1,016 87.8 13.0 1,026 96.8 11.4 Hip circumference, cm 1,002 99.0 12.3 1,007 100.9 11.2 Men and women 35–74 years. Table 2: Mean and standard deviation of blood pressure, blood test and anthropometric measurements. Total cardiovascular risk Fruits and vegetables food frequency per day. Alcohol intake expressed in grams. * Significant difference between means of men and women (t-test; p-value ≤ 0.05). Men and women aged 35–74 years. Total cardiovascular risk serie food (r = -0.05) and level of physical activity during leisure time (r = -0.19). Correlation of BMI to physical activity at work and other food items consumption was not statistically significant. Similar results were recorded using waist circumference as obesity indicator. Three percent of sample (48 men and 10 women) declared to have experienced a cerebrovascular or coro- nary event before entering the study. The TCR was calcu- lated for 806 women and 805 men. Mean value of TCR was 4% in women (standard deviation was 5%) and 11% in men (standard deviation was 11%). One percent of women and 16% of men were at high cardiovascular risk (TCR ≥ 20.0% in 10 years). Multivariate regression model confirmed that higher con- sumption of meat was associated with higher levels of BMI, while higher consumption of pulses, chocolate, sweets and rotisserie food was associated with lower levels of BMI (Table 4). Physical activity during leisure time had the strongest inverse impact on BMI (Table 4). Physical activity during leisure time was found to be statistically associated with all risk factors considered with the excep- tion of TC (Table 4); no association was found with phys- ical activity at work. In both men and women the mean value of TCR was asso- ciated with BMI classes (women: 3%, 5%, 7%; men: 8%, 11%, 13% for normal weight, overweight and obese respectively); similar results were recorded considering waist circumference (women: 3% and 6% for non-obese and obese respectively; men: 9% and 14%). In both men and women the mean value of TCR was asso- ciated with BMI classes (women: 3%, 5%, 7%; men: 8%, 11%, 13% for normal weight, overweight and obese respectively); similar results were recorded considering waist circumference (women: 3% and 6% for non-obese and obese respectively; men: 9% and 14%). Thirty percent of obese women versus 78% of normal weight women, and 15% of obese men versus 44% of nor- Table 3: Food frequency per week and alcohol intake. Page 6 of 11 (page number not for citation purposes) Table 3: Food frequency per week and alcohol intake. Total cardiovascular risk WOMEN MEN Median Mean Min Max Median Mean Min Max Meat * 3 3.4 0 8 3 3.7 0 8 Cheese 3 3.3 0 8 3 3.4 0 8 Fish 2 1.7 0 8 2 1.8 0 8 Pulses 2 2.2 0 8 2 2.1 0 8 Chocolate * 1 2.3 0 8 1 1.9 0 8 Fry food * 1 0.8 0 7 1 1.0 0 8 Sweet 1 1.6 0 8 1 1.7 0 8 Wholmeal food * 0 1.9 0 8 0 1.5 0 8 Rotisserie food 0 0.7 0 8 0 0.8 0 7 Sugar drink * 0 0.9 0 8 0 1.1 0 8 Fruits 2 2.5 0 8 2 2.4 0 8 Vegetables * 2 1.9 0 8 1 1.7 0 8 Alcohol intake* 0 8.1 0 83 20 24.6 0 239 Men and women aged 35–74 years. Fruits and vegetables food frequency per day. Alcohol intake expressed in grams. * Significant difference between means of men and women (t-test; p-value ≤ 0.05). Table 3: Food frequency per week and alcohol intake. WOMEN MEN Median Mean Min Max Median Mean Min Max Meat * 3 3.4 0 8 3 3.7 0 8 Cheese 3 3.3 0 8 3 3.4 0 8 Fish 2 1.7 0 8 2 1.8 0 8 Pulses 2 2.2 0 8 2 2.1 0 8 Chocolate * 1 2.3 0 8 1 1.9 0 8 Fry food * 1 0.8 0 7 1 1.0 0 8 Sweet 1 1.6 0 8 1 1.7 0 8 Wholmeal food * 0 1.9 0 8 0 1.5 0 8 Rotisserie food 0 0.7 0 8 0 0.8 0 7 Sugar drink * 0 0.9 0 8 0 1.1 0 8 Fruits 2 2.5 0 8 2 2.4 0 8 Vegetables * 2 1.9 0 8 1 1.7 0 8 Alcohol intake* 0 8.1 0 83 20 24.6 0 239 http://www.biomedcentral.com/1471-2296/9/53 BMC Family Practice 2008, 9:53 Table 4: Relation between food frequency, physical activity, alcohol intake and main cardiovascular risk factors. Total cardiovascular risk SBP DBP TC HDL-C Glycemia BMI Waist Std Coeff Std Coeff Std Coeff Std Coeff Std Coeff Std Coeff Std Coeff Meat 0.061* 0.052* 0.051* Cheese Fish -0.063* Pulses -0.051* -0.074 Chocolate 0.122* -0.050* -0.071 -0.077 Fry food -0.057* 0.056* Sweet -0.071** -0.056* -0.055* Wholmeal food -0.065** Rotisserie food -0.058* -0.072 Sugar drink -0.075 0.052* Fruits -0.078** Vegetables 0.062* Physical activity at work Leisure-time physical activity -0.049* -0.074 0.132* -0.066 -0.187* -0.195* Alcohol intake 0.083 0.073 0.146* 0.064* 0.078 Men and women aged 35–74 years. Variables on the top of the columns are considered as dependent variables. Food intake, physical activity and alcohol intake are considered as independent covariates in each model. The seven models are also age and sex adjusted. Std Coeff is the standardized coefficient of multivariate linear regression model. Significance: * p < 0.001, ** p < 0.01, * p < 0.05. All food frequencies are per week except for fruits and vegetables (per day). Alcohol intake expressed in grams. SBP: systolic blood pressure (mmHg); DBP: diastolic blood pressure (mmHg); HDL-C: HDL-cholesterol (mmol/l); glycaemia: fasting plasma glucose (mmol/l); BMI: body mass index (kg/m2); waist: waist circumference (cm). Table 4: Relation between food frequency, physical activity, alcohol intake and main cardiovascular risk factors. Table 4: Relation between food frequency, physical activity, alcohol intake and main cardiovascular risk factors. side and an increasingly sedentary lifestyle on the other side appear to be the main factors behind this situation. Many degenerative diseases are associated with obesity: diabetes, hypertension, ischemic heart disease, heart fail- ure, some types of cancers (endometrial, colorectal, kid- ney, gall bladder and breast cancer in post-menopausal women), and osteoarticular pathologies. Obesity has a rather complicated association with cardiovascular risk: some researchers argue that this association is indirect and mal weight men were at low cardiovascular risk (Figure 2). Four percent of obese women versus 0.3% of normal weight women, and 20% of obese men versus 10% of nor- mal weight men were at high cardiovascular risk. http://www.biomedcentral.com/1471-2296/9/53 Given the availability of data, it was worth studying die- tary habits, physical activity and relation of these with CVD risk factors, despite this went beyond the main aim of the study. Since cross-sectional design prevents deter- mination of causality and assessment of lifestyle changes' effects (diet, physical activity) over time, analyses of data focused on association of cardiovascular risk factors with food frequency and overweight/obesity, underlying differ- ences between men and women. Results showed that men had a greater tendency to be overweight compared to women, while obesity is a problem affecting both gen- ders. Compared to men, women seemed to care more about their diet, eating healthy food, such as wholemeal food and vegetables, more frequently than did men. On the other hand, men ate food rich in saturated fats such as meat, fried food and sugar drinks more frequently than did women. Persons with lower BMI ate sweets, rotisserie food and chocolate more frequently than did overweight persons and declared to eat away from home more fre- quently. This behaviour may indicate either that persons with high BMI or waist circumference knew their weight problems and were willing to solve them or that persons overweight/obese had the tendency to not declare their unhealthy dietary habits. Although fish, wholemeal food and vegetables consumption resulted not significant when adjusted for all other foods consumption, a more frequent consumption of these food items was found to be correlated with persons having a lower BMI or waist cir- cumference. Physical activity during leisure time was a better discriminator than physical activity at work for assessing the tendency to overweight or obesity; however, both types of physical activity were low or very low. Men made more high or very high physical activity during lei- sure time compared to women. Alcohol intake was found to be associated with most cardiovascular risk factors and, together with physical activity during leisure time, was one of the most important factors within the multivariate analyses. Alcohol consumption was higher for men than for women. All things considered, initiatives aiming at monitoring trends in national obesity and promoting prevention actions at local level are of strategic importance. In 1984, the Center for Disease Control and Prevention (CDC) of USA developed the Behavioral Risk Factor Surveillance System to collect monthly in all 50 states information on health risk behaviours and preventive health practices [26]. http://www.biomedcentral.com/1471-2296/9/53 In Europe, the European Physical Activity Surveil- lance System (EUPASS) has been implemented to contrib- ute to the establishment of an appropriate, cost-effective and feasible Community health monitoring system to be achieved by developing and testing a surveillance system, using physical activity, being a major indicator of public health, as its substantive focus and starting point [27]. In UK, the Counterweight programme developed and tested a weight management intervention programme in pri- mary care [28,29]. Italy still lacks a valid national system of surveillance to track overweight, obesity and other cardiovascular risk factors. For this reason, a pilot project was implemented with the aim of evaluating the feasibility of a national sur- veillance network for obesity and other cardiovascular risk factors. Forty GPs were engaged to collect their patients' data on dietary habits, physical activity, use of medica- tions, smoking habit, major cardiovascular risk factors level and anthropometric measurements using standard- ized methodologies. The completeness of collected information on routinely performed measurements and laboratory assays, as well as information collected through questionnaire such as die- tary habits, physical activity level and smoking habit was considered satisfactory. Younger persons were easy to enrol, in fact the homogeneity in age distribution foreseen in the protocol was not altered. The parameters used for the analysis of reliability of collected data referred to qual- ity controls performed in the epidemiological field, where the accuracy of measurements represents a crucial bond for reliability of estimates [23]. In this study, the analyses on the distribution of terminal digit of blood pressure readings and anthropometric measurements represented a valid self-monitoring tool for GPs using standardized methodologies. However, a good reliability of collected information was confirmed by the significance found in known associations (dietary habits, physical activity, BMI, waist circumference, cardiovascular risk factors) and by homogeneity of results with those obtained through OEC survey [11]. The level of 10-year TCR calculated through the CUORE Project function ranged in the risk category to be kept under control through the adoption of healthy lifestyle (risk category between 3.0% and 19.9%) for both men and women. The likelihood for normal weight persons to be low risk was almost four times higher compared to obese persons. It is not easy to establish to what extent patients included in this study are representative of the Italian patient pop- ulation. http://www.biomedcentral.com/1471-2296/9/53 http://www.biomedcentral.com/1471-2296/9/53 http://www.biomedcentral.com/1471-2296/9/53 BMC Family Practice 2008, 9:53 Given the availability of data, it was worth studying die- tary habits, physical activity and relation of these with CVD risk factors, despite this went beyond the main aim of the study. Since cross-sectional design prevents deter- mination of causality and assessment of lifestyle changes' effects (diet, physical activity) over time, analyses of data focused on association of cardiovascular risk factors with food frequency and overweight/obesity, underlying differ- ences between men and women. Results showed that men had a greater tendency to be overweight compared to women, while obesity is a problem affecting both gen- ders. Compared to men, women seemed to care more about their diet, eating healthy food, such as wholemeal food and vegetables, more frequently than did men. On the other hand, men ate food rich in saturated fats such as meat, fried food and sugar drinks more frequently than did women. Persons with lower BMI ate sweets, rotisserie food and chocolate more frequently than did overweight persons and declared to eat away from home more fre- quently. This behaviour may indicate either that persons with high BMI or waist circumference knew their weight problems and were willing to solve them or that persons overweight/obese had the tendency to not declare their unhealthy dietary habits. Although fish, wholemeal food and vegetables consumption resulted not significant when adjusted for all other foods consumption, a more frequent consumption of these food items was found to be correlated with persons having a lower BMI or waist cir- cumference. Physical activity during leisure time was a better discriminator than physical activity at work for assessing the tendency to overweight or obesity; however, both types of physical activity were low or very low. Men made more high or very high physical activity during lei- sure time compared to women. Alcohol intake was found to be associated with most cardiovascular risk factors and, together with physical activity during leisure time, was one of the most important factors within the multivariate analyses. Alcohol consumption was higher for men than for women. dependent on an increased prevalence of diabetes, hyper- tension and dysplipidemia, while others consider obesity as an independent risk factor for CVD [24,25]. dependent on an increased prevalence of diabetes, hyper- tension and dysplipidemia, while others consider obesity as an independent risk factor for CVD [24,25]. Discussion Obesity represents a threat to health which is peculiar, although not exclusive, to the so-called 'well-being socie- ties'. A constant increase in the availability of food on one Distribution of Total Cardiovascular Risk (TCR) by Body Mass Index categories Figure 2 Distribution of Total Cardiovascular Risk (TCR) by Body Mass Index categories. Men and women 35–69 years without previous cardiovascular disease. Distributio Figure 2 Distribution of Total Cardiovascular Risk (TCR) by Body Mass Index categories Figure 2 Distribution of Total Cardiovascular Risk (TCR) by Body Mass Index categories. Men and women 35–69 years without previous cardiovascular disease. Page 7 of 11 (page number not for citation purposes) http://www.biomedcentral.com/1471-2296/9/53 To ensure effective surveillance of prevention of obesity and CVD, the activity of GPs should be accompanied by other initiatives. As suggested by an increasing number of reports and as pointed out by Benjamin Caballero from The Johns Hopkins University also pointed out in his glo- bal epidemic of obesity overview, obesity is not only a dis- order of individual behaviour but it is highly influenced by socioeconomic environment [41]. This is why preven- tion needs to take place simultaneously at the individual level and at the population level. Good social support can help give people the emotional and practical resources they need, particularly for coping with difficult life transi- tions. Overall, the picture that emerged from this pilot project is positive, showing a population with mostly healthy, but certainly improvable, dietary habits (mean consumption of meat 3/4 times per week, fish and pulses twice per week, two portions of vegetables and fruits per day, small amount of fried food, acceptable mean level of alcohol intake per day) and a mean cardiovascular risk level not exceeding the threshold of high risk. The WHO Regional Office for Europe, within the Strategy for the Prevention and Control of Noncommunicable Diseases, on the occasion of the Copenhagen meeting held on 12 September 2006, indicated that effective interventions can include: control of the quality of consumer information about certain products, such as food, tobacco, alcohol; cooperation with industry through voluntary or enforced agreements to reduce levels of added salt, fat and sugars in manufactured foods; regulations within the built environ- ment, for example through health impact assessment of transport and urban development proposals in order to promote walking and cycling [42]. The objectives of the strategy should be combined with integrated action on risk factors and their underlying determinants across sec- tors, with efforts to strengthen health systems toward improved prevention and control. Obesity prevalence should not be underestimated and is probably due to food quantity rather than food quality, and to a mostly sedentary lifestyle. Monitoring body weight and risk factors levels represent a first step towards persuading the general population to implement preven- tive actions against the onset of diseases leading to wors- ening quality of life or even reduction in the years of life. Obesity has its roots in the common way of thinking, which must be changed through lifestyle interventions targeted to the general population. http://www.biomedcentral.com/1471-2296/9/53 BMC Family Practice 2008, 9:53 http://www.biomedcentral.com/1471-2296/9/53 kg/m2 for men, SBP 135 mmHg for women and 132 mmHg for men, HDL-C 1.5 mmol/l for women and 1.3 mmol/l for men). Mean values of TC, glycemia, hyperten- sion and hypercholesterolemia were higher for both gen- ders (OEC values: TC 5.5 mmol/l for women and 5.3 mmol/l for men, glycaemia 4.8 mmol/l for women and 5.2 mmol/l for men, hypertension 31% for women and 33% for men, hypercholesterolemia 25% for women and 21% for men) [11]. The opposite was found for smoking habit: OEC showed higher prevalence compared to this study (OEC values: 21% for women and 30% for men). Physical inactivity prevalence was found to be higher in this study, especially for women. Differences could be ascribed to the different period of screening (OEC survey was performed at the end of 1990s); this is particularly true for smoking prevalence, which has been changing after the introduction of the Italian anti-smoking legisla- tion in January 2005 banning smoking in public places. kg/m2 for men, SBP 135 mmHg for women and 132 mmHg for men, HDL-C 1.5 mmol/l for women and 1.3 mmol/l for men). Mean values of TC, glycemia, hyperten- sion and hypercholesterolemia were higher for both gen- ders (OEC values: TC 5.5 mmol/l for women and 5.3 mmol/l for men, glycaemia 4.8 mmol/l for women and 5.2 mmol/l for men, hypertension 31% for women and 33% for men, hypercholesterolemia 25% for women and 21% for men) [11]. The opposite was found for smoking habit: OEC showed higher prevalence compared to this study (OEC values: 21% for women and 30% for men). Physical inactivity prevalence was found to be higher in this study, especially for women. Differences could be ascribed to the different period of screening (OEC survey was performed at the end of 1990s); this is particularly true for smoking prevalence, which has been changing after the introduction of the Italian anti-smoking legisla- tion in January 2005 banning smoking in public places. determine to what extent this choice influenced the study success. Preventing overweight and obesity could avoid further increase of health care use and related costs: Liset van Dijk et al. study [40] evaluated GPs consultations in moderate and severely overweight persons in Netherlands and reported that obese persons, regardless of different gender, age, social status and lifestyle, are more likely to consult their GP. Page 9 of 11 (page number not for citation purposes) http://www.biomedcentral.com/1471-2296/9/53 As for cardiovascular risk factors comparison, mean levels of risk factors and prevalence of risk condi- tions were in line with results obtained through OEC sur- vey: mean BMI, SBP, HDL-C values were quite similar to those reported by OEC (BMI 26 kg/m2 for women and 27 Page 8 of 11 (page number not for citation purposes) Page 8 of 11 (page number not for citation purposes) References References 1. Kesteloot H, Sans S, Kromhout D: Evolution of all-causes and cardiovascular mortality in the age-group 75–84 years in Europe during the period 1970–1996. A comparison with worldwide changes. European Heart Journal 2002, 23(5):384-398. g p J ( ) 2. Kesteloot H, Sans S, Kromhout D: Dynamics of cardiovascular and all-cause mortality in Western and Eastern Europe between 1970 and 2000. European Heart Journal 2006, 27(1):107-113. ( ) 3. Poirier P, Giles TD, Bray GA, Hong Y, Stern JS, Pi-Sunyer FX, Eckel RH: Obesity and cardiovascular disease: pathophysiology, evaluation, and effect of weight loss. Arterioscler Thromb Vasc Biol 2006, 26(5):968-76. Conclusion ( ) 4. Eckel RH, Krauss RM: American Heart Association call to action: obesity as a major risk factor for coronary heart dis- ease. AHA Nutrition Committee. Circulation 1998, 97(21):2099-100. The active role of GPs in the process of prevention repre- sents a potential key to success. GPs have the primary and complete responsibility toward their patients, contribute to the development of individual assistance plan and rec- ommend specialist visits when necessary. This study has demonstrated the feasibility of a GPs surveillance network for obesity and CVD in Italy. It may support informative campaigns on how people can adopt healthy diet and increase daily physical activity; it also represents a good way to monitor overweight, obesity and CVD, and a great step toward assuring a good health in the future. ( ) 5. Department of Health: Choosing Health: Making Healthier Choices Easier. Cm 6374. The Stationery Office: London; 2004. 6. Italian Ministry of Health [http://www.ministerosalute.it/stiliVita/ stiliVita.jsp] 7. Healthy people 2010 [http://www.healthypeople.gov/] y p p [ p yp p g ] 8. Public Health programme 2003–2008 [http://ec.europa.eu/ health/ph_programme/programme_en.htm] 9. International Conference on Health Benefits of Mediterra- nean Style Diet [http://www.aprifel.com/_pdf/ MinutesEGEA_2005_1.pdf] 10. ISTAT: Condizioni di salute, fattori di rischio e ricorso ai servizi sanitari. [http://www.istat.it/salastampa/comunicati/ non_calendario/20070302_00/]. ] 11. Osservatorio Epidemiologico Cardiovascolare Italiano. Ital Heart J 2004, 5(Suppl 3):49S-92S. 11. Osservatorio Epidemiologico Cardiovascol Competing interests Heart J 2004, 5(Suppl 3):49S-92S. 12. Progetto regione Piemonte "Sorveglianza e prevenzione dell'obesità "Progetto 4-2" [http://www.epicentro.iss.it/territo rio/pdf/report_intermedio.pdf] The authors declare that they have no competing interests. http://www.biomedcentral.com/1471-2296/9/53 http://www.biomedcentral.com/1471-2296/9/53 CUORE Project http://www.cuore.iss.it, helps GPs to set up a data archive, to monitor cardiovascular risk and risk factors trends in patients over time, to keep a record of the prescribed therapies and lifestyle recommendations given to patients, to collect data related to obesity and over- weight, and to print out the patient's cardiovascular risk score, including some recommendations for lifestyle and therapy, if prescribed. The Observatory of Cardiovascular Risk (ORC), coordinated by the ISS, pools data collected by GPs through the cuore.exe, supports GPs with quality control, and disseminate results for monitoring TCR and risk factors by genders, age and geographical area and evaluating the efficacy of prevention action in primary care. Preliminary data of the ORC are under evaluation. Carlo Fedele Marulli, Massimo Mazzilli, Gerardo Medea, Giovanni Merlino, Antonio Metrucci, Raffaella Micheli, Vito Milazzo, Rosalba Murgia, Giuseppe Noberasco, Romano Paduano, Italo Paolini, Damiano Parretti, Domenico Pasculli, Salvatore Pasqualetto, Andrea Salvetti, Aurelio Sessa, Alfonso Toz- zoli, Cesare Filippo Urru, Paolo Valle, Domenico Valletta. Carlo Fedele Marulli, Massimo Mazzilli, Gerardo Medea, Giovanni Merlino, Antonio Metrucci, Raffaella Micheli, Vito Milazzo, Rosalba Murgia, Giuseppe Noberasco, Romano Paduano, Italo Paolini, Damiano Parretti, Domenico Pasculli, Salvatore Pasqualetto, Andrea Salvetti, Aurelio Sessa, Alfonso Toz- zoli, Cesare Filippo Urru, Paolo Valle, Domenico Valletta. http://www.biomedcentral.com/1471-2296/9/53 Little rigorous research study exist on how effective train- ing is in helping primary care teams to manage obesity more effectively [30]; nevertheless, several studies showed that GPs have a significant role to play in preventing and diagnosing weight problems and in improving initial counselling [31,32]. It is a fact that they are often con- sulted [33] and most patients believe that their GPs could help them lose weight [34]. A cross sectional telephone survey conducted in 2002 in France on 600 GPs reported that most of them regarded obesity as a disease and agreed that their role includes weight problem management but complain lack of time as a frequent problem [35], which is confirmed also by other studies [36,37], and generally known to be a significant barrier to preventive care in gen- eral practice [38,39]. In our pilot project, a random selec- tion of GPs was not feasible and it is now impossible to In Italy, beside population-based interventions imple- mented within the Project 'Guadagnare Salute', the CCM, in collaboration with ISS, has launched a national train- ing plan aimed at increasing GPs' awareness about the key role of cardiovascular prevention, in order to implement an Italian national surveillance system following the objectives of the pilot study under consideration. The new skills acquired through the training course will translate into health objectives for the general population and for individuals: lifestyle changes among individuals at high risk and general population; reduction in the frequency of risk conditions and in the mean level of single risk factors (an exhaustive session is dedicated to overweight, obesity, nutrition and physical activity). The software cuore.exe, downloadable free of charge from the website of the Page 9 of 11 (page number not for citation purposes) BMC Family Practice 2008, 9:53 http://www.biomedcentral.com/1471-2296/9/53 General Practitioners involved in the pilot study: Authors' contributions 13. Fontana G, Cristofori M, Cattaneo C, Minardi V, Bella A, Perra A e Binkin N, on behalf of Gruppo Epidemiologia in Azione 2007: Stato nutrizionale dei bambini delle scuole primarie nella provin- cia di Terni e fattori correlati. Ben Notiziario ISS September 2007, 20(9): [http://www.epicentro.iss.it/ben/2007/settembre/1.asp]. CD: performed statistical analysis, interpretation of data and made substantial contribution in drafting the manu- script. CL: made substantial contribution to acquisition of data. OB: was involved in revising the manuscript criti- cally. GR: was involved in revising the manuscript criti- cally. PC: was involved in drafting the manuscript and making language revision. FD: made contribution to acquisition of data. LP: was involved in revising the man- uscript critically. SG: made substantial contribution in conception and design, and revised the manuscript criti- cally. All authors read and approved the final manuscript. 20(9): [http://www.epicentro.iss.it/ben/2007/settembre/1.asp 14. Monitoraggio e prevenzione dell'obesità nelle donne in età fertile e nel bambino in Sardegna [http://www.epicentro.iss.it/ problemi/obesita/presentazioni_Sian/sessione3/Arras.pdf] 15. Il progetto di sorveglianza nutrizionale nella Regione Emilia Romagna: valutazione per la progettazione di interventi educativi nella scuola [http://www.epicentro.iss.it/problemi/obes ita/presentazioni_Sian/sessione3/Fabbri.pdf] 16. Uno studio multicentrico dello stato nutrizionale della popo- lazione infantile nel Veneto [http://www.epicentro.iss.it/prob lemi/obesita/presentazioni_Sian/sessione3/Stano.pdf] p p ] 17. Mezzani L, Patussi V, Rossi A, Russo R, Scafato E: Establishing an italian general practitioner brief intervention pilot project for problem drinkers. Subst Use Misuse 2007, 42(12):1979-89. Acknowledgements The Obesity Project Research Group Arch Intern Med 167(6):537-9. 2007 Mar 26 25. Sowers JR: Obesity as a cardiovascular risk factor. Am J Med 2003, 115(Suppl 8A):37S-41S. 26. Behavioral Risk Factors Surveillance System [http:// www.cdc.gov/BRFSS/] 27. Rütten A, Vuillemin A, Ooijendijk WT, Schena F, Sjöström M, Stahl T, Auweele Y Vanden, Welshman J, Ziemainz H: Physical activity monitoring in Europe. The European Physical Activity Sur- veillance System (EUPASS) approach and indicator testing. Public Health Nutr 2003, 6(4):377-84. ( ) 28. Hudson AJ: The counterweight programme: tackling obesity in primary care. Prof Nurse 2004, 19(9):484-6. 29. McQuigga M, Browna JE, Broomb J, Lawsc RA, Recklessd JPD, Noblee PA, Kumare S, McCombief EL, Leanf MEJ, Lyonsg FG, Frosth GS, Quinni MF, Barthj JH, Haynesk SM, Finerl N, Haslamm D, Rossn HM, Holeo DJ, Radziwoniko S: The Counterweight programme: Prevalence of CVD risk factors by body mass index and the impact of 10% weight change. Obesity Research & Clinical Practice 2008, 2:15-27. 30. Harvey EL, Glenny AM, Kirk SF, Summerbell CD: A systematic review of interventions to improve health professionals' management of obesity. Int J Obes Relat Metab Disord 1999, 23:1213-22. 31. Basdevant A, Laville M, Ziegler O: Recommendations for the diagnosis, the prevention and the treatment of obesity. Dia- betes Metab 2002, 28:146-50. 32. National Institutes of Health, National Heart Lung and Blood Institute: Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults. Bethesda MD: National Institutes of Health; 1998. 33. Cherry DK, Burt CW, Woodwell DA: National ambulatory med- ical care survey: 2001 summary. Adv Data 2003, 337:1-44. 34. Potter MB, Vu JD, Croughan-Minihane M: Weight management: what patients want from their primary care physicians. J Fam Pract 2001, 50:513-8. 35. Bocquie A, Verger P, Basdevant A, Andreotti G, Baretge J, Villani P, Paraponaris A: Overweight and Obesity: Knowledge, Atti- tudes, and Practices of General Practitioners in France. Obesity Research 2005, 13(4):787-795. y ( ) 36. Timmerman GM, Reifsnider E, Allan JD: Weight management practices among primary care providers. J Am Acad Nurses Pract 2000, 12:113-6. 37. Pratt CA, Nosiri UI, Pratt CB: Michigan physicians' perceptions of their role of managing obesity. Percept Mot Skills 1997, 84:848-50. 38. Ward JE, Gordon J, Sanson-Fisher RW: Strategies to increase preventive care in general practice. Med J Aust 1991, 154:523-526-31. 39. Roland MO, Bartholomew J, Courtenay MJ, Morris RW, Morrell DC: The "five minute" consultation: effect of time constraint on verbal communication. Acknowledgements The Obesity Project Research Group Br Med J 1986, 292:874-6. Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Page 11 of 11 (page number not for citation purposes) Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Publish with BioMed Central and every scientist can read your work free of charge J 40. van Dijk L, Otters HB, Schuit AJ: Moderately overweight and obese patients in general practice: a population based sur- vey. BMC Fam Pract 2006, 7:43. 41. Caballero B: The global epidemic of obesity: an overview. Epi- demiol Rev 2007, 29:1-5. 41. Caballero B: The global epidemic of obesity: an overview. Epi- demiol Rev 2007, 29:1-5. 42. Regional Committee for Europe: Gaining health: The European Strategy for the Prevention and Control of Noncommunica- ble Diseases. Copenhagen meeting . Acknowledgements The Obesity Project Research Group 18. Ciaralli F, Rovetta S, Rebella V, Borgia P, Molino N, Chini F, Zuccaro P, D'Urso A, Panà A, Gruppo di studio MMG: Feasibility of behav- ioural interventions for tobacco smoking cessation in rou- tine general practice. Ig Sanita Pubbl 2006, 62(6):653-64. Simona Giampaoli, Ovidio Brignoli, Gabriele Riccardi, Cinzia Lo Noce, Chi- ara Donfrancesco, Paola Ciccarelli, Francesco Dima, Luigi Palmieri. g p g ( ) 19. Bertozzi N, Bakken E, Bolognesi M, Castoldi F, Massarini M, Palazzi M, Pietrantoni L, Righi F, Vitali P: Promoting physical activity in overweight and obese patients: counseling in primary care from Italy (Cesena, 2002–2003). Ann Ig 2004, 16(6):745-51. General Practitioners involved in the pilot study: 20. y ( ) g ( ) 20. Bruno M, Marciello A, Perosa P, Quaglia M, Reina E, Rizzuto A, Mal- cangi U: Early diagnosis and prevention of renal disease: col- laboration between nephrologists and general practitioners. G Ital Nefrol 2007, 24(Suppl 38):80-2. Riccardo Agati, Franco Bagagli, Giuseppe Bernardi, Alberto Bozzani, Ovidio Brignoli, Bernardino Bruno, Luigi Canciani, Mario Celebrano, Giulio Corg- atelli, Antonio Pompeo Coviello, Gaetano D'Ambrosio, Vittorio Di Carlo, Simonetta Di Cristoforo, Giuliano Ermini, Alessandro Filippi, Bruno Franco Novelletto, Giuseppe Galli, Fabrizio Germini, Shafik Kurtam, Luciano Lippa, f ( pp ) 21. Palmieri L, Panico S, Vanuzzo D, et al.: per il Gruppo di ricerca del Progetto CUORE. La valutazione del rischio cardiovascolare globale assoluto: il punteggio individuale del Progetto CUORE. Ann Ist Super Sanità 2004, 40(4):. ( pp ) 21. Palmieri L, Panico S, Vanuzzo D, et al.: per il Gruppo di ricerca del Progetto CUORE. La valutazione del rischio cardiovascolare globale assoluto: il punteggio individuale del Progetto CUORE. Ann Ist Super Sanità 2004, 40(4):. Page 10 of 11 (page number not for citation purposes) BMC Family Practice 2008, 9:53 http://www.biomedcentral.com/1471-2296/9/53 http://www.biomedcentral.com/1471-2296/9/53/prepub 22. Giampaoli S, Palmieri L, Donfrancesco C, Panico S, Vanuzzo D, Pilotto L, Ferrario M, Cesana G, Mattiello A, on behalf of The CUORE Project Reseach Group: Cardiovascular risk assessment in Italy: the CUORE Project risk score and risk chart. Italian Journal of Public Health 2007, 4(2):102-109. Year 5 23. Evans A, Salomaa V, Kulathinal S, Asplund K, Cambien F, Ferrario M, Perola M, Peltonen L, Shields D, Tunstall-Pedoe H, Kuulasmaa K, for the MORGAM Project: MORGAM (an international pooling of cardiovascular cohorts). Int J Epidemiol 2005, 34:21-27. ) J p 24. Stamler J: Low risk–and the "No more than 50% " myth/ dogma. Pre-publication history p The pre-publication history for this paper can be accessed here: The pre-publication history for this paper can be accessed here:
https://openalex.org/W1868118740	http://pure-oai.bham.ac.uk/ws/files/22607106/Fintelman_et_al_Numerical_Investigation_Engineering_Applications_2015.pdf	English	null	A numerical investigation of the flow around a motorbike when subjected to crosswinds	Engineering applications of computational fluid mechanics	2,015	cc-by	8,831	Publisher Rights Statement: Publisher Rights Statement: This is an open-access article distributed under the terms of the Creative Commons Attribution License htt // ti /li /b /4 0/htt // ti /li /b /4 0/ hi h it g This is an open-access article distributed under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/, which permits un reproduction in any medium provided the original work is properly cited g This is an open-access article distributed under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. This is an open-access article distributed under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/, which permits un reproduction in any medium, provided the original work is properly cited. http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/, which permits unrestricted us reproduction in any medium, provided the original work is properly cited. Checked October 2015 A numerical investigation of the flow around a motorbike when subjected to crosswinds Fintelman, Danielle; Hemida, Hassan; Sterling, Mark; Li, Francois-Xavier DOI: 10.1080/19942060.2015.1071524 License: Creative Commons: Attribution (CC BY) Document Version Publisher's PDF, also known as Version of record Citation for published version (Harvard): Fintelman, D, Hemida, H, Sterling, M & Li, F-X 2015, 'A numerical investigation of the flow around a motorbike when subjected to crosswinds', Journal of Engineering Applications of Computational Fluid Mechanics, vol. 9, no. 1, pp. 528-542. https://doi.org/10.1080/19942060.2015.1071524 Citation for published version (Harvard): Fintelman, D, Hemida, H, Sterling, M & Li, F-X 2015, 'A numerical investigation of the flow around a motorbike when subjected to crosswinds', Journal of Engineering Applications of Computational Fluid Mechanics, vol. 9, no. 1, pp. 528-542. https://doi.org/10.1080/19942060.2015.1071524 Link to publication on Research at Birmingham portal Link to publication on Research at Birmingham portal Publisher Rights Statement: General rights l li General rights Unless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or the copyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposes permitted by law. •Users may freely distribute the URL that is used to identify this publication. •Users may freely distribute the URL that is used to identify this publication. •Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private study or non-commercial research. y p oad and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private mercial research y •User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and P •Users may not further distribute the material nor use it for the purposes of commercial gain. Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document. Where a licence is displayed above, please note the terms and conditions of the licence govern your use of th When citing, please reference the published version. Take down policy y ty of Birmingham exercises care and attention in making items available there are rare occasions when an item has bee or has been deemed to be commercially or otherwise sensitive. Take down policy While the University of Birmingham exercises care and attention in making items available there are rare occasions wh uploaded in error or has been deemed to be commercially or otherwise sensitive. this is the case for this document, please contact UBIRA@lists.bham.ac.uk providing details and we will remove access ely and investigate. If you believe that this is the case for this document, please contact UBIRA@lists.bham.ac.uk providing details and we the work immediately and investigate. Download date: 24. Oct. 2024 Date: 29 October 2015, At: 08:48 Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tcfm20 1. Introduction Researchers have used wind tunnel experiments to investigate and optimize the stability and aerodynamic performance of motorbikes and many of their individual components. For example, the eﬀect of handlebar fairing and windshield on the stability of a full-scale motorbike was investigated by Cooper (1983), while Bridges and Russell (1987) studied the eﬀect of a ‘topbox’ on the sta- bility of a motorbike. In the paper of Araki and Gotou (2001), the aerodynamic characteristics of diﬀerent motor- bikes were compared and an outline of the aerodynamic development of motorbikes by means of wind tunnels pre- sented. Whereas the three aforementioned studies focused on aerodynamic performance and stability, limited data is available on the forces experienced by motorbikes in cross- winds. Ubertini and Desideri (2002) experimentally mea- sured the aerodynamic forces and moments on a scooter and rider at diﬀerent yaw angles up to 10° (with the yaw angle deﬁned as the angle the crosswind makes relative to the direction of travel). It was observed that the aerody- namic drag force coeﬃcients increased by approximately 16% as a result of the increasing yaw angle. The side forces and yawing moments tended to increase linearly with the yaw angle. Although this research provided an insight into the aerodynamic eﬀects (albeit over a reduced range of yaw angles), due to the nature of the experiments, no infor- mation pertaining to the instantaneous ﬂow structures and pressure distributions were obtained. There are an estimated 200 million motorbikes around the world (Shuhei, 2006). The motorbike and rider experience diﬀerent aerodynamic forces and moments when travel- ling along roads. Scibor-Rylski and Sykes (1984) state that improving the aerodynamic performance of a rider is an important factor in reducing fuel consumption and improving motorbike maneuverability. In addition, cross- winds have the potential to severely inﬂuence the stability of the motorbike and rider, due to increased aerodynamic forces (Cheli, Bocciolone, Pezzola, & Leo, 2006). These forces include side forces, rolling, pitching, and yawing moments, in addition to the normal aerodynamic drag and lift forces. Despite the several reported fatal accidents due to the eﬀects of crosswinds, aerodynamic research relat- ing to crosswinds is rather limited (Carr, 2011; Donell, 2010; Gauger, 2013). A numerical investigation of the ﬂow around a motorbike when subjected to crosswinds D. Fintelmana∗, H. Hemidab, M. Sterlingb and F-X. Lia aSchool of Sport, Exercise and Rehabilitation Sciences, University of Birmingham, UK; bSchool of Civil Engineering, University of Birmingham, UK (Received 3 April 2013; ﬁnal version received 25 June 2015) A numerical investigation of the ﬂow around a motorbike when subjected to crosswinds D. Fintelmana∗, H. Hemidab, M. Sterlingb and F-X. Lia aSchool of Sport, Exercise and Rehabilitation Sciences, University of Birmingham, UK; bSchool of Civil Engineering, University of Birmingham, UK (Received 3 April 2013; ﬁnal version received 25 June 2015) Crosswinds have the potential to inﬂuence the stability and therefore the safety of a motorbike rider. Numerical computations using both delayed detached-eddy simulations (DDES) and Reynolds-Averaged Navier-Stokes (RANS) were employed to investigate the ﬂow around a motorbike subjected to crosswinds with yaw angles of 15, 30, 60 and 90 degrees. The Reynolds number was 2.2 million, based on the crosswind velocity and the height of the rider from the ground. The aerodynamic force coeﬃcients and ﬂow structures around the motorbike and rider were obtained and analysed. Although both DDES and RANS provided comparable overall aerodynamic forces, RANS failed to predict both the DDES surface pressures at the separation regions and the location and size of the main circulation region. The DDES results showed that the drag coeﬃcients decrease with increasing yaw angles, while the side force coeﬃcients signiﬁcantly increase. It was found that increasing yaw angles result in stronger vortex shedding around the windshield and helmet. Downloaded by [University of Birmingham] at 08:48 29 October 2015 Keywords: motorbike; crosswind; DDES; RANS; aerodynamic forces; ﬂow structures © 2015 The Author(s). Published by Taylor & Francis. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. D. Fintelman, H. Hemida, M. Sterling & F-X. Li To cite this article: D. Fintelman, H. Hemida, M. Sterling & F-X. Li (2015) A numerical investigation of the flow around a motorbike when subjected to crosswinds, Engineering Applications of Computational Fluid Mechanics, 9:1, 528-542, DOI: 10.1080/19942060.2015.1071524 © 2015 The Author(s). Published by Taylor & Francis Published online: 11 Sep 2015. Submit your article to this journal Article views: 75 View related articles View Crossmark data Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tcfm20 Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tcfm20 Download by: [University of Birmingham] Engineering Applications of Computational Fluid Mechanics, 2015 Vol. 9, No. 1, 528–542, http://dx.doi.org/10.1080/19942060.2015.1071524 Engineering Applications of Computational Fluid Mechanics, 2015 Vol. 9, No. 1, 528–542, http://dx.doi.org/10.1080/19942060.2015.1071524 © 2015 The Author(s). Published by Taylor & Francis. 1. Introduction In contrast, crosswind research on other road vehicles (e.g., cars and lorries) is common (Baker et al., 2009; Cheli, Belforte, Melzi, Sabbioni, & Tomasini, 2006; Guilmineau & Chometon, 2009; Hemida & Baker, 2010; Hemida & Krajnovi´c, 2009a; Sterling et al., 2010; Tsubokura et al., 2010; Wang, Xu, Zhu, Cao, & Li, 2013; Wang, Xu, Zhu, & Li, 2014). Given the impor- tance of motorbikes as a form of transport, further research is required in order to ensure that traﬃc regulations and ultimately the safety of the road network is as robust as possible. Corresponding author. Email: DMF144@bham.ac.uk Corresponding author. Email: DMF144@bham.ac.uk 529 Engineering Applications of Computational Fluid Mechanics Figure 1. Orthogonal views of the motorbike showing the aerodynamic forces and moments, velocity directions and yaw angle, β (the angle between the motorbike traveling direction, Ux, and the eﬀective crosswind, Ueﬀ). Downloaded by [University of Birmingham] at 08:48 29 October 2015 Figure 1. Orthogonal views of the motorbike showing the aerodynamic forces and moments, velocity directions and yaw angle, β (the angle between the motorbike traveling direction, Ux, and the eﬀective crosswind, Ueﬀ). Computational ﬂuid dynamics (CFD) already plays a signiﬁcant role in motorbike design (Angeletti, Sclafani, Bella, & Ubertini, 2003). For example, it has been used to investigate the three-dimensional ﬂow characteristics inside motorbike engines, which is relatively diﬃcult and rather expensive to obtain experimentally (Chu, Chang, Hsu, Chien, & Liu, 2008; Gentilli, Zanforlin, & Frigo, 2006). Furthermore, CFD has been used to optimize indi- vidual engineering components, to improve energy eﬃ- ciency and to determine the eﬀect of local geometry changes of the motorbike on the aerodynamic forces (Taka- hashi et al., 2009; Watanabe, Okubo, Iwasa, & Aoki, 2003). Sakagawa, Yoshitake, and Ihara (2005) investigated the airﬂow pattern around a motorbike with the standard k-ε Reynolds-Averaged Navier-Stokes (RANS) simulation for the no-crosswind case in order to optimize aerodynamic performance. In addition, they carried out simulations for the design of an engine cooling system. However, as the ﬂow around a motorbike is fully turbulent, and thus varies in both time and space, the results obtained from the RANS simulations lack information about the instantaneous ﬂow structures. 2. Motorbike model Downloaded by [University of Birmingham] The current research was carried out on a sportive Yamaha R1 motorbike with rider (see Figure 1). The Yamaha R1 is a popular motorbike which has been manufactured since 1998 (Yamaha, 2009). Figure 1 shows the geometry of the motorbike, including the nomenclature adopted in this current work. The coordinate system used in this paper is also shown, where x opposes the direction of travel, y is in the lateral direction and z is in the vertical direc- tion. The resultant of the negative motorbike velocity (Ux) with the crosswind velocity (Uy), yields the eﬀective cross- wind velocity (Ueﬀ), which acts at a yaw angle (β) relative to the motorbike’s direction of travel. The length, height and width of the motorbike with rider are 2.04, 1.35 and 0.68 m, respectively. The motorbike model maintains a high level of geometrical detail including the main motor- bike components, although small details such as cables and bolts are omitted. 4. Computational domain and boundary conditions A generalized computational domain was used in this investigation (see Figure 2). Two inlet and two outlet boundaries are used to simulate the crosswind conditions. At the inlet boundaries, the ﬂow has two components: one in the negative direction of travel, x, and one perpendicular to the direction of travel, y. The eﬀective crosswind veloc- ity was set constant at 25 m/s in all simulations. The lateral ﬂow velocity,Uy, and the frontal ﬂow, Ux, were dependent on the yaw angle of the crosswind and are expressed as: fd ≡1 −tanh ([8rd]3) (2) rd ≡ νt + ν Uij Uij κ2l2 (3) (2) (3) Downloaded by [University of Birmingham] at 08:48 29 October 2015 where νt is the kinematic eddy viscosity, ν is the kinematic viscosity, Uij is the velocity gradient, κ is the Kármán con- stant and rd is the ratio of the model length scale to the wall distance. In the region where rd ≪1 (fd = 1), the LES model is employed. Ux = cos (β)Ueﬀ Uy = sin (β)Ueﬀ, (4) The time derivatives were discretized using a second- order backward implicit scheme. The gradient and diver- gence terms were discretized using a second-order central diﬀerencing, except for the velocity divergence terms, for which the linear-upwind stabilized transport (LUST) scheme (a blend of 75% second-order linear scheme and 25% linear-upwind scheme) was used to optimize the bal- ance between accuracy and stability. The transient ‘pres- sure implicit with splitting of operator’ (PISO) algorithm was implemented in the simulations to decouple the pres- sure and velocity (Issa, 1986). (4) where β is the yaw angle. Four diﬀerent yaw angles were investigated: 15°, 30° and 60° and 90°. The dimensions of the computational domain are shown in Figure 2, in which H is the height of the rider from the ground (1.35 m). The dimensions in the x-direction and z-direction were taken as constant for all simulations, while the dimension in the y-direction was extended for large yaw angles. The total y-dimension was set as 20 H for yaw angles between 15° and 30°, and 33 H for angles between 60° and 90°. These distances from the motorbike surface to the exit plane were chosen to be large enough for the zero-pressure exit bound- ary condition to be applied without aﬀecting the ﬂow or pressure ﬁelds around the motorbike. 3. Numerical method The open source ﬁnite volume CFD package OpenFOAM (Version 2.1.1) was used to solve the incompressible ﬂow equations in all the simulations of this paper. The ﬂow around a motorbike subjected to crosswinds is dominated by highly three-dimensional turbulent ﬂow structures. In order to obtain information about these turbulent struc- tures in both time and space, time-dependent DDES based on Spalart-Allmaras modeling was used (Spalart et al., 2006). Although more computationally expensive than the RANS approach, DDES is more accurate and yields The aim of this study is to provide an improved understanding of the time-averaged and instantaneous ﬂow around a motorbike subjected to crosswinds with diﬀerent yaw angles using both delayed detached-eddy simulation (DDES) and RANS techniques. In the current work, the Reynolds number is 2.2 × 106, based on the eﬀective crosswind velocity and the height of the rider from the ground. The open source CFD package OpenFOAM was used to solve the ﬂow equations. D. Fintelman et al. 530 information that is unobtainable from RANS. DDES is a hybrid technique that blends the RANS approach with the large eddy simulation (LES) approach. In the near wall region a RANS model is applied, while for the detached ﬂow the LES approach is used. The two approaches are combined by means of a modiﬁed distance function instead of calculating the complete ﬂow pattern as a func- tion of time. The SST k-ω predicts the turbulent viscosity by a relationship of the turbulent kinetic energy, k, and the speciﬁc dissipation ω near the wall, and the free-stream ﬂow is solved for using a k-ε model. Separate transport equations are used for k and ω. On the other hand, the k-ε model uses a relation between the turbulent dissipa- tion ε and turbulent kinetic energy to predict the turbulent viscosity. lDDES ≡l −fd max (0, l −CDES) (1) (1) where l is the distance from the wall, CDES is an empirically-derived constant (0.65) and is the largest dimension of the grid cell in all three directions, = max (δx, δy, δz ). The function fd is deﬁned as 4. Computational domain and boundary conditions No-slip boundaries were applied on the surface of the motorbike, rider and Downloaded by [University of Birmingham In addition to DDES, a number of RANS computations have also been undertaken. The RANS simulations were performed using two turbulence models; the shear stress transport (SST) k-ω model (Menter, 1992) and the standard k-ε model (Launder & Spalding, 1974). The RANS equa- tions predict the time-averaged velocity and pressure ﬁelds Figure 2. Computational domain dimensions. Figure 2. Computational domain dimensions. 531 Engineering Applications of Computational Fluid Mechanics (a) (c) (b) rse mesh around motorbike: (a) surface mesh wheel, (b) surface mesh rider, (c) ground mesh. (a) (b) (b) (a) (b) (c) Figure 3. Coarse mesh around motorbike: (a) surface mesh wheel, (b) surface mesh rider, (c) ground mesh. Downloaded by [University of Birmingham] Figure 3. Coarse mesh around motorbike: (a) surface mesh wheel, (b) surface mesh rider, (c) ground mesh of the DDES simulations, two diﬀerent meshes were eval- uated: a coarse mesh and a ﬁne mesh. The coarse mesh consisted of 3.8 million cells for the small yaw angles ranging from 15° to 30° and 4.0 million cells for large yaw angles from 60° to 90°. The number of cells for the ﬁne mesh was about 28.2 million. The meshes were concentrated in a region of size 9.5 m in length, 2 m in width and 2.5 m in height around the motorbike through a reﬁnement box in the SnappyHexMesh utility, as shown in Figure 3(c). The reﬁnement box was used to increase the number of cells close to the motorbike and in the wake where large changes in the ﬂow physics were expected. In addition, the mesh was reﬁned in the direction of the wake ﬂow through the use of three additional reﬁnement boxes. For large yaw angles a reﬁnement box was set in ground surface. The ground surface was simulated as a smooth wall with a velocity identical to the motorbike’s speed (Ux = 25 m/s), representing correct relative move- ment between the motorbike and the ground. To give a tangential velocity similar to that of the ground for both wheels, they rotated with an angular velocity of 66.7 rad/s (front) and 64.5 rad/s (rear) in the simulations. The discrep- ancy in angular velocity between the wheels is a result of the diﬀerent wheel diameters. ground surface. 4. Computational domain and boundary conditions The ground surface was simulated as a smooth wall with a velocity identical to the motorbike’s speed (Ux = 25 m/s), representing correct relative move- ment between the motorbike and the ground. To give a tangential velocity similar to that of the ground for both wheels, they rotated with an angular velocity of 66.7 rad/s (front) and 64.5 rad/s (rear) in the simulations. The discrep- ancy in angular velocity between the wheels is a result of the diﬀerent wheel diameters. The surface mesh was created with OpenFOAM’s SnappyHexMesh utility. The surface mesh contains 91% hexahedra-type cells, while the remaining 9% consist pre- dominantly of polyhedral and prisms cells. Images of the surface mesh of the motorbike and rider are shown in Fig- ures 3(a) and 3(b). To investigate the grid independency D. Fintelman et al. 532 Cp, is deﬁned as Cp, is deﬁned as the lee-side region of the motorbike, whereas for small yaw angles the reﬁnement box included both the wake behind the motorbike and a part of the lee-side ﬂow. A standard wall function was used in both the DDES and the RANS simulations. The mean wall normal resolution y + value of the RANS and DDES simulations was approximately 100. CP = ¯p −p∞ 1 2ρ U2 eﬀ , (5) (5) where ¯p is the total time-averaged pressure, p∞is the free stream pressure, and ρ is the air density, deﬁned as 1.205 kg/m3. Figure 4 shows the surface Cp acting on the rider at a height of 1.12 m above the ground for a yaw angle of 15°. From Figure 4 it is clear that the DDES ﬁne and coarse mesh yield similar results with respect to this parameter. To obtain an accurate and stable solution in these sim- ulations, a constant time step of 5 × 10−5 seconds was used to achieve a Courant-Friederich-Lewy (CFL) number below 1. In order to achieve the required criteria for the CFL number in the ﬁne mesh, an even smaller time step of 1 × 10−5 seconds was used. Overall, a non-dimensional simulation time, t∗= t Ueﬀ/H , of 55.5 was solved for, where t is the simulation time. Furthermore, the time- averaged ﬂow was only acquired from the fully developed turbulent ﬂow and averaged over t∗= 18.5. The visualiza- tion software ENSIGHT 9.2 was used to visualize the ﬂow in this paper. 4. Computational domain and boundary conditions During the simulations the aerodynamic coeﬃcients are calculated at each time step. The time-averaged drag force, CD, side force, CS, lift force, CL, and rolling moment CRoll, coeﬃcients are deﬁned as: Downloaded by [University of Birmingham] at 08:48 29 October 2015 CD = FD 0.5 Aρ U2 eﬀ , CS = FS 0.5 Aρ U2 eﬀ , CL = FL 0.5 A ρ U2 eﬀ , CRoll = MR 0.5 Aρ L U2 eﬀ . (6) (6) The simulations were performed using 16 proces- sors in the Blue Birmingham Environment for Academic Research (Blue BEAR) computational facility. This facil- ity has a Scientiﬁc Linux 6.6 operating system. The total wall time of the DDES simulations was approximately 900 hours, while the RANS simulations took approximately 90 hours. FD, Fs, FL and MR are the drag force, side force, lift force and rolling moment, respectively. The characteris- tic area, A, is deﬁned as 0.75 m2, and L the wheel base of 1.5 m. The time-averaged values of the aerodynamic coeﬃcients of the coarse and ﬁne DDES simulations are averaged over the ﬁnal 2.0 × 104 time steps (t* = 18.5). In addition to the surface pressure, the aerodynamic force and moment coeﬃcients of the diﬀerent meshes are given in Table 1. It can be noted from Table 1 that the aerody- namic coeﬃcients obtained from the ﬁne mesh are in good agreement with the aerodynamic coeﬃcients obtained from the coarse mesh. This indicates that the coarse mesh resolu- tion is adequate to resolve the main ﬂow features and thus no further reﬁnement of the mesh is needed. All the results presented in the remaining sections of this paper are from the coarse mesh, unless otherwise stated. Downloaded by [University of Birmingham] 5.2. Time-averaged ﬂow Figure 4 shows the peak pressure coeﬃcients on both shoulders of the rider. The highest pressures are found at the windward (left) side of the rider, in particular around the shoulders (α = 290°) and neck (α = 317°). A low pressure area develops at the side of the left arm, indi- cating a region of ﬂow separation. At the left side of the helmet and around the neck, high pressure regions are present. Figure 4 shows a quite similar behavior for the RANS and DDES results. However, a large negative sur- face pressure at α = 220° is visible in both RANS results which is not present in the DDES results. This low pressure area is located on the windward side, at the trunk of the rider’s body between the arms and the legs. The disagree- ment is likely a result of the failure of the RANS models to accurately predict ﬂow separation. Downloaded by [University of Birmingham] at 08:48 29 October 2015 The time-averaged force coeﬃcients for the RANS and DDES simulations for the diﬀerent yaw angles are shown in Figure 6. The main aerodynamic components that inﬂuence the stability of a motorbike and rider are the side force and the rolling and yawing moment coef- ﬁcients. The CS values increase from about 0.42 to 2.08 with increasing yaw angle, caused by the increasing stag- nation pressure on the streamwise face and the reduction of the lee face pressure as a result of the recirculation region in the wake ﬂow for large yaw angles. The side force and rolling moment coeﬃcients appear to have an almost lin- ear increase with respect to the yaw angle for values below β ∼60°. Thereafter, the rate of change of these parame- ters decreases as a result of the ﬂow structures occurring in the wake of the motorbike and rider. The peak side force and rolling moment coeﬃcients are observed as occurring at a yaw angle of 90°, which possibly suggests this is Figure 5 shows the streamlines around the motorbike at a height of 1.12 m above the ground. The main diﬀer- ence between the two CFD approaches (RANS and DDES) is the prediction of the location of the core of the recir- culation region in the leeward side of the wake. 5.1. Mesh independence Two diﬀerent meshes – a coarse and ﬁne mesh – are evaluated to investigate the grid independence of the sim- ulations. For both meshes, the pressure coeﬃcient around the body of the rider is calculated. The pressure coeﬃcient, Figure 4. Pressure coeﬃcient along a line on the surface of the rider parallel to the ground at a height of 1.12 m, at a yaw angle of 15° for all CFD approaches and the time-averaged DDES simulation results. Figure 4. Pressure coeﬃcient along a line on the surface of the rider parallel to the ground at a height of 1.12 m, at a yaw angle of 15° for all CFD approaches and the time-averaged DDES simulation results. Engineering Applications of Computational Fluid Mechanics 533 Table 2. Time-averaged aerodynamic coeﬃcients for diﬀerent yaw angles of the DDES simulations. Table 2. Time-averaged aerodynamic coeﬃcients for diﬀerent yaw angles of the DDES simulations. Table 1. Time-averaged aerodynamic coeﬃcients for a coarse mesh and ﬁne mesh of the DDES simulations at 15° yaw angle. Table 2. Time-averaged aerodynamic coeﬃcients for diﬀerent yaw angles of the DDES simulations. β = 15° β = 30° β = 60° β = 90° CD 0.570 0.560 0.288 −0.148 CS 0.421 1.050 1.838 2.084 CL −0.045 0.030 0.317 0.412 CRoll −0.143 −0.389 −0.690 −0.761 CPitch 0.218 0.205 0.126 −0.070 CYaw −0.089 −0.178 −0.149 0.024 Coarse Fine CD 0.570 0.578 CS 0.421 0.420 CL −0.045 −0.027 CRoll −0.143 −0.144 CPitch 0.218 0.215 CYaw −0.089 −0.090 The time-averaged aerodynamic coeﬃcients of the DDES simulations for diﬀerent crosswind angles are given in Table 2. For a super-sportive motorbike such as the one under investigation, CD values between 0.45 and 0.6 are reported by Araki and Gotou (2001) of the Yamaha Motor Corporation. Ubertini and Desideri (2002) have reported an optimal drag area CDA of around 0.3 m2 for riders in an optimal aerodynamic position, e.g., ‘lying’ position. The DDES drag value at β = 15° is about 0.57 (0.43 m2), which is at the higher end of the range of the reported val- ues at β = 0°. This can be justiﬁed by the fact that the drag can increase up to about 16% when increasing the yaw angle from 0–10° (Ubertini & Desideri, 2002). 5.2. Time-averaged ﬂow The dynamic nature of a gust of wind might result in sudden changes of the yaw angle and will, as this study shows, likely lead to a change in drag and might there- fore aﬀect the motorbike speed and stability. The decrease in drag coeﬃcients under increasing yaw angles is also observed for other road vehicles, such as lorries (Baker, 1991; Hargreaves, Morvan, & Wright, 2006) and dou- ble decker buses (François, Delnero, Colman, Maranón, & Camocardi, 2009; Hemida & Krajnovi´c, 2009b). The drag force coeﬃcient is closely related to the pitching moment coeﬃcient in which the moment coeﬃcient grad- ually decreases with the yaw angle. Finally, the lift force coeﬃcients are relatively small and tend to increase with increasing yaw angles. The maximum lift force coeﬃcient is observed for the 90° crosswind condition. Figure 7 shows the pressure contour lines on the motor- bike and rider at diﬀerent crosswind yaw angles. It can be clearly seen that at β = 15° (Figure 7(a)), high pres- sure areas and stagnation points exist on the helmet (s1), windshield (s2) and shoulder (s3). These stagnation points appear to move towards the side of the body as the yaw angle increases. At large yaw angles (>30°), stagnation points s2 and s3 disappear and new stagnation points emerge at the side of the rider (s4) and at the side of the motorbike (s5). However, the stagnation point (s1) remains on the head with an increase in yaw angle. In Figure 8, the iso-surface of the time-averaged pres- sure coeﬃcient around the motorbike is shown for yaw angles of 15°, 30°, 60° and 90° with Cp values of −0.29, −0.36, −0.5 and −0.64, respectively. For a 15° yaw angle (Figure 8(a)), low pressure regions predominantly exist at the back of the rider. With increasing yaw angle, low pressure areas develop at the leeward side and front of the motorbike, in particular at the windshield. These low pressure regions contribute signiﬁcantly to the aero- dynamic forces. It can be observed in Figure 6 that the RANS and DDES results are similar for the CD. For large yaw angles, the RANS k-ε predicts 19% higher side forces than the two other approaches. 5.2. Time-averaged ﬂow Figure 5 shows that the core of the circulation region in the y- direction is located 0.035 m closer to the motorbike in both RANS simulations than that obtained with the time- averaged DDES results. The k-ε model showed the largest x-deviation from the location of the core in the DDES sim- ulation. In the k-ε and k-ω models, the core was 0.087 m and 0.057 m closer to the motorbike in the x-direction, respectively. (a) (a) (b) (c) Figure 5. Streamlines around the motorbike in a plane at a height of 1.12 m from the ground of the diﬀerent turbulence models. (c) (b) (c) (b) (c) Figure 5. Streamlines around the motorbike in a plane at a height of 1.12 m from the ground of the diﬀerent turbulence models. 534 D. Fintelman et al. 534 D. Fintelman et al. (a) (b) (c) (d) (e) (f) Figure 6. Aerodynamic force coeﬃcients of the RANS and DDES simulations for diﬀerent yaw angles. (a) (c) (a) (b) (c) (a) (b) (c) (f) (d) (e) (f) Downloaded by [University of Birmingham] at 08:48 29 October 2015 Downloaded by [University of Birmingham] at 08:48 29 October 2015 Figure 6. Aerodynamic force coeﬃcients of the RANS and DDES simulations for diﬀerent yaw angles. (Makowski & Kim, 2000). Similarly, there are discrep- ancies in the lift forces between the RANS and DDES approach at large yaw angles. This can be related to the failure of RANS approaches to accurately predict the sep- aration and reattachment regions. Therefore, the results of the DDES simulations have been used to explore the ﬂow around the motorbike in the time-averaged and instanta- neous ﬂow unless otherwise explicitly stated. the critical angle in terms of wind force and thus motor- bike stability. The rolling moment coeﬃcient values vary from −0.14 to −0.76 with increasing yaw angle. The CD values decrease from about 0.57 to −0.15 with increas- ing yaw angle, caused by the increased side-wind velocity. The maximum drag force coeﬃcient is observed for the small crosswind condition, i.e., β = 15°. The drag force coeﬃcient decreases to nearly zero for β = 90°, i.e., pure crosswind conditions. The decrease in drag might aﬀect the stability of the motorbike rider when subjected to a gust of wind. 5.2. Time-averaged ﬂow It is commonly observed in external vehicle aerodynamics that the RANS k-ε model over-predicts the turbulent kinetic energy in a ﬂow with strong anisotropic turbulence and non-equilibrium eﬀects, such as the ﬂow under investigation at large yaw angles Figures 9 and 10 show the time-averaged stream- lines for diﬀerent crosswind yaw angles on two diﬀerent planes parallel to the ground at heights of z/H = 0.8 and z/H = 0.4, respectively. The vortex cores of the main 535 Engineering Applications of Computational Fluid Mechanics (a) (b) (c) (d) Figure 7. Pressure contour lines of the DDES simulations on motorbike and rider at diﬀerent yaw angles. Stagnation points are shown on the head (s1), windshield (s2), left shoulder (s3), thigh (s4) and motorbike (s5). (a) (b) (b) (a) (b) (d) (c) (d) Downloaded by [University of Birmingham] Figure 7. Pressure contour lines of the DDES simulations on motorbike and rider at diﬀerent yaw angles. Stagnation points are shown on the head (s1), windshield (s2), left shoulder (s3), thigh (s4) and motorbike (s5). ﬂow are also shown. Diﬀerences in ﬂow behavior can be observed for the two planes. For the plane at z/H = 0.8 (Figure 9), at small crosswind yaw angles (β = 15–30°), recirculation areas are found close to the side of the rider (v1–v2). At larger yaw angles ( >30°), the ﬂow becomes more turbulent, resulting in an increase in the number of vortex cores (v3–v6). These vortex cores also move further away from the rider as a result of increased vortex shed- ding and ﬂow separation. This is paired with the increase of the size of the wake. For the plane at z/H = 0.4 in Figure 10, recirculation areas are mainly found at the back of the motorbike for β = 15–30° (v7–v8). At yaw angles >30°, there are large circulation areas at the leeward side of the motorbike (v9–v13), whereas at β = 60° the main vortex cores v9–v11 are centrally located in the wake, and at β = 90° the vortex cores v12 and v13 move more to the edges of the wake. This is predominantly caused by the ﬂow streams around the sides of the wheels of the motorbike. is based on the Eigen-analysis algorithm. The computed vortex cores of the time-averaged ﬂow around the motor- bike at diﬀerent crosswind yaw angles are shown in Figure 11. 5.2. Time-averaged ﬂow The recirculation regions shown in the time- averaged streamlines of the motorbike in Figures 9 and 10 are also found in the vortex cores. The large number of vor- tex core lines reveals the complex turbulent ﬂow around the motorbike. The ﬂow becomes more turbulent with increasing yaw angle, which is caused by the increased number of bluﬀbodies seen by the upcoming ﬂow. The low pressure regions at the back and head of the rider shown in Figure 8 are also present as vortex cores v14 and v15 respectively. Similarly, around the windshield a low pres- sure region develops which corresponds with vortex core v16 for yaw angles >30°. ﬂow are also shown. Diﬀerences in ﬂow behavior can be observed for the two planes. For the plane at z/H = 0.8 (Figure 9), at small crosswind yaw angles (β = 15–30°), recirculation areas are found close to the side of the rider (v1–v2). At larger yaw angles ( >30°), the ﬂow becomes more turbulent, resulting in an increase in the number of vortex cores (v3–v6). These vortex cores also move further away from the rider as a result of increased vortex shed- ding and ﬂow separation. This is paired with the increase of the size of the wake. For the plane at z/H = 0.4 in Figure 10, recirculation areas are mainly found at the back of the motorbike for β = 15–30° (v7–v8). At yaw angles >30°, there are large circulation areas at the leeward side of the motorbike (v9–v13), whereas at β = 60° the main vortex cores v9–v11 are centrally located in the wake, and at β = 90° the vortex cores v12 and v13 move more to the edges of the wake. This is predominantly caused by the ﬂow streams around the sides of the wheels of the motorbike. 5.3. Transient ﬂow The second invariant of the velocity gradient tensor, Q, is used to visualize the instantaneous ﬂow ﬁeld and to determine the positions of the vortex cores, where Q is The time-averaged ﬂow features are visualized using the vortex core technique (Banks & Singer, 1994), which D. Fintelman et al. (a) (b) (c) (d) re 8. Iso-surface around motorbike of the DDES turbulence model for the time-averaged ﬂow at diﬀerent yaw angles. The coeﬃcient essure has been set to: (a) Cp = −0.29, (b) Cp = −0.36, (c) Cp = −0.5, and (d) Cp = −0.64. 536 D. Fintelman et al. (b) (a) (b) (a) Downloaded by [University of Birmingham] at 08:48 29 October 2015 (c) (d) (d) (c) Downloaded by [University of Birmingham Figure 8. Iso-surface around motorbike of the DDES turbulence model for the time-averaged ﬂow at diﬀerent yaw angles. The coeﬃcient of pressure has been set to: (a) Cp = −0.29, (b) Cp = −0.36, (c) Cp = −0.5, and (d) Cp = −0.64. deﬁned as large yaw angles, the variations are increasingly signiﬁcant compared to small yaw angles, which are caused by strong vortex shedding. The largest force coeﬃcient variations are found in the lift forces. Q = −1 2 uij uji . (7) (7) Figure 12 shows the temporal development of the vor- tex structures at diﬀerent crosswind angles for Q = 8 × 103. Many vortices are born at the entire leeward side of the motorbike and rider surface. These vortices are convected downstream by the crosswind ﬂow. The vortex structures are stretched over a larger wake distance from the motor- bike with increasing yaw angles. However, at a yaw angle of 90°, the wake vortices remain closer to the motorbike and are smaller in size. The attachment and detachment of the ﬂow from the body contribute to high-frequency modes in the aerodynamic forces. The time histories of the force coeﬃcients are used to reveal the frequencies of the ﬂow around the motorbike. The power spectral densities (PSDs) of the time-varying signals are calculated to evaluate the distribution of the power of the signal over the frequencies. The PSDs are calculated by multiplying the Fourier transform of the time-varying signal by the conjugate of the Fourier trans- form. Figure 14 shows the PSDs plotted versus the Strouhal number, St, for diﬀerent yaw angles and force coeﬃcients. 5.3. Transient ﬂow The Strouhal number is deﬁned as St = f H Ueﬀ (8) The time histories of the force coeﬃcients at diﬀer- ent yaw angles are shown Figure 13. Large variations of the force coeﬃcients are found, which are caused by the instability and movement of the wake vortices that are dis- turbing the instantaneous motorbike surface pressure. For (8) where f is the frequency and H is the characteristic length, deﬁned as the height of the motorbike from the ground. An 537 Engineering Applications of Computational Fluid Mechanics (a) (a) (b) (c) (d) Figure 9. Time-averaged streamlines and the positions of the main vortex cores of the DDES simulations on a surface parallel to the ground at z/H = 0.8 and at diﬀerent crosswind yaw angles. (b) (d) (b) (a) (d) (c) (c) (d) Figure 9. Time-averaged streamlines and the positions of the main vortex cores of the DDES simulations on a surface parallel to the ground at z/H = 0.8 and at diﬀerent crosswind yaw angles. g y g (a) (b) (c) (d) Figure 10. Time-averaged streamlines and the positions of the main vortex cores of the DDES simulations on a surface parallel to the ground at z/H = 0.4 and at diﬀerent crosswind yaw angles. (a) (b) (a) (d) (c) (c) Figure 10. Time-averaged streamlines and the positions of the main vortex cores of the DDES simulation ground at z/H = 0.4 and at diﬀerent crosswind yaw angles. Figure 10. Time-averaged streamlines and the positions of the main vortex cores of the DDES simulations on a surface parallel to the ground at z/H = 0.4 and at diﬀerent crosswind yaw angles. D. Fintelman et al. 538 overview of the Strouhal numbers of the dominant peaks is given in Table 3. ﬂow structures on the surface of the motorbike and to the shear layer instabilities. The dominant low-frequency peaks in the PSD of the CD lie in the Strouhal number range 0.2 < St < 1.7. For the CS, the dominant frequen- cies are found in the range 0.1 < St < 2.1, while for the CL in the range 0.1 < St < 1.2. The energy of the domi- nant frequency in the lift force coeﬃcient is up to seven times higher than that of the other force coeﬃcients. 5.3. Transient ﬂow In all PSDs of the force coeﬃcients, the energy of the dominant peaks increases at higher yaw angles (Figure 14). Predominately low-frequency peaks are found in the PSDs. These low-frequency peaks account for vortex shedding in the wake of the motorbike. Furthermore, high-frequency peaks in the drag force coeﬃcient are mainly found in the 60° and 90° crosswind yaw angle situations. This can be explained by the attachment and detachment of the small Some of the dominant low frequency peaks appear mul- tiple times at diﬀerent yaw angles. The dominant peak, (a) (b) (c) (d) (e) (f) (g) (h) Vortex cores of the time-averaged velocity ﬁeld of the DDES simulations around the motorbike at d diﬀerent orthogonal views: side view and top view. (a) (b) (b) (c) (d) (d) (e) (f) (f) (e) (h) (g) (g) Figure 11. Vortex cores of the time-averaged velocity ﬁeld of the DDES simulations around the motorbike at diﬀerent crosswind yaw angles and diﬀerent orthogonal views: side view and top view. Engineering Applications of Computational Fluid Mechanics 539 (b) (c) (b) (a) (d) Figure 12. Instantaneous iso-surface of the velocity gradient tensor, Q, around the motorbike at diﬀerent crosswind angles of the DDES simulations, for Q = 8 × 103. (a) (c) (d) (c) Downloaded by [University of Birmingham] at 08:48 29 October 2015 Figure 12. Instantaneous iso-surface of the velocity gradient tensor, Q, around the motorbike at diﬀerent crosswind angles of the DDES simulations, for Q = 8 × 103. s iso-surface of the velocity gradient tensor, Q, around the motorbike at diﬀerent crosswind angles of the DDES 103. Figure 12. Instantaneous iso-surface of the velocity gradient tensor, Q, around the motorbike at diﬀerent c simulations, for Q = 8 × 103. (a) (b) (c) Figure 13. Time histories of the force coeﬃcients on the motorbike at diﬀerent yaw angles obtained from the DDES simulations: (a) drag force coeﬃcient, CD, (b) side force coeﬃcient, CS, and (c) lift force coeﬃcient, CL. (a) (b) (c) Figure 13. Time histories of the force coeﬃcients on the motorbike at diﬀerent yaw angles obtained from the DDES simulations: (a) drag force coeﬃcient, CD, (b) side force coeﬃcient, CS, and (c) lift force coeﬃcient, CL. which appears in the PSD of the CL at β = 30◦. 5.3. Transient ﬂow Another common frequency is St = 0.270, which appears as a dom- inant peak in the CD at β = 90◦and in the CL at yaw angles 15° and 60°, respectively. St = 0.162, in the PSD of the CS at β = 30◦and β = 60◦ can also found as the dominant peak in the PSD of the CL at β = 60◦and β = 90◦. This is also the case for the dom- inant peak St = 0.216 in the PSD of the CD at β = 60◦, 540 D. Fintelman et al. (a) (c) (b) Figure 14. Strouhal number versus power spectral density of the time-varying aerodynamic forces for diﬀerent yaw angles obtained from the DDES simulations: (a) PSD of drag force, (b) PSD of side force, and (c) PSD of lift force. 540 D. Fintelman et al. (a) (b) 540 D. Fintelman e (a) (b) D. Fintelman et al. 540 (b) Downloaded by [University of Birmingham] at 08:48 29 October 2015 (c) Figure 14. Strouhal number versus power spectral density of the time-varying aerodynamic forces for diﬀerent yaw angles obtained from the DDES simulations: (a) PSD of drag force, (b) PSD of side force, and (c) PSD of lift force. Figure 14. Strouhal number versus power spectral density of the time-varying aerodynamic forces for diﬀerent yaw angles obtained from the DDES simulations: (a) PSD of drag force, (b) PSD of side force, and (c) PSD of lift force. Table 3. Strouhal numbers of the two dominant peaks in the force coeﬃcient data for diﬀerent yaw angles of the DDES simulations. good agreement with those of the DDES simulations. How- ever, diﬀerences are observed in the location and size of the recirculation regions and associated surface pressures. The rolling moments, side forces and lift forces increase with the yaw angle, while the drag forces decrease. The peak side forces are almost four times higher than the peak drag forces. The DDES results showed that high sur- face pressures are predominantly found on the helmet and motorbike for almost all yaw angles. Main vortex cores are found behind the helmet, at the back of the rider and at the windshield, reﬂecting the highly turbulent ﬂow at these regions. In particular behind the windshield, strong vortex shading takes place for yaw angles >15°.The vortex shed- ding is found to be stronger at higher yaw angles. 5.3. Transient ﬂow From the results it could be suggested that optimization of the design of the motorbike, with focus on the windshield, could help to improve the safety of motorbike riders in crosswinds. The CFD results have provided valuable data for other numerical and experimental work. It is recommended that future work should attempt to reveal the eﬀects of gusts on motorbike stability, since it is hypothesized that these β = 15° β = 30° β = 60° β = 90° CD 0.486 0.324 0.216 0.270 1.458 0.540 0.810 1.674 CS 2.052 0.162 0.162 0.108 0.108 0.486 1.458 0.864 CL 0.270 0.216 0.162 0.162 1.242 0.432 0.270 0.864 References Hargreaves, D. M., Morvan, H. P., & Wright, N. G. (2006). CFD modeling of high-sided vehicles in cross-winds. The Fourth International Symposium on Computational Wind Engineering (CWE2006). Yokohama, Japan. Angeletti, M., Sclafani, L., Bella, G., & Ubertini, S. (2003). The role of cfd on the aerodynamic investigation of motorcycles. SAE Technical Paper 2003-01-0997. doi:10.4271/2003-01-0997 Hemida, H., & Baker, C. (2010). Large-eddy simulation of the ﬂow around a freight wagon subjected to a crosswind. Computers & Fluids, 39(10), 1944–1956. doi:10.1016/ j.compﬂuid.2010.06.026 Araki, Y., & Gotou, K. (2001). Development of aerodynamic characteristics for motorcycles using scale model wind tun- nel. SAE Technical Paper 2001-01-1851. doi:10.4271/2001- 01-1851 Hemida, H., & Krajnovi´c, S. (2009a). Exploring ﬂow structures around a simpliﬁed ICE2 train subjected to a 30° side wind using LES. Engineering Applications of Computational Fluid Mechanics, 3(1), 28–41. doi:10.1080/19942060.2009. 11015252 Downloaded by [University of Birmingham] at 08:48 29 October 2015 Baker, C., Cheli, F., Orellano, A., Paradot, N., Proppe, C., & Rocchi, D. (2009). Cross-wind eﬀects on road and rail vehicles. Vehicle System Dynamics, 47(8), 983–1022. doi:10.1080/00423110903078794 Hemida, H., & Krajnovi´c, S. (2009b). Transient simulation of the aerodynamic response of a double-deck bus in gusty winds. Journal of Fluids Engineering, 131(3), 031101. Baker, C. (1991). Ground vehicles in high cross winds part I: Steady aerodynamic forces. Journal of Fluids and Struc- tures, 5, 69–90. Issa, R. I. (1986). Solution of the implicitly discretised ﬂuid ﬂow equations by operator-splitting. Journal of Computational Physics, 62, 40–65. Banks, D. C., & Singer, B. A. (1994). Vortex tubes in turbulent ﬂows – Identiﬁcation, representation, reconstruction Insti- tute for Computer Applications in Science and Engineering. Hampton: NASA Langley Research Center. Launder, B. E., & Spalding, D. B. (1974). The numerical com- putation of turbulent ﬂows. Computer Methods in Applied Mechanics and Engineering, 3(2), 269–289. doi:http:// dx.doi.org/10.1016/0045-7825(74)90029-2 Downloaded by [University of Birmingham Bridges, P., & Russell, J. B. (1987). The eﬀect of topboxes on motorcycle stability. Vehicle System Dynamics, 16(5–6), 345–354. Makowski, F. T., & Kim, S.-E. (2000). Advances in external-aero simulation of ground vehicles using the steady RANS equa- tions. SAE Technical Paper, 2000-01-0484. doi:10.4271/ 2000-01-0484 Carr, A. (2011, 12 August 2013). Fatal motorcycle crash, wind gust causes Sherman wreck, from http://www.post-journal. com/page/content.detail/id/583501/Fatal-Motorcycle- Crash.html?nav = 5192 Menter, F. R. (1992). Improved two-equation k-omega turbulence models for aerodynamic ﬂows. NASA-TM-103975, NASA- TM-103975. Cheli, F., Belforte, P., Melzi, S., Sabbioni, E., & Tomasini, G. (2006). 6. Conclusion The work of this paper represents the ﬁrst numerical study investigating the time-averaged and instantaneous ﬂow around a motorbike subjected to crosswinds. DDES and RANS simulations were performed on the ﬂow around a motorbike subjected to crosswind with yaw angles ranging from 15° to 90°. Results showed that the time-averaged aerodynamic coeﬃcients of the RANS are in reasonably 541 Engineering Applications of Computational Fluid Mechanics com/dpp/news/local/fox_cities/fatal-motorcycle-crash-in- waupaca-co are likely to play a signiﬁcant role in determining safety criteria. are likely to play a signiﬁcant role in determining safety criteria. Gentilli, R., Zanforlin, S., & Frigo, S. (2006, 4–7 July 2006). Numerical and experimental analysis on a small GDI, strat- iﬁed charge, motorcycle engine. Paper presented at the Proceedings of the 8th Biennial Conference on Engineering Systems Design and Analysis, Turin, Italy. Disclosure statement No potential conﬂict of interest was reported by the authors. No potential conﬂict of interest was reported by the authors. Guilmineau, E., & Chometon, F. (2009). Eﬀect of side wind on a simpliﬁed car model: Experimental and numerical analysis. Journal of Fluids Engineering, 131(2), 021104. References Numerical-experimental approach for evaluating cross-wind aerodynamic eﬀects on heavy vehicles. Vehicle System Dynamics, 44, 791–804. doi:10.1080/004231106008 86689 Sakagawa, K., Yoshitake, H., & Ihara, E. (2005). Computa- tional ﬂuid dynamics for design of motorcycles (Numerical analysis of coolant ﬂow and aerodynamics). SEA Technical Paper, 2005-32-0033. doi:10.4271/2005-32-0033 Cheli, F., Bocciolone, M., Pezzola, M., & Leo, E. (2006, 4– 7 July 2006). Numerical and experimental approaches to investigate the stability of a motorcycle vehicle. Paper presented at the Proceedings of the 8th Biennial Confer- ence on Engineering Systems Design and Analysis, Turin, Italy. Scibor-Rylski, A. J., & Sykes, D. M. (1984). Road vehicle aerodynamics (2nd ed.). Plymouth: Pentech Press. Shuhei, A. (2006). Fuel cell powered motorcycles. Journal of the Society of Automotive Engineers of Japan, 60(1), 90–93. Spalart, P. R., Deck, S., Shur, M. L., Squires, K. D., Strelets, M. K., & Travin, A. (2006). A new version of detached-eddy simulation, resistant to ambiguous grid densities. Theoreti- cal and Computational Fluid Dynamics, 20, 181–195. y Chu, L.-M., Chang, M.-H., Hsu, H.-C., Chien, W.-T., & Liu, C.-H. (2008). Simulation and experimental measurement of ﬂow ﬁeld within four-stroke motorcycle engines. Jour- nal of the Chinese Society of Mechanical Engineers, 29(2), 149–158. Sterling, M., Quinn, A. D., Hargreaves, D. M., Cheli, F., Sab- bioni, E., Tomasini, G., . . . Morvan, H. (2010). A com- parison of diﬀerent methods to evaluate the wind induced forces on a high sided lorry. Journal of Wind Engineering and Industrial Aerodynamics, 98(1), 10–20. doi:10.1016/ j.jweia.2009.08.008 Cooper, K. R. (1983). The eﬀect of handlebar fairings on motorcycle aerodynamics. SAE Technical Paper, 830156. doi:10.4271/830156 Donell, S. (2010, 18 April). Wind blamed in fatal motorcy- cle accident, NBC Washington. Retrieved from http://www. nbcwashington.com/news/local/Weather-Blamed-in-Freak- Fatal-Motorcycle-Accident-91433499.html Takahashi, Y., Kurakawa, Y., Sugita, H., Ishima, T., & Obokata, T. (2009). CFD analysis of airﬂow around the rider of a motorcycle for rider comfort improvement. SAE Technical paper, 2009–01-1155. doi:10.4271/2009-01-1155 François, D., Delnero, J., Colman, J., Maranón, D. L. J., & Camo- cardi, M. (2009). Experimental determination of Stationary Aerodynamics loads on a double deck Bus. 11th Americas Conference on Wind Engineering, San Juan, Puerto Rico. Tsubokura, M., Nakashima, T., Kitayama, M., Ikawa, Y., Doh, D. H., & Kobayashi, T. (2010). Large eddy simulation on the unsteady aerodynamic response of a road vehi- cle in transient crosswinds. International Journal of Heat Gauger, W. (2013). Fatal motorcycle crash in Waupaca Co. Retrieved 14 August 2013, from http://www.fox11online. 542 D. References Fintelman et al. and Fluid Flow, 31(6), 1075–1086. doi:10.1016/j.ijheat- ﬂuidﬂow.2010.05.008 Wang, B., Xu, Y. L., Zhu, L. D., & Li, Y.-L. (2014). Crosswind eﬀect studies on road vehicle passing by bridge tower using computational ﬂuid dynamics. Engineering Applications of Computational Fluid Mechanics, 8(3), 330–344. Ubertini, S., & Desideri, U. (2002). Aerodynamic investigation of a scooter in the university of perugia wind tunnel facility. 2002-01-0254. Retrieved from http://dx.doi.org/10.4271/ 2002-01-0254 doi:10.4271/2002-01-0254 Watanabe, T., Okubo, T., Iwasa, M., & Aoki, H. (2003). Estab- lishment of an aerodynamic simulation system for motor- cycle and its application. Jsae Review, 24(2), 231–233. doi:10.1016/S0389-4304(03)00003-1 Wang, B., Xu, Y. L., Zhu, L. D., Cao, S.-Y., & Li, Y.-L. (2013). Determination of aerodynamic forces on stationary/ moving vehicle-bridge deck system under crosswinds using computational ﬂuid dynamics. Engineering Applications of Computational Fluid Mechanics, 7(3), 355–368. ( ) Yamaha. (2009). YZF-R1: History 1998–2010, from http://www. yamaha-motor.eu/designcafe/en/about-bikes/supersport/in- dex.aspx?view = article&id = 442650
https://openalex.org/W2920182657	https://dergipark.org.tr/tr/download/article-file/264840	Turkish	null	Eleştirel Düşünme ve Empatik Eğilim İlişkisi: Beden Eğitimi Öğretmeni Adayları Örneği	DergiPark (Istanbul University)	2,016	cc-by	9,995	Araştırma Makalesi The Relationship between Critical Thinking and Empathic Tendency: The Example of Preservice Physical Education Teachers Ezel Nur KORUR1,* Erman ÖNCÜ 2 Sonnur KÜÇÜK KILIÇ3 1 Ordu Üniversitesi, ezelnurkorur@odu.edu.tr 2 Karadeniz Teknik Üniversitesi, eoncu@ktu.edu.tr 3 Karadeniz Teknik Üniversitesi, sonnurkucukkilic@ktu.edu.tr * Corresponding Author: eoncu@ktu.edu.tr Article Info Abstract The purpose of this study was to examine preservice physical education teachers’ critical thinking skills and empathic tendencies according to some demographic variables, and to determine the relationship between their critical thinking skills and empathic tendencies. The study was conducted on 1006 preservice physical education teachers who were selected from ten universities. In the study, the California Critical Thinking Disposition Inventory and the Empathy Tendency Criteria were administered on the participants as data collection tools. Critical thinking skills of the participants were at the low level and empathic tendencies of the participants were at middle level. Whereas there was no significant difference according to gender, the education level of the parents of the participants and the years of study spent at the university can be said to be determining factors in the critical thinking and empathic skills of the pre-service physical education teachers. Received: June 28, 2016 Accepted: July 28,2016 Online: December 29, 2016 Keywords: Critical thinking, empathy, preservice physical education teacher Eleştirel Düşünme ve Empatik Eğilim İlişkisi: Beden Eğitimi Öğretmeni Adayları Örneği Makale Bilgisi Öz To cite this article: Korur, E. N., Öncü, E. & Küçük Kılıç, S. (2016). Eleştirel düşünme ve empatik eğitim ilişkisi: Beden eğitimi öğretmeni adayları örneği Journal of Computer and Education Research, 4 (8), 179- 204. DOI: https://doi.org/10.18009/jcer.87227 Araştırma Makalesi The Relationship between Critical Thinking and Empathic Tendency: The Example of Preservice Physical Education Teachers Ezel Nur KORUR1,* Erman ÖNCÜ 2 Sonnur KÜÇÜK KILIÇ3 1 Ordu Üniversitesi, ezelnurkorur@odu.edu.tr 2 Karadeniz Teknik Üniversitesi, eoncu@ktu.edu.tr 3 Karadeniz Teknik Üniversitesi, sonnurkucukkilic@ktu.edu.tr * Corresponding Author: eoncu@ktu.edu.tr Article Info Abstract The purpose of this study was to examine preservice physical education teachers’ critical thinking skills and empathic tendencies according to some demographic variables, and to determine the relationship between their critical thinking skills and empathic tendencies. The study was conducted on 1006 preservice physical education teachers who were selected from ten universities. In the study, the California Critical Thinking Disposition Inventory and the Empathy Tendency Criteria were administered on the participants as data collection tools. Critical thinking skills of the participants were at the low level and empathic tendencies of the participants were at middle level. Whereas there was no significant difference according to gender, the education level of the parents of the participants and the years of study spent at the university can be said to be determining factors in the critical thinking and empathic skills of the pre-service physical education teachers. Received: June 28, 2016 Accepted: July 28,2016 Online: December 29, 2016 Keywords: Critical thinking, empathy, preservice physical education teacher Eleştirel Düşünme ve Empatik Eğilim İlişkisi: Beden Eğitimi Öğretmeni Adayları Örneği Makale Bilgisi Öz To cite this article: Korur, E. N., Öncü, E. & Küçük Kılıç, S. (2016). Eleştirel düşünme ve empatik eğitim ilişkisi: Beden eğitimi öğretmeni adayları örneği Journal of Computer and Education Research, 4 (8), 179- 204. DOI: https://doi.org/10.18009/jcer.87227 Araştırma Makalesi The Relationship between Critical Thinking and Empathic Tendency: The Example of Preservice Physical Education Teachers Ezel Nur KORUR1,* Erman ÖNCÜ 2 Sonnur KÜÇÜK KILIÇ3 The purpose of this study was to examine preservice physical education teachers’ critical thinking skills and empathic tendencies according to some demographic variables, and to determine the relationship between their critical thinking skills and empathic tendencies. The study was conducted on 1006 preservice physical education teachers who were selected from ten universities. In the study, the California Critical Thinking Disposition Inventory and the Empathy Tendency Criteria were administered on the participants as data collection tools. Critical thinking skills of the participants were at the low level and empathic tendencies of the participants were at middle level. Whereas there was no significant difference according to gender, the education level of the parents of the participants and the years of study spent at the university can be said to be determining factors in the critical thinking and empathic skills of the pre-service physical education teachers. To cite this article: Korur, E. N., Öncü, E. & Küçük Kılıç, S. (2016). Eleştirel düşünme ve empatik eğitim ilişkisi: Beden eğitimi öğretmeni adayları örneği Journal of Computer and Education Research, 4 (8), 179- 204. DOI: https://doi.org/10.18009/jcer.87227 To cite this article: Korur, E. N., Öncü, E. & Küçük Kılıç, S. (2016). Eleştirel düşünme ve empatik eğitim ilişkisi: Beden eğitimi öğretmeni adayları örneği Journal of Computer and Education Research, 4 (8), 179- 204. DOI: https://doi.org/10.18009/jcer.87227 Eleştirel Düşünme ve Empatik Eğilim İlişkisi: Beden Eğitimi Öğretmeni Adayları Örneği Makale Bilgisi Öz Geliş: 28 Haziran 2016 Bu araştırmanın amacı, beden eğitimi öğretmeni adaylarının eleştirel düşünme ve empatik eğilimlerini bazı değişkenlere göre incelemek ve aralarındaki ilişkiyi belirlemektir. Çalışma grubunu, on üniversitede öğrenim gören toplam 1006 beden eğitimi öğretmeni adayı oluşturmuştur. Araştırmada veri toplama aracı olarak ‘Kaliforniya Eleştirel Düşünme Eğilimi Ölçeği’ ile ‘Empatik Eğilim Ölçeği’ kullanılmıştır. Araştırma sonuçları, katılımcıların eleştirel düşünme becerilerinin düşük; empatik eğilimlerinin ise orta seviyede olduğunu göstermektedir. Katılımcıların eleştirel düşünme ve empatik eğilimleri cinsiyet değişkenine göre anlamlı bir şekilde farklılaşmazken; sınıf, anne ve baba eğitim durumu değişkenlerine göre ise anlamlı bir şekilde farklılaşmaktadır. Kabul: 28 Temmuz 2016 Yayın: 29 Aralık 2016 Anahtar kelimeler: Eleştirel düşünme, empati, beden eğitimi öğretmeni adayı Eleştirel Düşünme ve Empatik Eğilim İlişkisi: Beden Eğitimi Öğretmeni Adayları Örneği Makale Bilgisi Öz Geliş: 28 Haziran 2016 Bu araştırmanın amacı, beden eğitimi öğretmeni adaylarının eleştirel düşünme ve empatik eğilimlerini bazı değişkenlere göre incelemek ve aralarındaki ilişkiyi belirlemektir. Çalışma grubunu, on üniversitede öğrenim gören toplam 1006 beden eğitimi öğretmeni adayı oluşturmuştur. Araştırma Makalesi The Relationship between Critical Thinking and Empathic Tendency: The Example of Preservice Physical Education Teachers Araştırmada veri toplama aracı olarak ‘Kaliforniya Eleştirel Düşünme Eğilimi Ölçeği’ ile ‘Empatik Eğilim Ölçeği’ kullanılmıştır. Araştırma sonuçları, katılımcıların eleştirel düşünme becerilerinin düşük; empatik eğilimlerinin ise orta seviyede olduğunu göstermektedir. Katılımcıların eleştirel düşünme ve empatik eğilimleri cinsiyet değişkenine göre anlamlı bir şekilde farklılaşmazken; sınıf, anne ve baba eğitim durumu değişkenlerine göre ise anlamlı bir şekilde farklılaşmaktadır. Kabul: 28 Temmuz 2016 Yayın: 29 Aralık 2016 Anahtar kelimeler: Eleştirel düşünme, empati, beden eğitimi öğretmeni adayı Eleştirel Düşünme ve Empatik Eğilim İlişkisi: Beden Eğitimi Öğretmeni Adayları Örneği Eleştirel Düşünme ve Empatik Eğilim İlişkisi: Beden Eğitimi Öğretmeni Adayları Örneği Öz Bu araştırmanın amacı, beden eğitimi öğretmeni adaylarının eleştirel düşünme ve empatik eğilimlerini bazı değişkenlere göre incelemek ve aralarındaki ilişkiyi belirlemektir. Çalışma grubunu, on üniversitede öğrenim gören toplam 1006 beden eğitimi öğretmeni adayı oluşturmuştur. Araştırmada veri toplama aracı olarak ‘Kaliforniya Eleştirel Düşünme Eğilimi Ölçeği’ ile ‘Empatik Eğilim Ölçeği’ kullanılmıştır. Araştırma sonuçları, katılımcıların eleştirel düşünme becerilerinin düşük; empatik eğilimlerinin ise orta seviyede olduğunu göstermektedir. Katılımcıların eleştirel düşünme ve empatik eğilimleri cinsiyet değişkenine göre anlamlı bir şekilde farklılaşmazken; sınıf, anne ve baba eğitim durumu değişkenlerine göre ise anlamlı bir şekilde farklılaşmaktadır. Bu araştırmanın amacı, beden eğitimi öğretmeni adaylarının eleştirel düşünme ve empatik eğilimlerini bazı değişkenlere göre incelemek ve aralarındaki ilişkiyi belirlemektir. Çalışma grubunu, on üniversitede öğrenim gören toplam 1006 beden eğitimi öğretmeni adayı oluşturmuştur. Araştırmada veri toplama aracı olarak ‘Kaliforniya Eleştirel Düşünme Eğilimi Ölçeği’ ile ‘Empatik Eğilim Ölçeği’ kullanılmıştır. Araştırma sonuçları, katılımcıların eleştirel düşünme becerilerinin düşük; empatik eğilimlerinin ise orta seviyede olduğunu göstermektedir. Katılımcıların eleştirel düşünme ve empatik eğilimleri cinsiyet değişkenine göre anlamlı bir şekilde farklılaşmazken; sınıf, anne ve baba eğitim durumu değişkenlerine göre ise anlamlı bir şekilde farklılaşmaktadır. * Bu çalışma, Karadeniz Teknik Üniversitesi Eğitim Bilimleri Enstitüsü Beden Eğitimi ve Spor Anabilim Dalı’nda yüksek lisans tezi olarak kabul edilmiş ayrıca çalışmanın özeti, 13. Uluslararası Spor Bilimleri Kongresi (7-9 Kasım 2014, Konya)’nde sözlü bildiri olarak sunulmuştur. * Bu çalışma, Karadeniz Teknik Üniversitesi Eğitim Bilimleri Enstitüsü Beden Eğitimi ve Spor Anabilim Dalı’nda yüksek lisans tezi olarak kabul edilmiş ayrıca çalışmanın özeti, 13. Uluslararası Spor Bilimleri Kongresi (7-9 Kasım 2014, Konya)’nde sözlü bildiri olarak sunulmuştur. Introduction In today’s modern education philosophy, the skills of empathy and critical thinking count among the most essential skills that teachers should possess. These skills carry more significance for the physical education teachers who have a difficult profession of teaching physical skills. The purpose of this research is to examine the critical thinking and empathic tendencies of the physical education teacher candidates as per with various factors and determine the connection between them. The workgroup of the research consists of 417 female and 589 male, 1006 in total, physical education teacher candidates studying Physical Education and Sports Coaching. 269 of the teacher candidates in the scope of the research are in the first year, 277 of them are in the second year, 262 of them are in the third year and 198 of them are in the fourth year. ‘California Critical Thinking Disposition Inventory (Kökdemir, 2003) and ‘Empathy Tendency Criteria ‘(Dökmen, 1998) were used as the data collecting tools in the research. Descriptive statistics, t test, ANOVA, Tukey post-hoc test and correlation analysis were used in the data analysis. According to the evidence gathered from the research, critical thinking skills of the participants were at the low level and empathic tendencies of the participants were at middle level. Whereas there was no significant difference according to gender, the education level of the parents of the participants and the years of study spent at the university can be said to be determining factors in the critical thinking and empathic skills of the pre-service physical education teachers. The most striking result of the study was that the participants having parents with university degrees had lower empathic tendency and critical thinking skills than the ones whose parents did not graduate from a university. The results of the study showed that there was a direct correlation between the critical skills and the empathic tendencies of the participants. 180 180 Giriş Bilgiyi edinme ve kullanmada değişme ve gelişmelerin hızlı bir şekilde yaşandığı günümüzde eğitimin amacı, hür ve bilimsel düşünme gücüne sahip aynı zamanda yapıcı ve yaratıcı, kişilik ve teşebbüse değer veren verimli, toplumsal gelişimi sağlayacak bireyler yetiştirmektir. Bu bağlamda bireylere, özellikle de eğitim yoluyla, üst düzey düşünme gerektiren bazı bilişsel becerilerin kazandırılması hedeflenmektedir (MEB, 2007). Eleştirel düşünme ve empati becerisi de bu beceriler arasında yer almaktadır. Bu bağlamda, toplumda yer alan her bir bireyin, teknoloji ve bilgiyle baş edebilmesi için eleştirel düşünebilmesi ve düşüncelerini başkalarıyla paylaşarak etkili bir iletişim sağlayabilmesi için de empati becerisine sahip olması gerektiğini söyleyebiliriz. Paul (1993) eleştirel düşünmeyi, belirli bir düşünme alanına uygun aynı zamanda belirli bir düşünme şekli oluşturan, disiplinli bir şekilde yönlendirilmiş düşünme olarak ifade etmiştir. Change, eleştirel düşünmeyi “gerçekleri analiz etme, düşünceleri üretip düzenleme, düşünceleri savunma, karşılaştırma yapma, çıkarımlar yapma, savları değerlendirme ve sorun çözme becerilerinin bir bütünü” olarak tanımlamıştır (Akt: Doğanay & Ünal, 2006). 1990 yılında Amerikan Psikoloji Derneğinin (APA) öncülüğünde, çok sayıda kuramcı Delphi Projesi kapsamında eleştirel düşünmeye ilişkin "Bireyin ne yapacağına ve neye inanacağına öz düzenleyici olarak karar vermesi için ölçütlerin gerçeklerine yönelik çözümleyici, değerlendirmeye yönelik bilinçli yargılarda bulunması ve bu yargıları ifade etmesi" biçiminde bir tanım yapılmıştır (APA, 1990). Beden eğitiminde eleştirel düşünme ise, bir hareket hakkında savunucu ve mantıklı kararlar alabilmek için kullanılan yansıtıcı düşünmedir (McBride & Cleland, 1998). Beden eğitiminde yalnızca kılavuzla gösterilen bir hareketin yapılması veya verilen dönüte tepki verilmesi hareket ve bilişsel gelişim alanında sınırlılıklar yaratabilir. Bu bağlamda, harekete yönelik etkinliklerle bilişsel zorlukların etkileşimde bulunma gerekliliği beden eğitimi alanında eleştirel düşünmeyi son derece önemli kılmaktadır (Saçlı, 2008). Beden eğitimi derslerinde öğrencilerden eleştirel düşünme eğilimlerini kendi kendilerine kazanmaları ve kullanmalarını bekleyemeyiz. Bu bağlamda bu imkânı sağlayacak olan bu becerilere sahip beden eğitimi öğretmenleridir. 181 181 Eğitimin önemli fonksiyonlarından biri de toplumda iletişim kurma becerisine sahip, duyarlı, anlayışlı bireyler yetiştirilmesine yöneliktir. Bu durum empati kavramının kazandırılmasıyla mümkündür. Herkes çok iyi iletişim kuramayabilir ama en azından çevresiyle ilişkilerinde bazı bilgileri öğrenerek iletişim çatışmaları yaşamadan hayatını kaliteli bir şekilde sürdürebilir (Saygın, 2005). Bu bağlamda Dökmen (2003)’e göre empati, bir kişinin kendini karşısındaki kişinin yerine koyarak, olaylara onun bakış açısı ile bakması ve o kişinin duygu ve düşüncelerini doğru olarak anlaması ve bu durumu iletmesi sürecidir. Rogers, bir kişinin kendisini karşısındaki kişinin yerine koyarak olaylara onun bakış açısıyla bakması, o kişinin duygularını ve düşüncelerini doğru olarak anlaması, hissetmesi ve bu durumu ona iletmesi sürecine “empati” adını vermiştir (Akt: Dökmen, 2003: 135). Giriş “İnsanları hem ruhen, hem fikren ve bedenen olgunlaştıran, geliştiren, en az yorgunluğa karşılık en çok fayda sağlayan spor ortamında” (Açak, 2006: 1), iletişim kurabilmek de çok önemlidir. Hem fiziksel beceri öğretimi, hem spor öğretimi gibi zor bir görevi olan beden eğitimi öğretmenlerinin iletişim kuramadığı yerde başarının zor olacağı söylenebilir. Bu yüzden öğrencileri dinlemek ve gerçek problemin ne olduğunu tespit edip tanımladıktan sonra empati yapmak çok önemlidir (Rosenfeld & Wilder, 1990). Türk Milli Eğitim’in genel amaçları arasında gelişen ve değişen dünyaya ayak uydurabilmek için, düşünen ve düşüncelerini problem durumlarında kullanabilen, hızlı ve etkili karar verebilen, insan ilişkileri kuvvetli bireyler yetiştirilmesi yer almaktadır. Konu ile ilgili yurt içinde yapılan çalışmalar incelendiğinde, öğretmen adaylarının eleştirel düşünme eğilimlerinin incelendiği (Akar, 2007; Akbıyık, 2002; Akdere, 2012; Aybek, 2007; Bal, 2011; Cengiz, 2004; Deniz, 2009; Erdoğan, 2012; Göbel, 2013; Gülveren, 2007; Hazer, 2011; Kızıltaş, 2011; Kürüm, 2002; Özelçi, 2012; Tufan, 2008; Yeşilpınar 2011; Zayif, 2008) ve beden eğitimi öğretmeni adaylarının eleştirel düşünme eğilimlerinin incelendiği (Saçlı, 2008; 2013) aynı zamanda öğretmen adaylarının empatik eğilimlerinin incelendiği (Akbulut, 2010; Ata, 2010; Balcı, 2012; Çelik, 2008; Erçoban, 2003; Sargın, 1993; Yavuz, 2006; Yılmaz, 2011; Yiğiter, 2008) ve beden eğitimi öğretmeni adaylarının empatik eğilimlerinin incelendiği (Arslanoğlu, 2012; Erkmen, 2007) ve son olarak eleştirel düşünme ve empatik eğilimin birlikte incelendiği (Ekinci, 2009) çalışmalar mevcuttur. Diğer yandan yurt dışında yapılan çalışmalar incelendiğinde üniversite öğrencilerinin empatik eğilimlerinin incelendiği (Myyry & Helkema, 2001; Rohner, 2003); üniversite öğrencilerinin eleştirel düşünme eğilimlerinin incelendiği (Hager & Kaye, 1991; Jones & Ratcliff, 1993; McDonough, 1997); antrenörlük 182 bölümü öğrencilerinin eleştirel düşünme eğilimlerinin incelendiği (Leaver-Dunm ve diğ., 2002; Loken, 2005; Swiger, 2005); öğretmen adayları (Kong, 2001) ve öğretmenlerin eleştirel düşünme eğilimlerinin incelendiği (McBride & Knight, 1993); beden eğitimi öğretmenliği bölümünde öğrenim gören öğrenciler (McBride ve diğ., 1998) ve beden eğitimi öğretmenlerinin eleştirel düşünme eğilimlerinin incelendiği çalışmalar mevcuttur (Howarth, 1996). Fakat yapılan ulusal ve uluslararası literatür taramasında beden eğitimi öğretmeni adaylarının eleştirel düşünme ve empatik eğilimleri arasındaki ilişkiyi konu alan bir çalışmaya rastlanmamıştır. Milli Eğitim Bakanlığı’nın 2004–2005 yılında gerçekleştirdiği reform ile düşünme becerilerinin gelişimine önem veren çağdaş yaklaşımlar temel alınmış ve öğrenciyi merkeze alan bu yaklaşımla; farkındalıklarını bilen, kendini gerçekleştiren, işbirliğine ve grup çalışmasına istekli, öğrenmeyi öğrenen, düşünme ve etkili iletişim becerisi kazanan, teknolojiyi zamanı ve enerjiyi etkin kullanan bireyleri hedefleyen programlar geliştirilmeye çalışılmıştır (Avşar ve Temel, 2008’den aktaran Certel, Çatıkkaş ve Yalçınkaya, 2011: 75). Giriş Geliştirilen programların yanı sıra, öğrenme ve öğretme ortamlarının öğrencinin yaratıcı ve eleştirel düşünme becerilerini geliştirecek biçimde düzenlenmesi gerektiği ve bunu gerçekleştirme sorumluluğu taşıyan öğretmenlerin de gerek hizmet öncesi gerek hizmet süresince bu becerileri kazanma konusunda eğitilmeleri gerektiği belirtilmektedir (Seferoğlu & Akbıyık, 2006). Bu bağlamda temel amacı bireyin fiziksel, duygusal, bilişsel, devinişsel ve sosyal yönden gelişimine katkıda bulunmak, yaşam boyu fiziksel etkinliklere katılımını sağlamak olan beden eğitimi programıyla ulaşılması beklenen temel becerilerin başında eleştirel düşünme ve yaratıcı düşünme gelmekte ve aday beden eğitimi öğretmenlerinden de öğrencilerin eleştirel düşünme ve duygusal zekâ düzeylerini geliştirebilmeleri için, öncelikle kendilerinin bu beceri ve yeteneklere sahip olmaları beklenmektedir (Avşar & Temel, 2008’den aktaran Certel, Çatıkkaş & Yalçınkaya, 2011: 75). Dolayısıyla, bu araştırmanın beden eğitimi öğretmeni adaylarının bu becerilere hizmet öncesinde ne derece sahip olduklarının belirlenmesi, eleştirel düşünme ve empatik eğilimin bazı değişkenlere göre nasıl farklılaştığının belirlenmesi ve aralarında bir ilişkinin olup olmadığının ortaya çıkarılması noktasında önemli olduğu düşünülmektedir. Literatür taramasında eleştirel düşünme ve empatik eğilimin, cinsiyet değişkenine göre (Aktaş & Ünlü, 2013; Büküm, 2013; Dökmen, 1987; Göbel, 2013; Karademir, 2013; Kökdemir, 2003; Küçük & Uzun, 2013; Leaver & Dunm, 2002; Loken, 2005; Manger ve diğ., 2008; McDonough, 183 Erdoğan, 2012; Karakoyun, 2011; Kızıltaş, 2011; Kürüm, 2002; Leaver & Dunm, 2002; Saçlı, 2008; Yılmaz, 2011), sınıf değişkenine göre (Beşoluk & Önder, 2010; Çetinkaya, 2011; Ekinci, 2009; Ercoşkun, 2005; Erdoğan, 2012; Jones & Ratclift, 1993; Karalı, 2012; Kızıltaş, 2011; Küçük & Uzun, 2013; Kürüm, 2003; Leaver & Dunm, 2002; Özelç, 2012; Öztürk, 2006; Saçlı, 2008; Swiger, 2005), anne ve baba eğitim durumu değişkenlerine göre (Bal, 2011; Ekinci, 2009; Erdoğan, 2012; Gülveren, 2007; Karademir, 2013; Kürüm, 2002; Önder, 2010; Özelç, 2012; Tümkaya & Aybek, 2008) incelendiği görülmüştür. Beden eğitimi öğretmeni adayları üzerinde eleştirel düşünme ve empatik eğilimin birlikte yürütüldüğü bir çalışmaya rastlanılmadığından ve eleştirel düşünme ile empatik eğilimin cinsiyet, yaş, sınıf, anne ve baba eğitim durumu değişkenlerinden etkileneceği düşünüldüğünden bağımsız değişkenler oluşturulurken literatürde etkisi görülmüş bağımsız değişkenler tercih edilmiştir. Araştırma sonunda beden eğitimi öğretmeni yetiştiren kurumların yetiştirdikleri öğretmenlerin, mesleğe başlamadan önce, lisans eğitimleri esnasında eleştirel düşünme ve empatik eğilim becerisini kazanmış olmalarını sağlamaları ve bu becerileri okulda olduğu kadar, günlük yaşamlarında da kullanan bireyler olmalarını sağlayabilmelidirler. Bu bağlamda eksik yönlerini fark edip telafi etmeleri, beden eğitimi öğretmeni adaylarının empatik ve eleştirel düşünme eğilim düzeylerini görerek eksikliklerini geliştirmeleri beklenmektedir. Bu anlamda öğrencilerin yaşamında önemli etkileri olan öğretmenlerin eleştirel düşünebilmeleri ve empati kurabilmelerinin öneminin göz ardı edilemeyeceği söylenebilir. Giriş Buradan yola çıkılarak bu araştırmanın amacı; beden eğitimi öğretmeni adaylarının eleştirel düşünme ve empatik eğilimlerini bazı değişkenlere göre incelemek ve aralarındaki ilişkiyi belirlemektir. Bu amaç kapsamında aşağıdaki araştırma sorularına yanıt aranmıştır: 1. Beden eğitimi öğretmeni adaylarının eleştirel düşünme becerileri ve empatik eğilimleri ne düzeydedir? 2. Beden eğitimi öğretmeni adaylarının eleştirel düşünme becerileri ve empatik eğilimleri; cinsiyet, sınıf, anne ve baba eğitim durumu değişkenlerine göre anlamlı bir farklılık göstermekte midir? 3. Beden eğitimi öğretmeni adaylarının eleştirel düşünme becerileri ile empatik eğilimleri arasında anlamlı bir ilişki var mıdır? 184 Yöntem Bu araştırmada; olgular hakkında sistemli ve düzenli bilgilere ulaşmamızı sağlayan geniş gruplar üzerinde yürütülen, gruptaki bireylerin bir olgu veya olayla ilgili olarak görüşlerinin, tutumlarının alındığı, olgu ve olayların kendi koşulları içinde ve olduğu gibi betimlenmeye çalışıldığı tarama (survey) yöntemi (Arslanoğlu, 2013; Karakaya, 2009; Karasar, 2005) ve veri toplama tekniği olarak da bu model içerisinde sıklıkla yararlanılan anket tekniği kullanılmıştır (Nachmias & Nachmias, 1996). Araştırma Grubu Araştırma Grubu Araştırma grubunu oluşturan beden eğitimi öğretmeni adaylarına ilişkin demografik bilgiler aşağıda sunulmuştur (Tablo 1). Araştırma grubunu oluşturan beden eğitimi öğretmeni adaylarına ilişkin demografik bilgiler aşağıda sunulmuştur (Tablo 1). Tablo 1. Öğretmen adaylarına ait demografik bilgilerin dağılımı Tablo 1. Öğretmen adaylarına ait demografik bilgilerin dağılımı n Yüzde (%) Cinsiyet Kadın 417 41.5 Erkek 589 58.5 Sınıf 1.Sınıf 269 26.7 2.Sınıf 277 27.5 3.Sınıf 262 26.0 4.Sınıf 198 19.7 Anne Eğitim Durumu Okur-yazar değil 162 16.1 Okur-yazar 108 10.7 İlkokul 352 35.0 Ortaokul 121 12.0 Lise 177 17.6 Üniversite 86 8.5 Baba Eğitim Durumu Okur-yazar değil 54 5.4 Okur-yazar 70 7.0 İlkokul 265 26.3 Ortaokul 192 19.1 Lise 261 25.9 Üniversite 164 16.3 Üniversite Karadeniz Teknik Üniversitesi 120 11.9 Atatürk Üniversitesi 71 7.1 Akdeniz Üniversitesi 74 7.4 100. Yıl Üniversitesi 141 14.0 Balıkesir Üniversitesi 92 9.1 Anadolu Üniversitesi 79 7.9 Kocaeli Üniversitesi 118 11.7 Erzincan Üniversitesi 96 9.5 İnönü Üniversitesi 125 12.4 Hacettepe Üniversitesi 90 8.9 185 Araştırmanın çalışma grubunu, 2013-2014 eğitim-öğretim yılı bahar yarıyılında, Karadeniz Teknik Üniversitesi (N=120), Atatürk Üniversitesi (N=71), Akdeniz Üniversitesi (N=74), 100. Yıl Üniversitesi (N=141), Balıkesir Üniversitesi (N=92), Anadolu Üniversitesi (N=79), Kocaeli Üniversitesi (N=118), Erzincan Üniversitesi (N=96), İnönü Üniversitesi (N=125) ve Hacettepe Üniversitesi (N=90)’ne bağlı Beden Eğitimi ve Spor Öğretmenliği Bölümlerinde öğrenim gören 417 (%41.5) kadın ve 589 (%58.5) erkek toplam 1006 beden eğitimi öğretmeni adayı oluşturmaktadır. Araştırma kapsamındaki öğretmen adaylarının; 269 (%26.8)’u 1.sınıfta, 277 (%27.5)’si 2.sınıfta, 262 (%26.0)’si 3.sınıfta ve 198 (%19.7)’i 4.sınıfta öğrenim görmektedir. Katılımcılardan 162 (%16.1)’si anne eğitim durumunu okur-yazar değil, 108 (%10.7)’i okur-yazar, 352 (%35.0)’si ilkokul, 121 (%12.0)’i ortaokul, 177 (%17.6)’si lise, 86 (%8.5)’sı ise üniversite olarak belirtmiştir. Öğretmen adaylarından 54 (%5.4)’ü baba eğitim durumunu okur-yazar değil, 70 (%7.0)’i okur-yazar, 265 (%26.3)’i ilkokul, 192 (%19.1)’si ortaokul, 261 (%25.9)’i lise, 164 (%16.3)’ü ise üniversite olarak belirtmiştir. Araştırmanın çalışma grubu oluşturulurken, basit seçkisiz örnekleme yöntemi kullanılmıştır. Türkiye’nin beş farklı bölgesinde, beden eğitimi ve spor öğretmenliği eğitimi veren üniversiteler içerisinden basit seçkisiz örnekleme yöntemi ile seçilen on farklı üniversite çalışma grubunu oluşturmuştur. Empatik Eğitilim Ölçeği (EEÖ) Empatik Eğitilim Ölçeği (EEÖ) Empatik Eğilim Ölçeği, empatinin duygusal bileşenini ve bireylerin günlük yaşamdaki empati kurma potansiyellerini ölçmek amacıyla Dökmen (1988) tarafından geliştirilmiştir. Tek boyutlu bir yapıya sahip ve 20 ifadeden oluşan EEÖ’nin seçenekleri, 5’li Likert (1=Tamamen Aykırı, 2=Oldukça Aykırı, 3=Kararsızım, 4=Oldukça Uygun ve 5=Tamamen Uygun) tipinde olup bireylerin o maddedeki görüşe ne ölçüde katıldıklarını belirtmektedir. Ölçekten alınacak en düşük puan 20, en yüksek puan 100’dür. Puanın yüksek olması, empatik eğilimin yüksek olduğunu; düşük olması empatik eğilimin düşük olduğunu göstermektedir. Belirlenmiş bir norm olmadığından, bu çalışma kapsamında grubun aritmetik ortalamaları karşılaştırılarak puanlar yorumlanmıştır. EEÖ, Dökmen (1988) tarafından 70 kişilik bir öğrenci grubuna, üç hafta arayla iki defa uygulanmıştır. Bu uygulamadan elde edilen korelasyon katsayısı 0.82’dir. Ölçek yarılama yöntemiyle hesaplanan güvenirlik katsayısı 0.86, Cronbach Alpha güvenirlik katsayısı ise 0.72’dir. EEÖ’nin geçerlik çalışması yine Dökmen (1988) tarafından yapılmıştır. 24 kişilik bir denek grubunun EEÖ ve Edwards Kişisel Tercih Envanteri’nin “Duyguları Anlama” boyutundan aldıkları puanlar arasındaki korelasyon katsayısı 0.68 olarak bulunmuştur. Bu araştırma kapsamında toplanan veriler üzerinde hesaplanan Cronbach Alpha iç tutarlık katsayısı ise 0.70 olarak tespit edilmiştir. Verilerin Toplanması Veri Toplama Araçları Veri Toplama Araçları Araştırmada, Kaliforniya Eleştirel Düşünme Eğilimi Ölçeği ve Empatik Eğilim Ölçeği veri toplama aracı olarak kullanılmıştır. Kaliforniya Eleştirel Düşünme Eğilimi Ölçeği (KEDEÖ) Bu ölçek 1990 yılında Amerikan Felsefe Derneği’nin düzenlediği Delphi projesinin bir sonucu olarak ortaya çıkmıştır. Ölçeğin Türkçe’ye uyarlama çalışması Kökdemir (2003) tarafından 913 öğrenci üzerinde yapılmıştır. Ölçeğin kuramsal olarak belirlenmiş ve psikometrik olarak test edilmiş 7 alt boyutu ve 75 maddesi bulunmaktadır. KEDEÖ’nin alt boyutlarının iç tutarlık katsayıları 0.60 ile 0.78 arasında değişirken, toplam puan için ölçeğin iç tutarlık katsayısı 0.90 olarak hesaplanmıştır (Kökdemir, 2003). Madde toplam puan korelasyonu ve temel bileşenler analizi sonucunda ölçek 75 maddeden 51 maddeye, 7 boyuttan 6 boyuta (Analitiklik 10 madde; Açık Fikirlilik 12 madde; Meraklılık 9 madde; Kendine Güven 7 madde; Doğruyu Arama 7 madde ve Sistematiklik 6 madde) indirilmiştir. Yeni ölçeğin alt boyutlarının iç tutarlık katsayıları 0.61 ile 0.78 arasında değişirken, ölçeğin iç tutarlık katsayısı ise 0.88 olarak belirlenmiştir. Her bir alt boyut için puanı 40’tan düşük olan 186 bireylerin o boyuttaki eleştirel düşünme eğilimlerinin düşük; puanı 50’den yüksek olan bireylerin ise o boyuttaki eleştirel düşünme eğilimlerinin yüksek olduğu söylenebilir. Dolayısıyla ölçek bir bütün olarak değerlendirildiğinde puanı 240’tan az olan kişilerin genel eleştirel düşünme eğilimlerinin düşük, puanı 300’den fazla olanların ise bu eğilimlerin yüksek olduğu düşünülmektedir. Ölçek maddelerine ait seçenekler 6’lı Likert’tir. Bu araştırma kapsamında toplanan veriler üzerinde toplam ölçek için Cronbach Alpha iç tutarlık katsayısı 0.83 olarak hesaplanmıştır. Verilerin Toplanması Araştırmada kullanılan veri toplama araçları, 2013-2014 Eğitim-Öğretim Yılı Bahar Yarıyılında, uygulama yapılacak üniversitelerin ilgili birimlerinden gerekli izinlerin alınmasını takiben çalışma grubunu oluşturan öğretmen adaylarına ders esnasında uygulanmıştır. Yönergede, araştırmanın amacı ile ilgili gerekli açıklamalar yapılarak veri toplama aracının doldurulması hakkında detaylı bilgiler verilmiştir. Veri toplama aracı, öğrenciler tarafından gönüllü olarak doldurulmuştur. Araştırmacı tarafından toplanan anket 187 formları (1030), kontrol edilerek eksik veya yanlış doldurulanlar araştırma dışında tutulmuştur. Daha sonra anketlerin geçerli ve kabul edilebilir nitelikte olanları, değerlendirilmek üzere kodlanarak bilgisayar ortamına aktarılmıştır. Verilerin değerlendirilmesinde istatistikî yöntem olarak; frekans, yüzdelik dağılım, aritmetik ortalama, standart sapma; parametrik testlerden bağımsız gruplar için t-testi, tek yönlü varyans analizi (ANOVA), Tukey HSD çoklu karşılaştırma ve Pearson Korelasyon testleri kullanılmıştır. Verilerin parametrik testlerin ön şartlarını sağlayıp sağlamadığına Skewness ve Kurtosis (verilerin normal dağılım durumu) değerleri ve Levene (varyansların eşitliği) testi sonuçları incelenerek karar verilmiştir (Büyüköztürk, 2008). Verilerin çarpıklık (KEDEÖ=0.58, EEÖ=0.08) ve basıklık (KEDEÖ=0.10, EEÖ=0.11) değerleri ile Levene testi (sınıf değişkeni için: pKEDEÖ=0.34, pEEÖ=0.60; anne eğitim durumu değişkeni için:pKEDEÖ=0.26, pEEÖ=0.12; baba eğitim durumu değişkeni için: pKEDEÖ=0.99, pEEÖ=0.10) sonuçları incelendiğinde verilerin parametrik testlerin ön şartlarını sağladığı söylenebilir. Verilerin analizinde, anlamlılık düzeyi 0.05 olarak alınmıştır. Çalışma kapsamındaki verilerin analizinde korelasyon katsayısının, mutlak değer olarak 0.00-0.30 arasında olması düşük düzeyde; 0.30-0.70 arasında olması orta düzeyde ve 0.70-1.00 arasında olması yüksek düzeyde bir ilişki olduğu şeklinde ifade edilebilir (Büyüköztürk, 2008). Bulgular Çalışma kapsamındaki katılımcıların, KEDEÖ’nden aldıkları toplam puanların aritmetik ortalaması 202.11 ve standart sapması 23.35’tir. KEDEÖ puanları faktör bazında incelendiğinde en düşük puan ortalamasının Doğruyu Arama boyutunda (23.10), en yüksek puan ortalamasının ise Açık Fikirlilik boyutunda (44.26) olduğu görülmektedir. Çalışma kapsamında yer alan beden eğitimi öğretmeni adaylarının toplam ölçekten aldıkları en düşük puan 128 iken en yüksek puan ise 303’tür. Katılımcıların, EEÖ’nden aldıkları toplam puanların aritmetik ortalaması 67.87 ve standart sapması da 8.58’dir. Katılımcıların toplam ölçekten aldıkları en düşük puan 39 iken en yüksek puan ise 96’dır (Tablo 2). Tablo 2. Ölçek puanlarının dağılımı Madde Sayısı Ort ± Ss Minimum Maksimum KEDEÖ 51 202.11 ± 23.35 128.00 303.00 Analitiklik 10 43.79 ± 7.23 14.00 60.00 Açık Fikirlilik 12 44.26 ± 9.43 19.00 72.00 Meraklılık 9 38.59 ± 6.90 13.00 54.00 Kendine Güven 7 28.39 ± 5.24 7.00 42.00 188 Tablo 2. Ölçek puanlarının dağılımı Madde Sayısı Ort ± Ss Minimum Maksimum KEDEÖ 51 202.11 ± 23.35 128.00 303.00 Analitiklik 10 43.79 ± 7.23 14.00 60.00 Açık Fikirlilik 12 44.26 ± 9.43 19.00 72.00 Meraklılık 9 38.59 ± 6.90 13.00 54.00 Kendine Güven 7 28.39 ± 5.24 7.00 42.00 Doğruyu Arama 7 23.10 ± 5.06 9.00 42.00 Sistematiklik 6 23.98 ± 4.90 9.00 36.00 EEÖ 20 67.87 ± 8.58 39.00 96.00 Doğruyu Arama 7 23.10 ± 5.06 9.00 42.00 Sistematiklik 6 23.98 ± 4.90 9.00 36.00 EEÖ 20 67.87 ± 8.58 39.00 96.00 Tablo 3’te katılımcıların KEDEÖ ve EEÖ’nden aldıkları puanların cinsiyet değişkenine göre t-testi sonuçları sunulmuştur. Analizler, cinsiyet değişkenine göre hem KEDEÖ hem de EEÖ puanlarının anlamlı bir şekilde farklılaşmadığını göstermektedir. Tablo 3. Cinsiyet değişkenine göre t-testi sonuçları a o y ğ ş g ç Cinsiyet Kadın (N=417) Erkek (N=589) Ort ± Ss Ort ± Ss t p KEDEÖ 202.26 ± 24.76 202.00 ± 22.32 0.17 0.87 Analitiklik 44.01 ± 7.37 43.64 ± 7.14 0.78 0.43 Açık Fikirlilik 44.82 ± 10.07 43.86 ± 8.93 1.56 0.12 Meraklılık 38.44 ± 7.24 38.70 ± 6.65 0.58 0.56 Kendine Güven 28.22 ± 5.55 28.51 ± 5.00 0.86 0.39 Doğruyu Arama 23.16 ± 5.08 23.05 ± 5.05 0.33 0.74 Sistematiklik 23.61 ± 4.82 24.24 ± 4.95 1.99 0.05 EEÖ 68.25 ± 9.02 67.60 ± 8.25 1.17 0.24 Tablo 4’te katılımcıların KEDEÖ ve EEÖ’nden aldıkları puanların sınıf değişkenine göre ANOVA sonuçları sunulmuştur. Analizler, sınıf değişkenine göre KEDEÖ puanlarının anlamlı bir şekilde farklılaştığını göstermektedir. Bulgular Katılımcıların toplam ölçek [F(3,1002)=10.85, p<0.05], Analitiklik [F(3,1002)=7.78, p<0.05], Açık Fikirlilik [F(3,1002)= 3.08, p<0.05], Meraklılık [F(3,1002)=12.89, p<0.05], Kendine Güven [F(3,1002)=6.42, p<0.05] ve Sistematiklik [F(3,1002)=4.73, p<0.05] alt boyutlarından aldıkları puanların ortalamaları arasındaki fark, istatistiksel açıdan anlamlı bulunmuştur. Alt ölçeklerden alınan puanlar incelendiğinde Doğruyu Arama boyutu dışındaki alt boyutlarda 1.sınıfta öğrenim gören katılımcıların ortalama puanlarının daha düşük olduğu görülmektedir. Farklılaşmanın hangi grup lehine olduğunu tespit etmek amacıyla yapılan Post Hoc Tukey HSD testi sonucu incelendiğinde, katılımcıların toplam ölçekten aldıkları ortalama puanlara göre 1. sınıfta öğrenim gören katılımcıların ortalama puanlarının, 2, 3 ve 4. sınıfta öğrenim gören katılımcıların ortalama puanlarından anlamlı düzeyde düşük olduğu söylenebilir. Analizler, sınıf değişkenine göre EEÖ puanlarının da anlamlı bir şekilde farklılaştığını göstermektedir. Katılımcıların EEÖ’den [F(3,1002)=8.23, p<0.05] aldıkları puanların ortalamaları arasındaki fark, istatistiksel açıdan anlamlı bulunmuştur. Yapılan Post Hoc Tukey HSD testi sonucu incelendiğinde 1. sınıfta öğrenim göre katılımcıların puanlarının, 2, 3 ve 4. sınıfta öğrenim gören katılımcıların puanlarından anlamlı bir şekilde düşük olduğu söylenebilir. 189 Tablo 4. Sınıf değişkenine göre ANOVA sonuçları Sınıf 1.Sınıf (N=269) 2.Sınıf (N=277) 3.Sınıf (N=262) 4.Sınıf (N=198) Ort ± Ss Ort ± Ss Ort ± Ss Ort ± Ss F p KEDEÖ 195.95 ± 22.42 203.71 ± 24.36 202.48 ± 22.64 207.75 ± 22.35 10.85 0.00* Analitiklik 42.04 ± 7.49 44.50 ± 7.39 44.05 ± 6.92 44.84 ± 6.68 7.78 0.00* Açık Fikirlilik 43.51 ± 8.40 44.99 ± 9.85 43.33 ± 9.24 45.48 ± 10.22 3.08 0.03* Meraklılık 36.53 ± 6.59 38.78 ± 6.88 39.32 ± 6.86 40.16 ± 6.79 12.89 0.00* Kendine Güven 27.40 ± 5.01 28.26 ± 5.48 28.77 ± 5.04 29.43 ± 5.23 6.42 0.00* Doğruyu Arama 23.32 ± 4.77 22.90 ± 5.59 23.10 ± 4.97 23.07 ± 4.81 0.32 0.81 Sistematiklik 23.15 ± 4.78 24.27 ± 5.10 23.91 ± 4.85 24.77 ± 4.70 4.73 0.00* EEÖ 65.87 ± 8.34 68.57 ± 8.88 67.91 ± 8.30 69.56 ± 8.36 8.23 0.00* p<0.05 Tablo 4. Sınıf değişkenine göre ANOVA sonuçları Tablo 5’te katılımcıların KEDEÖ ve EEÖ’nden aldıkları puanların anne eğitim durumu değişkenine göre ANOVA sonuçları sunulmuştur. Analizler, anne eğitim durumu değişkenine göre KEDEÖ puanlarının anlamlı bir şekilde farklılaştığını göstermektedir. Katılımcıların toplam ölçek [F(5,1000)=9.46, p<0.05], Analitiklik [F(5,1000)=9.37, p<0.05], Açık Fikirlilik [F(5,1000)=3.38, p<0.05], Meraklılık [F(5,1000)=6.24, p<0.05] ve Sistematiklik [F(5,1000)=5.76, p<0.05] alt boyutlarından aldıkları puanların ortalamaları arasındaki fark, istatistiksel açıdan anlamlı bulunmuştur. Bulgular Alt ölçeklerden alınan puanlar incelendiğinde Kendine Güven ve Doğruyu Arama boyutları dışındaki alt boyutlarda anne eğitim durumunu üniversite olarak belirten katılımcıların ortalama puanlarının daha düşük olduğu görülmektedir. Farklılaşmanın hangi grup lehine olduğunu tespit etmek amacıyla yapılan Post Hoc Tukey HSD testi sonucu incelendiğinde, katılımcıların toplam ölçekten aldıkları ortalama puanlara göre, anne eğitim durumunu üniversite olarak belirten katılımcıların ortalama puanlarının, anne eğitim durumunu okur-yazar değil, ilkokul, ortaokul ve lise olarak belirten katılımcıların ortalama puanlarından anlamlı düzeyde düşük olduğu söylenebilir. Analizler, anne eğitim durumu değişkenine göre EEÖ puanlarının da anlamlı bir şekilde farklılaştığını göstermektedir. Katılımcıların EEÖ’den [F(5,1000)=3.49, p<0.05] aldıkları puanların ortalamaları arasındaki fark, istatistiksel açıdan anlamlı bulunmuştur. Yapılan Post Hoc Tukey HSD testine göre anne eğitim durumu Üniversite olan katılımcıların ortalama puanlarının, anne eğitim durumu ilkokul ve ortaokul olan katılımcıların ortalama puanlarından anlamlı bir şekilde düşük olduğu söylenebilir. 190 190 Tablo 5. Anne eğitim durumu değişkenine göre ANOVA sonuçları Eğitim Durumu* 1 (N=162) 2 (N=108) 3 (N=352) 4 (N=121) 5 (N=177) 6 (N=86) Ort ± Ss Ort ± Ss Ort ± Ss Ort ± Ss Ort ± Ss Ort ± Ss F p KEDEÖ 203.35 ± 195.43 ± 207.29 ± 201.57 ± 200.26 ± 191.55 ± 9.46 0.00* Analitiklik 44.69 ± 6.96 41.93 ± 7.46 45.24 ± 6.90 43.90 ± 7.41 42.77 ± 7.34 40.51 ± 6.60 9.37 0.00* Açık Fikirlilik 44.54 ± 10.00 43.51 ± 8.60 45.58 ± 9.48 42.59 ± 9.09 44.08 ± 9.73 41.97 ± 8.22 3.38 0.00* Meraklılık 39.31 ± 6.86 36.48 ± 6.73 39.59 ± 6.86 39.16 ± 7.03 37.93 ± 6.73 36.39 ± 6.45 6.24 0.00* Kendine 28.34 ± 5.26 27.87 ± 5.39 28.57 ± 5.14 28.87 ± 5.47 28.46 ± 5.13 27.63 ± 5.27 0.87 0.50 Doğruyu 22.59 ± 4.94 22.79 ± 5.20 23.54 ± 5.11 23.08 ± 5.21 22.93 ± 5.00 22.98 ± 4.79 0.99 0.42 Sistematiklik 23.88 ± 5.82 22.85 ± 4.61 24.77 ± 4.63 23.97 ± 4.34 24.09 ± 4.92 22.08 ± 4.43 5.76 0.00* EEÖ 67.71 ± 7.90 67.37 ± 8.72 68.92 ± 8.82 68.57 ± 8.72 67.13 ± 8.66 65.04 ± 7.58 3.49 0.00* * p<0.05 1: Okur-Yazar Değil, 2: Okur-Yazar, 3: İlkokul, 4: Ortaokul, 5: Lise, 6: Üniversite Tablo 5. Anne eğitim durumu değişkenine göre ANOVA sonuçları 1: Okur-Yazar Değil, 2: Okur-Yazar, 3: İlkokul, 4: Ortaokul, 5: Lise, 6: Üniversite Tablo 6’da katılımcıların KEDEÖ ve EEÖ’nden aldıkları puanların baba eğitim durumu değişkenine göre ANOVA sonuçları sunulmuştur. Analizler, baba eğitim durumu değişkenine göre KEDEÖ puanlarının anlamlı bir şekilde farklılaştığını göstermektedir. Katılımcıların toplam ölçek [F(5,1000)=4.53, p<0.05], Analitiklik [F(5,1000)=3.51, p<0.05], Açık Fikirlilik [F(5,1000)=2.96, p<0.05], Meraklılık [F(5,1000)=2.36, p<0.05] ve Sistematiklik [F(5,1000)=2.92, p<0.05] alt boyutlarından aldıkları puanların ortalamaları arasındaki fark, istatistiksel açıdan anlamlı bulunmuştur. Farklılaşmanın hangi grup lehine olduğunu tespit etmek amacıyla yapılan Post Hoc Tukey HSD testi sonucu incelendiğinde, katılımcıların toplam ölçekten aldıkları ortalama puanlara göre, baba eğitim durumunu üniversite olarak belirten katılımcıların puanlarının, baba eğitim durumunu ilkokul ve ortaokul olarak belirten katılımcıların puanlarından anlamlı düzeyde düşük olduğu söylenebilir. Analizler, baba eğitim durumu değişkenine göre EEÖ puanlarının da anlamlı bir şekilde farklılaştığını göstermektedir. Katılımcıların EEÖ’den [F(5,1000)=5.20, p<0.05] aldıkları puanların ortalamaları arasındaki fark, istatistiksel açıdan anlamlı bulunmuştur. Yapılan Post Hoc Tukey HSD testine göre baba eğitim durumu Üniversite olan katılımcıların ortalama puanlarının, baba eğitim durumu ilkokul, ortaokul ve lise olan katılımcıların ortalama puanlarından anlamlı bir şekilde daha düşük olduğu söylenebilir. 191 Tablo 6. 1: Okur-Yazar Değil, 2: Okur-Yazar, 3: İlkokul, 4: Ortaokul, 5: Lise, 6: Üniversite Baba eğitim durumu değişkenine göre ANOVA sonuçları Eğitim Durumu 1 (N=54) 2 (N=70) 3 (N=265) 4 (N=192) 5 (N=261) 6 (N=164) Ort ± Ss Ort ± Ss Ort ± Ss Ort ± Ss Ort ± Ss Ort ± Ss F p KEDEÖ 195.69 ± 22.58 198.01 ± 23.89 205.07 ± 22.78 205.10 ± 23.64 202.67 ± 23.42 196.80 ± 22.56 4.5 3 0.00* Analitiklik 42.88 ± 6.77 42.83 ± 8.09 44.56 ± 7.37 44.83 ± 7.09 43.73 ± 6.67 42.17 ± 7.52 3.5 1 0.00* Açık Fikirlilik 41.94 ± 9.07 43.70 ± 8.89 45.86 ± 9.16 44.48 ± 9.94 43.91 ± 9.63 42.98 ± 8.97 2.9 6 0.01* Meraklılık 37.52 ± 6.81 38.30 ± 6.53 39.04 ± 7.05 39.45 ± 7.45 38.65 ± 6.60 37.25 ± 6.48 2.3 6 0.04* Kendine Güven 27.61 ± 5.96 28.49 ± 5.32 27.97 ± 5.22 28.48 ± 5.39 29.14 ± 5.00 28.01 ± 5.09 1.8 4 0.10 Doğruyu Arama 22.71 ± 5.00 22.02 ± 4.87 23.52 ± 5.08 23.21 ± 5.32 22.99 ± 4.94 23.04 ± 5.00 1.1 0 0.36 Sistematiklik 23.04 ± 5.86 22.68 ± 5.56 24.12 ± 4.75 24.66 ± 4.69 24.26 ± 4.92 23.35 ± 4.55 2.9 2 0.01* EEÖ 65.69 ± 7.59 67.07 ± 7.93 68.17 ± 8.95 69.27 ± 8.26 68.74 ± 8.95 65.43 ± 7.70 5.2 0 0.00* p<0.05 1: Okur-Yazar Değil, 2: Okur-Yazar, 3: İlkokul, 4: Ortaokul, 5: Lise, 6: Üniversite Tablo 6. Baba eğitim durumu değişkenine göre ANOVA sonuçları 1: Okur-Yazar Değil, 2: Okur-Yazar, 3: İlkokul, 4: Ortaokul, 5: Lise, 6: Üniversite Katılımcıların KEDEÖ ve EEÖ’nden aldıkları puanlar arasındaki korelasyon katsayıları, eleştirel düşünme becerisi ve empatik eğilim arasında pozitif yönde, düşük ve orta düzeylerde anlamlı bir ilişki olduğunu göstermektedir (Tablo 7). Tablo 7. KEDEÖ ve EEÖ puanları arasındaki korelasyon sonuçları Empatik Eğilim Ölçeği Puanları n r KEDEÖ 1006 0.52* Analitiklik 1006 0.41 Açık Fikirlilik 1006 0.16 Meraklılık 1006 0.41 Kendine Güven 1006 0.38 Doğruyu Arama 1006 0.14 Sistematiklik 1006 0.43 *p<0.01 Tartışma ve Sonuç Tablo 7. KEDEÖ ve EEÖ puanları arasındaki korelasyon sonuçları Tartışma ve Sonuç Beden eğitimi öğretmeni adaylarının eleştirel düşünme becerileri ve empatik eğilimlerini bazı değişkenlere göre incelemek ve aralarındaki ilişkiyi belirlemek amacıyla yürütülen bu çalışmadan elde edilen bulgulara göre beden eğitimi öğretmeni adaylarının eleştirel düşünme eğilimlerinin düşük seviyede olması literatürdeki bazı çalışmalarla paralellik göstermektedir (Akdere, 2012; Erdoğan, 2012; Göbel, 2013). Diğer yandan aday öğretmenlerin eleştirel düşünme eğiliminin yüksek düzeyde olduğu (Karademir, 2013; Karalı, 2012; Yeşilpınar, 2011) az sayıda çalışmalar dışında eleştirel düşünme seviyesinin orta ve düşük düzeyde olduğu sonucuna ulaşılmıştır. Paul (2005), günümüzde birçok 192 192 üniversitede, eğitimcilerin, eleştirel düşünme becerisi açısından eksik olduğunu, özellikle, eleştirel düşünme ile yaratıcı düşünme, problem çözme, karar verme ya da iletişim arasındaki yakın ilişkiden habersiz olunduğunu ifade etmiştir. Derslerde öğrencilerin eleştirel düşünmelerini geliştiren öğretim yöntemlerinin kullanımında yetersiz kalındığını ileri sürmektedir (Saçlı, 2008). Diğer yandan Korkmaz (2009), yaptığı çalışmada, eğitim fakültesinde verilen eğitimin, öğrencilerin eleştirel düşünme eğilim ve düzeylerine yeterince katkı sağlamadığını saptamış, bunun nedeni olarak ise eğitimin çok teorik kalması, öğrenme- öğretme ortamlarında öğrencilerin genellikle pasif olması, değerlendirme faaliyetlerinde genellikle çoktan seçmeli soruların tercih edilmesi, analiz, sentez ve değerlendirme becerisi gerektiren etkinliklere az yer verilmesi ve benzeri gibi uygulamaların olabileceğini belirtilmiştir. Bu araştırma kapsamında, beden eğitimi ve spor öğretmenliği programlarında öğrenim gören öğrencilerin düşük eleştirel düşünme becerilerine sahip olmalarının yukarıda belirtilen durumların beden eğitimi ve spor öğretmenliği bölümlerinde öğrenim gören öğretmen adayları için de geçerli olduğu şeklinde yorumlanabilir. Beden eğitimi öğretmeni adaylarının empatik eğilimlerinin orta seviyede olması Deniz (2009), Dev (2010) ve Ekinci (2009) tarafından yapılan çalışmalarla benzerlik göstermektedir. Arslanoğlu (2012), Ata (2010), Yılmaz & Akyel (2008) tarafından yapılan çalışmalarda empatik eğilim düzeyi yüksek olarak tespit edilmiştir. Bu sonuçlara göre, empatinin etkili öğretimin ve iletişimin önemli bir özelliği olduğu dikkate alındığında beden eğitimi öğretmeni adaylarının, empatik eğilimlerinin orta seviyede olması onların etkili öğretim ve iletişim yapamadıkları anlamına gelebilir. Empati becerisi öğretilebilir bir kavram olduğundan beden eğitimi öğretmenliği programlarında çeşitli öğretim teknikleriyle dâhil edilmesi gerektiği düşünülebilir. Cinsiyet değişkenine göre katılımcıların KEDEÖ puanları incelendiğinde kadın ve erkek katılımcıların toplam ölçek ve ölçek alt faktörlerinden aldıkları puanların ortalamaları arasındaki fark, istatistiksel açıdan anlamlı bulunmamıştır. Bu bulguya paralel olarak bazı çalışmalarda cinsiyete göre farklılık tespit edilmemiştir (Akar, 2007; Çekiç, 2007; Erdoğan, 2012; Göbel, 2013; Karademir, 2013; Karalı, 2012; Küçük & Uzun, 2013; Kürüm, 2002; Leaver- Dunn & diğ., 2002; Loken, 2005; McDonough, 1997; Saçlı, 2008; Tokyürek, 2001). Tartışma ve Sonuç Bu durum Karalı (2012), Kızıltaş (2011), Saçlı (2008) ve Zayif (2008) tarafından yapılan çalışmalar ile benzerlik göstermektedir. Öte yandan sınıf düzeyi arttıkça öğrencilerin eleştirel düşünme becerilerinde azalma olan çalışmalar da mevcuttur (Akar, 2007; Gülveren, 2007). Diğer taraftan eleştirel düşünme eğiliminin sınıf değişkenine göre farklılığın tespit edilmediği çalışmalar da mevcuttur (Beşoluk & Önder, 2010; Ekinci, 2009; Erdoğan, 2012; Küçük & Uzun, 2013; Leaver-Dunn & diğ. 2002). Korkmaz (2009)’a göre, bu durum hem lise düzeyinde, hem de özellikle yükseköğrenim düzeyinde verilen derslerin, derslerin veriliş biçimlerinin ve ölçme değerlendirme süreçlerinin, öğrencilerin eleştirel düşünme eğilim ve düzeylerine yeterince katkı sağlamamasından kaynaklanmış olabilir. Çetinkaya (2011), araştırmalar sonucunda çıkan bulgulara göre üniversite eğitimi almış meraklı, kendine güvenen ve sistematik düşünebilen bireyler olmaları beklendiğini öğrencilerden, daha analitik, açık fikirli, doğruyu arama eğiliminde, dolayısıyla bu farklılığın sebebinin öğretmen, öğretim süreci, dersler, içerik vb. açıdan çok kapsamlı bir biçimde araştırılması gerektiğini söylemiştir. Bu araştırma kapsamında düşünüldüğünde drama dersi gibi eleştirel düşünme eğiliminin gelişimine katkı sağlayacak derslerin sonraki dönemlerde görülmeye başlanması, öğretmen adaylarının eleştirel düşünmeleri üzerindeki farklılıkları açıklar nitelikte olabilir. Ayrıca Çıkrıkçı (1996)’ya göre eleştirel düşünme, olgunlaşma ve yaşantı zenginliğiyle artmakta ve zihinsel yeteneğin önemli bir boyutunu oluşturmaktadır (Ay & Akgöl, 2008). Bu açıdan bakıldığında sınıf düzeyi büyüdükçe olgunlaşmanın ve yaşantı zenginliğinin arttığı ve bu durumun da eleştirel düşünme eğilimini olumlu etkilediği söylenebilir. öğrenim gören katılımcıların ortalama puanları, 2. 3. ve 4. sınıfta öğrenim gören katılımcıların ortalama puanlarından daha düşüktür. Bu durum Karalı (2012), Kızıltaş (2011), Saçlı (2008) ve Zayif (2008) tarafından yapılan çalışmalar ile benzerlik göstermektedir. Öte yandan sınıf düzeyi arttıkça öğrencilerin eleştirel düşünme becerilerinde azalma olan çalışmalar da mevcuttur (Akar, 2007; Gülveren, 2007). Diğer taraftan eleştirel düşünme eğiliminin sınıf değişkenine göre farklılığın tespit edilmediği çalışmalar da mevcuttur (Beşoluk & Önder, 2010; Ekinci, 2009; Erdoğan, 2012; Küçük & Uzun, 2013; Leaver-Dunn & diğ. 2002). Korkmaz (2009)’a göre, bu durum hem lise düzeyinde, hem de özellikle yükseköğrenim düzeyinde verilen derslerin, derslerin veriliş biçimlerinin ve ölçme değerlendirme süreçlerinin, öğrencilerin eleştirel düşünme eğilim ve düzeylerine yeterince katkı sağlamamasından kaynaklanmış olabilir. Çetinkaya (2011), araştırmalar sonucunda çıkan bulgulara göre üniversite eğitimi almış meraklı, kendine güvenen ve sistematik düşünebilen bireyler olmaları beklendiğini öğrencilerden, daha analitik, açık fikirli, doğruyu arama eğiliminde, dolayısıyla bu farklılığın sebebinin öğretmen, öğretim süreci, dersler, içerik vb. açıdan çok kapsamlı bir biçimde araştırılması gerektiğini söylemiştir. Bu araştırma kapsamında düşünüldüğünde drama dersi gibi eleştirel düşünme eğiliminin gelişimine katkı sağlayacak derslerin sonraki dönemlerde görülmeye başlanması, öğretmen adaylarının eleştirel düşünmeleri üzerindeki farklılıkları açıklar nitelikte olabilir. Tartışma ve Sonuç Diğer taraftan Aktaş & Ünlü (2013), Ay & Akgöl (2008), Beşoluk & Önder (2010), Gülveren (2007), Hazer (2011), Kızıltaş (2011), Kökdemir (2003), Özelçi (2012), Uluçınar (2012), Torun (2011) ve Zayif (2008) tarafından yapılan çalışmalarda cinsiyete göre farklılık tespit edilmiştir. Bu 193 çalışmalardan Kızıltaş (2011) ve Torun (2011) tarafından yapılan çalışmada erkeklerin lehine sonuç bulunmuş ve bu bulgu erkek öğretmenlerin kadın öğretmenlere göre daha yüksek düzeyde meraklılık eleştirel düşünme eğilimine sahip oldukları şeklinde yorumlanmıştır. Beşoluk & Önder (2010), Hazer (2011), Uluçınar (2012) ve Zayif (2008) tarafından yapılan çalışmalarda kadınların lehine sonuç bulunmuştur. Bu bulguya göre de kadınların temkinli düşünme eğilimlerinin erkeklere oranla daha yüksek olduğu görülmüştür. Cinsiyet değişkenine göre anlamlı bir farklılığın tespit edilememesi, çalıştığımız örneklemde yer alan beden eğitimi öğretmeni adayları için eleştirel düşünmenin cinsiyet değişkeninden bağımsız bir eğilim olduğu söylenebilir. Araştırmadan elde edilen bulgulara göre katılımcıların EEÖ’den aldıkları puanlar, cinsiyet değişkenine göre anlamlı bir farklılık göstermemektedir. Bu sonuç literatürdeki bazı çalışmalarla paralellik göstermektedir (Balcı, 2012; Büküm, 2013; Ercoşkun, 2005; Erkmen, 2007; Dökmen, 1987; Manger ve diğ., 2001; Tanrıdağ, 1992; Yiğiter, 2008). Literatür incelendiğinde, cinsiyetin empatik beceri üzerindeki etkisi ile ilgili farklı araştırma bulgularının olduğu görülmektedir (Akbulut, 2010; Arslanoğlu, 2012; Ekinci, 2009; Genç & Kalafat, 2008; Karakoyun, 2011). Bu bulguların birçoğunda farklılığın önemli ve kadınların lehine olduğu sonucuna ulaşılmıştır. Erkmen (2007) beden eğitimi ve spor yüksekokulu öğrencileri ile yaptığı çalışma sonucunda empatik eğilim ile cinsiyet arasında istatistiksel olarak anlamlı bir fark bulmamasına rağmen kadınların empatik eğilim puan ortalamasının erkeklerden daha yüksek olduğunu belirtmiştir. Bu bulgular kadınların erkeklere oranla olaylara daha çok duygusal tepki gösterdikleri genel kabulüne de uygundur (Dökmen, 2003). Genel olarak kadınların lehine sonuçlanan empatik eğilimin, bu çalışma kapsamında kadın ve erkek katılımcılar arasında, cinsiyete göre farklılık oluşturmaması, beden eğitimi ve spor öğretmenliği programında okuyan kadın öğrencilerin androjen kimliğe sahip olmalarından kaynaklandığı düşünülebilir. Araştırma sonuçları, birçok çalışmada (Gülveren, 2007; Jones & Ratclift, 1993; Karalı, 2012; Kızıltaş, 2011; Kürüm, 2002; Özelçi, 2012; Saçlı, 2008; Swiger, 2005; Uluçınar, 2012; Zayif, 2008) elde edilen sonuçlara paralel olarak eleştirel düşünme eğiliminin sınıf değişkeni açısından farklılaştığını göstermektedir. Bununla birlikte alt ölçeklerden alınan puanlar incelendiğinde ‘Doğruyu Arama ve ‘Sistematiklik’ boyutları dışındaki alt boyutlarda 1. sınıfta öğrenim gören katılımcıların ortalama puanlarının daha düşük olduğu görülmektedir. Katılımcıların toplam ölçekten aldıkları ortalama puanlara göre 1. sınıfta 194 öğrenim gören katılımcıların ortalama puanları, 2. 3. ve 4. sınıfta öğrenim gören katılımcıların ortalama puanlarından daha düşüktür. Tartışma ve Sonuç Ayrıca Çıkrıkçı (1996)’ya göre eleştirel düşünme, olgunlaşma ve yaşantı zenginliğiyle artmakta ve zihinsel yeteneğin önemli bir boyutunu oluşturmaktadır (Ay & Akgöl, 2008). Bu açıdan bakıldığında sınıf düzeyi büyüdükçe olgunlaşmanın ve yaşantı zenginliğinin arttığı ve bu durumun da eleştirel düşünme eğilimini olumlu etkilediği söylenebilir. Beden eğitimi öğretmeni adaylarının empatik eğilim düzeylerinin sınıf değişkeni açısından aldıkları puanların ortalamaları arasındaki fark göstermesi Ekinci (2009) ve Genç & Kalafat (2008) tarafından yapılan çalışmaların sonucuyla benzerlik göstermektedir. Buna göre 1.sınıfta öğrenim gören katılımcıların puanları, 2. 3. ve 4. sınıfta öğrenim gören katılımcıların puanlarından daha düşüktür. Ercoşkun (2005) yapmış olduğu çalışmada empatik eğilim düzeyinin sınıf değişkenine göre anlamlı bir şekilde farklılaşmadığını ancak en yüksek ortalama değerin 4. sınıf ve en düşük ortalama değerin 1. sınıf öğrencilerine ait olduğu sonucunu bulmuştur. Bu durumun, empatik eğilimin gelişimine katkı sağladığı düşünülen eğitsel oyunlar ve drama gibi derslerin daha sonraki dönemlerde okutulmaya başlanmasından kaynaklandığı düşünülebilir. 195 195 Araştırmadan elde edilen bir diğer sonuç, anne eğitim durumuna göre beden eğitimi öğretmeni adaylarının KEDEÖ puanlarının anlamlı bir şekilde farklılaştığını göstermektedir. Alt ölçeklerden alınan puanlar incelendiğinde ‘Doğruyu Arama’ ve ‘Kendine Güven’ boyutları dışındaki alt boyutlarda anne eğitim durumunu ‘üniversite’ olarak belirten katılımcıların ortalama puanlarının daha düşük olduğu görülmektedir. Katılımcıların toplam ölçekten aldıkları ortalama puanlara göre, anne eğitim durumunu ‘üniversite’ olarak belirten katılımcıların puanları, anne eğitim durumunu ‘okur-yazar değil’, ‘ilkokul’, ‘ortaokul’ ve ‘lise’ olarak belirten katılımcıların puanlarından daha düşüktür. Gülveren (2007) tarafından yapılan çalışmada da benzer sonuçlara ulaşılmış ve anne eğitim düzeyi arttıkça bu becerilerin artmadığı belirtilmiştir. Literatürde anne eğitim durumu değişkenine göre farklılığın tespit edildiği ve anlamlı farklılığın anne eğitim durumu üniversite olan katılımcıların lehine olduğu çalışmalar da mevcuttur (Bal, 2011; Karademir, 2013; Karalı, 2012; Kürüm, 2002; Özelçi, 2012; Tümkaya & Aybek, 2008). Glaser (1985), kişilerin düşünme eylemlerine ilişkin tutumlarına da önem vermektedir. Buna göre göre bireyler gerekli bilgi ve beceriye sahip olsalar dahi tutumları eleştirel düşünmeyi desteklemezse, eleştirel düşünmeyi harekete geçirecek ivme eksik kalacaktır (Gürkaynak, Üstel & Gülgöz, 2008). Dolayısıyla anne eğitim durumu üniversite olan katılımcıların eleştirel düşünme düzeylerinin düşük çıkma nedeni olarak annelerinin hâlihazırda eleştirel düşünme becerisine sahip olmadıkları ya da düşük düzeyde sahip oldukları; sahip olsalar bile bu düşünme eğilimini harekete geçirecek tutuma sahip olmadıkları gösterilebilir. Bulgular dâhilinde anne eğitim durumu değişkenine göre EEÖ puanlarının anlamlı bir şekilde farklılaştığı tespit edilmiştir. Anne eğitim durumu üniversite olan katılımcıların ortalama puanları, anne eğitim düzeyi ilkokul ve ortaokul olan katılımcıların ortalama puanlarından daha düşüktür. Tartışma ve Sonuç Bu durum Ekinci (2009) tarafından yapılan çalışmanın sonuçlarıyla zıtlık göstermiştir. Araştırmacı anne eğitim düzeyi “okur-yazar olmayanların” puan ortalamasının, anne eğitim düzeyleri “ilkokul”, “lise” ve “üniversite” olanlardan düşük olduğu sonucuna ulaşmıştır. Diğer yandan Kalliopuska (1984) tarafından yapılan çalışmanın sonucunda farklı sosyal sınıflara mensup annelerin empati kurma becerileri arasında farklılık olmadığı tespit edilmiştir. Her ne kadar eğitim düzeyi yüksek bireylerin empati kurma becerilerinin de yüksek olacağı düşünülse de araştırma sonuçları eğitim durumunun empati becerisi üzerinde etkili olmadığını göstermektedir. 196 Araştırmanın bir diğer sonucu, baba eğitim durumu değişkenine göre eleştirel düşünme eğilimi düzeyinin farklılaştığıdır. Bu durum literatürdeki bazı çalışmalarla (Bal, 2011; Ekinci, 2009; Karademir, 2013; Kürüm, 2002; Özelçi, 2012) paralellik göstermektedir. Katılımcıların toplam ölçekten aldıkları ortalama puanlara göre, baba eğitim durumunu ‘üniversite’ olarak belirten katılımcıların puanları, baba eğitim durumunu ‘okur-yazar olmayan’ ve ‘ilkokul’ olarak belirten katılımcıların puanlarından çok daha düşüktür. Bu durum baba eğitim durumunu üniversite olarak belirten katılımcıların sayısının daha az olmasından kaynaklanıyor olabilir. Ayrıca Kürüm (2002), yapmış olduğu çalışmada bulduğu farklılığı yükseköğretim ve ortaöğretim lehine bulmuştur. Literatürde baba eğitim durumuna göre eleştirel düşünme eğilimi düzeyinin farklılaşmadığı çalışmalar da mevcuttur (Beşoluk & Önder, 2010; Erdoğan, 2012; Karalı, 2012; Tümkaya & Aybek, 2008). Anne eğitim durumuna bezer şekilde baba eğitim durumunu üniversite olarak belirten katılımcıların eleştirel düşünme eğilimlerinin düşük çıkma nedeni olarak katılımcıların babalarının hâlihazırda eleştirel düşünme becerisine sahip olmadıkları ya da düşük düzeyde sahip oldukları; sahip olsalar bile bu düşünme eğilimini harekete geçirecek tutuma sahip olmadıkları gösterilebilir. Araştırma kapsamında katılımcıların EEÖ’n den aldıkları puanların baba eğitim durumu değişkenine göre anlamlı bir şekilde farklılaştığı sonucuna ulaşılmıştır. Baba eğitim durumu üniversite olan katılımcıların ortalama puanları, baba eğitim düzeyi okur-yazar değil, ilkokul, ortaokul ve lise olan katılımcıların ortalama puanlarından çok daha düşüktür. Araştırma kapsamında katılımcıların EEÖ’n den aldıkları puanların baba eğitim durumu değişkenine göre anlamlı bir şekilde farklılaştığı sonucuna ulaşılmıştır. Baba eğitim durumu üniversite olan katılımcıların ortalama puanları, baba eğitim düzeyi okur-yazar değil, ilkokul, ortaokul ve lise olan katılımcıların ortalama puanlarından çok daha düşüktür. Literatürde empatik eğilim ile baba eğitim düzeyi arasında anlamlı farklılığın olmadığı çalışmalar da mevcuttur (Ercoşkun, 2005; Genç & Kalafat, 2008). Diğer yandan Kalliopuska (1984) orta sınıfa mensup babaların, diğer sosyal sınıflara mensup babalara oranla daha fazla empati kurma becerisine sahip olduğunu tespit etmiştir. Bunun nedeni olarak verilen eğitimin empati becerisini geliştirmeye yönelik olmadığı daha çok teorik bilgi ve ezbere dayalı bir eğitim sistemi içerisinde bireylerin yetiştirilmeye çalışıldığı gösterilebilir. Tartışma ve Sonuç Ayrıca empatik eğilimin eğitim durumundan bağımsız olduğu ve bu durumun kişilik yapısından kaynaklanabileceği gibi verilen eğitimin empati becerisini geliştirmeye yönelik olması gerektiği söylenebilir. Araştırmadan elde edilen diğer bir sonuç katılımcıların eleştirel düşünme eğilimi ve empatik eğilim ölçeklerinde aldıkları puanlar arasındaki korelasyon katsayılarının pozitif yönde, düşük ve orta düzeylerde anlamlı bir ilişki olduğudur. Araştırma sonucunda eleştirel 197 düşünme eğilimi arttıkça empatik eğilim düzeyinin de arttığı sonucuna ulaşılmıştır. Bu durum Paul (1993)’un insanları eleştirel düşünme açısından üç gruba ayırması durumuyla ilişkili olabilir. Bu bağlamda birinci grubu düşünmeyle hiç ilgilenmeyenler oluşturmaktadır. İkinci grupta, eleştirel düşünme becerilerine sahip ancak bunu bencil biçimde kullananlar bulunmaktadır. Bu gruptakiler iyi düşünebilirler, ancak başkalarına karşı dürüst değillerdir. Genelde kendi düşünme süreçlerini değerlendirmezler aksine karşı olduklarının düşünme süreçlerini değerlendirir ve onların hatalarını ararlar. Üçüncü grupta ise, güçlü eleştirel düşünenler bulunmaktadır. Bu gruptakiler eleştirel düşünme becerilerini etik sorumluluk çerçevesinde yerine getirirler. Bu tip insanlar iyi düşünürler ve başkalarına karşı dürüsttürler. Bu durumda eleştirel düşünme eğilimi yüksek ve üçüncü grupta bulunan bireylerin başkalarına karşı dürüst ve karşılarındaki kişilerin de düşünme süreçleriyle ilgilenmelerinden dolayı empati eğilimlerinin de artacağı sonucuna varılabilir. Beden eğitimi öğretmeni adaylarının eleştirel düşünme eğilimleri düşük seviyededir. Beden eğitimi öğretmeni adaylarının eleştirel düşünme eğilimleri kadın ve erkek katılımcılar için, herhangi bir şekilde değişiklik göstermemiştir. Birinci sınıfta öğrenim gören beden eğitimi öğretmeni adaylarının eleştirel düşünme eğilimleri, ikinci, üçüncü ve dördüncü sınıfta öğrenim gören adaylardan daha düşüktür. Anne eğitim durumu üniversite olan beden eğitimi öğretmeni adaylarının eleştirel düşünme eğilimleri, anne eğitim durumu okur- yazar değil, ilkokul, ortaokul ve lise olan adaylardan daha düşüktür. Baba eğitim durumu üniversite olan beden eğitimi öğretmeni adaylarının eleştirel düşünme eğilimleri, baba eğitim durumu ilkokul ve ortaokul olan adaylardan daha düşüktür. Beden eğitimi öğretmeni adaylarının empatik eğilimleri orta seviyededir. Kadın ve erkek beden eğitimi öğretmeni adaylarının empatik eğilimleri bu araştırma kapsamında cinsiyetlerine göre birbirlerinden farklı değildir. Birinci sınıfta öğrenim gören beden eğitimi öğretmeni adaylarının empatik eğilimleri, ikinci, üçüncü ve dördüncü sınıfta öğrenim gören adaylardan daha düşüktür. Anne eğitim durumu üniversite olan beden eğitimi öğretmeni adaylarının empatik eğilimleri, anne eğitim düzeyi ilkokul ve ortaokul olan adaylardan daha düşüktür. Baba eğitim durumu üniversite olan beden eğitimi öğretmeni adaylarının empatik eğilimleri, baba eğitim düzeyi ilkokul, ortaokul ve lise olan adaylardan daha düşüktür. Beden eğitimi öğretmeni adaylarının eleştirel düşünme eğilimleri arttıkça empatik eğilimleri artmaktadır. Öneriler 198 Çalışma sonucunda öğretmen adaylarının eleştirel düşünme düzeylerinin düşük, empatik eğilimlerinin ise orta düzeyde olduğu tespit edilmiştir. Tartışma ve Sonuç Bu bağlamda beden eğitimi öğretmeni adaylarının eleştirel düşünme ve empatik eğilimlerini yükseltecek çalışmalar planlanabilir. Böylelikle bu çalışmalar öğretmen adaylarının karşılaşabilecekleri problemlerle başa çıkma ve ileride öğretmenlik yapacağı öğrencileri ile iletişim problemi yaşamamaları açısından onlara katkı sağlayabilir. Eleştirel düşünebilen ve empatik eğilim gösterebilen bireyler yetiştirmek için eğitim ve öğretim programlarının yeniden planlanması ve olanakların arttırılmasıyla bu becerilerin artacağı söylenebilir. Üniversiteler beden eğitimi öğretmeni adaylarının eleştirel düşünme ve empatik eğilim düzeylerini arttıracak yaratıcı drama, rol oynama gibi etkinlikler düzenleyebilir ve bu sayede öğretmen adaylarının mesleğe başladıklarında bu becerilere sahip olarak kendi öğrencilerine uygulamaları sağlanabilir. Üniversitelerde görev yapan öğretim elemanlarının empatik ve eleştirel düşünme eğilim düzeyleri yetiştirecekleri öğrencilerin de empatik ve eleştirel düşünme eğilim düzeylerini etkileyebileceğinden, öğretim elemanlarının çalışma grubunu oluşturduğu araştırmalar planlanabilir. Araştırmadan elde edilen sonuçlarla literatürdeki sonuçlar bazı değişkenler açısından farklılık gösterdiğinden bu konular üzerine güncel çalışmalar tekrarlanarak yapılabilir. Eleştirel düşünme ve empati eğilimi ile ilgili alana özel ölçme araçları geliştirilebilir. Kaynaklar Açak, M. (2006). Beden Eğitimi Öğretmeninin El Kitabı. İstanbul: Morpa Yayınları. Akar, Ü. (2007). Öğretmen Adaylarının Bilimsel Süreç Becerileri ve Eleştirel Düşünme Beceri Düzeyleri Arasındaki İlişki. (Yayınlanmamış yüksek lisans tezi). Afyon Kocatepe Üniversitesi/Sosyal Bilimler Enstitüsü, Afyonkarahisar. Akbıyık, C. (2002). Eleştirel Düşünme Eğilimleri ve Akademik Başarı. (Yayınlanmamış yüksek lisans tezi). Hacettepe Üniversitesi/Sosyal Bilimler Enstitüsü, Ankara. Akbulut, E. (2010). Sınıf Öğretmenlerinin Empatik Eğilim Düzeyleri. (Yayınlanmamış yüksek lisans tezi). Sakarya Üniversitesi/Sosyal Bilimler Enstitüsü, Sakarya. Akdere, N. (2012). Türkiye’de Öğretmen Adaylarının Eleştirel Düşünme Becerileri, Eleştirel Düşünme Öğretimine Yönelik Tutumları ve Öz Yeterlik Seviyeleri. (Yayınlanmamış doktora tezi). Orta Doğu Teknik Üniversitesi/Sosyal Bilimler Enstitüsü, Ankara. Aktaş, G. S. & Ünlü, M. (2013). Critical thinking skills of teacher candidates of elementary mathematics. Procedia-Social and Behavioral Sciences, 93, 831-835. APA. (1990). The Delphi report. Retrieved from https://assessment.trinity.duke.edu/documents/Delphi_Report.pdf p y p p p Arslanoğlu, C. (2012). Farklı Bölümlerde Öğrenim Gören Beden Eğitimi ve Spor Yüksekokulu Öğrencilerinin Empatik Eğilim Düzeyleri ve Saldırganlık İlişkisinin İncelenmesi. (Yayınlanmamış doktora tezi). Gazi Üniversitesi/Eğitim Bilimleri Enstitüsü, Ankara. 199 Arslanoğlu, İ. (2013). Sosyal bilimlerde metod ve araştırma yöntemleri. w3.gazi. edu.tr/~iarslan/arastek.doc adresinden 24 Şubat 2014 tarihinde alınmıştır. Ata, S. (2010). Okul Öncesi Öğretmenlerinin Bağlanma Stilleri, Duygusal Zekâ Düzeyleri ve Empatik Eğilimleri Arasındaki İlişkinin İncelenmesi. (Yayınlanmamış yüksek lisans tezi). Muğla Üniversitesi/Sosyal Bilimler Enstitüsü, Muğla. Ay, Ş. & Akgöl, H. (2008). Eleştirel düşünme gücü ile cinsiyet, yaş ve sınıf düzeyi. Kuramsal Eğitimbilim, 1 (2), 65-75. ğ Aybek, B. (2007). Eleştirel düşünmenin öğretiminde öğretmenin rolü. Bilim, Eğitim ve Düşünce Dergisi, 7 (2). Bal, M. (2011). Okul Öncesi Öğretmenlerinin Problem Çözme Becerileri ile Eleştirel Düşünme Eğilimleri Arasındaki İlişkinin İncelenmesi. (Yayınlanmamış yüksek lisans tezi). Abant İzzet Baysal Üniversitesi/Sosyal Bilimler Enstitüsü, Bolu. Balcı, E. (2012). Sınıf Öğretmenlerinin Empatik Eğilim Düzeyleri (Beşiktaş İlçesi Örneği). (Yayınlanmamış yüksek lisans tezi). Yeditepe Üniversitesi/Sosyal Bilimler Enstitüsü, İstanbul. Beşoluk, Ş. & Önder, İ. (2010). Öğretmen adaylarının öğrenme yaklaşımları, öğrenme stilleri ve eleştirel düşünme eğilimlerinin incelenmesi. İlköğretim Online, 9 (2), 679-693. Büküm, S. (2013). Din Görevlileri ile Öğretmenlerin Empatik Eğilim ve Beceri Düzeylerinin Karşılaştırılması (Fethiye Örneği). (Yayınlanmamış yüksek lisans tezi). Süleyman Demirel Üniversitesi/Sosyal Bilimler Enstitüsü, Isparta. y p üyüköztürk, Ş. (2008). Sosyal Bilimler İçin Veri Analizi El Kitabı. Ankara: Pegem Akademi Cengiz, G. E. (2004). Üniversite Öğrencilerine Yönelik Eleştirel Düşünme Etkinliklerinin Eleştirel Düşünme Eğilimi ve Becerileri Açısından Değerlendirilmesi. (Yayınlanmamış yüksek lisans tezi). Ankara Üniversitesi/Sosyal Bilimler Enstitüsü, Ankara. Certel, Z., Çatıkkaş, F. & Yalçınkaya, M. (2011). Beden eğitimi öğretmen adaylarının duygusal zekâ ile eleştirel düşünme eğilimlerinin incelenmesi. Kaynaklar Selçuk Üniversitesi Beden Eğitimi ve Spor Bilim Dergisi, 13(1), 74-81. Çekiç, S. (2007). Matematik Öğretmenliği Lisans Öğrencilerinin Eleştirel Düşünme Gücü Düzeylerinin Bazı Değişkenlere Göre Düzenlenmesi. (Yayınlanmamış yüksek lisans tezi). Dokuz Eylül Üniversitesi/Eğitim Bilimleri Enstitüsü, İzmir. Çelik, E. (2008). Okul Öncesi Eğitim Öğretmenlerinin Empatik Eğilimlerinin Bazı Değişkenler Açısından İncelenmesi. (Yayınlanmamış yüksek lisans tezi). Selçuk Üniversitesi/Sosyal Bilimler Enstitüsü, Konya. y Çetinkaya, Z. (2011). Türkçe öğretmen adaylarının eleştirel düşünmeye ilişkin görüşlerinin belirlenmesi. Ahi Evran Üniversitesi Eğitim Fakültesi Dergisi, 12 (3) , 93-108. Deniz, E. (2009). Öğretmen Adaylarının Eleştirel Düşünme Beceri Düzeyleri Üzerine Bir İnceleme. (Yayınlanmamış yüksek lisans tezi). Hacettepe üniversitesi/Sosyal Bilimler Enstitüsü, Ankara. Dev, N. (2010). İlköğretim Okullarında Görev Yapan Yönetici ve Öğretmenlerin Empatik Beceriler Açısından Karşılaştırılması (Kartal İlçesi Örneği). (Yayınlanmamış yüksek lisans tezi). Yeditepe Üniversitesi/Sosyal Bilimler Enstitüsü, İstanbul. Doğanay, A. & Ünal, F. (2006). İçerik Türlerine Dayalı Öğretim. Ankara: Nobel Yayınevi Dökmen, Ü. (1987). Yüz ifadeleri konusunda verilen eğitimin duygusal yüz ifadelerini teşhis becerisi ve iletişim çatışmalarına girme eğilimi üzerindeki etkisi. Psikoloji Dergisi, 6 (21), 75-79. Dökmen, Ü. (1988). Empatinin yeni bir modele dayanılarak ölçülmesi ve psiko-drama ile geliştirilmesi. Ankara Üniversitesi Eğitim Bilimleri Fakültesi Dergisi, 21 (1), 155-190. Dökmen, Ü. (1988). Empatinin yeni bir modele dayanılarak ölçülmesi ve psiko-drama ile geliştirilmesi. Ankara Üniversitesi Eğitim Bilimleri Fakültesi Dergisi, 21 (1), 155-190. 200 200 Dökmen, Ü. (2003). İletişim Çatışmaları ve Empati. İstanbul: Sistem Yayıncılık. Ekinci, Ö. (2009). Öğretmen Adaylarının Empatik ve Eleştirel Düşünme Eğilimlerinin İncelenmesi. (Yayınlanmamış yüksek lisans tezi). Çukurova Üniversitesi/Sosyal Bilimler Enstitüsü, Adana. Ercoşkun, M. H. (2005). Sınıf Öğretmenliği Öğrencilerinin Empatik Eğilimlerinin Çeşitli Değişkenler Açısından İncelenmesi. (Yayınlanmamış yüksek lisans tezi). Atatürk Üniversitesi/Sosyal Bilimler Enstitüsü, Erzurum. Erçoban, S. (2003). İlköğretim II. Kademesindeki Branş Öğretmenlerinin Empatik Eğilim Düzeylerinin Çeşitli Değişkenler Açısından İncelenmesi. (Yayınlanmamış yüksek lisans tezi). Uludağ Üniversitesi/Sosyal Bilimler Enstitüsü, Bursa. Erdoğan, İ. (2012). Din Kültürü ve Ahlak Bilgisi Öğretmen Adaylarının Eleştirel Düşünme Eğilim Düzeyleri Üzerine Bir İnceleme. (Yayınlanmamış yüksek lisans tezi). Necmettin Erbakan Üniversitesi/Eğitim Bilimleri Enstitüsü, Konya. Erkmen, G. (2007). Selçuk Üniversitesi Beden Eğitimi ve Spor Yüksekokulu’nda Öğrenim Gören Öğrencilerin Empatik Eğilimlerinin Sporda Tercih Ettikleri Lider Davranışları ile Karşılaştırılması. (Yayınlanmamış yüksek lisans tezi). Selçuk Üniversitesi/Sağlık Bilimleri Enstitüsü, Konya. Genç, S. Z. & Kalafat, T. (2008). Öğretmen adaylarının demokratik tutumları ile empatik becerilerinin değerlendirilmesi üzerine bir araştırma. Sosyal Bilimler Dergisi, 19, 211- 222. Göbel, D. Ş. (2013). Sınıf Öğretmenlerinin Eleştirel Düşünme Becerisi Öğretimi Yeterlilikleri ve Uygulamaları. (Yayınlanmamış yüksek lisans tezi). Sakarya Üniversitesi/Eğitim Bilimleri Enstitüsü, Sakarya. Gülveren, H. (2007). Kaynaklar Eğitim Fakültesi Öğrencilerinin Eleştirel Düşünme Becerileri ve Bu Becerileri Etkileyen Eleştirel Düşünme Faktörleri. (Yayınlanmamış doktora tezi). Dokuz Eylül Üniversitesi/Eğitim Bilimleri Enstitüsü, İzmir. ğ Gürkaynak, İ., Üstel, F. & Gülgöz, S. (2008). Eleştirel Düşünme. ERG Raporları. İstanbul: Artpres. Hager, P. & Kaye, M. (1991). Critical thinking ability and teacher effectiveness. Higher Education Research and Development, 10(2), 177-186. Howarth, K. (1996). Qualitative Study of the Teaching of Thinking Skills in Physical Education. (Unpublished doctoral thesis). Temple University, USA. Jones, E. & Ratcliff, G. (1993). Critical thinking skills for college students. Retrieved from http://findarticles.com/p/articles/mi_hb3387/is_n5_v67/ai_n28677196 adresinden 17 Nisan 2014 tarihinde alınmıştır. Karademir, A. Ç. (2013). Öğretmen Adaylarının Sorgulama ve Eleştirel Düşünme Becerilerinin Öğretmen Öz Yeterlik Düzeyine Etkisi. (Yayınlanmamış doktora tezi). Adnan Menderes Üniversitesi/Sosyal Bilimler Enstitüsü, Aydın. Karakaya, İ. (2009). Bilimsel Araştırma Yöntemleri. Abdurrahman Tanrıöğen (Ed.), Bilimsel Araştırma Yöntemleri içinde (s. 55-84). Ankara: Anı Yayıncılık. Karakoyun, Y. (2011). İlköğretim Öğretmenlerinin, Kulüp Faaliyetlerine İlişkin Tutumları ile Empatik Eğilimleri Arasındaki İlişki (Bağcılar İlçesi Örneği). (Yayınlanmamış yüksek lisans tezi). Yeditepe üniversitesi/Sosyal Bilimler Enstitüsü, İstanbul. 201 Karalı, Y. (2012). Eğitim Fakültesi Öğrencilerinin Eleştirel Düşünme Eğilimleri (İnönü Üniversitesi Örneği). (Yayınlanmamış yüksek lisans tezi). İnönü Üniversitesi/Eğitim Bilimleri Enstitüsü, Malatya. Karasar, N. (2005). Bilimsel Araştırma Yöntemleri (14. Baskı). Ankara: Nobel Yayın ve Dağıtım. Kızıltaş, Y. (2011). Sınıf Öğretmeni Adaylarının ve Sınıf Öğretmenlerinin Eleştirel Düşünce Eğilimlerinin Çeşitli Değişkenler Açısından İncelenmesi (Van İli Örneği). (Yayınlanmamış yüksek lisans tezi). Yüzüncü Yıl Üniversitesi/Sosyal Bilimler Enstitüsü, Van. opuska, M. (1984). Empathy in adults and social class. Psychological Reports, 55, 131-134 Kong, S. L. (2001). Critical thinking dispositions of pre-service teachers in singapore: a preliminary ınvestigation, http://www.aare.edu.au/ adresinden 17 Nisan 2014 tarihinde alınmıştır. Korkmaz, Ö. (2009). Eğitim fakültelerinin öğrencilerin eleştirel düşünme eğilim ve düzeylerine etkisi. Türk Eğitim Bilimleri Dergisi, 7 (4), 879-902. y ğ g Kökdemir, D. (2003). Belirsizlik Durumlarında Karar Verme ve Problem Çözme. Ankara: Kültür ve Sosyal Dayanışma Vakfı Yayınları. Küçük, P. D. & Uzun, Y. B. (2013). Müzik öğretmeni adaylarının eleştirel düşünme eğilimleri. Kırşehir Eğitim Fakültesi Dergisi, 14 (1). 327-345. Kürüm, D. (2002). Öğretmen Adaylarının Eleştirel Düşünme Gücü. (Yayınlanmamış yüksek lisans tezi). Anadolu Üniversitesi/Eğitim Bilimleri Enstitüsü, Eskişehir. Leaver-Dunn, D., Harrelson, G. L., Martin, M. & Wyatt, T. (2002). Critical-thinking predisposition among undergraduate athletic training students. Journal of Athletic Training, 37 (4), 147-151. g ( ) Loken, M. L. (2005). Critical Thinking Abilities of Undergraduate Entry-Level Athletic Training Students. (Unpublished doctoral thesis). The University of South Dakota, USA. Manger, T., Eikeland, O. J. & Asbjornsen A. Kaynaklar (2001). Effects of social-cognitive training on on students empathy. Swiss Journal of Psychology, 60 (2), 82-88. McBride, R. E. & Cleland, F. (1998). Critical thinking in physical education. Journal of Physical Education, Recreation and Dance, 69 (7), 42-52. McBride, R. & Knight, S. (1993). Identifying teacher behaviors during critical thinking tasks. Clearing House, 66(6), 374-378. McDonough, M. (1997). An Assessment of Critical Thinking at The Community College Level. (Unpublished doctoral thesis). Teachers College, Columbia University. Milli Eğitim Bakanlığı. (2007). Çağdaş öğretmen profili. http://oyegm.meb.gov.tr/ yet/yayinler/ulusal/Cagdas/Ogretmen-Profili.htm adresinden 26 Aralık 2007 tarihinde alınmıştır. Myyry, L. & Helkama, K. (2001). University students value priorities and emotional empathy. Educational Psychology, 21(1), 25-40. Nachmias, C. F. & Nachmias D. (1996). Research Methods in The Social Sciences. (5th Edition). New York: St. Martin's Press. Nosich, M. G. (2012). Eleştirel Düşünme ve Disiplinler Arası Eleştirel Düşünme Rehberi. (Çeviren: B. Aybek) Ankara: Anı Yayıncılık. 202 202 Özelçi, Y. S. (2012). Eleştirel Düşünme Tutumuna Etki Eden Faktörler: Sınıf Öğretmeni Adayları Üzerine Bir Çalışma. (Yayınlanmamış doktora tezi). Adnan Menderes Üniversitesi/Sosyal Bilimler Enstitüsü, Aydın. y y Paul, R. (1993). Critical thinking. Retrieved from http://outopia.org/teach/resources/CritThink1.pdf Paul, R. W & Elder, L. (2002). Critical Thinking: Tools for Taking Charge of Your Professional And Personal Life. Prentice Hall, Upper Saddle River, NJ. Rohner R. P. (2003). Perceived parental acceptance and emotional empathyamong university students in Korea. Journal of Cross-Cultural Psychology, 34(6), 723-735. Rosenfeld, L. & Wilder, L. (1990). Communication Effectively. In Aplied Sport Psychology: Personal Growth to Peak Performance. Mountain View, CA: Mayfield. Saçlı, F. (2008). Beden Eğitimi ve Spor Öğretmenliği Antrenörlük ve Rekreasyon Programlarında Öğrenim Gören Öğrencilerin Eleştirel Düşünme Düzeylerinin Saptanması ve Karşılaştırılması. (Yayınlanmamış yüksek lisans tezi). Hacettepe Üniversitesi/Sağlık Bilimleri Enstitüsü, Ankara. Saçlı, F. (2013). Yaratıcı drama eğitiminin aday beden eğitimi öğretmenlerinin eleştirel düşünme becerileri ve eğilimleri üzerine etkisi. (Yayınlanmamış doktora tezi). Hacettepe Üniversitesi/Sağlık Bilimleri Enstitüsü, Ankara. Sargın, N. (1993) Rehber öğretmenlerin empati düzeylerinin geliştirilmesine ilişkin bir model önerisi. (Yayınlanmamış doktora tezi). Dokuz Eylül Üniversitesi/Sosyal Bilimler Enstitüsü, İzmir. Saygın, O. (2005). İnsan İlişkilerinde 44’lük İletişim. İstanbul: Bilge Matbaacılık. Seferoğlu, S. S. & Akbıyık, C. (2006). Eleştirel düşünme ve öğretimi. Hacettepe Üniversitesi Eğitim Fakültesi Dergisi, 30, 193-200. Shin, S., Ha, J., Shin, K. & Davis, M. K., (2006). Critical thinking ability of associate, baccalaureate and RN-BSN senior students in Korea. Nursing Outlook, 34 (5), 328- 333. Swiger, W. S. (2005). Kaynaklar Correlation between Critical Thinking Skills and Decision Making Skills in Athletic Training and The Examination of Critical Thinking Differences Across The Curriculum. (Unpublished doctoral thesis). School of Physical Education, West Virginia University, Morgantown, WV. Tanrıdağ, Ş. R. (1992). Ankara’daki Ruh Sağlığı Hizmetlerinde Çalışan Personelin Empatik Eğilim ve Empatik Beceri Düzeylerinin Çeşitli Değişkenler Açısından İncelenmesi. (Yayımlanmamış doktora tezi). Hacettepe Üniversitesi/Sosyal Bilimler Enstitüsü, Ankara. Tokyürek, T. (2001). Öğretmen Tutumlarının Öğrencilerin Eleştirel Düşünme Becerilerine Etkisi. (Yayınlanmamış yüksek lisans tezi). Sakarya Üniversitesi/Sosyal Bilimler Enstitüsü, Sakarya. Torun, N. (2011). Fen ve Teknoloji Öğretmenlerinin Eleştirel Düşünme Eğilimleri ile Duygusal Zekâ Düzeyleri Arasındaki İlişki. (Yayınlanmamış yüksek lisans tezi). Çukurova üniversitesi/Sosyal Bilimler Enstitüsü, Adana. Tufan, D. (2008). Öğretmen Adaylarının Eleştirel Düşünme Becerileri: Orta Doğu Teknik Üniversitesi Yabancı Dil Öğretmenliği Örneği. (Yayınlanmamış yüksek lisans tezi). Orta Doğu Teknik Üniversitesi/Sosyal Bilimler Enstitüsü, Ankara. Tümkaya, S. & Aybek, B. (2008). Üniversite öğrencilerinin eleştirel düşünme eğilimlerinin sosyo-demografik özellikler açısından incelenmesi. Çukurova Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 17 (2), 387-402. 203 Uluçınar, U. (2012). Öğretmen Adaylarının Eleştirel Düşünme Eğilimlerinin Demokratik Değerlerini Yordama Düzeyi. (Yayınlanmamış yüksek lisans tezi). Osmangazi Üniversitesi/Eğitim Bilimleri Enstitüsü, Eskişehir. Yavuz, Ç. (2006). Rehberlik ve Psikolojik Danışmanlık Öğrencilerinde Öznel İyi Olma, Psikiyatrik Belirtiler ve Bazı Kişilik Özellikleri: Karşılaştırmalı Bir Çalışma. (Yayınlanmamış yüksek lisans tezi). İstanbul Üniversitesi/Sosyal Bilimler Enstitüsü, İstanbul. Yeşilpınar, M. (2011). Sınıf Öğretmenlerinin ve Öğretmen Adaylarının Eleştirel Düşünmenin Öğretimine Yönelik Yeterliklerine İlişkin Görüşleri. (Yayınlanmamış yüksek lisans tezi). Çukurova Üniversitesi/Sosyal Bilimler Enstitüsü, Adana. Yılmaz, N. (2011). Okul Öncesi Öğretmenlerinin İletişim Becerileri, Problem Çözme Becerileri ve Empatik Eğilim Düzeyleri. (Yayımlanmamış yüksek lisans tezi). Muğla Üniversitesi/Eğitim Bilimleri Enstitüsü, Muğla. ğ ğ Yılmaz, İ. & Akyel Y. (2008). Beden eğitimi öğretmen adaylarının empatik eğilim düzeylerinin çeşitli değişkenler açısından incelenmesi. Ahi Evran Üniversitesi Kırşehir Eğitim Fakültesi Dergisi (KEFAD), 9 (3), 27–33. Yiğiter, K. (2008). Kocaeli İlinde Görevli İlk ve Orta Öğrenim Kurumlarında Çalışan Beden Eğitimi ve Spor Öğretmenlerinin Empati Becerilerinin İncelenmesi. (Yayınlanmamış yüksek lisans tezi). Sakarya Üniversitesi/Sosyal Bilimler Enstitsü, Sakarya. Zayif, K. (2008). Öğretmen Adaylarının Eleştirel Düşünme Eğilimleri. (Yayınlanmamış yüksek lisans tezi). Abant İzzet Baysal Üniversitesi/Sosyal Bilimler Enstitüsü, Bolu. 204 204
https://openalex.org/W2626929802	http://real.mtak.hu/55224/1/004.2017.029.pdf	English	null	Heart rate, heart rate variability, faecal glucocorticoid metabolites and avoidance response of dairy cows before and after changeover to an automatic milking system	Acta veterinaria Hungarica	2,017	cc-by	6,643	Acta Veterinaria Hungarica 65 (2), pp. 301–313 (2017) DOI: 10.1556/004.2017.029 Acta Veterinaria Hungarica 65 (2), pp. 301–313 (2017) DOI: 10.1556/004.2017.029 HEART RATE, HEART RATE VARIABILITY, FAECAL GLUCOCORTICOID METABOLITES AND AVOIDANCE RESPONSE OF DAIRY COWS BEFORE AND AFTER CHANGEOVER TO AN AUTOMATIC MILKING SYSTEM Viktor JURKOVICH1, Fruzsina Luca KÉZÉR2,3, Ferenc RUFF4, Mikolt BAKONY1, Margit KULCSÁR5 and Levente KOVÁCS2,3 1Department of Animal Hygiene, Herd Health and Veterinary Ethology, University of Veterinary Medicine, István utca 2, H-1078 Budapest, Hungary; 2MTA–SZIE Large Animal Clinical Research Group, Üllő Dóra-major, Hungary; 3Institute of Animal Husbandry, Faculty of Agricultural and Environmental Science, Szent István University, Gödöllő, Hungary; 4Department of Methodology, Hungarian Central Statistical Office, Budapest, Hungary; 5Department and Clinic of Reproduction, University of Veterinary Medicine, Budapest, Hungary This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium for non-commercial purposes, provided the original author and source are credited. Corresponding author; E-mail: jurkovich.viktor@univet.hu; Phone: 0036 (1) 478-4242; Fax: 0036 (1) 478-4243 Corresponding author; E-mail: jurkovich.viktor@univet.hu; Phone: 0036 (1) 478-4242; Fax: 0036 (1) 478-4243 This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium for non-commercial purposes, provided (Received 22 December 2016; accepted 4 April 2017) The heart rate variability (HRV) parameters of dairy cows were monitored during parlour (PARL) and the later installed automatic (AMS) milking on a small-scale commercial dairy farm in Hungary. The aim of the study was to assess stress in relation to the type of milking and the frequency of human interaction. Parlour milking involved regular moving and crowding of the animals with fre- quent human interaction, which were much less frequent in automatic milking. The first phase of the study was conducted prior to the changeover [n = 27] and the second two months afterwards [n = 19 (of the cows from the first phase)]. Heart rate (HR) was recorded by the Polar RS800 CX recording system. HRV pa- rameters indicative of sympathovagal balance were calculated for periods of lying and standing in the barn, waiting before milking and milking, respectively. Morn- ing and evening faecal glucocorticoid concentrations were also measured. Fear of humans was tested by an avoidance distance test. Baseline HRV parameters showed no difference (P > 0.05) between the two systems. In the periods before, during and after milking a higher sympathetic tone was detected in cows in the PARL phase. Mean faecal glucocorticoid concentrations were higher at the time of parlour milking. The avoidance distance did not differ between the two phases. The results suggest that automatic milking might be less stressful for cows than Corresponding author; E-mail: jurkovich.viktor@univet.hu; Phone: 0036 (1) 478-4242; Fax: 0036 (1) 478-4243 0236-6290/$ 20.00 © 2017 The Authors 302 302 JURKOVICH et al. parlour milking, possibly due to the shorter duration of restraint after milking and the less frequent human interaction. parlour milking, possibly due to the shorter duration of restraint after milking and the less frequent human interaction. Key words: Animal welfare, dairy cow, heart rate variability, cortisol, automatic milking Automatic milking systems (milking robot, AMS) were first installed in the early 1990s in the Netherlands, and are now used in more than 8,000 dairies worldwide (Jacobs and Siegford, 2012a). Compared to parlour milking, automatic systems have the inevitable advantage of voluntary milking; however, it takes time for the animals to get adapted to a newly installed automatic milking system. A novel environment and isolation from mates during the time of milking can cause stress in cows (Rushen et al., 2001; Svennersten-Sjaunja and Pettersen, 2008; Jacobs and Siegford, 2012a). Acta Veterinaria Hungarica 65, 2017 Animals and husbandry The study was performed on a small-scale (80 lactating cows) commercial dairy farm in Hungary (GPS position: N 47.6926935, E 19.6080115) prior to and following a changeover in the milking system from a conventional herringbone milking parlour to a milking robot. In the first study period (March 2013), when the parlour milking system (PARL) was in operation, 27 clinically healthy Hol- stein-Friesian dairy cows [age = 4.8 ± 1.6 years, parity = 2.6 ± 0.5, days in milk (DIM) = 213.3 ± 36.4, body condition score (BCS) = 2.6 ± 0.5, milk yield = 23.1 ± 9.9 kg] without any clinical or lameness issues in the previous two months were examined. In the second part of the study (May 2013) the same animals were in- volved (BCS = 3.1 ± 0.5, milk yield = 18.3 ± 9.1) two months after the AMS had been installed. In the latter period, data from only 19 out of the 27 animals could be recorded. The data of the other eight animals deriving from the first study pe- riod were excluded from the analysis due to calving or being unable to adjust to robotic milking. Parlour milking. The animals were kept in a free-stall barn bedded with rubber mats in two separated production groups (lower yielding and high- yielding, 35–40 animals each). All the experimental cows were chosen from the low-producing group, being in the last trimester of lactation, to avoid any possi- ble disturbing factors near the early lactation that may affect the HRV parame- ters. Total mixed ration (TMR) was fed twice a day and water was available ad libitum. The cows were milked twice a day (05:00 and 17:00 o’clock) in a 2 × 5 stall herringbone milking parlour. A single person was responsible for milking the cows, including moving them from and back to the barn. It was observed that the milker often used aversive handling methods (shouting and sometimes hit- ting) when forcing the animals to move faster. The animals usually spent a rela- tively long time (on average 25 min with the extremes of 1 and 53 min) in the holding pen outside the milking parlour. Automatic milking. The conventional milking parlour was changed to a 1- stall milking robot (DeLaval VMS, DeLaval International AB, Tumba, Sweden). STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING 303 The aim of the present study was to compare the stress load of milking in a parlour system with frequent and sometimes aversive handling with that in an automatic milking system with significantly less human interaction. (Received 22 December 2016; accepted 4 April 2017) The variability of time intervals between two consecutive heartbeats (R–R interval) is a reliable indicator of the changes in the parasympathetic tone (von Borell et al., 2007) related to the stress response (Por- ges, 2003). In farm animals, the parasympathetic branch of the autonomic nerv- ous system plays a key role in regulating heart rate (HR) in response to stress (Hopster and Blokhuis, 1994). Reduced vagal tone was found in cows subjected to waiting after parlour milking with non-voluntary exit (Kovács et al., 2013) or during milking in a novel milking environment (Sutherland et al., 2012). Heart rate variability (HRV-) based studies investigating the stress load involved in milking have reported inconsistent results of either higher (Gygax et al., 2008), lower (Hopster et al., 2002) or similar (Hagen et al., 2005) sympathetic tone in cows milked in robotic units compared to conventional systems. The adrenocortical responses to stress can be assessed on the basis of the concentration of cortisol metabolites in biological samples (Möstl and Palme, 2002; Cook, 2012). Potentially painful veterinary procedures such as disbudding or rectal examination induce increased cortisol secretion (Nakao et al., 1994; Stafford and Mellor, 2011). Rough handling also results in higher plasma cortisol concentrations (Bauer et al., 2012). The plasma cortisol concentration is also in- fluenced by the time of day (Möstl and Palme, 2002), the procedure of blood sampling (Negrão et al., 2004) and non-aversive events (Manteca, 1998). A di- rect relationship between faecal glucocorticoid metabolites, blood cortisol, and adrenal activity has been demonstrated in dairy cattle (Morrow et al., 2002; Möstl and Palme, 2002). The quality of handling during milking is also important, as fear of people can be a major source of stress (Rushen and de Pasillé, 2015). Based on the study of Windschnurer et al. (2008), the avoidance distance is a good indicator (with respect to high inter-observer reliability and repeatability) of the human–animal relationship. STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING Acta Veterinaria Hungarica 65, 2017 Recording and analysis of HR and HRV Heart rate was recorded using a Polar RS800 CX recording system, in- cluding two adhesive electrodes and a Polar H3 transmitter (Polar Oy Kempele, Finland). During PARL, HR monitors were fitted on the animals after the morn- ing milking (between 06:30 and 08:00 o’clock) and recordings lasted until re- turning from the evening milking (between 20:00 and 21:00 o’clock). During the AMS period, the electrodes were attached in the morning (between 06:30 and 08:00 o’clock) and removed after a 24-h recording session. The electrodes were positioned and fixed as described in our earlier studies (Kovács et al., 2013, 2014). The heart rate was recorded continuously after a 2-h acclimatisation pe- riod. In parallel with the HR recordings, the starting and end points of undis- turbed bouts of lying and standing in the barn (without feeding or ruminating) were recorded using a stopwatch, while the duration of waiting before milking and of the main phase of milking was determined from video recordings (Canon Legria HF M36, CANON, Japan). HR and HRV parameters were calculated for the periods of (1) undisturbed lying and (2) standing in the barn, (3) the period of waiting in the holding pen/ waiting area before milking (PREMILK), and (4) the period from the start of ud- der preparation until exiting the milking parlour/milking unit (MILK). The R–R signal streams were analysed in 3-min time windows (Task Force of ESC- NASPE, 1996). Longer data streams were subdivided into several segments. Taking into account the circadian variation in HR and HRV (Kovács et al., 2016), the data of one occasion of robotic milking per cow – which was closest in time to the time of milking in the parlour system – was included in the analy- sis. Taking into account the effects of physical activity (Frondelius et al., 2015; Kovács et al., 2015b), cardiac parameters calculated for the PREMILK and MILK periods were compared to those calculated for periods of undisturbed standing. The Kubios HRV software (version 2.0, University of Kuopio, Finland) was used for HRV analysis (Tarvainen and Niskanen, 2008). Ectopic heartbeats and artefacts were corrected as described in an earlier study (Kovács et al., 2015a). For computing frequency-domain HRV, R–R intervals were subjected to Fast Fourier Transformation of power spectrum analysis (Akselrod et al., 1981). Animals and husbandry The groups and the area of the barn were united, but no other elements of the housing or the feeding regime were changed. Semi-forced cow traffic was in op- eration, which meant that cows entered the feeding area through a one-way gate, and from there – based on the time passed since the last milking – they could ei- ther reach the resting area or the waiting area of the milking unit through a selec- Acta Veterinaria Hungarica 65, 2017 304 JURKOVICH et al. tion gate. After milking, cows returned to the feeding area, and re-entered the resting area through the selection gate. The cows spent a short time (1 to 15 min- utes) in the waiting area of the AMS. Acta Veterinaria Hungarica 65, 2017 STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING 305 graphical representation of the correlation between successive R–R intervals, where each interval in the time series (RRi+1) is plotted against its successor (RRi), the Poincaré plot was used. We calculated the standard deviation 1 (SD1), which strongly reflects vagal tone, and the sympathetic measure SD2/SD1 (the ratio between standard deviation 2 and SD1) as described in earlier reviews on humans (Task Force of ESC-NASPE, 1996) and on farm animals (von Borell et al., 2007). Analysis of faecal samples Faecal samples were collected twice a day for the analysis of cortisol me- tabolites (Möstl and Palme, 2002). About 50 to 100 g of faeces was obtained manually from the rectum once in the morning (immediately after fixing the HR monitors, between 06:30 and 08:30 o’clock) and once in the evening (in PARL: before removing the heart rate monitors, between 20:00 and 21:00 o’clock; in AMS: in the same time period, when animals were standing in the barn). Sam- ples were stored on ice and frozen at −18 °C within 2 h after collection until analysis, as described by Möstl et al. (2002). For the extraction of the faecal glu- cocorticoid metabolites, samples were thawed at room temperature, stirred, and 0.5 g of faeces was dispersed in 5 ml of 80% methanol (Palme and Möstl, 1997). The dispersion was vortexed for 30 min and centrifuged at +4 °C at 3000 rpm for 20 min (Möstl et al., 2002). The supernatant was transferred into 1.5-ml Eppen- dorf tubes and stored at −18 °C until further analysis. A H3 homemade radioim- munoassay was carried out to determine the concentration of faecal glucocorti- coid metabolites following the description of Csernus (1982). All samples were analyzed in duplicate. Intra-assay and inter-assay coefficients of variation were calculated. Recording and analysis of HR and HRV Spectral parameters included the normalised power of the high-frequency band (HF) and the LF/HF ratio [the ratio between the low frequency (LF) component and HF], the latter of which reflects on the sympathovagal balance. The recom- mendations of von Borell et al. (2007) were considered by setting the limits of the spectral components as follows: LF: 0.05–0.20 Hz, HF: 0.20–0.58 Hz. For Avoidance distance Fear of humans was tested by an avoidance distance test according to the Welfare Quality protocol (Welfare Quality, 2009). At the feed bunk, the animals were approached at a speed of one step per sec and a step length of approxi- mately 60 cm with the arm held overhand in an angle of approximately 45° from the body, until the animals showed signs of withdrawal or until touching the muzzle. We defined withdrawal when the animal moved back, turned their head to the side, or pulled back their head trying to get out of the feeding rack. In case of withdrawal the avoidance distance was estimated (the distance between the hand and the muzzle at the moment of withdrawal) with a resolution of 10 cm (200 to 10 cm possible). If withdrawal took place at a distance less than 10 cm, the test result was still 10 cm. If the muzzle could be touched, an avoidance dis- tance of 0 cm was recorded. Animals were retested 30 min later and the average of the two measurements were used for statistical analysis. Acta Veterinaria Hungarica 65, 2017 306 JURKOVICH et al. Statistical analysis All statistical analyses were performed in the R-3.0.2 statistical environ- ment (R Core Team, 2015). For reducing the number of statistical comparisons between PARL and AMS, cardiac response parameters were calculated as area under the curve (AUC), which represents both the magnitude of the response and the changes occurring over time (Fekedulegn et al., 2007). Areas under the re- sponse curves were determined for each individual in the PREMILK and MILK periods utilising a trapezoid method described by Lay et al. (1996) as follows: p g p y y AUC = Σ [(Pn + Pn+1)/2 × m – BASELINE], p g p y y AUC = Σ [(Pn + Pn+1)/2 × m – BASELINE], where ‘P’ is the HRV parameter at a given time point, ‘m’ is the time in minutes between the two P values and ‘baseline’ is the mean value of recorded HR and HRV measures during the undisturbed standing position in the barn, be- fore milking. The data were then tested for constant variance (modified robust Brown- Forsythe Levene-type test) and the Shapiro–Wilk test was used for testing nor- mality of data for both PARL and AMS periods. For the comparison of cardiac autonomic responses between the PARL and AMS periods, the Friedman rank sum test were used. Comparisons between the PARL and AMS periods for faecal cortisol, avoidance distance as well as HR and HRV parameters calculated for ly- ing and standing were made by the paired t-test (for data of normal distribution) and the Wilcoxon non-parametric test (in case of non-normal distribution). The level of significance was set at 0.05 in all cases. We found no difference in the avoidance distance between the study peri- ods (PARL = 23.4 ± 5.7 cm; AMS = 37.1 ± 9.7 cm; P = 0.204). STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING 307 Table 1 Heart rate and heart rate variability parameters measured in the barn, during lying and standing1 Lying Standing Item2 Indicative of sympathetic/ vagal tone PARL3 AMS4 P value PARL AMS P value HR (min–1) 66.2 ± 1.0 66.2 ± 1.1 0.997 78.1 ± 3.5 71.0 ± 1.3 0.335 HF (n.u.) Vagal tone 46.1 ± 2.5 49.7 ± 2.5 0.387 24.8 ± 4.6 18.6 ± 2.5 1.000 LF/HF Symp./ parasymp. balance 1.9 ± 0.5 1.4 ± 0.2 0.952 5.6 ± 0.9 7.7 ± 1.5 0.593 SD1 (ms) Vagal tone 16.4 ± 1.8 14.5 ± 1.5 0.349 13.9 ± 6.1 9.2 ± 1.1 0.108 SD2/SD1 Symp. tone 2.9 ± 0.4 3.4 ± 1.2 0.324 6.2 ± 0.9 6.6 ± 0.6 0.660 1Descriptive statistics are based on means ± SEM of non-transformed data; 2HR = heart rate (beats/min); HF = the high-frequency component [where n.u. = normalised units); SD1 = standard deviation of instantaneous R–R variability measured from axis 1 in the Poincaré plot; SD2/SD1 = the ratio between SD2 (standard deviation of long-term continuous R–R variability measured from axis 2 in the Poincaré plot) and SD1; 3PARL = the period of conventional milking; 4AMS = the pe- riod of automatic milking system Table 1 Results Baseline HR and HRV parameters – calculated for lying and standing – showed no differences when comparing the two systems (Table 1). In the PREMILK period, the AUC of HR, SD1/SD2 and LF/HF were higher, while the AUC of HF was lower in PARL compared to AMS, indicating a higher sympathetic tone in parlour milking (Table 2). Comparing the MILK pe- riod, the HR did not differ between PARL and AMS; however, the AUC values of SD2/SD1 and LF/HF were higher, while HF was lower in case of parlour milking, again indicating a higher sympathetic tone. Concentrations of faecal corticosteroids were significantly higher during the parlour milking period, both in the morning and in the evening samples (Ta- ble 3). We found no difference in the avoidance distance between the study peri- ods (PARL = 23.4 ± 5.7 cm; AMS = 37.1 ± 9.7 cm; P = 0.204). Acta Veterinaria Hungarica 65, 2017 STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING Acta Veterinaria Hungarica 65, 2017 Table 3 Faecal corticosteroid concentrations during the study periods1 Corticosteroid concentrations (ng/g) Study period2 Morning Evening PARL 54.5 ± 3.1a 58.2 ± 4.3A AMS 20.6 ± 1.8b 19.0 ± 1.9B 1Descriptive statistics are based on means ± SEM of non- transformed data. Different superscripts in the same col- umn refer to significant differences between cortisol lev- els (a,bP < 0.05, A,BP < 0.01); 2PARL = the period of con- ventional milking, AMS = the period of automatic milk- ing system Faecal corticosteroid concentrations during the study periods1 Table 2 Areas under the curves (AUC) for heart rate and heart rate variability parameters calculated for premilking and milking periods1 Lying Standing Item2 Indicative of sympathetic/ vagal tone PARL3 AMS4 P value PARL AMS P value HR (min–1) 66.2 ± 1.0 66.2 ± 1.1 0.997 78.1 ± 3.5 71.0 ± 1.3 0.335 HF (n.u.) Vagal tone 46.1 ± 2.5 49.7 ± 2.5 0.387 24.8 ± 4.6 18.6 ± 2.5 1.000 LF/HF Symp./ parasymp. balance 1.9 ± 0.5 1.4 ± 0.2 0.952 5.6 ± 0.9 7.7 ± 1.5 0.593 SD1 (ms) Vagal tone 16.4 ± 1.8 14.5 ± 1.5 0.349 13.9 ± 6.1 9.2 ± 1.1 0.108 SD2/SD1 Symp. tone 2.9 ± 0.4 3.4 ± 1.2 0.324 6.2 ± 0.9 6.6 ± 0.6 0.660 1Descriptive statistics are based on means ± SEM of non-transformed data; 2HR = heart rate (beats/min); HF = the high-frequency component [where n.u. = normalised units); SD1 = standard deviation of instantaneous R–R variability measured from axis 1 in the Poincaré plot; SD2/SD1 = the ratio between SD2 (standard deviation of long-term continuous R–R variability measured from axis 2 in the Poincaré plot) and SD1; 3PARL = the period of conventional milking; 4AMS = the pe- riod of automatic milking system Areas under the curves (AUC) for heart rate and heart rate variability parameters calculated for premilking and milking periods1 Acta Veterinaria Hungarica 65, 2017 308 308 JURKOVICH et al. Acta Veterinaria Hungarica 65, 2017 STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING 309 Baseline HR and HRV parameters calculated for lying and standing did not differ between the two milking systems. Earlier studies found elevated (Gygax et al., 2008) or similar (Hagen et al., 2005) HR during lying posture in AMS com- pared to conventional milking systems; however, both studies reported on higher basal sympathetic tone in AMS. Hopster et al. (2002) found no difference be- tween HR and HRV measured during standing in the barn in either parlour or ro- botic milking systems. We suppose that stress related to the procedure of milking in the conventional system is acute, considering that between two milkings the cows are left undisturbed in the barn, without human presence, free to rest or per- form their daily activities, just as in the robotic milking system. y j g y In the period of PREMILK, there were significant differences between the two systems regarding all parameters. AUCs calculated for HR, frequency do- main and Poincaré measures of HRV indicated lower vagal and higher sympa- thetic activity in parlour milking. Hopster et al. (2002) also reported lower HR in the waiting area in AMS, compared to PARL. Stress was presumably a result of aversive handling methods (Pajor et al., 2000) or crowding in the holding area (Morgan and Tromborg, 2007; Kovács et al., 2013). During the MILK period, HRV parameters reflected lower vagal tone and higher sympathetic activity in cows in PARL compared to AMS. Waiting for ex- iting the milking stall after the milking has finished is stressful for cows (Kovács et al., 2013). In PARL, it took an inevitably longer time (4.8 ± 1.0 min), due to the individual differences in milk let-down speed, compared to AMS, where it was considerably shorter (0.3 ± 0.15 min). Earlier results investigating the changes of HR and HRV during milking in AMS or PARL are inconsistent. Gygax et al. (2008) reported an increased stress level (higher HR and lower vagal tone) in AMS, Hopster et al. (2002) found lower HR and adrenaline levels in AMS, whereas Hagen et al. (2005) found no difference in HR and HRV between AMS and PARL. Their results suggest that there are multiple factors that can influence the level of stress experienced during milking, which are needed to be identified and compared in further stress as- sessment studies. Discussion In this study, we have compared the cardiac parameters, faecal glucocorti- coid concentrations and avoidance distance of cows milked in a parlour and fre- quently experiencing aversive handling (shouting and sometimes hitting during moving them to the milking parlour), to that of the same cows two months after parlour milking had been changed to a milking robot and human interaction had become less frequent. The handling method during PARL was the way the milker handled the animals, it was not an experimental condition. Although Hopster et al. (2002) studied the effects of AMS 4 weeks after the changeover, we have found that an 8-week interval is more appropriate to exclude not only the effects of transition to the new milking system (Weiss et al., 2004, 2005; Jacobs and Siegford, 2012b) but also the possible long-term effects of rough handling. Con- sidering that cows can remember the site of the preferred feed for at least 6 weeks (Kovalčik and Kovalčik, 1986) and positive handling had prolonged ef- fects on the avoidance reaction for 8 weeks (Schmied et al., 2008), we supposed that unpleasant experiences might as well have a longer-term influence on behav- iour. Physical activity, space allowance, milking frequency, parity and milk yield have a considerable impact on HRV (Kovács et al., 2015b). None of these factors changed significantly between the first and the second study period. The experi- mental animals were in mid- and late lactation and the difference in milk yield between the study periods was not significant (P = 0.132). Metabolic disorders that might influence behaviour or the level of stress are rare in this period of the lactation (Ingvartsen et al., 2003; Brydl et al., 2015). Season has only moderate importance in the interpretation of HRV parameters (Kovács et al., 2015b). Pregnancy-related weight gain in the given period is not as intensive as in the dry period and the body condition scores of the experimental animals did not differ significantly (P = 0.08). Acta Veterinaria Hungarica 65, 2017 Acta Veterinaria Hungarica 65, 2017 STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING Considering that in the present study the basal HRV parameters did not differ between the two study periods, the changes in HRV measured dur- ing PREMILK and MILK must have arisen from differences in the milking sys- tem and in the way of handling. Faecal cortisol concentrations determined twice a day adequately represent the level of stress experienced in the preceding 24 h (Möstl and Palme, 2002). In the present study, faecal cortisol concentrations were higher in the period of PARL than in AMS. Others have found no difference in cortisol concentrations measured in plasma (Hopster et al., 2002; Abeni et al., 2005), milk (Gygax et al., 2006) or faeces (Hopster et al., 2002). Considering that basal HRV parameters did not differ between the two study periods, the increase in hypothalamic– pituitary–adrenal (HPA) activity in PARL can be linked to the effects of human Acta Veterinaria Hungarica 65, 2017 310 JURKOVICH et al. interaction (Bauer et al., 2012) or the differences in the type of milking (Gorewit et al., 1992). The avoidance distance did not differ in PARL and AMS in our study. The cows seem to remember and avoid those stockpersons who have treated them roughly (Munksgaard et al., 1997, 2001), and it was shown that the avoidance distance increases on farms where rough handling is frequent (Waiblinger et al., 2003; Rushen and de Pasillé., 2015). Cows can distinguish people and seem to generalise previous experience, especially if the people working around them wear clothing of the same colour (Munksgaard et al., 1997, 2001). In our study, the milker was not present at the time of the avoidance distance testing, and the persons performing the test were unfamiliar to cows and wore clothing that dif- fered from that of the farm workers. The avoidance distance was relatively short in both the PARL and AMS periods. Two explanations may be assumed: (1) The negative experiences during handling the animals in relation to milking had not been generalised to other situations. (2) The type of avoidance distance testing was not sensitive enough to measure fine differences between the same animals at two time points, especially if we hope for an improvement of already small values. A slightly different approach (MacKay et al., 2014) may be more sensi- tive in such situations. STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING Nevertheless, it is not common that farms employ only a single person for all the tasks involved in milking, including the herding of the animals. Conclusions In this study, parlour milking caused greater stress in dairy cows than ro- botic milking. The difference between conventional and automatic milking pri- marily arises from the duration of restraint after milking and the frequency and quality of animal handling. Increasing our knowledge on the nature of stressors involved in the process of milking – e.g. the quality of the milking routine, the attitude of milkers, the effects of isolation from herdmates, novelty in the envi- ronment, etc. – would help identify the most critical elements of parlour milking and improve animal welfare in conventional milking systems. References Abeni, F., Calamari, L., Calza, F., Speroni, M., Bertoni, G. and Pirlo, G. (2005): Welfare assess- ment based on metabolic and endocrine aspects in primiparous cows milked in a parlor or with an automatic milking system. J. Dairy Sci. 88, 3542–3552. h an automatic milking system. J. Dairy Sci. 88, 3542–3552 Akselrod, S., Gordon, D., Ubel, F. A., Shannon, D. C., Berger, A. C. and Cohen, R. J. (1981): Power spectrum analysis of heart rate fluctuation: a quantitative probe of beat-to-beat car- diovascular control. Science 213, 220–222. Bauer, J. M., Kegley, E. B., Richeson, J. T., Galloway, D. L., Hornsby, J. A. and Reynolds, J. L. (2012): Impact of different handling styles (good vs. adverse) on growth performance, be- havior, and cortisol concentrations in beef cattle. Discovery 13, 3–10. von Borell, E., Langbein, J., Després, G., Hansen, S., Leterrier, C., Marchant-Forde, J., Minero, M., Mohr, E., Prunier, A. and Veissier, I. (2007): Heart rate variability as a measure of autonomic regulation of cardiac activity for assessing stress and welfare in farm animals: a review. Physiol. Behav. 92, 293–316. Brydl, E., Jurkovich, V., Tirián, A., Kovács, P. and Könyves, L. (2015): Incidence rate of subclini- cal metabolic disorders in Hungarian dairy herds during the last two decades [in Hungar- ian, with English abstract]. Magy. Allatorvosok 137, Suppl 1, 280–290. Cook, N. J. (2012): Minimally invasive sampling media and the measurement of corticosteroids as biomarkers of stress in animals. Can. J. Anim. Sci. 92, 227–259. Csernus, V. (1982): Antibodies of high affinity and specificity for RIA determination of progester- one, testosterone, estradiol-17b and cortisol. In: Görög, S. (ed) Advances in Steroid Analy- sis. Akadémiai Kiadó, Budapest, Hungary. pp 171–177. Fekedulegn, D. B., Andrew, M. E., Burchfiel, C. M., Violanti, J. M., Hartley, T. A., Charles, L. E. and Miller, D. B. (2007): Area under the curve and other summary indicators of repeated waking cortisol measurements. Psychosom. Med. 69, 651–659. g y Frondelius, L., Järvenranta, K., Koponen, T. and Mononen, J. (2015): The effects of body posture and temperament on heart rate variability in dairy cows. Physiol. Behav. 139, 437–441. p y y y Gorewit, R. C., Svennersten, K., Butler, W. R. and Uvnäs-Moberg, K. (1992): Endocrine responses in cows milked by hand and machine. J. Dairy Sci. 75, 443–448. Gygax, L., Neuffer, I., Kaufmann, C., Hauser, R. and Wechsler, B. STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING 311 Acknowledgements The study and the publication was supported by the Hungarian Ministry of Hu- man Resources (grant no.: 9878-3/2015/FEKUT, 9878-3/2016/FEKUT and 11475-4/2016/ FEKUT). Acta Veterinaria Hungarica 65, 2017 Acta Veterinaria Hungarica 65, 2017 References (2006): Milk cortisol concentra- tion in automatic milking systems compared with auto-tandem milking parlors. J. Dairy Sci. 89, 3447–3454. Gygax, L., Neuffer, I., Kaufmann, C., Hauser, R. and Wechsler, B. (2008): Restlessness behaviour, heart rate and heart-rate variability of dairy cows milked in two types of automatic milking systems and auto-tandem milking parlours. Appl. Anim. Behav. Sci. 109, 167–179. y g p pp , Hagen, K., Langbein, J., Schmied, C., Lexer, D. and Waiblinger, S. (2005): Heart rate variability in dairy cows – influences of breed and milking system. Physiol. Behav. 85, 195–204. Hopster, H. and Blokhuis, H. J. (1994): Validation of a heart-rate monitor for measuring a stress- response in dairy cows. Can. J. Anim Sci. 74, 465–474. Hopster, H., Bruckmaier, R. M., van der Werf, J. T. N., Korte, S. M., Macuhova, J., Korte-Bouws, G. and van Renen, C. G. (2002): Stress responses during milking; comparing conventional and automatic milking in primiparous dairy cows. J. Dairy Sci. 85, 3206–3216. Ingvartsen, K. L., Dewhurst, R. J. and Friggens, N. C. (2003): On the relationship between lacta- tional performance and health: is it yield or metabolic imbalance that cause production dis- eases in dairy cattle? A position paper. Livest. Prod. Sci. 83, 277–308. Jacobs, J. A. and Siegford, J. M. (2012a): The impact of automatic milking systems on dairy cow management, behavior, health and welfare. J. Dairy Sci. 95, 2227–2247. Jacobs, J. A. and Siegford, J. M. (2012a): The impact of automatic milking systems on dairy cow management, behavior, health and welfare. J. Dairy Sci. 95, 2227–2247. Jacobs, J. A. and Siegford, J. M. (2012b): Lactating dairy cows adapt quickly to being milked by t ti ilki t J D i S i 95 1575 1584 Jacobs, J. A. and Siegford, J. M. (2012b): Lactating dairy cows adapt quickly to being milked by an automatic milking system. J. Dairy Sci. 95, 1575–1584. Acta Veterinaria Hungarica 65, 2017 312 JURKOVICH et al. Kovács, L., Bakony, M., Tőzsér, J. and Jurkovich, V. (2013): Changes in heart rate variability of dairy cows during conventional milking with nonvoluntary exit. J. Dairy Sci. 96, 7743–7747. Kovács, L., Kézér, F. L., Bakony, M., Hufnagel, L., Tőzsér, J. and Jurkovich, V. (2015b): Associa- tions between heart rate variability parameters and housing- and individual-related vari- ables in dairy cows using canonical correspondence analysis. PLoS ONE 10, e0145313. Kovács, L., Kézér, F. L., Jurkovich, V., Kulcsár-Huszenicza, M. References and Tőzsér, J. (2015a): Heart rate variability as an indicator of chronic stress caused by lameness in dairy cows. PLoS ONE 10, e0134792. Kovács, L., Kézér, F. L., Ruff, F. and Szenci, O. (2016): Cardiac autonomic activity has a circadian rhythm in summer but not in winter in non-lactating pregnant dairy cows. Physiol. Behav. 155, 56–65. Kovács, L., Tőzsér, J., Szenci, O., Póti, P., Kézér, F. L., Ruff, F., Gábriel-Tőzsér, Gy., Hoffmann, D., Bakony, M. and Jurkovich, V. (2014): Cardiac responses to palpation per rectum in lac- tating and nonlactating dairy cows. J. Dairy Sci. 97, 6955–6963. Kovalčik, K. and Kovalčik, M. (1986): Learning ability and memory testing in cattle of different ages. Appl. Anim. Behav. Sci. 15, 27–29. Lay, D. C., Friend, T. H., Randel, R. D., Jenkins, O. C., Neuendorff, D. A., Kapp, G. M. and Bushong, D. M. (1996): Adrenocorticotropic hormone dose response and some physiologi- cal effects of transportation on pregnant Brahman cattle. J. Anim. Sci. 74, 1806–1811. p p g Manteca, X. (1998): Neurophysiology and assessment of welfare. Meat Sci. 49, Suppl. 1, S205– S218. MacKay, J. R., Haskell, M. J., Deag, J. M. and van Reenen, K. (2014): Fear responses to novelty in testing environments are related to day-to-day activity in the home environment in dairy cattle. Appl. Anim. Behav. Sci. 152, 7–16. Morgan, K. N. and Tromborg, C. T. (2007): Sources of stress in captivity. Appl. Anim. Behav. Sci. 102, 262–302. Morrow, C. J., Klover, E. S., Verkerk, G. A. and Mathews, L. R. (2002): Fecal glucocorticoid me- tabolites as a measure of adrenal activity in dairy cattle. Gen. Comp. Endocrinol. 126, 229–241. Möstl E and Palme R (2002): Hormones as indicators of stress Domest Anim Endocrinol 23, Morrow, C. J., Klover, E. S., Verkerk, G. A. and Mathews, L. R. (2002): Fecal glucocorticoid me- tabolites as a measure of adrenal activity in dairy cattle. Gen. Comp. Endocrinol. 126, 229–241. Möstl, E. and Palme, R. (2002): Hormones as indicators of stress. Domest. Anim. Endocrinol. 23, 67–74. tabolites as a measure of adrenal activity in dairy cattle. Gen. Comp. Endocrinol. 126, 229–241. Möstl, E. and Palme, R. (2002): Hormones as indicators of stress. Domest. Anim. Endocrinol. 23, 67–74. Möstl, E., Maggs, J. L., Schrötter, G., Besenfelder, U. and Palme, R. (2002): Measurement of corti- sol metabolites in faeces of ruminants. Vet. Res. Comm. 26, 127–139. Munksgaard, L., de Pasille, A. STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING 313 Rushen, J. and de Passillé, A. M. (2015): The importance of good stockmanship and its benefits to animals. In: Grandin, T. (ed) Improving Animal Welfare: A Practical Approach. Second edition. CABI, Wallingford, UK. pp. 125–138. Rushen, J., Munksgaard, L., Marnet, P. G. and de Passillé, A. M. (2001): Human contact and the effects of acute stress on cows at milking. Appl. Anim. Behav. Sci. 73, 1–14. Schmied, C., Boivin, X. and Waiblinger, S. (2008): Stroking different body regions of dairy cows: Effects on avoidance and approach behavior toward humans. J. Dairy Sci. 91, 596–605. Stafford, K. J. and Mellor, D. J. (2011): Addressing the pain associated with disbudding and de- horning in cattle. Appl. Anim. Behav. Sci. 135, 226–231. Sutherland, M. A., Rogers, A. R. and Verkerk, G. A. (2012): The effect of temperament and re- sponsiveness towards humans on the behavior, physiology and milk production of multipa- rous dairy cows in a familiar and novel milking environment. Physiol. Behav. 107, 329–337. sponsiveness towards humans on the behavior, physiology and milk production of multipa- rous dairy cows in a familiar and novel milking environment. Physiol. Behav. 107, 329–337. Svennersten-Sjaunja, K. M. and Petterson, G. (2008): Pros and cons of automatic milking in Europe. J. Anim. Sci. 86, Suppl. 1, 37–46. Tarvainen, M. P. and Niskanen, J. P. (2008): Kubios HRV Version 2.0 User’s Guide. Department of Physics, University of Kuopio, Kuopio, Finland. Task Force of the European Society of Cardiology, North American Society of Pacing and Electro- physiology (1996): Heart rate variability: standards of measurement, physiological inter- pretation, and clinical use. Circulation 93, 1043–1065. p , Waiblinger, S., Menke, C. and Fölsch, D. W. (2003): Influences on the avoidance and approach behaviour of dairy cows towards humans on 35 farms. Appl. Anim. Behav. Sci. 84, 23–39. Weiss, D., Helmreich, S., Möstl, E., Dzidic, A. and Bruckmaier, R. M. (2004): Coping capacity of dairy cows during the change from conventional to automatic milking. J. Anim. Sci. 82, 563–570. Weiss, D., Möstl, E. and Bruckmaier, R. M. (2005): Physiological and behavioural effects of change over from conventional to automatic milking in dairy cows with and without previ- ous experience. Vet. Med. Czech 50, 253–261. Welfare Quality (2009): Welfare Quality Assessment Protocol for Cattle. Welfare Quality Consor- tium, Lelystad, The Netherlands. Windschnurer, I., Schmied, C., Boivin, X. and Waiblinger, S. Acta Veterinaria Hungarica 65, 2017 References M., Rushen, J., Thodberg, K. and Jensen, M. B. (1997): Discrimina- tion of people by dairy cows based on handling. J. Dairy Sci. 80, 1106–1112. Munksgaard, L., de Passillé, A. M., Rushen, J., Herskin, M. S. and Kristensen, A. M. (2001): Dairy cows’ fear of people: social learning, milk yield and behaviour at milking. Appl. Anim. Behav. Sci. 73, 15–26. Nakao, T., Sayo, T., Moriyoshi, M. and Kawata, K. (1994): Plasma cortisol response in dairy cows to vaginoscopy, genital palpation per rectum and artificial insemination. Zbl. Vet-med. A 41, 16–21. Negrão, J. A., Porcionato, M. A., de Passille, A. M. and Rushen, J. (2004): Cortisol in saliva and plasma of cattle after ACTH administration and milking. J. Dairy Sci. 87, 1713–1718. p g y Pajor, E. A., Rushen, J. and de Passillé, A. M. (2000): Aversion learning techniques to evaluate dairy cattle handling practices. Appl. Anim. Behav. Sci. 69, 89–102. Palme, R. and Möstl, E. (1997): Measurement of cortisol metabolites in faeces of sheep as a pa- rameter of cortisol concentration in blood. Int. J. Mammal. Biol. 62, 192–197. Porges, S. W. (2003) The polyvagal theory: phylogenetic contributions to social behavior. Physiol. Behav. 79, 503–513. R Core Team (2015): R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. URL: https://www.R-project.org/. Acta Veterinaria Hungarica 65, 2017 STRESS ASSESSMENT IN DAIRY COWS IN RELATION TO THE TYPE OF MILKING (2008): Reliability and inter-test rela- tionship of tests for on-farm assessment of dairy cows’ relationship to humans. Appl. Anim. Behav. Sci. 114, 37–53. Acta Veterinaria Hungarica 65, 2017
https://openalex.org/W2594242949	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0172770&type=printable	English	null	Evaluation of 90 day repeated dose oral toxicity and reproductive/developmental toxicity of 3'-hydroxypterostilbene in experimental animals	PloS one	2,017	cc-by	10,536	RESEARCH ARTICLE Editor: David L. McCormick, IIT Research Institute, UNITED STATES Received: October 21, 2016 Accepted: February 9, 2017 Published: March 3, 2017 Copyright: © 2017 Majeed et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. gross and histopathological changes. In conclusion, the results suggested a No-Observed- Adverse-Effect-Level (NOAEL) of 200 mg/kg bwt/day in rats after oral administration, imply- ing 3’-HPT did not exhibit any toxicity under the study conditions employed. Data Availability Statement: All relevant data are within the paper. Funding: This work was sponsored by Sami Labs Limited. There was no external funding involved. The funder provided support in the form of salaries for authors of both Sami Labs Limited and ClinWorld (P) Limited [MM, SB, AP, SN and NVS], but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of all authors affiliated to Sami Labs Evaluation of 90 day repeated dose oral toxicity and reproductive/developmental toxicity of 3’-hydroxypterostilbene in experimental animals Muhammed Majeed1, Sarang Bani2, Sankaran Natarajan3, Anjali Pandey2, Naveed S4* 1 Sami Labs Limited, Bangalore, Karnataka, India, 2 Biological Research Department, Sami Labs Limited, Bangalore, Karnataka, India, 3 Research and Development Center, Sami Labs Limited, Bangalore, Karnataka, India, 4 ClinWorld, ClinWorld Private Limited, Bangalore, Karnataka, India a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 * naveed@clinworld.org * naveed@clinworld.org * naveed@clinworld.org Abstract 3’-Hydroxypterostilbene (3’-HPT) is one of the active constituents of Sphaerophysa salsula and Pterocarpus marsupium. Despite many proposed therapeutic applications, the safety profile of 3’-HPT has not been established. The present work investigated 90 day repeated oral dose and reproductive (developmental) toxicity of 3’-HPT as a test substance in rats as per OECD guidelines. 90 day toxicity was conducted in sixty Sprague Dawley rats of each sex (120 rats), grouped into six dosage groups of 0 (control), 0 (control recovery), 20 (low dose), 80 (mid dose), 200 (high dose) and 200 (high dose recovery) mg/kg bwt/day (body weight/day) respectively. For the reproductive toxicity study forty Wistar rats of each sex (80 rats) divided into four dosage groups received 0 (vehicle control), 20 (low dose), 100 (mid dose) and 200 (high dose) mg/kg bwt/day of 3’-HPT respectively for a period of two weeks while pre-mating, mating, on the day before sacrifice, in females during pregnancy and four days of lactation period. Results showed no significant differences in body weight, food intake, absolute organ weight, haematology, with no adverse effects (toxicity) on biochemi- cal values nor any abnormal clinical signs or behavioural changes were observed in any of the control/treatment groups, including reproductive and developmental parameters, gross and histopathological changes. In conclusion, the results suggested a No-Observed- Adverse-Effect-Level (NOAEL) of 200 mg/kg bwt/day in rats after oral administration, imply- ing 3’-HPT did not exhibit any toxicity under the study conditions employed. OPEN ACCESS Citation: Majeed M, Bani S, Natarajan S, Pandey A, S N (2017) Evaluation of 90 day repeated dose oral toxicity and reproductive/developmental toxicity of 3’-hydroxypterostilbene in experimental animals. PLoS ONE 12(3): e0172770. doi:10.1371/journal. pone.0172770 Editor: David L. McCormick, IIT Research Institute, UNITED STATES Since pterostilbene is widely recognized for its anti-cancer and anti-diabetic properties, it is important to evaluate the toxi- cological role of its metabolites, as 3’-Hydroxypterostilbene has been found be one of the uri- nary metabolites of pterostilbene in mice [3]. Stilbene compounds like (trans-) resveratrol and piceatannol are widely distributed in nature. In view of their extensive range of biological activities such as cardioprotection and cancer chemopreventive properties, they have become of special interest to biologists and chemists. The trans-resveratrol-like natural stilbenoids have shown remarkable anti-cancer activity higher than resveratrol (Fig 1). 3’-HPT for exam- ple, is reported to be 50 to 97 times more potent than trans-resveratrol in eliciting significant apoptotic action on tumour cell lines and is markedly more active than pterostilbene in induc- ing apoptosis in sensitive and multi drug-resistant (MDR) leukemia cells [4]. Based on struc- ture activity relationships, the importance of 3,5-dimethoxy motif in conferring pro-apoptic activity on leukemia cells has been recognized [5]. Pterostilbene and 3’-HPT possess the neces- sary 3,5-dimethoxy motif. Among these, 3’-HPT is an interesting natural compound which may be useful in treating different types of haematological malignancies as it is reported to possess a ‘privileged’ structure [4] motif of 3,5-dimethoxy group in addition to incorporating a piceatannol skeleton. 3’-HPT is a trans-stilbene molecule containing two methoxy groups in one ring and two hydroxy groups in the other ring. Recent research also revealed that 3’-HPT enhances the SIRT1 polypeptide activity and inhibits adipogenesis more effectively than res- veratrol or its natural methoxylated derivative pterostilbene [6]. The antioxidant activity of 3’- HPT is approximately 30 times more potent than resveratrol, 70 times more than vitamin E, 80 times more than vitamin C and 90 times than BHT (butylated hydroxytoluene, the com- monly used synthetic antioxidant) [7]. An important factor that hampers the utility of resveratrol as a dietary supplement is its questionable stability on storage [8]. In evaluating the structure-activity correlation, 3’-HPT possesses the skeleton structure of piceatannol, a known and active metabolite of resveratrol. The dimethoxy-substitution seems to confer relative stability on the molecule [9, 10]. 3’-HPT has been reported to be an excellent anti-proliferative and anti-apoptotic compound [5] and less toxic on normal haemopoietic stem cells than on leukemia and lymphoma cells [4]. In spite of the beneficial pharmacological and health promoting properties of 3’-HPT, there is a lacuna in information on its safety (toxicity) profile. Introduction 3’-Hydroxypterostilbene (3’-HPT) is a naturally occurring stilbenoid molecule present in the plants Sphaerophysa salsula [1], Pterocarpus marsupium and also in Honey Bee propolis (bee glue) [2]. Sphaerophysa salsula (Pall.) DC (Leguminosae) is a shrub widely distributed in the Central-Asia and in the northwest of China. It is used as a folk medicine for the treatment of hypertension in China [1]. 3’-HPT is an analogue of another natural stilbenoid pterostilbene 1 / 14 PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity Fig 1. Stilbenoid analogues of resveratrol. doi:10.1371/journal.pone.0172770.g001 or ClinWorld are articulated in the ’author contributions’ section. Competing interests: The authors have declared that no competing interests exist. This does not alter our adherence to PLOS ONE policies on sharing data and materials. Fig 1. Stilbenoid analogues of resveratrol. Fig 1. Stilbenoid analogues of resveratrol. doi:10.1371/journal.pone.0172770.g001 doi:10.1371/journal.pone.0172770.g001 doi:10.1371/journal.pone.0172770.g001 which belongs to a class of compounds known as phytoalexins. Since pterostilbene is widely recognized for its anti-cancer and anti-diabetic properties, it is important to evaluate the toxi- cological role of its metabolites, as 3’-Hydroxypterostilbene has been found be one of the uri- nary metabolites of pterostilbene in mice [3]. Stilbene compounds like (trans-) resveratrol and piceatannol are widely distributed in nature. In view of their extensive range of biological activities such as cardioprotection and cancer chemopreventive properties, they have become of special interest to biologists and chemists. The trans-resveratrol-like natural stilbenoids have shown remarkable anti-cancer activity higher than resveratrol (Fig 1). 3’-HPT for exam- ple, is reported to be 50 to 97 times more potent than trans-resveratrol in eliciting significant apoptotic action on tumour cell lines and is markedly more active than pterostilbene in induc- ing apoptosis in sensitive and multi drug-resistant (MDR) leukemia cells [4]. Based on struc- ture activity relationships, the importance of 3,5-dimethoxy motif in conferring pro-apoptic activity on leukemia cells has been recognized [5]. Pterostilbene and 3’-HPT possess the neces- sary 3,5-dimethoxy motif. Among these, 3’-HPT is an interesting natural compound which may be useful in treating different types of haematological malignancies as it is reported to possess a ‘privileged’ structure [4] motif of 3,5-dimethoxy group in addition to incorporating a piceatannol skeleton. 3’-HPT is a trans-stilbene molecule containing two methoxy groups in one ring and two hydroxy groups in the other ring. Recent research also revealed that 3’-HPT enhances the SIRT1 polypeptide activity and inhibits adipogenesis more effectively than res- veratrol or its natural methoxylated derivative pterostilbene [6]. The antioxidant activity of 3’- HPT is approximately 30 times more potent than resveratrol, 70 times more than vitamin E, 80 times more than vitamin C and 90 times than BHT (butylated hydroxytoluene, the com- monly used synthetic antioxidant) [7]. which belongs to a class of compounds known as phytoalexins. PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 Chemicals and analysis of 3’-HPT The chemicals and solvents used throughout the study were of analytical grade. The test item 3’- HPT was provided by Sami Labs Ltd., (Bengaluru, Karnataka, India). The 3’- HPT as pro- vided was obtained through the steps disclosed in the US patent no. US 9458075 B1 [11]. The 3’-HPT as obtained had the following specifications and structural analysis was per- formed using IR spectrum, proton NMR, carbon NMR and mass spectrum. Analysis using HPLC indicated that the compound obtained was 98.31% w/w pure and was off white in color. Microbiological profile revealed negative for total yeast, E. coli, Salmonella spp., Staphylococcus aureus, Pseudomonas aeruginosa and Enterobacteriaceae when tested. Obtained compound was stable at 25 ± 2˚C and 60 ± 5% RH conditions when tested for a period of three years. Compound was also found stable when tested at accelerated conditions 40 ± 2˚C and 75 ± 5% RH for a period of 6 months. Its molecular formula, the molecular weight being C16H16O4 and 272.30 respectively, having the CA index name: 1,2-Benzenediol, 4-[(1E)-2-(3,5-dimethoxy- phenyl)ethenyl]- (9CI). Hence, the current study targeted at a 2 / 14 PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity 90 day repeated dose oral toxicity as well as a reproductive/developmental toxicity of 3’- HPT was performed. Ethics Both 90 day repeated dose oral toxicity and reproductive/developmental toxicity studies were performed in strict accordance with the recommendation of the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA) guidelines for laboratory ani- mal facility, and as per OECD guideline 408, 421 respectively [12, 13]. The individual protocol was approved by the Institutional Animal Ethics Committee (IAEC) of Bioneeds (Registration No. 969/bc/06/CPCSEA) and Dabur Research Foundation (DRF) (Registration No.64/PO/br/ s/99/CPCSEA) respectively. The animals were euthanized using carbon dioxide asphyxiation, and all efforts were made to minimize suffering. Animals For reproductive/developmental toxicity study—young healthy adult Wistar rats aged 12–13 weeks at the time of mating as sourced from Dabur Research Foundation (DRF) animal facility and for 90 day repeated dose oral toxicity study- Sprague Dawley (SD) rats aged 6–8 weeks sourced from Bioneeds animal facility were used. These were housed in a controlled environ- ment of temperature (22 ± 3˚C) and relative humidity (RH) (50–60%), with photo periods of 12/12 h light-dark cycles. Provided with pelleted diet (Golden Feeds, New Delhi, India), fil- tered water ad libitum. Housing in cages: For reproductive/developmental toxicity study, five animals of same sex per cage before mating, two animals per cage of each sex during mating and pregnant females were individually housed with soft nesting material. For 90 day repeated dose oral toxicity study, two animals of same sex per cage were housed during acclimatization and study period. Dosage groups for repeated dose 90 day oral toxicity study in SD rats. Groups Dose (mg/kg/d) Concentration (mg/ml) No. of Rats/group, Sex (n) Male Female NG1 Vehicle Control (VC) (Corn oil) 0 NA 10 10 NG2 Control Recovery (CR) (Corn oil) 0 NA 10 10 NG3 Low dose (LD) 20 2 10 10 NG4 Mid dose (MD) 80 8 10 10 NG5 High dose (HD) 200 20 10 10 NG6 High dose Recovery (HDR) 200 20 10 10 NA- Not Applicable; n = number. doi:10.1371/journal.pone.0172770.t001 Table 1. Dosage groups for repeated dose 90 day oral toxicity study in SD rats. Groups Dose (mg/kg/d) Concentration (mg/ml) No. of Rats/group, Sex (n) Male Female NG1 Vehicle Control (VC) (Corn oil) 0 NA 10 10 NG2 Control Recovery (CR) (Corn oil) 0 NA 10 10 NG3 Low dose (LD) 20 2 10 10 NG4 Mid dose (MD) 80 8 10 10 NG5 High dose (HD) 200 20 10 10 NG6 High dose Recovery (HDR) 200 20 10 10 NA- Not Applicable; n = number. doi:10 1371/journal pone 0172770 t001 able 1. Dosage groups for repeated dose 90 day oral toxicity study in SD rats. 90 day repeated dose oral toxicity study Experimental design. The dose formulations, vehicle were administered to the treatment and recovery group animals (Table 1). Dose formulations, vehicle were administered by oral gavage once daily, to the respective groups, at the same time each day (varying by ± 2 h) for a period of 90 days with the actual volume calculated based on the recent body weight recorded at weekly intervals. There was no administration of test item formulations / vehicle during the 28-day recovery period in order to permit evaluation of the persistence, reversibility or delayed occurrence of toxic effects. Clinical observations. All treated, control rats were observed twice (morning and after- noon) daily for clinical signs and mortality for a total period of 90 days. The body weight of each rat was recorded prior to treatment on day 1, weekly thereafter and at end of the 90 day study. Group mean body weights were calculated. The food consumption of all rats in each cage was recorded from day of commencement of treatment and average weekly consumption was calculated. During post-treatment period, food consumption was recorded weekly for recovery group rats. Neurobehavioural assessment. The neurological examination was carried out on all the animals in twelfth week of the study. The neurobehavioural assessment (functional examina- tion) of sensory reactivity to different types of auditory, visual and proprioceptive stimuli (landing foot splay, righting reflex, and gait), assessment of grip strength and motor activity assessment were recorded (Table 2). At the end of treatment all surviving animal were fasted overnight with water given ad libitum. g g Ophthalmological examination. Ophthalmological examination was carried out at the beginning and at termination of the study for all the groups. Ophthalmological examination. Ophthalmological examination was carried out at the beginning and at termination of the study for all the groups. Haemotology and clinical chemistry. Blood samples were collected in EDTA tubes for determination of haematological parameters (Table 3) and heparin tubes for clinical chemistry (Table 4). Haemotology and clinical chemistry. Blood samples were collected in EDTA tubes for determination of haematological parameters (Table 3) and heparin tubes for clinical chemistry (Table 4). Necropsy/Gross pathology and organ weights. On completion of 90 days and 118 days in case of treatment and recovery periods respectively, all surviving rats were fasted overnight, subjected to complete necropsy, weighed before exsanguination, and gross pathological exami- nation was performed. Dosing The test item 3’- HPT was freshly prepared and daily administered orally after suspending it in corn oil (mg/kg/bwt). 3 / 14 PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity 90 day repeated dose oral toxicity study Experimental design. The dose formulations, vehicle were administered to the treatment and recovery group animals (Table 1). Dose formulations, vehicle were administered by oral gavage once daily, to the respective groups, at the same time each day (varying by ± 2 h) for a period of 90 days with the actual volume calculated based on the recent body weight recorded at weekly intervals. There was no administration of test item formulations / vehicle during the 28-day recovery period in order to permit evaluation of the persistence, reversibility or delayed occurrence of toxic effects. Clinical observations. All treated, control rats were observed twice (morning and after- noon) daily for clinical signs and mortality for a total period of 90 days. The body weight of each rat was recorded prior to treatment on day 1, weekly thereafter and at end of the 90 day study. Group mean body weights were calculated. The food consumption of all rats in each cage was recorded from day of commencement of treatment and average weekly consumption was calculated. During post-treatment period, food consumption was recorded weekly for recovery group rats. Neurobehavioural assessment. The neurological examination was carried out on all the animals in twelfth week of the study. The neurobehavioural assessment (functional examina- tion) of sensory reactivity to different types of auditory, visual and proprioceptive stimuli (landing foot splay, righting reflex, and gait), assessment of grip strength and motor activity assessment were recorded (Table 2). At the end of treatment all surviving animal were fasted overnight with water given ad libitum. Ophthalmological examination. Ophthalmological examination was carried out at the beginning and at termination of the study for all the groups. Haemotology and clinical chemistry. Blood samples were collected in EDTA tubes for determination of haematological parameters (Table 3) and heparin tubes for clinical chemistry (Table 4). Necropsy/Gross pathology and organ weights. On completion of 90 days and 118 days in case of treatment and recovery periods respectively, all surviving rats were fasted overnight, subjected to complete necropsy, weighed before exsanguination, and gross pathological exami- nation was performed. The organs thus collected (Table 5) from all animals were trimmed of adherent tissue, fat as appropriate and weighed wet as soon as possible to avoid drying. Histopathology. Histopathological examination was performed on tissues of all animals Table 1. PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity Table 2. Effect of 90 day oral administration of 3’-HPT on functional observation battery (neurobehavioural assessment) tests in male and female SD rats. Groups Functional observation battery Neurobehavioural response Dose (mg/kg/d) Visual Response Auditory Response Response to Proprioceptive Stimuli Gait Landing Foot splay (cm) Righting Reﬂex a b c d Locomotor Activity Grasping strength Male Female NG1 (VC) 0 0 0 0 3.4 ± 0.9 3.8 ± 0.9 0 0 0 0 0 0 NG2 (CR) 0 0 0 0 3.9 ± 0.7 4.3 ± 1.2 0 0 0 0 0 0 NG3 (LD) 20 0 0 0 3.6 ± 1.0 3.6 ± 0.9 0 0 0 0 0 0 NG4 (MD) 80 0 0 0 3.5 ± 0.8 4.0 ± 0.8 0 0 0 0 0 0 NG5 (HD) 200 0 0 0 4.1 ± 0.9 4.1 ± 0.9 0 0 0 0 0 0 NG6 (HDR) 200 0 0 0 3.5 ± 0.6 3.9 ± 1.0 0 0 0 0 0 0 n = 10; 0: Normal; a: on back; b: body tilted; c: head tilted; d: dropped. Values are in Mean ± SEM. doi:10.1371/journal.pone.0172770.t002 ation of 3’-HPT on functional observation battery (neurobehavioural assessment) tests in male and female Table 3. Effect of 90 day oral administration of 3’-HPT on haematology data of male and female SD rats. Parameters Groups / Dose (mg/ kg/d) NG1 (VC) / (0) NG2 (CR) / (0) NG3 (LD) / (20) NG4 (MD) / (80) NG5 (HD) / (200) NG6 (HDR) / (200) Sex Total Leucocyte Count (103 cells/μl) M 14.9 ± 2.8 15.6 ± 3.2 14.4 ± 3.1 14.9 ± 2.3 15.5 ± 2.7 17.7 ± 1.5 F 10.0 ± 2.9 9.5 ± 2.0 8.5 ± 1.3 8.7 ± 2.5 11.2 ± 2.9 9.1 ± 2.8 Total Erythrocyte count (106 cells/μl) M 7.5 ± 0.3 7.5 ± 0.4 7.5 ± 0.4 7.3 ± 0.8 7.7 ± 0.5 8.0 ± 0.6 F 6.6 ± 0.4 7.0 ± 0.8 7.0 ± 0.6 6.6 ± 0.6 6.8 ± 0.7 6.7 ± 0.6 Haemoglobin (g/dl) M 13.2 ± 0.6 13.1 ± 0.6 13.2 ± 0.8 13.0 ± 0.5 12.9 ± 0.7 13.7 ± 1.0 F 12.4 ± 0.8 13.0 ± 1.2 13.1 ± 1.2 12.5 ± 1.4 12.8 ± 1.2 12.6 ± 0.8 Haematocrit (%) M 42.9 ± 1.9 42.5 ± 2.2 42.9 ± 2.3 41.1 ± 4.9 42.7 ± 2.7 44.5 ± 3.2 F 38.1 ± 2.2 39.8 ± 3.8 39.9 ± 3.2 38.0 ± 4.1 38.8 ± 3.6 38.4 ± 2.2 Mean corpuscular volume (fL) M 57.4 ± 2.1 56.4 ± 1.9 57.0 ± 1.6 55.9 ± 0.9 55.6 ± 1.9 55.8 ± 2.4 F 57.6 ± 1.3 56.6 ± 1.6 57.0 ± 2.0 57.1 ± 2.3 57 ± 2.1 57.2 ± 3.4 Mean corpuscular haemoglobin (pg) M 17.8 ± 0.6 17.4 ± 0.5 17.6 ± 0.8 17.9 ± 2.9 16.9 ± 0.6 17.1 ± 0.8 F 18.8 ± 0.8 18.6 ± 0.9 18.8 ± 1.1 18.7 ± 0.9 18.8 ± 0.8 18.8 ± 1.2 Mean corpuscular haemoglobin concentration (g/dl) M 30.9 ± 0.4 30.9 ± 0.4 30.8 ± 1.1 32.1 ± 5.6 30.3 ± 0.4 30.7 ± 0.9 F 32.6 ± 0.7 32.8 ± 0.2 33.0 ± 1.2 32.8 ± 0.4 32.9 ± 0.6 32.9 ± 0.7 Platelet Count (103 cells/μl) M 688.7 ± 123.7 624.0 ± 85.5 633.5 ± 100.1 618.3 ± 169.6 714.4 ± 83.8 703.8 ± 61.9 F 641.2 ± 81.3 659.6 ± 109.4 621.8 ± 81.3 563.6 ± 144.6 639.5 ± 96.0 593.8 ± 120.4 Blood clotting time (Sec) M 72.9 ± 16.9 70.3 ± 14.3 66.7 ± 8.3 67.1 ± 18.2 60.7 ± 13.6 67.9 ± 15.7 F 55.8 ± 22.5 62.9 ± 18.8 51.0 ± 11.5 53.5 ± 14.5 57.2 ± 18.2 59.4 ± 11.2 Neutrophils (%) M 15.6 ± 5.5 20.6 ± 7.0 19.2 ± 5.4 16.1 ± 4.7 17.2 ± 4.8 17.3 ± 5.9 F 20.4 ± 4.9 21.4 ± 5.1 17.6 ± 6.3 20.7 ± 5.8 20.9 ± 6.6 23.3 ± 6.1 Lymphocytes (%) M 84.2 ± 5.8 81.0 ± 6.3 80.7 ± 5.4 83.7 ± 5.0 82.4 ± 5.2 82.6 ± 5.9 F 79.4 ± 4.7 78.2 ± 5.5 82.0 ± 6.6 79.0 ± 6.2 78.8 ± 6.8 76.6 ± 6.1 Eosinophils (%) M 0.1 ± 0.3 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.1 ± 0.3 0.1 ± 0.3 F 0.0 ± 0.0 0.1 ± 0.3 0.1 ± 0.3 0.3 ± 0.5 0.0 ± 0.0 0.0 ± 0.0 Monocytes (%) M 0.1 ± 0.3 0.1 ± 0.3 0.1 ± 0.3 0.2 ± 0.6 0.3 ± 0.7 0.0 ± 0.0 F 0.2 ± 0.6 0.3 ± 0.7 0.3 ± 0.7 0.2 ± 0.6 0.3 ± 0.7 0.1 ± 0.3 n = 10; M, male; F, female. 90 day repeated dose oral toxicity study The organs thus collected (Table 5) from all animals were trimmed of adherent tissue, fat as appropriate and weighed wet as soon as possible to avoid drying. Histopathology. Histopathological examination was performed on tissues of all animals from the control and the high dosage group, sacrificed at scheduled termination. The tissues / organs (Mesenteric lymph node, thymus, trachea, lungs, heart, thyroid, stomach, small intestine, large intestine (colon), liver, spleen, kidneys, adrenals, epididymides / uterus, testes / ovaries and Histopathology. Histopathological examination was performed on tissues of all animals from the control and the high dosage group, sacrificed at scheduled termination. The tissues / organs (Mesenteric lymph node, thymus, trachea, lungs, heart, thyroid, stomach, small intestine, large intestine (colon), liver, spleen, kidneys, adrenals, epididymides / uterus, testes / ovaries and 4 / 14 PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 n = 10; M, male; F, female. Values are in Mean ± SD. PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity Table 4. Effect of 90 day oral administration of 3’-HPT on biochemical data of male and female SD rats. Parameter Groups / Dose (mg/ kg/d) NG1 (VC) / (0) NG2 (CR) / (0) NG3 (LD) / (20) NG4 (MD) / (80) NG5 (HD) / (200) NG6 (HDR) / (200) Sex Glucose (mg/dl) M 112.70 ± 18.18 109.40 ± 18.90 105.70 ± 15.09 100.39 ± 16.39 101.80 ± 12.84 98.70 ± 12.28 F 107.40 ± 18.65 114.50 ± 14.18 106.90 ± 10.80 105.10 ± 19.93 99.00 ± 13.78 103.10 ± 12.27 Total Cholesterol (mg/dl) M 42.10 ± 6.47 40.20 ± 7.94 37.80 ± 8.99 43.00 ± 9.76 38.30 ± 9.51 41.10 ± 7.71 F 57.50 ± 12.84 58.10 ± 8.52 50.70 ± 6.57 48.00 ± 8.94 47.40 ± 9.49 48.50 ± 7.81 Creatinine (mg/dl) M 0.86 ± 0.15 0.78 ± 0.04 0.77 ± 0.07 0.82 ± 0.09 0.77 ± 0.12 0.78 ± 0.14 F 0.98 ± 0.06 0.99 ± 0.03 0.98 ± 0.13 0.97 ± 0.18 0.93 ± 0.13 0.96 ± 0.08 Total Protein (mg/dl) M 5.62 ± 0.88 5.62 ± 0.38 5.55 ± 0.25 5.28 ± 0.44 5.64 ± 0.52 5.69 ± 0.37 F 6.85 ± 0.35 6.99 ± 0.43 6.78 ± 0.43 7.01 ± 0.52 7.15 ± 0.39 7.00 ± 0.49 Albumin (mg/dl) M 3.51 ± 0.22 3.31 ± 0.17 3.28 ± 0.25 3.28 ± 0.19 3.38 ± 0.17 3.47 ± 0.28 F 3.90 ± 0.27 3.93 ± 0.47 3.89 ± 0.23 4.12 ± 0.27 4.03 ± 0.18 3.96 ± 0.36 Triglycerides (mg/dl) M 60.05 ± 13.71 68.70 ± 14.00 65.00 ± 14.41 69.40 ± 20.66 50.40 ± 12.37 52.00 ± 12.34 F 69.30 ± 14.85 71.70 ± 21.15 55.90 ± 20.37 59.40 ± 13.33 55.90 ± 21.12 52.50 ± 16.22 Urea Nitrogen (mg/dl) M 13.42 ± 2.42 14.64 ± 1.65 14.62 ± 2.76 15.35 ± 3.24 16.03 ± 1.90 16.01 ± 1.89 F 18.33 ± 4.23 19.91 ± 5.21 15.75 ± 2.51 15.60 ± 2.29 17.54 ± 2.50 19.27 ± 3.79 Total bilirubin (mg/dl) M 0.33 ± 0.08 0.36 ± 0.07 0.33 ± 0.07 0.30 ± 0.08 0.32 ± 0.09 0.29 ± 0.07 F 0.37 ± 0.13 0.39 ± 0.12 0.32 ± 0.04 0.35 ± 0.07 0.33 ± 0.11 0.30 ± 0.05 Alanine aminotransferase (U/L) M 77.50 ± 20.46 67.60 ± 19.08 60.30 ± 14.65 61.20 ± 12.78 54.20± 18.08 53.30± 10.63 F 73.00 ± 16.15 78.00 ± 14.73 74.40 ± 18.07 65.30 ± 10.06 56.10± 10.65 57.40± 7.83 Aspartate aminotransferase (U/L) M 166.60 ± 17.52 168.40 ± 15.09 163.90 ± 25.05 149.00 ± 23.49 139.30± 9.17 140.00± 12.82 F 172.70 ± 20.34 169.40 ± 16.71 171.60 ± 17.76 167.30 ± 19.45 147.50± 20.64 143.20± 21.47 High density lipoprotein (mg/ dl) M 22.67 ± 1.95 22.56 ± 3.64 21.00 ± 5.49 24.07 ± 3.50 21.37 ± 4.80 27.43 ± 5.01 F 27.85 ± 6.20 26.63 ± 4.87 26.60 ± 4.58 27.31 ± 4.80 30.82 ± 6.10 31.59 ± 4.69 Alkaline phosphatase (U/L) M 143.10 ± 18.79 152.70 ± 22.07 151.80 ± 25.68 170.66 ± 23.42 144.70 ± 33.46 160.15 ± 21.45 F 144.44 ± 19.94 139.84 ± 27.01 145.68 ± 29.32 154.78 ± 31.17 163.13 ± 19.77 159.60 ± 22.95 Low density lipoprotein (mg/ dl) M 17.26 ± 2.37 13.86 ± 2.29 14.95 ± 3.80 16.47 ± 2.39 19.94 ± 4.67 17.69 ± 3.54 F 16.36 ± 3.03 14.95 ± 7.12 15.60 ± 3.30 15.00 ± 3.40 12.14 ± 2.10 14.55 ± 3.16 Calcium (mg/dl) M 9.26 ± 0.77 8.77 ± 1.12 9.09 ± 0.90 9.18 ± 0.86 9.09 ± 0.90 8.87 ± 0.96 F 9.14 ± 0.71 8.91 ± 1.09 17.51 ± 26.89 8.92 ± 0.97 8.79 ± 0.84 9.31 ± 1.12 Phosphorous (mg/dl) M 3.97 ± 0.62 3.75 ± 0.56 3.67 ± 0.57 4.94 ± 0.90 4.66 ± 1.40 4.55 ± 1.64 F 4.19 ± 0.59 4.14 ± 0.32 4.31 ± 0.51 4.20 ± 0.69 4.76 ± 0.58 4.72 ± 0.93 Sodium (mmol/L) M 141.40 ± 4.38 140.60 ± 4.67 140.20 ± 3.19 139.20 ± 7.11 140.80 ± 1.93 140.60 ± 4.72 F 141.30 ± 2.67 139.80 ± 4.87 140.90 ± 3.73 140.80 ± 3.74 141.50 ± 3.63 141.60 ± 2.59 Potassium (mg/dl) M 4.28 ± 0.57 4.18 ± 0.70 4.14 ± 0.67 4.11 ± 0.28 3.94 ± 0.39 4.10 ± 0.34 F 4.19 ± 0.34 4.11 ± 0.52 4.18 ± 0.47 4.16 ± 0.68 4.11 ± 0.28 4.06 ± 0.43 n = 10; M, male; F, female. Values are in Mean ± SD; Table 4. Effect of 90 day oral administration of 3’-HPT on biochemical data of male and female SD rats. brain) from all animals, were collected and preserved in 10% neutral buffered formalin with the eyes preserved in Davidson’s fluid. Reproductive/Developmental toxicity study Experimental design. The animal groups (Table 6) of each sex were dosed once daily for a period of two weeks while pre-mating, mating, on the day before sacrifice and in females n = 10; M, male; F, female. Values are in Mean ± SD; * Statistically signiﬁcant (P<0.05) compared with the VC group. doi:10.1371/journal.pone.0172770.t004 PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 6 / 14 n = 10; M, male; F, female. Values are in Mean ± SD; * Statistically signiﬁcant (P<0.05) compared with the VC group. brain) from all animals, were collected and preserved in 10% neutral buffered formalin with the eyes preserved in Davidson’s fluid. brain) from all animals, were collected and preserved in 10% neutral buffered formalin with the eyes preserved in Davidson’s fluid. Values are in Mean ± SD. Table 3. Effect of 90 day oral administration of 3’-HPT on haematology data of male and female SD rats. PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 5 / 14 Dosage groups for reproduction/developmental toxicity study in Wistar rats. Groups Dose (mg/kg/d) Concentration (mg/ml) No. of Rats/group, Sex (n) Male Female RG1 Vehicle Control (VC) (Corn oil) 0 NA 10 10 RG2 Low dose (LD) 20 2 10 10 RG3 Mid dose (MD) 100 10 10 10 RG4 High dose (HD) 200 20 10 10 NA- Not Applicable; n = number. doi:10.1371/journal.pone.0172770.t006 3 hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity Table 5. Effect of 90 day oral administration of 3’-HPT on absolute organ weight (mg) of male and female SD rats. Organs Groups/ Dose (mg/kg/ d) NG1 (VC) / (0) NG2 (CR) / (0) NG3 (LD) / (20) NG4 (MD) / (80) NG5 (HD) / (200) NG6 (HDR) / (200) Sex Liver M 8751.4 ± 1070.6 9867.3 ±1420.2 8931.7 ± 1279.6 9008.1 ± 1295.5 7630.4 ± 3601.3 7998.8 ± 1561.2 F 6817.0 ± 714.1 6701.5 ± 1258.3 6529.3 ± 663.9 6908.9 ± 386.7 6891.1 ± 383.8 7217.3 ± 1138.9 Kidneys M 1986.9 ± 317.4 2422.9 ± 383.4 1766.6 ± 617.8 1893.5 ± 242.1 1869.2 ± 198.8 2088.2 ± 288.1 F 1447.3 ± 175.6 1493.7 ± 173.6 1294.7 ± 160.2 1437.7 ± 94.5 1449.1 ± 157.6 1515.6 ± 198.4 Adrenals M 43.5 ± 8.7 53.1 ± 6.8 49.4 ± 11.0 49.0 ± 9.3 52.1 ± 14.6 54.4 ± 7.7 F 64.9 ± 8.9 64.0 ± 14.2 64.9 ± 13.4 67.3 ± 16.2 68.2 ± 8.1 65.9 ± 14.9 Testes / Ovaries M 2951.2 ± 343.0 3095.5 ± 375.3 3060.6 ± 359.0 3125.8 ± 316.1 2918.9 ± 248.2 2940.6 ± 543.8 F 101.4 ± 41.0 107.0 ± 18.3 94.5 ± 14.7 101.0 ± 20.5 85.1 ± 17.3 107.4 ± 18.8 Epididymides / Uterus M 1302.1 ± 160.9 1393.2 ± 235.6 1179.6 ± 117.0 1189.0 ± 211.2 1208.1 ± 147.4 1331.6 ± 245.0 F 1156.1 ± 156.2 1122.2 ± 255.7 1119.9 ± 194.9 1149.4 ± 114.0 1109.5 ± 254.5 1082.7 ± 194.2 Thymus M 294.8 ± 72.3 363.9 ± 42.0 317.2 ± 61.5 260.7 ± 96.4 277.3 ± 102.7 385.7 ± 89.6 F 365.7 ± 91.1 301.6 ± 81.0 360.9 ± 76.3 309.5 ± 62.3 339.4 ± 60.0 272.9 ± 72.5 Spleen M 1141.1 ± 163.5 1092.2 ± 128.5 1154.7 ± 222.9 1003.7 ± 158.1 917.9 ± 120.6 956.9 ± 148.0 F 928.3 ± 98.9 864.7 ± 152.1 790.3 ± 81.6 901.7 ± 181.2 917.6 ± 188.2 844.2 ± 188.0 Brain M 1844.1 ± 171.0 1986.3 ± 88.8 1869.0 ± 102.5 1859.7 ± 179.1 1810.6 ± 109.0 1920.2 ± 170.3 F 1862.4 ± 104.6 1817.7 ± 103.8 1826.3 ± 108.9 1819.3 ± 161.6 1885.4 ± 92.5 1839.8 ± 132.1 Heart M 1010.0 ± 154.1 1174.0 ± 117.7 990.6 ± 130.5 973.2 ± 136.0 926.3 ± 102.8 1103.2 ± 130.3 F 831.5 ± 63.4 814.4 ± 128.2 812.4 ± 99.8 845.1 ± 60.0 833.8 ± 114.7 795.2 ± 64.1 n = 10; M, male; F, female; values in Mean ± S.D. Reproductive/Developmental toxicity study Experimental design. The animal groups (Table 6) of each sex were dosed once daily for a period of two weeks while pre-mating, mating, on the day before sacrifice and in females 6 / 14 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity during pregnancy and four days of lactation period. Mating ratio of 1:1 (one male to one female), was allowed by housing them in single cages, subsequently the female animals were caged individually upon confirmation of pregnancy. Clinical observations. All animals were observed once daily throughout the period of study for clinical signs of toxicity. Rats were observed for clinical signs like changes in skin and fur, eyes, breathing, mucous membrane, occurrence of secretions and excretions, changes in gait, posture and presence of any abnormal behaviour (data not included). All animals were weighed on the first day of dosing, subsequently weekly once and at study termination. During pregnancy, females were weighed on day 0, 7, 14 and 21, 24 h of parturition and on day 4 post partum. Pups which were alive were counted, sexed and weighed within 24 hours of parturi- tion and on day 4 post partum. The quantity of food consumed by the animals (g/rat/day (d)) of the treatment group was measured during in-life phase (pre-mating, pregnancy and Table 5. Effect of 90 day oral administration of 3’-HPT on absolute organ weight (mg) of male and female SD rats. Organs Groups/ Dose (mg/kg/ d) NG1 (VC) / (0) NG2 (CR) / (0) NG3 (LD) / (20) NG4 (MD) / (80) NG5 (HD) / (200) NG6 (HDR) / (200) Sex Liver M 8751.4 ± 1070.6 9867.3 ±1420.2 8931.7 ± 1279.6 9008.1 ± 1295.5 7630.4 ± 3601.3 7998.8 ± 1561.2 F 6817.0 ± 714.1 6701.5 ± 1258.3 6529.3 ± 663.9 6908.9 ± 386.7 6891.1 ± 383.8 7217.3 ± 1138.9 Kidneys M 1986.9 ± 317.4 2422.9 ± 383.4 1766.6 ± 617.8 1893.5 ± 242.1 1869.2 ± 198.8 2088.2 ± 288.1 F 1447.3 ± 175.6 1493.7 ± 173.6 1294.7 ± 160.2 1437.7 ± 94.5 1449.1 ± 157.6 1515.6 ± 198.4 Adrenals M 43.5 ± 8.7 53.1 ± 6.8 49.4 ± 11.0 49.0 ± 9.3 52.1 ± 14.6 54.4 ± 7.7 F 64.9 ± 8.9 64.0 ± 14.2 64.9 ± 13.4 67.3 ± 16.2 68.2 ± 8.1 65.9 ± 14.9 Testes / Ovaries M 2951.2 ± 343.0 3095.5 ± 375.3 3060.6 ± 359.0 3125.8 ± 316.1 2918.9 ± 248.2 2940.6 ± 543.8 F 101.4 ± 41.0 107.0 ± 18.3 94.5 ± 14.7 101.0 ± 20.5 85.1 ± 17.3 107.4 ± 18.8 Epididymides / Uterus M 1302.1 ± 160.9 1393.2 ± 235.6 1179.6 ± 117.0 1189.0 ± 211.2 1208.1 ± 147.4 1331.6 ± 245.0 F 1156.1 ± 156.2 1122.2 ± 255.7 1119.9 ± 194.9 1149.4 ± 114.0 1109.5 ± 254.5 1082.7 ± 194.2 Thymus M 294.8 ± 72.3 363.9 ± 42.0 317.2 ± 61.5 260.7 ± 96.4 277.3 ± 102.7 385.7 ± 89.6 F 365.7 ± 91.1 301.6 ± 81.0 360.9 ± 76.3 309.5 ± 62.3 339.4 ± 60.0 272.9 ± 72.5 Spleen M 1141.1 ± 163.5 1092.2 ± 128.5 1154.7 ± 222.9 1003.7 ± 158.1 917.9 ± 120.6 956.9 ± 148.0 F 928.3 ± 98.9 864.7 ± 152.1 790.3 ± 81.6 901.7 ± 181.2 917.6 ± 188.2 844.2 ± 188.0 Brain M 1844.1 ± 171.0 1986.3 ± 88.8 1869.0 ± 102.5 1859.7 ± 179.1 1810.6 ± 109.0 1920.2 ± 170.3 F 1862.4 ± 104.6 1817.7 ± 103.8 1826.3 ± 108.9 1819.3 ± 161.6 1885.4 ± 92.5 1839.8 ± 132.1 Heart M 1010.0 ± 154.1 1174.0 ± 117.7 990.6 ± 130.5 973.2 ± 136.0 926.3 ± 102.8 1103.2 ± 130.3 F 831.5 ± 63.4 814.4 ± 128.2 812.4 ± 99.8 845.1 ± 60.0 833.8 ± 114.7 795.2 ± 64.1 n = 10; M, male; F, female; values in Mean ± S.D. doi:10.1371/journal.pone.0172770.t005 Table 6. 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity lactation) of the study. The food consumed was recorded for each animal per dose group on weekly basis. Necropsy, gross pathology, organ weights and histopathology. The rats to be sacrificed at term (males on day 32 and females on day 54) were anaesthetized, weighed and exsangui- nated. During necropsy, the animals were examined visually (macroscopically) for any abnor- malities or pathological changes. The organs showing macroscopic lesions were preserved and further subjected to microscopic examination. Special attention was paid to the reproductive organs to detect abnormalities, if any. The number of corpora lutea and implantation sites were recorded. The reproductive organs were removed (ovaries, testes, uterus and epididymi- des), examined, the weight of these organs were recorded as absolute values and subsequently placed in Bouin’s fixative for histopathological examination. Histopathological and gross lesions examination was carried out on the preserved tissues of vehicle control (RG1) and high dose (RG4) group. Statistical analysis The data (body weight, food intake, organ weights, haematology and clinical chemistry) gener- ated from the 90 day repeated dose oral toxicity and reproductive/developmental toxicity study was subjected to statistical analysis using GraphPad prism version 5.00 (2007), GraphPad Software, USA. The statistical analysis of the 90 day repeated dose oral toxicity experimental data was carried out using one way ANOVA with Dunnet’s post test done for different treat- ment groups comparing with the control group data. The unpaired ‘t’-test was done for control recovery and high dose recovery group data. All analysis and comparisons were evaluated at 5% (P<0.05) level. The statistical analysis of the reproductive/developmental toxicity experi- mental data was carried out for differences among treated / control groups using one way ANOVA. doi:10 1371/journal pone 0172770 t005 t of 90 day oral administration of 3’-HPT on absolute organ weight (mg) of male and female SD rats. Table 5. Effect of 90 day oral administration of 3’-HPT on absolute organ weight (mg) of male and female Table 5. Effect of 90 day oral administration of 3’-HPT on absolute organ weigh n = 10; M, male; F, female; values in Mean ± S.D. during pregnancy and four days of lactation period. Mating ratio of 1:1 (one male to one female), was allowed by housing them in single cages, subsequently the female animals were caged individually upon confirmation of pregnancy. Clinical observations. All animals were observed once daily throughout the period of study for clinical signs of toxicity. Rats were observed for clinical signs like changes in skin and fur, eyes, breathing, mucous membrane, occurrence of secretions and excretions, changes in gait, posture and presence of any abnormal behaviour (data not included). All animals were weighed on the first day of dosing, subsequently weekly once and at study termination. During pregnancy, females were weighed on day 0, 7, 14 and 21, 24 h of parturition and on day 4 post partum. Pups which were alive were counted, sexed and weighed within 24 hours of parturi- tion and on day 4 post partum. The quantity of food consumed by the animals (g/rat/day (d)) of the treatment group was measured during in-life phase (pre-mating, pregnancy and Table 6. Dosage groups for reproduction/developmental toxicity study in Wistar rats. Groups Dose (mg/kg/d) Concentration (mg/ml) No. of Rats/group, Sex (n) Male Female RG1 Vehicle Control (VC) (Corn oil) 0 NA 10 10 RG2 Low dose (LD) 20 2 10 10 RG3 Mid dose (MD) 100 10 10 10 RG4 High dose (HD) 200 20 10 10 NA- Not Applicable; n = number. doi:10.1371/journal.pone.0172770.t006 Table 6. Dosage groups for reproduction/developmental toxicity study in Wistar rats. PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 7 / 14 7 / 14 PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity Fig 2. Body weight changes on oral administration of 3’-HPT measured over 13 weeks (90 days) in male and female SD rats. D1: Day one; W1, W5, W9 and W13: Week one, five, nine and thirteen. Data in Mean ± SD (n = 10/group). d i 10 1371/j l 0172770 002 Fig 2. Body weight changes on oral administration of 3’-HPT measured over 13 weeks (90 days) in male and female SD rats. D1: Day one; W1, W5, W9 and W13: Week one, five, nine and thirteen. Data in Mean ± SD (n = 10/group). doi:10.1371/journal.pone.0172770.g002 doi:10.1371/journal.pone.0172770.g002 comparable to the value of vehicle control (NG1: 7.00 ± 0.51 g/rat/day) at the end of week thir- teen. However, no significant differences in mean average food consumption were observed in both males and females in any treatment groups compared to that of the respective control groups. Neurobehavioural assessment. Functional observational battery tests such as visual response, auditory response, gait, landing foot splay, righting reflex including grasping strength and locomotor activity measures performed (Table 2) on all test item treated group rats up to the highest dose (200 mg/kg bwt/d) did not reveal any treatment related abnormalities. Ophthalmic examinations. No treatment-related ophthalmic findings were observed in control animals or treatment animals of both sexes (data not included). Haematology and clinical chemistry. The haematological data (Table 3) of the control and treated rats on repeated oral treatment for 90 days did not cause any significant changes in any of the haematological parameters at all the doses tested. However, in clinical chemistry a significantly lower (P<0.05) Alanine aminotransferase (ALT) and Aspartate aminotransferase (AST) in high dose (NG5) and high dose recovery (NG6), in both male and female groups dur- ing treatment was observed with all the other biochemical parameters remaining unaffected (Table 4). Necropsy/Gross pathology, organ weights and histopathology. No gross lesions during necropsy were observed in any of the animals in treatment and recovery groups, indicating that the test item did not cause any lesions systemically (data not included). None of the ani- mals in treatment and recovery groups showed any variations in mean value of the absolute organ weights (Table 5) compared to respective control groups indicating that the test item did not had any adverse effects on organ weights. 90 day repeated oral toxicity The 90 day repeated oral toxicity in SD rats was evaluated with the 3’-HPT being administered orally in corn oil (vehicle) to the treatment group animals at different doses per kg bwt/day except control group (NG1) and control recovery (NG2) groups (Table 1). The vehicle and 3’- HPT administration to control recovery and high dose recovery (NG6) group animals respec- tively was stopped after 90 days of administration and observed for another 28 days to evaluate the effects during recovery period. Changes in body weight and food consumption. The body weights of the rats were mon- itored from day one till the end of thirteen weeks (90 days) (Fig 2). At the end of first week the mean body weights of both male and female rats were found to have increased following treat- ment. The final mean body weights in the male rats were higher than that in the female rats. The overall changes in body weight among the groups within both male and female rats were found to be not statistically significant. Food consumption was monitored from the first week for all groups NG1 to NG6 till the end of thirteen weeks (90 days) and was found to be compa- rable across all the test groups (data presented in S1 Fig). A slight decrease (not statistical sig- nificant) in the mean average food consumption was observed in the high dose group (NG5: 7.96 ± 1.70 g/rat/day) as compared to the vehicle control group (NG1: 11.16 ± 1.59 g/rat/day) at the end of first week in males (S1A Fig). A similar trend (not statistically significant) was observed in females (S1B Fig), in the mid dose group (NG4) with the mean average food con- sumption being 6.93 ± 0.99 g/rat/day at end of thirteenth week compared within the group at 9.09 ± 0.67 g/rat/day at end of first week, but the thirteenth week value of NG4 group was PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 8 / 14 PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 Reproductive/Developmental toxicity study To evaluate the reproductive/developmental toxicity, control animals were administered only with corn oil, while the other treatment groups were administered with the test item i.e. 3’- HPT at different dosages per kg bwt/d (Table 6). Body weight changes and food consumption. Males and females (during pre-mating period) exhibited an increase in body weight during the study period with the female body weight not being affected by the administration of 3’-HPT during the gestation and lactation period and was comparable with the control group (data not included). The changes in food consumption (g/rat/d) of both sexes during pre-mating was not statistically significant and comparable to the control group and that of the females (g/rat/d) during the gestation, lacta- tion period was comparable to the control group (data not included). Mean body weight of both male and female pups remained unaffected during lactation period (Table 7). Necropsy/Gross pathology, organ weights and histopathology. The animals were evalu- ated via gross visual observation, the male and female genitals and other organs with gross lesions were subjected to microscopic examination. Histopathology of the tissues of control and high dose groups did not show any test item related abnormality across all groups. The ovaries and epididymides did not show any test item related abnormality. However, the testes from the high dose group showed a mild increase in the number of seminiferous tubules with a mean value of 1.6% as against the control mean value of 0.3%. One uterus from the high dose group showed pronounced infiltration of endometrial tissues by eosinophils. The absolute organ weight of males (testes and epididymides) and females (ovaries and uterus) were not affected and found to be comparable with the control group (Table 8) after oral exposure to 3’- HPT. The mean number of implantations, mean number of corpora lutea (Table 9), number of pregnancies, number of dams littered, gestation length and number of days of pregnancy (Table 10) remained unaffected by the daily oral exposure of 3’-HPT. Females across groups Table 8. Absolute organ weights (in g) of animals upon 3’-HPT administration in reproductive/ developmental toxicity study of 3’-HPT in Wistar rats. Group Dose (mg/kg/d) Lactation period Day 1 Day 4 Male Female Male Female RG1 (VC) 0 27.8 ± 4.02 31.0 ± 1.93 36.1 ± 4.90 37.2 ± 3.40 RG2 (LD) 20 25.8 ± 3.04 30.7 ± 2.74 29.5 ± 5.44 37.7 ± 6.11 RG3 (MD) 100 21.8 ± 4.15 23.1 ± 5.32 29.4 ± 7.51 30.3 ± 8.08 RG4 (HD) 200 38.7 ± 6.33 24.7 ± 3.45 40.9 ± 8.00 26.7 ± 4.71 Values are in Mean ± SEM. doi:10.1371/journal.pone.0172770.t007 Table 8. Absolute organ weights (in g) of animals upon 3’-HPT administration in reproductive/ developmental toxicity study of 3’-HPT in Wistar rats. Table 7. Body weight of male and female pups (in g) during lactation in reproductive/developmental toxicity study of 3’-HPT in Wistar rats. Group Dose (mg/kg/d) Lactation period Day 1 Day 4 Male Female Male Female RG1 (VC) 0 27.8 ± 4.02 31.0 ± 1.93 36.1 ± 4.90 37.2 ± 3.40 RG2 (LD) 20 25.8 ± 3.04 30.7 ± 2.74 29.5 ± 5.44 37.7 ± 6.11 RG3 (MD) 100 21.8 ± 4.15 23.1 ± 5.32 29.4 ± 7.51 30.3 ± 8.08 RG4 (HD) 200 38.7 ± 6.33 24.7 ± 3.45 40.9 ± 8.00 26.7 ± 4.71 Values are in Mean ± SEM. Table 7. Body weight of male and female pups (in g) during lactation in reproductive/developmental tox None of the animals in any of the groups showed any histological changes during evaluation, indicating that the test item did not cause any histological changes in the tissues (data not included). 9 / 14 PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity Reproductive/Developmental toxicity study To evaluate the reproductive/developmental toxicity, control animals were administered only with corn oil, while the other treatment groups were administered with the test item i.e. 3’- HPT at different dosages per kg bwt/d (Table 6). Body weight changes and food consumption. Males and females (during pre-mating period) exhibited an increase in body weight during the study period with the female body weight not being affected by the administration of 3’-HPT during the gestation and lactation period and was comparable with the control group (data not included). The changes in food consumption (g/rat/d) of both sexes during pre-mating was not statistically significant and comparable to the control group and that of the females (g/rat/d) during the gestation, lacta- tion period was comparable to the control group (data not included). Mean body weight of both male and female pups remained unaffected during lactation period (Table 7). Necropsy/Gross pathology, organ weights and histopathology. The animals were evalu- ated via gross visual observation, the male and female genitals and other organs with gross lesions were subjected to microscopic examination. Histopathology of the tissues of control and high dose groups did not show any test item related abnormality across all groups. The ovaries and epididymides did not show any test item related abnormality. However, the testes from the high dose group showed a mild increase in the number of seminiferous tubules with a mean value of 1.6% as against the control mean value of 0.3%. One uterus from the high dose group showed pronounced infiltration of endometrial tissues by eosinophils. The absolute organ weight of males (testes and epididymides) and females (ovaries and uterus) were not affected and found to be comparable with the control group (Table 8) after oral exposure to 3’- HPT. The mean number of implantations, mean number of corpora lutea (Table 9), number of pregnancies, number of dams littered, gestation length and number of days of pregnancy (Table 10) remained unaffected by the daily oral exposure of 3’-HPT. Females across groups Table 7. Body weight of male and female pups (in g) during lactation in reproductive/developmental toxicity study of 3’-HPT in Wistar rats. Reproductive/Developmental toxicity study Group Dose (mg/kg/d) Male Female Testes Epididymides Ovaries Uterus RG1 (VC) 0 3.03 ± 0.09 1.19 ± 0.06 0.17 ± 0.01 0.60 ± 0.04 RG2 (LD) 20 2.9 ± 0.1 1.21 ± 0.04 0.18 ± 0.01 0.62 ± 0.04 RG3 (MD) 100 3.0 ± 0.13 1.24 ± 0.05 0.17 ± 0.01 0.65 ± 0.07 RG4 (HD) 200 3.1 ± 0.08 1.3 ± 0.04 0.17 ± 0.00 0.58 ± 0.04 Values are in Mean ± SEM. doi:10.1371/journal.pone.0172770.t008 nimals upon 3’-HPT administration in reproductive/ developmental toxicity study of 3’-HPT in Wistar Table 8. Absolute organ weights (in g) of animals upon 3’-HPT administration in reproductive/ developmental toxicity study of 3’-HPT in Wistar rats. e organ weights (in g) of animals upon 3’-HPT administration in reproductive/ developmental toxicity study o Table 8. Absolute organ weights (in g) of animals upon 3’-HPT administration in reproductive/ development rats PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 10 / 14 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity Table 9. Corpora lutea and implantation sites in numbers in reproductive/ developmental toxicity study in Wistar rats. Group Dose (mg/kg/d) No. of Corpora lutea (CL) No. of implantations (IMP) Successful implantations (%) Total (Mean) Total (Mean) RG1 (VC) 0 11.63 11.63 100 RG2 (LD) 20 11.14 10.29 92.31 RG3 (MD) 100 10.57 10.86 102.7 RG4 (HD) 200 12.00 11.00 91.67 doi:10.1371/journal.pone.0172770.t009 orpora lutea and implantation sites in numbers in reproductive/ developmental toxicity study in Wistar rats. Table 9. Corpora lutea and implantation sites in numbers in reproductive/ developmental toxicity study Table 9. Corpora lutea and implantation sites in numbers in reproductive/ deve showed normal pregnancy (in days). No dose dependent changes were observed in pre-coital interval, percentage of pre and post-implantation loss in all treatment groups (Table 10). doi:10.1371/journal.pone.0172770.t010 N = number; M, male; F, female; Pup weight (in g). $ p g ( g) $ Values in Mean ± SD; $ Values in Mean ± SD; $ $ $ $ Values in Mean ± SEM. 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity have since been associated with this compound and its close analogues (pterostilbene and 3’- HPT) have been the subject of numerous studies [2]. However, 3’-HPT was found markedly more active than the above compounds, displaying a low toxicity on normal haemopoietic stem cells [4]. 3’-HPT has been evidenced to have anti-adipogenic, anti-inflammatory, anti- oxidant, Histone Deacetylase (HDAC), and Sirtuin 1 (SIRT1) inhibitory activity. Collectively, the increased anti-oxidant, anti-adipogenic activity, cell selectivity and COX-2 specificity of 3’- HPT was found in support of its activity in preventing cardiovascular and pulmonary diseases. Coupled with HDAC and SIRT1 inhibitory activity, 3’-HPT has been found to demonstrate protective effects from neurodegenerative diseases and select cancers [14]. In spite of the several beneficial and therapeutic properties proposed, studies evaluating the toxicity of the 3’-HPT have not been reported. Hence the current study was proposed. y p y p p In current 90 day repeated oral toxicity study, all the animals were observed for health sta- tus, clinical signs of toxicity and mortality; weekly detailed veterinary examination (data not included) and subjected to neurological assessments. Body weight changes observed in both the sexes at 200 mg/kg bwt/d especially in the males (NG5) during the first week was consid- ered as non adverse as this was not a common observation in other groups. Also the slight decrease in the food consumption observed in the high dose group (NG5) at the end of first week in males and in the mid dose group (NG4) in females at the end of thirteenth week was not statistically significant for both males and females and was comparable across the groups in the subsequent weeks in case of males, and in case of females comparable across other treat- ment groups including vehicle control group (NG1) inferring that the 3’-HPT treatment did have any statistically significant effect on the food consumption in rats and was unrelated to body weight of both sex. The functional observational battery tests related to the neurobeha- vioral assessment were not affected as no treatment related abnormalities were observed even at the highest dose (NG5) of 200 mg/kg bwt/d. The same was true with the ophthalmic exami- nations with no changes observed in all the groups. The haematological data was comparable at all doses tested with no significant changes observed in other clinical chemistry parameters estimated in all the groups. PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 Discussion 3’-HPT an analogue of resveratrol is found naturally in plants such as Sphaerophysa salsula [1], Pterocarpus marsupium, and Honey Bee propolis (bee glue) [2]. The health benefits of stilbe- noid compounds such as resveratrol are well established. Various pharmacological activities Table 10. Summary of effects of 3’-HPT administration on reproduction/development in Wistar rats. Groups / Dose (mg/kg/d) VC (RG1) / (0) LD (RG2) / (20) MD (RG3) / (100) HD (RG4) / (200) Observations Pairs started (N) 10 10 10 10 Females showing evidence of copulation (N) 10 09 09 10 Females achieving pregnancy (conﬁrmed at necropsy) (N) 08 08 07 07 Pre-coital interval$ (Days) 4.6 ± 0.43 3.11 ± 0.57 3.67 ± 2.49 4.20 ± 3.31 Gestation length$ (Days) 22 ± 0.43 22.14 ± 0.35 22 ± 0.00 21.14 ± 0.35 Pregnancy 21 days (N) 0 0 0 0 Pregnancy = 22 days (N) 06 06 07 06 Pregnancy 23 days (N) 02 01 0 01 Dams littered (N) 08 07 07 07 Dams dead at gestation (N) 0 01 0 0 Dams dead at Lactation (N) 0 0 0 0 Dams with total litter loss on day 4 of lactation (N) 0 01 01 0 Corpora lutea/Group (mean) 11.63 11.14 10.57 12.00 No. of Implantations/Group (mean) 11.63 10.29 10.86 11.00 Live pups/dam at birth (mean) 10.75 10 9.43 9.71 Live pups/dam at day 4 (mean) 10.13 09 07 9.14 Sex ratio (M:F) at birth 0.85 0.79 0.89 1.31 No. of live pups (M+F) at day 4 81 63 49 64 Pup weight at day 1$ $ (M) 27.8 ± 4.02 25.8 ± 3.04 21.8 ± 4.15 38.7 ± 6.33 Pup weight at day 1$ $ (F) 31.0 ± 1.93 30.7 ± 2.74 23.1 ± 5.32 24.7 ± 3.45 Pup weight at day 4$ $ (M) 36.1 ± 4.9 29.5 ± 5.44 29.4 ± 7.51 40.9 ± 8.00 Pup weight at day 4$ $ (F) 37.2 ± 3.4 37.7 ± 6.11 30.3 ± 8.08 26.7 ± 4.71 No. of pups dead at birth (N) 0 0 02 06 Pre-implantation loss (%) 0 7.69 -2.7 8.33 Post-implantation loss (%) 6.45 2.78 13.16 10.39 Table 10. Summary of effects of 3’-HPT administration on reproduction/development in Wistar rats. Summary of effects of 3’-HPT administration on reproduction/development in Wistar rats. HPT administration on reproduction/development in Wistar rats. PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 11 / 14 Supporting information S1 Fig. Effect of oral administration of 3’-HPT on food consumption (average) in SD rats measured over 90 days. (A) Male SD rats (B) Female SD rats. Data in Mean ± SD (n = 10/ group). (EPS) Acknowledgments We thank Dabur Research Foundation and Bioneeds, India, for conducting reproductive/ developmental toxicity and 90 day repeated oral toxicity studies respectively. 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity period suggested the absence of any adverse effects or dose associated changes on the body weight, food consumption, mean body weight of pups following repeated oral dosing of 3’- HPT. The histopathology of the tissues (genitals) of all high dose (RG4) and control (RG1) groups lacked any abnormalities owing to the exposure to the test item. A mild increase in the mean value of the number of seminiferous tubules of the testes in high dose group (RG4) compared to the controls (RG1) was not a common observation in other dosage groups; infiltration of endometrial tissues in one of the high dose group observed indicated an allergic reaction. However, as there was no change in the absolute organ weight of males (testes and epididymi- des) and females (ovaries and uterus) which indicated that the findings were incidental and not dose dependent, suggesting that the changes were not toxicologically relevant to the test item. The lack of maternal toxicity with no changes in number of pregnancies, number of dams littered, gestation length, number of days of pregnancy, mean number of implantations and corpora lutea being comparable across the groups with no significant differences observed in percent of successful implantations between control and treatment groups was further evi- dent with majority of the females across groups showing normal pregnancy equal to 22 days. Also no changes in the pre-coital interval, percentage of pre and post-implantation loss in all treatment groups showed lack of dose dependent changes. From the results obtained in the present reproductive/developmental toxicity study, 3’-HPT was found to be safe at the doses of 20, 100 & 200 mg/kg bwt in Wistar rats when administered orally. To conclude the study was performed as per OECD guideline 408, 421 respectively in rats and based on the above results, which showed no overt toxic changes in both 90 day and reproductive toxicity study, it could be suggested that a repeated oral exposure to test item 3’-hydroxypterostilbene, up to 200 mg/kg bwt/d was found to be safe and may be treated as ‘No Observed Adverse Effect Level’ (NOAEL) under the test conditions employed. The lower values observed in the Alanine aminotransferase (ALT) and Aspartate aminotransferase (AST) in high dose (NG5) and high dose recovery (NG6), in both male and female groups during treatment were shown to be of non-pathological impor- tance, however, there were no changes observed in the AST and ALT concentrations at the end of recovery period, hence considered to be non adverse as the decrease in serum AST and ALT concentrations has not been shown to be a pathologically important phenomenon which was unlikely to be caused by the oral exposure to 3’-HPT. Necropsy, gross pathology and histo- pathological examination of tissues and organs did not reveal any pathological changes. In treatment and recovery groups, the mean absolute organ weight of male and female rats at nec- ropsy observed for liver, kidneys, brain, spleen, heart, thymus testes/ovaries, epididymides/ uterus, and adrenals, did not show any statistically significant variations in absolute organ weights suggesting that the test item did not have any evident toxicological effects upon administration. The reported results showed that the No-Observed-Adverse-Effect-Level (NOAEL) of 3’-hydroxypterostilbene in SD rats, following repeated 90 day oral route adminis- tration was found to be 200 mg/kg bwt. In reproductive/developmental toxicity study the animal groups of each sex were dosed for a period of two weeks while pre-mating, mating, on the day before sacrifice and in females during gestation and four days of lactation period. Increase in the body weight of male and female rats during pre-mating period and the comparable weights of the female body weights with the control group during the gestation period, the food consumption being comparable among both sexes during pre-mating and in females during gestation to control groups, with the mean body weights of the male and female pups remaining unaffected during lactation 12 / 14 References 1. Ma Z, Li X, Li N, Wang J. Stilbenes from Sphaerophysa salsula. Fitoterapia. 2002; 73(4): 313–5. PMID: 12234574 2. Nagabhushanam K, Thomas SM, Prakash L, Majeed M. Stilbenols-a class of privileged structures. PharmaChem. 2009; 8(10): 2–4. (ISSN:1720-4003). 3. Shao X, Chen X, Badmaev V, Ho CT, Sang S. Structural identification of mouse urinary metabolites of pterostilbene using liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spec- trom. 2010; 24(12): 1770–8. doi: 10.1002/rcm.4579 PMID: 20499322 4. Tolomeo M, Grimaudo S, Cristina AD, Roberti M, Pizzirani D, Meli M, et al. Pterostilbene and 30-hydro- xypterostilbene are effective apoptosis-inducing agents in MDR and BCR-ABL-expressing leukemia cells. Int J Biochem Cell Biol. 2005; 37(8): 1709–26. doi: 10.1016/j.biocel.2005.03.004 PMID: 15878840 5. Roberti M, Pizzirani D, Simoni D, Rondanin R, Baruchello R, Bonora C, et al. Synthesis and biological evaluation of resveratrol and analogues as apoptosis-inducing agents. J Med Chem. 2003; 46(16): 3546–54. doi: 10.1021/jm030785u PMID: 12877593 6. Majeed M, Reinhardt J, Nagabhushanam K, Thomas SM, Jayaraman K, inventors; Sami Labs Limited, assignee. Orally bioavailable stilbenoids—compositions and therapeutic applications thereof. United States patent US 9050284 B2. 2015 Jun 9. 7. Venkateswarlu S, Satyanarayana B, Sureshbabu CV, Subbaraju GV. Synthesis and Antioxidant Activity of 4-[2-(3, 5-Dimethoxyphenyl) ethenyl]-1, 2-benzenediol, a Metabolite of Sphaerophysa salsula. Biosci Biotechnol Biochem. 2003; 67(11): 2463–6. PMID: 14646210 8. Monagas M, Herna´ndez-Ledesma B, Go´mez-Cordove´s C, Bartolome´ B. Commercial dietary ingredi- ents from Vitis vinifera L. leaves and grape skins: Antioxidant and chemical characterization. J Agric Food Chem. 2006; 54(2): 319–27. doi: 10.1021/jf051807j PMID: 16417286 9. Gill MT, Bajaj R, Chang CJ, Nichols DE, McLaughlin JL. 3, 3’, 5’-Tri-O-methylpiceatannol and 4, 3’, 5’- tri-O-methylpiceatannol: improvements over piceatannol in bioactivity. J Na Prod. 1987; 50(1):36–40. 10. Roupe KA, Remsberg CM, Ya´ñez JA, Davies NM. Pharmacometrics of stilbenes: Seguing towards the clinic. Curr Clin Pharmacol. 2006; 1(1):81–101. PMID: 18666380 11. Majeed M, Nagabhushanam K, Ganjihal S, inventors; Sami Labs Limited, assignee. Process for the manufacture of 3’-Hydroxy pterostilbene. United States Patent US 9458075 B1. 2016 Oct 4. 12. OECD. 1998. OECD Guidelines for the Testing of Chemicals, Section 4. Test No. 408: Repeated Dose 90-Day Oral Toxicity Study in Rodents. OECD Publishing. 13. OECD. 1995. OECD Guidelines for the Testing of Chemicals, Section 4. Test No. 421: Reproduction/ Developmental Toxicity Screening Test. OECD Publishing. 14. Takemoto JK, Remsberg CM, Davies NM. Pharmacologic Activities of 3’-Hydroxypterostilbene: Cyto- toxic, Anti-Oxidant, Anti-Adipogenic, Anti-Inflammatory, Histone Deacetylase and Sirtuin 1 inhibitory activity. Author Contributions Conceptualization: MM. Investigation: SB AP. Investigation: SB AP. Resources: SN MM. Validation: SB AP. Visualization: NVS. PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 13 / 14 3’-hydroxypterostilbene: 90 day oral, reproductive/developmental toxicity PLOS ONE \| DOI:10.1371/journal.pone.0172770 March 3, 2017 References J Pharm Pharm Sci. 2015; 18(4): 713–27. PMID: 26626255 14 / 14
https://openalex.org/W2614555109	https://europepmc.org/articles/pmc5437569?pdf=render	English	null	Exploring the role of socioeconomic factors in the development and spread of anti-malarial drug resistance: a qualitative study	Malaria journal	2,017	cc-by	12,026	© The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Exploring the role of socioeconomic factors in the development and spread of anti‑malarial drug resistance: a qualitative study Philip Emeka Anyanwu, John Fulton, Etta Evans and Timothy Paget Anyanwu et al. Malar J (2017) 16:203 DOI 10.1186/s12936-017-1849-1 Anyanwu et al. Malar J (2017) 16:203 DOI 10.1186/s12936-017-1849-1 Malaria Journal Open Access Correspondence: philip.anyanwu@sunderland.ac.uk University of Sunderland, Pasteur Building, Sciences Complex, City Campus, Sunderland SR1 3SD, UK Background Considering the quick and widespread of previous cases of resistance to anti- malarial drugs from the SEA to Africa, the African region stands at risk of spread of artemisinin resistance [22, 23]. Some features of the Nigerian healthcare system that contribute to the persistence of the disease burden in this population (by constituting a major barrier to early and effective treatment) include the lack of universal and equitable access to primary healthcare as this is the level where conditions like malaria infection are managed [8]; and the current government policy of payment for healthcare in all public health facilities at the point of use [8, 9]. It is established that socioeconomic factors are key drivers of health disparities between and within coun- tries. As a socioeconomic issue, the high level of poverty in Nigeria, as with most malaria endemic countries, is an important factor that reinforces the persistent malaria burden in Nigeria. Importantly, the mechanism behind the development and spread of anti-malarial drug resistance is a com- plex one with multiple factors in play—such as drug use practices/behaviours, drug half-life, transmission inten- sity, clone multiplicity, parasite density, host immunity, within-host dynamics and the genetic basis of drug resist- ance [24]. Nevertheless, anti-malarial drug use practices/ behaviours of presumptive treatment, drug overuse, use of sub-therapeutic doses, non-adherence to the treat- ment regime, among others, remain very important driv- ers of drug resistance as they can affect some of the other factors [24]. In addition, unlike most of the other causal factors to anti-malarial drug resistance, drug use prac- tices can be controlled through changes in human behav- iours and activities [25]. In order to encourage changes in human behaviours around the use of anti-malarial drugs, there is need for a good understanding of the factors that play important roles in determining these behaviours, and how they interact with one another in influencing these behaviours. Indeed, a strong relationship exists between poverty and malaria [10]. This relationship is evident in the fact that most malaria endemic countries are also among the poorest countries of the world. Poverty contributes to the malaria burden as it has the ability to affect integral aspects of malaria treatment-seeking behaviours [10], including access to preventive measures and treatment— in relation to affordability, acceptability and availability— [11], and adherence to treatment. Background the use of artemisinin-based combination therapy (ACT) [15]. There has been significant improvement in access, availability and affordability of ACT in malaria endemic regions [14]. g Malaria is a parasitic infection caused by species of Plas- modium parasites [1]. Although a global health issue, its burden is unevenly distributed to the disadvantage of the developing countries of the world [2, 3]. Africa is still the leading region in terms of malaria burden. This region accounts for majority (80%) of the malaria cases as well as malaria-associated deaths (90%) globally [1]; with chil- dren under 5 years of age at most risk of malaria infec- tion. Nigeria, a middle-income country in sub-Saharan Africa, is among the countries with the highest malaria burden in the world, with 32% of global malaria mortal- ity occurring in this country and 97% of the population at risk of malaria infection [4]. Some factors that drive the persistence of malaria in the Nigerian population are: the availability of conducive environment for vector breeding [5]; the Nigerian healthcare system; and the high level of poverty in Nigeria [6], where about 46% of the population live below the World Bank poverty line of $1.25 per day [7]. ACT is presently the World Health Organization (WHO) recommended first-line treatment for uncom- plicated malaria [1]. This recommendation has been adopted by most malaria-endemic countries, like Nige- ria, where artemether-lumefantrine is the first-line treat- ment for uncomplicated malaria cases [4, 16]. In relation to previously used anti-malarial drugs, artemisinin drugs are very effective in parasite clearance and can relief the malaria symptoms faster [17, 18]. However, the develop- ment of resistance to ACT (which has been confirmed in five countries of Southeast Asia (SEA): Cambodia, Lao PDR, Myanmar, Thailand and Viet Nam [1, 19, 20] poses a major threat to its efficacy and use as first-line treat- ment. This also threatens the sustainability of the present success in malaria control [21]. Similar to the artemisinin class drugs, the development of resistance to most of the previously used anti-malarial drugs (such as chloroquine, sulfadoxine-pyrimethamine, mefloquine) originated from South East Asia (SEA). Nev- ertheless, the burden and effects of resistance are usually borne more by the African region that accounts for most of the global malaria cases [1]. Abstract Background: Malaria remains a global health issue with the burden unevenly distributed to the disadvantage of the developing countries of the world. Poverty contributes to the malaria burden as it has the ability to affect integral aspects of malaria control. There have been renewed efforts in the global malaria control, resulting in reductions in the global malaria burden over the last decade. However, the development of resistance to artemisinin-based combi- nation therapy threatens the sustainability of the present success in malaria control. Anti-malarial drug use practices/ behaviours remain very important drivers of drug resistance. This study adopted a social epidemiological stance in exploring the underlying socioeconomic factors that determine drug use behaviours promoting anti-malarial drug resistance. Methods: A qualitative approach, involving the use of interviews, was used in this inquiry to explore the existing anti-malarial drug use practices in the Nigerian population; and the different socioeconomic factors influencing the behaviours. Results: The significant malaria treatment behaviours influenced by socioeconomic factors in this study were the practice of ‘mixing’ drugs for malaria treatment, presumptive treatment, sharing of malaria treatment course, and the use of anti-malaria monotherapies. All the rural dwellers in this study reported they have mixed drugs for malaria treatment. When symptoms were experienced, socio-economic factors, like type of settlement, income level and occupation, tended to determine the treatment behaviour and, therefore, informed and determined the experience of the illness. Discussion: Social and economic contexts can influence behaviours as they contribute in shaping norms and in creating opportunities that promote certain behaviours. As shown in this study, income level and type of settlement, as structural factors, affect the decision on where to seek malaria treatment and whether or not a malaria diagnostic test will be used prior to treatment. One of the dangers of using the mixed anti-malarial drugs is that it offers a safe route for the sale of expired and fake anti-malarial drugs as the mixed drugs are not sold or dispensed in their original packets. Conclusions and recommendations: Population-wide improvements in income, education, environmental and structural conditions of rural dwellers in malaria-endemic settings will encourage behavioural change on how anti- malarial drugs are used. Keywords: Malaria, Drug resistance, Socioeconomic factors, Drug use, Treatment seeking, Behaviour, Antimalarial drugs, Education *Correspondence: philip.anyanwu@sunderland.ac.uk University of Sunderland, Pasteur Building, Sciences Complex, City Campus, Sunderland SR1 3SD, UK Anyanwu et al. Malar J (2017) 16:203 Page 2 of 15 Background Presently, there have been renewed efforts in the global malaria control with several organizations and non-endemic countries increasingly getting involved in the fight against malaria [12, 13]. These efforts have resulted in reduction in the global malaria burden over the last decade [14]. The current achievement is mostly attributed to the increase in malaria research funding, and scale-up of interventions against malaria, including insecticides-treated nets (ITNs), indoor residual spraying (IRS), rapid diagnostic testing (RDT), and importantly, Evidently, most of the studies on anti-malarial drug resistance have been conducted from parasitologi- cal [26–29] and pharmacological [21, 24] perspectives. These studies, while adopting a biomedical approach, Page 3 of 15 Anyanwu et al. Malar J (2017) 16:203 Anyanwu et al. Malar J (2017) 16:203 have mostly looked at the molecular, biological and/or pharmacokinetic aspects of anti-malarial drug resist- ance. Although they have contributed to the fight against resistance to anti-malarial drugs by providing evidence on the mechanisms involved, the persistent develop- ment and spread of resistance to anti-malarial drugs by the Plasmodium parasites despite these efforts calls for additional approaches in studying this important pub- lic health issue. In order to have a wider picture of the issue of anti-malarial drug resistance, there is need for a psychosocial perspective that will explore the underlying causes of human behaviours that promote the develop- ment and spread of resistance. This will help to protect the efficacy of malaria treatments by informing effec- tive behavioural change campaigns. Currently, there is no existing qualitative enquiry that focused specifically on exploring the interactions between socioeconomic factors, anti-malarial drug use behaviours and drug resistance. of the states, making a total of four (4) study sites—Abuja municipal (Urban); Zuba (rural); Owerri municipal (urban) and Ihitte-uboma (Rural). Nigeria is a sub-Saharan African country with a popu- lation of about 177 million people [33], and about 53% of the population living in the rural areas [34]. The coun- try has a land mass of 923,768 square kilometres and a tropical climate with two seasons: the wet and the dry seasons. The annual rainfall ranges between 550 mm in some part of the north (mainly in the fringes of Sahara desert) to 4000 mm (in the coastal region around Niger delta area in the south). The temperature in Nigeria oscil- lates between 25 and 40 °C.h The climatic and geographic features of Nigeria make it a conducive site for high malaria transmission. Background All the states in Nigeria are holoendemic for malaria, and trans- mission is usually year round. However, the wet or rainy season is the peak malaria transmission season in Nige- ria. In addition, the duration of the peak transmission season differs in different locations with the areas in the north having shorter duration of transmission (approxi- mately 3 months) than the south (almost year round) [5]. Plasmodium falciparum is the most prevalent (97%) malaria parasite specie in Nigeria [5]. Consequently, a social epidemiological stance was adopted in this study. As a sub-discipline of epidemiol- ogy, social epidemiology is a perspective on health that is concerned with the connection between social sys- tems and population health [30]. It deals with the social distribution and determinants of states of health [31]. By adopting a psychosocial approach to health, social epidemiologists see health outcomes as not necessarily products of biological factors alone, but also of socio- structural forces acting around individuals [32]. In addi- tion, evidences from existing studies have shown that health-related behaviours are not randomly distributed in nature but are socially and economically patterned [31]. In malaria treatment-seeking, socioeconomic struc- tures can constrain choices of where to seek treatment, affect knowledge level and perception about malaria, and encourage behaviours that might be formed as cop- ing strategies to cushion the effects of these forces. This study explored the underlying socioeconomic factors that determine drug use behaviours promoting anti-malarial drug resistance. Apart from the geographical and climate change dif- ferences between the northern and southern Nigeria; as with most countries in the world, like England, the north–south divide exists in Nigeria with regards to predominant cultures, religion, educational level, type of occupation and level of urbanization. As such, to get a northern and southern perspective of the issue under study, Imo state in the south and Abuja in the north were selected for this study. Other factors, like participants and researchers’ health and safety, were considered in choosing these sites. Study design A qualitative approach, involving the use of interviews, was used in this inquiry to explore the existing anti- malarial drug use practices in the Nigerian population; and the different socioeconomic factors influencing the behaviours. This study is part of a larger study on anti- malarial drug resistance in Nigeria. Fifteen in-depth, face-to-face interviews were conducted with household heads (thirteen), a drug vendor and a pharmacy assistant from the four locations this study covered. Research aimh The aim of this study is to explore the anti-malarial drug use behaviours of people from different socioeconomic backgrounds. Study settingh The inclusion criteria for participants in this study were: This study was conducted in two areas, Imo state and Abuja in the southern and northern parts of Nigeria, respectively (see Fig. 1 showing a map of Nigeria with the study areas highlighted in green). The interviewees were selected from a rural and an urban area from each • • Participants must be at least 18 years of age. • • Participants must be resident in Nigeria. Anyanwu et al. Malar J (2017) 16:203 Page 4 of 15 Fig. 1 Map of Nigeria with study areas highlighted in green Fig. 1 Map of Nigeria with study areas highlighted in green • • Participants must be household heads who bear the financial cost of malaria treatment for themselves and members of their household. The reason for this is to ensure that those who can provide accurate information on the financial cost of malaria treat- ment to their household are recruited. This enabled the researcher to probe or elicit further explanations on some behaviours or experiences identi- fied in previous interviews. This strategy was also impor- tant in determining when saturation was achieved in this study. In all, fifteen (15) individual in-depth interviews were conducted with fifteen participants. i Prior to the conduct of each interview, the interviewer made a brief first contact with each participant as an ice- breaker strategy to improve acceptance, trust and com- munication. It was during these first meetings that the participant information sheet was given to the partici- pants and verbally explained. Also, during this contact, the researcher made some observations like the willing- ness or reluctance of the individuals to participate in the study with the aim of excluding those who show reluc- tance and are uncomfortable all through the meeting. All individuals approached agreed to this first meeting and were included in this study afterwards as they all met these criteria. Subsequently, the interviewer and the par- ticipants agreed on a convenient day, time and venue for the interview. • • In addition to the above, a drug vendor and a phar- macy assistant from the study sites were interviewed to help interpret the findings from the participants. To ensure that participants who can recall and pro- vide relevant information to answer the research ques- tion were selected, a systematic theoretical sampling technique [35] was used in selecting participants for this study. Study settingh This technique involved continuous sampling and data collection until saturation was achieved; that is, the point when no new information or category, in relation to the issue under study, was emerging from subsequent interviews [36]. Consequently, the data collection and analysis were simultaneously conducted with each inter- view transcribed and analysed prior to the conduct of the next interview. With this strategy, emerging codes were considered and incorporated into subsequent interviews. A written informed consent was obtained from par- ticipants prior to the interview. In addition to this, a confirmatory verbal informed consent was obtained Anyanwu et al. Malar J (2017) 16:203 Page 5 of 15 The thematic analysis method, as outlined by Braun and Clarke [37] was used in analysing the interview data. This involved firstly, a comprehensive reading and re- reading of all the data set (transcript). In the second step, relevant pieces and expressions from the transcript were labelled or coded. Coding, in this study, was conducted at a descriptive level with the aim of summarizing the main concept of each statement from the interview thereby retaining the information from the statement. The third step involved further review of the codes for patterns and linking related codes together to form categories. The grouping into categories was based on not just exact or close similarities; but codes that had something in com- mon or are linked in some ways, that might not seem so at the face value, were also grouped together as a cate- gory [38]. The final step was about developing key themes from the categories. The themes were developed around reported practices and behaviours in seeking malaria treatment that were connected. To achieve this, the cat- egories were considered for relevance and how they con- nect and interact with one another. Categories that were similar or informed others were combined to build up a key theme. For the meaning of some key terms used in this study, see Table 1. and audio recorded at the beginning of each interview. All participants consented to the interview being audio recorded.i Confidentiality and anonymity were assured to all par- ticipants. These were maintained all through the study by assigning identification codes to each participant and by storing the interview data securely in a password pro- tected storage device. Socio‑demographic description of participantsh The socio-demographic characteristics of the partici- pants are shown in Table 2. Overall, 53.3% of the par- ticipants were females. The mean age of the participants was 40 years. The most common occupation of the par- ticipants was farming. Almost half (46.7%) of the par- ticipants reported they earn below the current Nigerian minimum wage of eighteen thousand Naira (₦18,000) (about £64) per month. Seventy three percent (73%) of the participants had at least some secondary education. Urban dwellers constituted 53.3% of the participants, see Table 2. Study settingh Prior to the commencement of each interview, participants were informed of their right to partake or not, and to withdraw or ask the interviewer to stop the interview at any point. Participants were also given the contact detail of the researchers and the Uni- versity of Sunderland Research Ethics Committee for any subsequent concerns. A 1-week period from the day of the interview was given for the participants to contact the researchers if they decide to withdraw their interview account from the study. The interviews were conducted by the first author whose familiarity of the study population and previous research experiences in Nigeria were very important in the recruitment and conduct of the interviews. The first author, having lived in Nigeria for more than 20 years, had a grounded knowledge and understanding of the local languages and culture of the study population. Furthermore, the interview questions were centred on participants’ experiences and behaviours in using anti- malarial drugs. Prior to the main interview, the interview schedule was piloted with two participants of similar characteristics with the main participants. The inter- views were held at community venues, which partici- pants selected and were already familiar with. In all, the interview time was between 30–35 min. Interviews were conducted in English and two Nigerian languages (Hausa and Igbo languages). Socio-demographic data were col- lected using a brief questionnaire administered to each participant at the end of the interview. Issues of validity and reliability All the interview records were transcribed verbatim from audio to a written form by the interviewer on the same day they were conducted. This helped to ensure that the memory and observations during the interviews were retained. Where necessary, the interview transcripts were translated from local language to English. Lincoln and Guba [39] argued that approaches to the validity and reliability of quantitative research were not appropriate for qualitative studies and they suggested credibility, transferability, dependability, and confirmabil- ity as more appropriate criteria. Credibly is ensuring that Table 1 Key terms used and their meaning in the context of this study Terms Meanings Pharmacy Used to refer to the registered pharmacy shops that are usually found in the cities. These shops are usually owned or run by a trained pharmacist. Some pharmacies in Nigeria are however run by pharmacy attendants or nurses Chemist/drug vendors Used to refer to drug vendors or patent medicine vendors in the study setting. These are usually traders with little or no pharmaceutical trainings who either hawk or sell drugs in a small shop. They are mostly common in the rural and suburban areas in Nigeria. The term chemist and drug vendors were used interchangeably in this study Table 1 Key terms used and their meaning in the context of this study Terms Meanings Pharmacy Used to refer to the registered pharmacy shops that are usually found in the cities. These shops are usually owned or run by a trained pharmacist. Some pharmacies in Nigeria are however run by pharmacy attendants or nurses Chemist/drug vendors Used to refer to drug vendors or patent medicine vendors in the study setting. These are usually traders with little or no pharmaceutical trainings who either hawk or sell drugs in a small shop. They are mostly common in the rural and suburban areas in Nigeria. The term chemist and drug vendors were used interchangeably in this study Table 1 Key terms used and their meaning in the context of this study Used to refer to the registered pharmacy shops that are usually found in the cities. These shops are usually owned or run by trained pharmacist. Some pharmacies in Nigeria are however run by pharmacy attendants or nurses Page 6 of 15 Page 6 of 15 Anyanwu et al. Malar J (2017) 16:203 Table 2 Socio-demographic description of participants Characteristics Number (%) Total no. of participants 15 (100) Household heads [13] Drug vendor [1] Pharmacy attendant [1] Sex Males 7 (46.7) Females 8 (53.3) Age Mean age 40 Mode age 31 Relationship status Single 3 (20) Married 6 (40) Widowed 6 (40) Educational level No formal education 1 (6.7) Some primary school education 3 (20) Some secondary school education 2 (13.3) O’level/SSCE holder 2 (13.3) Tertiary education/degree 4 (26.7) Post graduate degree 3 (20) Income level Below ₦18,000 (£63) 7 (46.7) ₦18,000 to ₦50,000 (£64 to £177) 1 (6.7) ₦50,001 to ₦100,000 (£178 to £355) 2 (13.3) ₦100,001 to ₦300,000 (£356 to £1065) 3 (20) Above ₦300,000 (£1065) 2 (13.3) Type of settlement Urban 7 (46.7) Owerri in Imo state [2] Abuja Municipal city [5] Rural 8 (53.3) Nkwumeato community in Imo state [4] Zuba community in Abuja [4] Table 2 Socio-demographic description of participants summarizing their accounts during the interview and asking them for a validation of this) and from other par- ticipants (through constant comparative method to vali- date the experiences of other participants). Furthermore, the interviewing researcher’s familiarity with the study setting was very important in establishing a good com- munication with the participants, and understanding the social context of the study population—language, culture, social structure, and healthcare system. y Dependability refers to the internal consistency. It is recognized that in qualitative research, the theoreti- cal position of the researcher can affect the direction of the analysis. This factor is recognized, but this study was an exploration of relatively concrete behaviour and the dependability largely was achieved through the quality of the data analysis. Thematic analysis [37] was used as the method of analysis and a clear process was followed, open to scrutiny. The process of analysis was carried out by one of the researchers and then reviewed by the other researchers. Table 1 Key terms used and their meaning in the context of this study All researchers were involved in the discus- sions and decisions in the analysis of the data—for exam- ple, in deciding which categories constitute a key theme from the analysis.i Indeed, confirmability, which refers to the objectivity of the study, can be complex in qualitative research as the researcher will bring his or her particular theoretical position. The key issue here is whether the data analysis was logical and consistent, and if the researchers were transparent in all processes; this was a guiding principle in the study design. Transferability equates to external validity, and again in qualitative research this can be dif- ficult as samples are small and often not representative. Generalization is about: can the findings be transferred to another setting or is the presentation of the data clear enough to allow the reader to do this? The sample was small and examined inhabitants of only two Nigerian areas, however it did identify strategies and behaviours which allow comparison and in this sense transferability was achieved. the findings of a qualitative study are credible and actu- ally reflect the situation under investigation. Merriam and Tisdell [40] posed the question “How congruent are the findings with reality?”. This was addressed through the use of the respondent’s own words to back up claims and by ensuring the outline of the categories was reflective of the interview transcripts. Shenton [41] also suggests that a wider range of respondents can add to the credibility of the study. Respondents were chosen from a range of socio-economic groups and from rural and urban set- tings. A range of views was, therefore, represented and differences and similarities among the respondents were highlighted in the report. Another approach was vali- dation of the data from the participants themselves (by the findings of a qualitative study are credible and actu- ally reflect the situation under investigation. Merriam and Tisdell [40] posed the question “How congruent are the findings with reality?”. This was addressed through the use of the respondent’s own words to back up claims and by ensuring the outline of the categories was reflective of the interview transcripts. Shenton [41] also suggests that a wider range of respondents can add to the credibility of the study. Respondents were chosen from a range of socio-economic groups and from rural and urban set- tings. Table 1 Key terms used and their meaning in the context of this study A range of views was, therefore, represented and differences and similarities among the respondents were highlighted in the report. Another approach was vali- dation of the data from the participants themselves (by Resultsh The social structures existing within societies and their socioeconomic indicators are usually based on charac- teristics that are valued in the society—educational level, income level, type of settlement, occupation, employ- ment status, among others. As indicators of socio-struc- tural position within the society, an individual’s level of socioeconomic measures can determine the likelihood of healthy or unhealthy exposures, behaviours and out- comes [42]. Several key themes emerged from the analysis of the interview data. The themes were developed around the anti-malarial drug use behaviours reported, and their Anyanwu et al. Malar J (2017) 16:203 Page 7 of 15 Page 7 of 15 associated socioeconomic factors. The socioeconomic factors identified in this study—income level, type of settlement, educational level and occupational/job type- are all related to one another with some (for example, educational and income levels) acting as determinants to the others. The significant malaria treatment behav- iours influenced by socioeconomic factors in this study were ‘mixing’, presumptive treatment of malaria, the use of drug vendors, sharing of drugs, and the use of anti- malarial monotherapies.h of the patient like socioeconomic factors (measures of educational level, level of income and wealth, type of settlement and occupational type) as well as socio-cul- tural factors (like beliefs, norms, and constructions). The structures in this study were identified from the participants’ interview accounts as well as the socio- demographic questionnaires completed by all the participants. These structural factors influence the ‘process’. The ‘process’ represents practices and behaviours of the participants in treating malaria, which are prod- ucts of the interaction with the structures. This includes reported behaviours like drug ‘mixing’ for malaria treat- ment, presumptive treatment, non-adherence to treat- ment guideline, use of monotherapies, among others. These behaviours were influenced by the structural fac- tors, which have the ability to limit people’s choices, how they seek malaria treatment and how they use anti- malarial drugs. The Donabedian model, which was initially developed to serve as an evaluative framework for quality of health- care [43], has been widely used in studying other aspects of healthcare [44]. It assumes three main components in studying health outcomes: Structure, Process and Outcomes [45–47]. These three domains work in uni- directional way with each domain influencing the next. This was an appropriate way of organising the categories established through the analysis of the interviews. In applying the Donabedian model to this study (see Fig. Knowledge and perception on malaria A key issue was the level of knowledge of malaria, which influences preventive and treatment seeking behaviours. Some of the perceptions about cause of malaria dem- onstrated by the participants (like the perception that consumption of fatty diets as well as spiritual attacks as a cause of malaria) were products of social construc- tion from the different communities. All the participants identified the link between mosquitoes and malaria and they also were aware of some indirect causal factors, such as: dirty environment, living in dilapidated hous- ing condition, stagnant waters around the house, among others. However, these indirect causal factors were perceived as direct causes of malaria by many of the participants. On experiencing malaria like symptoms, there are several facilities one can seek treatment in Nigeria: hos- pitals/clinics, health centres, chemists/drug vendors, pharmacies and herbal healers. These facilities usually offer different treatment options ranging from a labo- ratory diagnostic test prior, to presumptive treatment based on presenting symptoms. Those living in urban areas tended to use the pharmacy or hospital as the first point of contact whereas in the rural areas pharmacies and hospitals were less common as “chemists” were eas- ier to access hence more used. With a large proportion (53%) of the Nigerian population living in rural areas, the chemists/drug vendors serves as the first point of contact for malaria treatment for many. The pharma- cies provide professional and more expensive services whereas the drug vendors often lack the knowledge and skill which was reflected on both advice and cost of ser- vices in these facilities. The chemist/drug vendor was the most reported first point of seeking malaria treat- ment among participants whose monthly income was below the ₦18,000 (£63) minimum wage. Even when other sources were closely located, the chemists/drug vendors remained the most commonly used source for ‘What I know about malaria is that, some people say taking too much of oil -oil is one of the causes of malaria, palm oil. Another thing I know that causes it is like, for example, some people might be asked not to eat something like ‘egusi’, ‘egusi’ causes malaria too’ (Female, 49 years old, rural area). As malaria is a common disease, which all the partici- pants had experienced, knowledge of malaria symptoms was satisfactory among the participants. Malaria treatment‑seeking behaviours When symptoms were experienced, socio-economic factors, like type of settlement, income level and type of job, tended to determine the treatment behaviours and, therefore, informed and determined the experience of the illness. This was in terms of first point of contact for treatment and choice of anti-malarial drugs; as these involved money, socioeconomic status was a deter- minant. Conversely, other behaviours, like presump- tive treatment, were reported across all socioeconomic groups in this study. Socioeconomic factors As earlier stated, socioeconomic factors like income level, type of settlement and educational levels constitute structures that can enhance or endanger a population’s health. These factors were reported to affect the way malaria treatment is sought and how anti-malarial drugs are used. For instance, educational level affects one’s abil- ity to read and understand written instructions on how to administer anti-malarial drugs. At one end of the continuum many of the participants saw malaria infection as life threatening; however, others saw it as a very common disease which is easy to treat hence not a life-threatening or serious infection. ‘Well, maybe because I am a Nigerian, I don’t see it as a life threatening disease. But I know that eh, when you travel abroad, they don’t want to even hear that you have malaria. Malaria to us or to me personally, is like having headache. Am serious about that’ (Female, 40 years old, urban area). ‘Yes. But even if they write it, did I go to school? What I know is that way they have shown me to take it, that’s the way I follow’ (Female, rural area). ‘Let me say the poor condition that some of us we are living in because we don’t have access to eh this eh hospitals –the so called hospitals they are talking about because we don’t have money’ (Male, 32 years old, rural area). Structure ‘Some people say it’s a spiritual attack. You under- stand. Because it was very mysterious -my mouth was very bitter but the symptoms were malaria symptoms. I don’t know if anything passed through the malaria’ (Female, 33 years old, urban area). ‘Some people say it’s a spiritual attack. You under- stand. Because it was very mysterious -my mouth was very bitter but the symptoms were malaria symptoms. I don’t know if anything passed through the malaria’ (Female, 33 years old, urban area). Resultsh 2 of an adapted Donabedian model showing study findings), the ‘structure’ represents all the fac- tors that affect the context in which malarial treatment is sought. These factors control how patients and pro- viders act in seeking and providing malaria treatment respectively. This includes background characteristics Importantly, the process subsequently determines the ‘outcome’ which can be in the form of treatment fail- ure, drug resistance, effective treatment or death. Most of the behaviours identified through the interviews that constitute the ‘process’ have the tendency to encourage the development and/or spread of anti-malarial drug resistance. Death Fig. 2 Adapted Donabedian model showing study findings Fig. 2 Adapted Donabedian model showing study findings Fig. 2 Adapted Donabedian model showing study findings Anyanwu et al. Malar J (2017) 16:203 Page 8 of 15 Page 8 of 15 Anyanwu et al. Malar J (2017) 16:203 Knowledge and perception on malaria I may –I might have had it, but I may not know it was malaria –I think I have told you that I have never gone for any malaria lab test to test me, to justify what is really my prob- lem’ (Male, 32 years old, rural area). ‘… hospital is not far from me, from where I leave… why I don’t go there is that those two chemists, they ‘help’ me…. That’s why I go to theirs. So if I don’t have money they help me. If it’s the government hospital, everybody comes from different places to work. They will say pay us before we give you the drugs’ (Female, 31 years old, rural area). ‘… hospital is not far from me, from where I leave… why I don’t go there is that those two chemists, they ‘help’ me…. That’s why I go to theirs. So if I don’t have money they help me. If it’s the government hospital, everybody comes from different places to work. They will say pay us before we give you the drugs’ (Female, 31 years old, rural area). Another reported behaviour with regards to presump- tive treatment and the use of malaria diagnostic test prior to treatment was the practice of using previous results of malaria diagnostic test and the prescription received to treat subsequent episodes with similar symptoms. This was reported by two participants from the rural areas. The participants were fully aware of the difference in standard between the pharmacies and the chemists, but both availability and low cost determined the use of the chemists. ‘… sometimes, I go to the health centre. There was a time I went there and they ran a test and said its malaria and typhoid. They treated and gave me some drugs and also directed me to nearby phar- macy to buy more drugs and add to the ones they gave me and take them all. After the treatment, when next I have malaria, I go to the pharmacy and buy the same drugs. Its not all the times that I have the money to run a test’ (Male, 55 years old, rural area). Knowledge and perception on malaria ‘The way we look at chemist in Nigeria is that chem- ist is meant for, you know, people that are in the vil- lages, you know that cannot afford (laughs) because in that chemist you see all sort of, manner of drugs… you see fake drugs there’ (Female, 44 years old, urban area). Furthermore, when symptoms were severe or when seeking malaria treatment for children, most of the par- ticipants would use hospital or pharmacies as the first point of treatment. In all, only one participant said she always used a diag- nostic test to confirm she has malaria prior to her use of anti-malarial drugs. The other participants who also use laboratory diagnostic test prior to malaria treatment were not consistent as they also reported they practiced presumptive treatment. These were mostly participants who earned more than ₦100,000 per month. For the high-income earners in this study who practice presump- tive treatment, affordability was not reported as a reason for this behaviour. The main reasons for this practice include: confident they can diagnose malaria and the per- ception of presumptive treatment as time-saving route to malaria treatment.f Knowledge and perception on malaria There were, however, some cultural and superstitious constructs and explanations for malaria infection; three participants said they believed some evil spiritual powers can cause malaria. This perception was reported in relation to the cause of malaria as well as the manifestation of the disease: Anyanwu et al. Malar J (2017) 16:203 Page 9 of 15 treatment cost, lack of access to hospitals and cultural factors. These three participants were rural dwellers and earned below the minimum wage of ₦18,000 (£63) per month. malaria treatment by those of low socioeconomic sta- tus. The seeming trust in the chemist/drug vendors by the rural dwellers was not based on their ability to offer more effective treatment than the hospitals, health centre or pharmacies. Rather, it was because, unlike the other sources, the chemist/drug vendors are more available and accessible through the provision of credit facility, cheaper alternatives, and convenient locations and out-of-hours services. malaria treatment by those of low socioeconomic sta- tus. The seeming trust in the chemist/drug vendors by the rural dwellers was not based on their ability to offer more effective treatment than the hospitals, health centre or pharmacies. Rather, it was because, unlike the other sources, the chemist/drug vendors are more available and accessible through the provision of credit facility, cheaper alternatives, and convenient locations and out-of-hours services. ‘…I don’t know whether some of the sicknesses I expe- rience sometimes are malaria. I may –I might have had it, but I may not know it was malaria –I think I have told you that I have never gone for any malaria lab test to test me, to justify what is really my prob- lem’ (Male, 32 years old, rural area). ‘…I don’t know whether some of the sicknesses I expe- rience sometimes are malaria. I may –I might have had it, but I may not know it was malaria –I think I have told you that I have never gone for any malaria lab test to test me, to justify what is really my prob- lem’ (Male, 32 years old, rural area). ‘…I don’t know whether some of the sicknesses I expe- rience sometimes are malaria. Presumptive diagnosis and treatment of malaria Presumptive treatment of malaria, which involves treat- ing suspected malaria cases based on presenting symp- toms rather than a confirmatory diagnostic test [48], was the commonest approach and was found in all socioeco- nomic groups. Those who mostly use the hospital for treatment, who were also the highest earners with the highest educational levels, reported the use of labora- tory testing to confirm diagnosis more than those who use the chemist/drug vendors or pharmacies. The phar- macies tended to implement treatment on the basis of presumptive diagnosis; and chemist/drug vendors almost entirely implemented treatment on the basis of presump- tive diagnosis. Similarly, affordability was not always the key issue in decisions on the type of anti-malarial drug to get among different socioeconomic groups. The pharmacy attendant reported that most of the customers at the urban area are not really concerned about price but on the effectiveness of the drug, hence ACT was reported as the most sold class of anti-malarial drugs by the pharmacy attendant. This was however different in the rural areas and the low socio-economic groups; as would be expected, afford- ability was an important issue for these groups. The cost of ACT was reported as out of reach by many rural and urban dwellers who earned below the national minimum An example of this is with the three participants who reported they have never gone for a malaria diagnos- tic test before (neither microscopy laboratory nor RDT) as they all used the chemist for most malaria treat- ments. The most common reasons given by these par- ticipants for this practice were their inability to afford the Anyanwu et al. Malar J (2017) 16:203 Page 10 of 15 Page 10 of 15 ‘When I buy and give him and he takes and gets bet- ter and says he will not take anymore, I throw the rest away. Because when you keep it, do you know the day the sickness will come? By the time the sickness come, and u give it to him, it might bring another sickness for him because you don’t know if the drug has expired’ (Female, 31 years old, rural area). wage. As a result, monotherapies and non-ACT, like SP (which are cheaper and less effective in parasite clear- ance), were reportedly the most used anti-malarial drugs by low-income earners and rural dwellers. Use of monotherapies Generally, the knowledge of the names of anti-malarial drugs used was poor among the participants. Nevertheless, most of those who could identify the names of the anti- malarial drugs they use reported artesunate as the most used anti-malarial drug, (five from the rural area reported using this drug compared to two from urban sites). Addi- tionally, the drug vendor and pharmacy shop assistant that were interviewed also reported the sale of monotherapies (like artesunate) and non-artemisinin based combinations (like sulfadoxine-pyrimethamine (SP). The drug vendor in the rural area described a higher rate of monotherapy sale than the pharmacy attendant in the urban area. The major reason given for this by the drug vendor was the inability of the customers in the rural areas to afford the price for the artemisinin-based combination therapy, which is the first-line treatment for malaria in Nigeria. ‘There are times they will give me and I will take it and take it and take it, if the thing (malaria) clears, I will throw away the remaining. Even if its injec- tion, I will not go to complete it. Because the thing (malaria) has cleared’ (Male, 55 years old, rural area). ‘Yes. If they give me drugs, I finish them. I don’t leave them because I used my money to buy them. Any drug they give me in the hospital I must take all of them. I don’t miss it. I don’t miss it as -are my not the one that will bear the pains? How can I miss it? If I don’t take it I will feel the sign in my body of course’ (Female, 35 years old, rural area). Presumptive diagnosis and treatment of malaria ‘Like if severe, most of the people do not have enough money to treat themselves. They may tell you they need drugs of two hundred naira or one hundred and fifty naira. At times I have to use Laridox (sulf- adoxine-pyrimethamine)’ (Male Drug Vendor, rural area). Sharing of anti-malarial course among two or more people was also reported. One of the three participants who reported this behaviour was of high income (above ₦300,000 (£1065) per month) and high educational level (postgraduate education). The difference between this participant’s behaviour and that of the other two (of income level below ₦18,000 (£63) and some primary school education) is that for the former, it was a one-off act because she could not buy a new course for a member of her household who got ill in the middle of the night; while for the later, it was reported as a recurrent behav- iour. In most cases of sharing anti-malarial drugs with others, all parties involved end up not getting a complete treatment course. Adherence to treatment guideline As an important factor in the development of drug resist- ance, adherence (in terms of completion of treatment course and administration of the proper dose at the cor- rect time, in accordance with the WHO malaria treat- ment guidelines [49] was considered in this study. Many participants reported non-adherence at least once in the last 6 months. Generally, reports of poor adherence were more common among participants of lower level of education and income level of below ₦50,000 (£177). However, some low-income participants, who reported a moderate level of adherence, saw anti-malarial drugs as expensive, so when they do buy them, they tend to complete the treatment as they do not want to waste their (expensive) treatment, and so wants value for their money. Mixingh Participants’ experiences of the outcomes were differ- ent depending on reported practices and behaviours like the type of anti-malarial drug used, adherence to rec- ommendation on how to administer the drugs, amongst others. On their last malaria treatment experience, treat- ment failure was reported by three participants who used the mixed anti-malarial drugs. Only one of the partici- pants who used a formal health facility during her last malaria episode reported treatment failure. ‘No I haven’t taken those packaged as complete course. The ones they mix, that’s the ones they buy for me. That’s the one our hand can get to. We only buy the ones we can afford’ (Female, age not known, rural area).i ‘If you have up to five hundred naira or more, we can even mix Artesunate or Paracetamol or any analges- tic, Panadol, Amoxil or Septrin’ (Male Drug Vendor, from rural area). ‘If you have up to five hundred naira or more, we can even mix Artesunate or Paracetamol or any analges- tic, Panadol, Amoxil or Septrin’ (Male Drug Vendor, from rural area). “… you just go to the chemist. Because the way I used to even treat my own, when it happens like that, there’s one chemist that is near our house, I will meet the woman and she will just mix drugs and give me. When she gives me I will just give them. They will take. Their body will be ok. When it is ok for like two days, before three days again, they will still com- plain again. Before you know it, you carry they to the hospital” (Female, 31 years old, rural area). The anti-malarial drugs bought from the pharmacies were mostly sold as a complete course in their original packets. As the mixed drugs are sold in units of ‘one mixed dose’, a very important factor in determining how many doses of the mixture a customer gets is the amount they can afford. The customers can decide to buy a day’s dose or 2 days’ dose and so on. While the complete course of anti- malarial drugs like artemether-lumefantrine was reported to be about ₦800 (£2.84); ₦100 (£0.35 p) will get you a day dose of mixed anti-malarial drugs if you cannot afford ₦800. Mixingh The concept and practice of ‘mixing’ was first mentioned in the initial interviews with participants from Zuba (rural area) in Abuja, and then subsequently in most of the interviews from other rural areas. To clearly under- stand the practice of mixing and how it is done, a drug vendor from one of the communities in the rural areas and a pharmacy attendant from an urban area were interviewed. They described mixing as the practice of combining different drugs to treat malaria. It is done by including, for example, a dose from each of the included While ideas of cost effectiveness encouraged adher- ence in some low income earners, less commonly (mentioned only by two participants), it contributed to some non-adherence practices like stopping the treat- ment to save the drug for future use. The practice of stopping to save drugs for future use was not common as most of the participants were wary of this practice as it was seen as risky and there was a fear of taking expired drugs. Anyanwu et al. Malar J (2017) 16:203 Page 11 of 15 Page 11 of 15 an anti-malarial course; resistance to anti-malarial drugs by the Plasmodium parasite; or death. drugs to form a mixed dose. There is no standard in terms of the number or types of drugs used in the mix- ture for malaria treatment. Some participants in this study discussed the outcomes of their malaria treatment experiences. Through the interview data, outcomes like death could not be explored since most participants were interviewed based on their own malaria treatment seeking behaviour. Although drug resistance was not directly measured/reported in this study, it is worth mentioning that malaria treatment fail- ure can be an indicator of drug resistance [24].f ‘Almost all the people here normally mix drugs. They may come and say I need a malaria drug of fifty naira or hundred naira. … I don’t have much money to buy a complete one’ (Male Drug Vendor, from rural area). Mixing was described as a very handy and flexible cop- ing strategy by the low income earners. Also, all the par- ticipants from the rural areas reported they mixed. The highest level of education attained by most of those who mixed was some primary education. Mixingh Another important issue on the ‘mixing’ practice is that these drugs are not sold in their original packets as they are either cut up into bits or removed from the blister pack while being dispensed. Hence the customers/patients have little or no information about the individual drugs that constitute the mixture nor their expiry dates. “The difficulty I encounter sometimes is that there are sometimes that malaria will hold me down, and if I go and tell the chemists that malaria is disturb- ing me, they will mix medicine for me and tell me how I will take them. Even after taking them, it will be like I didn’t take any medicine” (Male, 55 years old, rural area). Nevertheless, on their last malaria episode, recovery or effective treatment was mostly reported in line with the practices of seeking treatment from a formal health facility and adhering to the recommended dose and com- pleting the course. As the formal health facilities was reportedly used last time by mostly those from urban areas, many of those who reported effective treatment, without prior failure, were urban dwellers. However, one participant from rural area who used a hospital last time reported an outcome of effective treatment with no prior treatment failure. ‘The last time I used anti-malarial drugs, I bought it from the chemist. They mixed the drugs for me. They mixed that of two days. … how they mix is, it depends on how many days’ dose you want them to mix for you’ (Male, 55 years old, rural area). Discussion Th l The results of this study indicate that socioeconomic factors play important role in determining anti-malar- ial drug use behaviours that promote drug resistance. To explicitly interpret the data on the malaria treat- ment experiences and behaviours of the participants, and to create a holistic insight on how these contribute to anti-malarial drug resistance, the Donabedian model was adapted and used in this inquiry (see Fig. 2). The structure, mainly but not exclusively socioeconomic fac- tors, determined the treatment seeking and drug use behaviours (process), which could in turn determine the outcome. The public health danger of the sale of counterfeited drugs is that some of them contain inadequate quanti- ties of the active ingredient which can encourage the development and spread of drug resistance by exposing the Plasmodium parasites to sub-therapeutic doses [58]. This creates a serious threat to vulnerable populations and jeopardises the progress and investments in malaria control [21]. Modelling analyses have also demonstrated that administration of under-dose of artemisinin plays an important part in the spread of artemisinin resistance [59]. Evidently, social and economic contexts can influence behaviours as they contribute in shaping norms and in cre- ating opportunities that promote certain behaviours [31]. Indeed, physical, social and economic environments can place constraints on an individual’s choices as well as that of a population. As shown in this study, income level and type of settlement, as structural factors, affect the decision on where to seek malaria treatment and whether or not a malaria diagnostic test will be used prior to treatment. In line with existing malaria studies in the Nigerian popula- tion [8, 50, 51], the chemist/drug vendors and pharmacy were the major sources of malaria treatment. While some level of good practice and quality of care exists in the phar- macies which are usually owned or managed by a trained pharmacist or health practitioner, the quality of care and practice in the chemists, which are owned and managed mostly by traders with little or no pharmaceutical train- ing, remains a big public health concern in Nigeria. The chemists, which are the cheapest and most used source for malaria treatment in Nigeria are also the most likely place to get counterfeit anti-malarial drugs [52, 53]. In addition, the practice of mixing will have similar outcome in terms of encouraging the development and spread of anti-malarial drug resistance. Outcome As earlier stated there are four likely outcomes from the reported practices and behaviours indicated under the ‘process’. These include recovery—which can be an indi- cator of an effective malaria treatment; treatment fail- ure—in the form of lack of response or worsening of the clinical manifestation of the infection after administering “so I rushed to meet that doctor… he said its malaria. … he started giving me treatment. I didn’t even go to the farm for one week. I stayed home and took all the drugs when I took all the treatments I got myself” (Female, 31 years old, rural area). Anyanwu et al. Malar J (2017) 16:203 Anyanwu et al. Malar J (2017) 16:203 Page 12 of 15 Page 12 of 15 As there are no detailed anti-malarial drug resist- ance surveillance programmes in Nigeria, some of the reported cases of treatment failure, which can be drug resistance cases, go unreported and not recorded or fol- lowed up. The lack of adequate care and surveillance for this potential cases of drug resistance can encourage further spread of resistant Plasmodium parasites in the population. The study by Onwujekwe et al. [53] on the quality of anti-malarial drugs provided by public and private healthcare providers in Southeast Nigeria found that 37% of the anti-malarial drugs tested did not meet the United States Pharmacopoeia (USP) specifications for the amount of active ingredients as they either lacked the active drug ingredient or contained sub-therapeu- tic quantities of it. Populations, like the Pailin province of Cambodia in SEA where artemisinin resistance have been confirmed have some similarity with the Nigerian population in terms of the existence of widely available cheap counterfeit anti-malarial drugs [57].h Discussion Th l With mixing, the patients can buy less or more number of doses than recommended. One of the dangers of using the mixed anti-malarial drugs is that it offers a safe route for the sale of expired and fake anti-malarial drugs as the mixed drugs are not sold or dispensed in their original packets. Another danger from this practice is the possibility that the interaction of the different drugs used in the mixture can alter the efficacy of these drugs when administered. fi Furthermore, the low prevalence in this study, of some practices that are known to encourage the development of drug resistance, like sharing of malaria treatment course, stopping treatment to save drugs for future use, among others [60, 61], was not necessarily because of the main dangers associated with these practices. For instance, the practice of stopping treatment to save drugs was low among the participants because those more likely to do this (people of low socioeconomic status) used mostly mixed drugs for malaria treatment as they usually do not know the expiry date of the anti-malarial drugs. Hence, they expressed their fear that saving the drugs might mean taking expired anti-malarial drugs in the future. Also, it is possible that this practice of stop- ping treatment to save drugs for future use, which pro- motes the use of incomplete treatment course, could have been higher than reported if the participants had knowledge of the expiry dates of the anti-malarial drugs, especially those that use the mixed anti-malarial drugs. Previous studies have shown that the cost of the origi- nal drugs will influence the circulation of counterfeit anti-malarial drugs [54–56]. These counterfeit anti- malarial drugs are usually sold cheaper than the origi- nal ones; hence those in the low socioeconomic gradient constitute the target market for the counterfeiters. As long as ACT remains unaffordable to most Nigerians, the chemist will remain the main source of malaria treat- ment, especially for the poor, and the circulation of fake anti-malarial drugs will flourish. Anyanwu et al. Malar J (2017) 16:203 Page 13 of 15 One of the behaviours that cut across different levels of socioeconomic measures in this study was presumptive treatment. Presumptive treatment has been identified as a major drive to drug over-use which promotes the devel- opment and spread of resistance to anti-malarial drugs by the Plasmodium parasites [2]. Ethics approval and consent This study was approved by the University of Sunderland Research Ethics Committee. Prior to each interview, a written informed consent was obtained from participants. Discussion Th l Over-use of anti-malarial drug has been reported to increase the incident of altering within-host ecology of drug resistant parasites; and remov- ing the drug-sensitive competitors has a significant impact on transmission potentials of drug resistant parasites [24]. are all coping strategies by the less well-off who cannot afford the recommended route to malaria treatment by the WHO. To achieve a sustainable improvement in health, cam- paigns should be targeted at social structures in the soci- ety that produce diseases [31]. In malaria campaigns, there is need to broaden the scope of anti-malarial drug resistance control strategies to include strategies tar- geted at improving the socioeconomic status of people in malaria endemic areas. Population-wide improvements in income, education, environmental and structural con- ditions of rural dwellers in malaria endemic settings, like Nigeria, will encourage behavioural change on how anti- malarial drugs are used. Finally, the WHO recommended that the use of arte- misinin monotherapies for treatment of uncomplicated malaria cases be banned as their use promotes the devel- opment of resistance to artemisinin by the Plasmodium parasites. Nigeria was actually listed in the latest WHO malaria report [1] as one of the countries that have stopped the use of oral artemisinin monotherapies; how- ever, these drugs (especially Artesunate) are still widely used and readily available in Nigeria and this threatens the efficacy of artemisinin both as monotherapies and in combination with other partner drugs. The wide use of monotherapies is another feature Nigeria shares with some of the countries in the Great Mekong region where artemisinin resistance has been confirmed. Governments of malaria endemic countries should ensure existence and enforcement of policies and regu- lations on anti-malarial drug distribution and dispens- ing. However, this can only successfully reduce the rate of anti-malarial drug abuse if the existing underlying factors that promote these behaviours, like poverty, poor access to healthcare services, poor level of educa- tion, social constructions, are first addressed. There is a need for the Nigerian government to review its policy on payment at the point of use for primary healthcare services as this constitutes a barrier to the use of for- mal health facilities that provides better quality malaria treatment. Limitations of the study One of the limitations of this study is that it did not further explore the role of socio-cultural factors in determining malaria treatment-seeking and drug use behaviours. Although some socio-cultural factors were identified in relation to malaria treatment, these were not further explored as the main focus of this study was on the contributory role of socioeconomic factors in deter- mining treatment behaviours that promote anti-malarial drug resistance. Conclusions and recommendations Socioeconomic factors are major variables in determin- ing behaviours and practices in malaria treatment. Most of the anti-malarial drug use behaviours that can pro- mote resistance, which were reported in this study, were more among people of lower income level, low level of education, and of rural settlements. Behaviours and practices like mixing drugs for malaria treatment, use of monotherapies, saving anti-malarial drugs for future use, sharing of malaria drug/treatment course, among others Availability of data and materials The datasets during and/or analysed during the current study available from the corresponding author on reasonable request. The datasets during and/or analysed during the current study available from the corresponding author on reasonable request. Authors’ contributions ll h b d All authors contributed to this study in the following areas: PEA; JF; EE and TP designed the study. Data collection and interviewing were conducted by PEA. PEA; JF; EE and TP analysed and interpreted the data. Manuscript writing up was done by PEA and JF. All authors read and approved the final manuscript. Competing interests h h d l h Also, this study is limited in terms of generalizability of the findings as it used a small sample and examined inhabitants of only two Nigerian areas. Nevertheless, the diversity of the participants, in terms of socio-demo- graphic characteristics, ensures the transferability of the findings. Competing interests The authors declare that they have no competing interests. Funding The study was funded by the Corresponding Author as part of his doctoral studies. There was no external funding for this study. The study was funded by the Corresponding Author as part of his doctoral studies. There was no external funding for this study. Acknowledgements We acknowledge all our study participants and the support by Emeka Okigbo Nwosu and our other contacts from Nigeria. Competing interests The authors declare that they have no competing interests. References 27. Bustamante C, Batista CN, Zalis M. Molecular and biological aspects of antimalarial resistance in Plasmodium falciparum and Plasmodium vivax. Curr Drug Targets. 2009;10:279–90. 1. WHO. World malaria report 2015. Geneva: World Health Organization; 2015. 28. Noedl H, Se Y, Schaecher K, Smith BL, Socheat D, Fukuda MM. Evidence of artemisinin-resistant malaria in western Cambodia. N Engl J Med. 2008;359:2619–20. 2. WHO. World malaria report 2013. Geneva: World Health Organization; 2013. 3. Miura Y. Malaria as a cause of poverty: poverty as a contributor to malaria. 2013. https://repository.wlu.edu/handle/11021/26009. Accessed 7 June 2016. 29. Phyo AP, Nkhoma S, Stepniewska K, Ashley EA, Nair S, McGready R, et al. Emergence of artemisinin-resistant malaria on the western border of Thailand: a longitudinal study. Lancet. 2012;379:1960–6. 4. Federal Ministry of Health Nigeria. National policy on malaria diagnosis and treatment. Abuja, Nigeria; 2011. 30. Cwikel J. Social epidemiology: strategies for public health activism. New York: Columbia University Press; 2006. j g 5. Mouzin E. Focus on Nigeria. Geneva: World Health Organization; 2012. 31. Berkman LF, Kawachi I, Glymour MM. Social epidemiology. Oxford: Oxford University Press; 2014. p. 641. 6. WHO. World malaria report: 2012. Geneva: World Health Organization; 2011. 7. The World Bank. Country Profile: Nigeria. 2014. http://www.worldbank. org/en/country/nigeria. Accessed 6 July 2016. 32. Krieger N. A glossary for social epidemiology. J Epidemiol Commun Health. 2001;55:693–700. 33. The World Bank. Urban Populatio: Nigeria. 2016. http://data.worldbank. org/indicator/SP.URB.TOTL.IN.ZS. Accessed 6 Jul 2016. 8. Onwujekwe O, Uzochukwu B. Socio-economic and geographic differen- tials in costs and payment strategies for primary healthcare services in Southeast Nigeria. Health Policy. 2005;71:383–97. org/indicator/SP.URB.TOTL.IN.ZS. Accessed 6 Jul 2016. 34. The World Bank. Data: Nigeria. 2016. http://data.worldbank.org/country/ nigeria. Accessed 7 June 2016. g y 9. Federal Ministry of Health, Nigeria (FMOH). Health policy of Nigeria. Abuja, Nigeria; 2001. 35. Glaser B, Strauss AL. Discovery of grounded theory: strategies for qualita- tive research. Chicago: Aldine Trans; 1999. p. 284. 10. Chuma JM, Thiede M, Molyneux CS. Rethinking the economic costs of malaria at the household level: evidence from applying a new analytical framework in rural Kenya. Malar J. 2006;5:76. 36. Glaser BG, Strauss AL. The discovery of grounded theory: Strategies for qualitative research [Internet]. Transaction Publishers. 2009. https://books. google.co.uk/books?hl=en&lr=&id=rtiNK68Xt08C&oi=fnd&pg=PP1&d q=The+discovery+of+grounded+theory+Strategies+for+qualitative +research+London&ots=UVzWUo0E2Q&sig=3tgFjzbZ-xiH1LgUHRyzPi- G67Yc. Accessed 7 June 2016. 11. Chuma JM, Okungu V, Molyneux CS. Barriers to prompt and effective malaria treatment among the poorest population in Kenya. Malar J. 2010;9:144. 12. Feachem RG. References Shrinking the malaria map: a guide on malaria elimina- tion for policy makers. 1st ed. San Francisco: Global Health Group, UCSF Global Health Sciences; 2009. p. 66. 37. Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol. 2006;3:77–101. 13. Roll Back Malaria. Global Malaria Action Plan. Geneva: World Health Organization. 2008. http://citeseerx.ist.psu.edu/viewdoc/summary? doi=10.1.1.430.2580. Accessed 7 June 2016. 38. Guest G, MacQueen KM, Namey EE. Applied thematic analysis. Thousand Oaks: SAGE Publications; 2011. p. 321. 39. Lincoln YS, Guba EG. Naturalistic Inquiry. 1st ed. Beverly Hills: SAGE Publi- cations, Inc.; 1985. p. 416. 14. WHO. 10 fact sheets about malaria. Geneva: World Health Organization; 2016. http://who.int/mediacentre/factsheets/fs094/en/. Accessed 6 Jul 2016. 40. Merriam SB, Tisdell EJ. Qualitative research: a guide to design and imple- mentation. New York: Wiley; 2015. 15. Murhandarwati EEH, Fuad A, Wijayanti MA, Bia MB, Widartono BS, Lobo NF, et al. Change of strategy is required for malaria elimination: a case study in Purworejo District, Central Java Province, Indonesia. Malar J. 2015;14:318. 41. Shenton AK. Strategies for ensuring trustworthiness in qualitative research projects. Educ Inf. 2004;22:63–75. 42. Lynch J, Kaplan G. Social epidemiology. In: Berkman LF, Kawachi I, Gly- mour MM. New York: Oxford University Press; 2014. mour MM. New York: Oxford University Press; 2014. 43. Donabedian A. Evaluating the quality of medical care. Milbank Q. 2005;83:691–729. 16. Ezenduka CC, Ogbonna BO, Ekwunife OI, Okonta MJ, Esimone CO. Drugs use pattern for uncomplicated malaria in medicine retail outlets in Enugu urban, southeast Nigeria: implications for malaria treatment policy. Malar J. 2014;13:243. 44. Ancker JS, Kern LM, Abramson E, Kaushal R. The Triangle Model for evalu- ating the effect of health information technology on healthcare quality and safety. J Am Med Inform Assoc. 2012;19:61–5. 17. Ridley RG. Medical need, scientific opportunity and the drive for antima- larial drugs. Nature. 2002;415:686–93. 45. Ammenwerth E, Ehlers F, Hirsch B, Gratl G. HIS-Monitor: an approach to assess the quality of information processing in hospitals. Int J Med Inf. 2007;76:216–25. 18. White NJ. The treatment of malaria. N Engl J Med. 1996;335:800–6. 19. Lim P, Alker AP, Khim N, Shah NK, Incardona S, Doung S, et al. Pfmdr1 copy number and arteminisin derivatives combination therapy failure in falciparum malaria in Cambodia. Malar J. 2009;8:11. 46. Kelley E, Hurst J. Health care quality indicators project. 2006. http://www. oecd-ilibrary.org/social-issues-migration-health/health-care-quality- indicators-project_440134737301. Accessed 7 June 2016. 20. Enserink M. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Received: 13 September 2016 Accepted: 9 May 2017 Received: 13 September 2016 Accepted: 9 May 2017 Anyanwu et al. Malar J (2017) 16:203 Page 14 of 15 References Signs of drug resistance rattle experts, trigger bold plan. Sci- ence. 2008;322:1776. 47. Kunkel S, Rosenqvist U, Westerling R. The structure of quality systems is important to the process and outcome, an empirical study of 386 hospi- tal departments in Sweden. BMC Health Serv Res. 2007;7:104. 21. Nayyar GM, Breman JG, Newton PN, Herrington J. Poor-quality antima- larial drugs in southeast Asia and sub-Saharan Africa. Lancet Infect Dis. 2012;12:488–96. 48. CDC. Treatment of malaria (Guidelines for Clinicians). Center for Disease Control and Prevention. 2013. http://www.cdc.gov/malaria/resources/ pdf/clinicalguidance.pdf. Accessed 6 June 2016. 22. Newman RD. Learning to outwit malaria. Bull World Health Organ. 2011;89:10–1. 23. Tanner M, Savigny D. Malaria eradication back on the table Geneva: World Health Organization. 2008. http://www.who.int/bulletin/vol- umes/86/2/07-050633/en/. Accessed 7 June 2016. 49. WHO. Guidelines for the treatment of treatment. 3rd ed. Geneva: World Health Organization. 2015. http://apps.who.int/iris/bitstr eam/10665/162441/1/9789241549127_eng.pdf. Accessed 6 June 2016. 24. Huijben S. Experimental studies on the ecology and evolution of drug- resistant malaria parasites. 2010 https://www.era.lib.ed.ac.uk/han- dle/1842/3945. Accessed 7 June 2016. 50. Onwujekwe O. Inequities in healthcare seeking in the treatment of commu- nicable endemic diseases in Southeast Nigeria. Soc Sci Med. 2005;61:455–63 51. Salako LA, Brieger WR, Afolabi BM, Umeh RE, Agomo PU, Asa S, et al. Treatment of childhood fevers and other illnesses in three rural Nigerian communities. J Trop Pediatr. 2001;47:230–8. 25. Pinder RJ, Berry D, Sallis A, Chadborn T. Antibiotic prescribing and behav- iour change in healthcare settings: literature review and behavioural analysis. 2015. https://spiral.imperial.ac.uk/handle/10044/1/22194. Accessed 7 June 2016. 52. Kaur H, Clarke S, Lalani M, Phanouvong S, Guérin P, McLoughlin A, et al. Fake anti-malarials: start with the facts. Malar J. 2016;15:1. 53. Onwujekwe O, Uzochukwu B, Dike N, Uguru N, Nwobi E, Shu E. Malaria treatment perceptions, practices and influences on provider behaviour: comparing hospitals and non-hospitals in south-east Nigeria. Malar J. 2009;8:246. 26. Anderson TJC, Nair S, Nkhoma S, Williams JT, Imwong M, Yi P, et al. High heritability of malaria parasite clearance rate indicates a genetic basis for artemisinin resistance in Western Cambodia. J Infect Dis. 2010;201:1326–30. Anyanwu et al. Malar J (2017) 16:203 Page 15 of 15 Page 15 of 15 59. White NJ, Pongtavornpinyo W, Maude RJ, Saralamba S, Aguas R, Step- niewska K, et al. Hyperparasitaemia and low dosing are an important source of anti-malarial drug resistance. Malar J. 2009;8:253. 54. Bate R, Coticelli P, Tren R, Attaran A. References Antimalarial drug quality in the most severely malarious parts of Africa—a six country study. PLoS ONE. 2008;3:e2132. 60. Anyanwu P, Fulton J, Paget T, Evans E. Socioeconomic determinants of antimalarial drug use behaviours: a systematic review. J Commun Public Health Nurs. 2016;2:2–9. http://www.omicsonline.org/open-access/ socioeconomic-determinants-of-antimalarial-drug-use-behaviours- asystematic-review-jcphn-1000123.pdf. Accessed 6 Sep 2016. 55. Bate R, Jin GZ, Mathur A. Does price reveal poor-quality drugs? Evidence from 17 Countries. National Bureau of Economic Research. 2011. http:// www.nber.org/papers/w16854. Accessed 7 June 2016. 56. Newton PN, McGready R, Fernandez F, Green MD, Sunjio M, Bruneton C, et al. Manslaughter by fake artesunate in Asia—will Africa be next? PLoS Med. 2006;3:e197. 57. Dondorp AM, Yeung S, White L, Nguon C, Day NPJ, Socheat D, et al. Arte- misinin resistance: current status and scenarios for containment. Nat Rev Microbiol. 2010;8:272–80. 58. Omole MK, Onademuren OT. A survey of antimalarial drug use prac- tices among urban dwellers in Abeokuta, Nigeria. Afr J Biomed Res. 2013;13:1–7. 59. White NJ, Pongtavornpinyo W, Maude RJ, Saralamba S, Aguas R, Step- niewska K, et al. Hyperparasitaemia and low dosing are an important source of anti-malarial drug resistance. Malar J. 2009;8:253. 61. Watsierah CA, Jura WG, Raballah E, Kaseje D, Abong’o B, Ouma C. Knowl- edge and behaviour as determinants of anti-malarial drug use in a peri- urban population from malaria holoendemic region of western Kenya. Malar J. 2011;10:99. 60. Anyanwu P, Fulton J, Paget T, Evans E. Socioeconomic determinants of antimalarial drug use behaviours: a systematic review. J Commun Public Health Nurs. 2016;2:2–9. http://www.omicsonline.org/open-access/ socioeconomic-determinants-of-antimalarial-drug-use-behaviours- asystematic-review-jcphn-1000123.pdf. Accessed 6 Sep 2016. Anyanwu et al. Malar J (2017) 16:203 Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries • Our selector tool helps you to ﬁnd the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and we will help you at every step:
https://openalex.org/W3170467238	https://opus.lib.uts.edu.au/bitstream/10453/150884/2/1-s2.0-S0031940621000626-main.pdf	English	null	Development and internal validation of prognostic models for recovery in patients with non-specific neck pain presenting in primary care	Physiotherapy	2,021	cc-by	8,033	Abstract This model has potential to inform primary care clinicians about a patient’s individual prognosis after treatment, but external validation is required before clinical use can be recommended. Conclusion A post-treatment prognostic model for disability was successfully developed and internally validated. This model has potential to inform primary care clinicians about a patient’s individual prognosis after treatment, but external validation is required before clinical use can be recommended. © 2021 The Authors. Published by Elsevier Ltd on behalf of Chartered Society of Physiotherapy. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Development and internal validation of prognostic models for recovery in patients with non-speciﬁc neck pain presenting in primary care Roel W. Wingbermühle a,b,∗, Alessandro Chiarotto b,c, Emiel van Trijffel a,d, Bart Koes b,e, Arianne P. Verhagen b,f, Martijn W. Heymans g a SOMT University of Physiotherapy, Amersfoort, The Netherlands b Department of General Practice, Erasmus MC, University Medical Center, Rotterdam, The Netherlands c Department of Health Sciences, Faculty of Science, Vrije Universiteit Amsterdam, Amsterdam Movement Sciences, The Netherlands d Vrije Universiteit Brussels, Experimental Anatomy Research Department, Department of Physiotherapy, Human Physiology and Anatomy, Faculty of Physical Education and Physiotherapy, Brussels, Belgium e Center for Muscle and Joint Health, University of Southern Denmark, Odense M, Denmark f University of Technology Sydney, Level 4 Building 7, Sydney, Australia g Department of Epidemiology and Data Science, VU University Medical Center, Amsterdam, The Netherlands Roel W. Wingbermühle a,b,∗, Alessandro Chiarotto b,c, Emiel van Trijffel a,d, Bart Koes b,e, Arianne P. Verhagen b,f, Martijn W. Heymans g a SOMT University of Physiotherapy, Amersfoort, The Netherlands b Department of General Practice, Erasmus MC, University Medical Center, Rotterdam, The Netherlands c Department of Health Sciences, Faculty of Science, Vrije Universiteit Amsterdam, Amsterdam Movement Sciences, The Netherlands d Vrije Universiteit Brussels, Experimental Anatomy Research Department, Department of Physiotherapy, Human Physiology and Anatomy, Faculty of Physical Education and Physiotherapy, Brussels, Belgium e Center for Muscle and Joint Health, University of Southern Denmark, Odense M, Denmark f University of Technology Sydney, Level 4 Building 7, Sydney, Australia g Department of Epidemiology and Data Science, VU University Medical Center, Amsterdam, The Netherlands Physiotherapy 113 (2021) 61–72 Physiotherapy 113 (2021) 61–72 Abstract Objectives Development and internal validation of prognostic models for post-treatment and 1-year recovery in patients with neck pain in primary care. D i bjectives Development and internal validation of prognostic models for post-treatment and 1-year recovery in mary care. Design Prospective cohort study. Setting Primary care manual therapy practices. Setting Primary care manual therapy practices. Participants Patients with non-speciﬁc neck pain of any duration (n = 1193). pants Patients with non-speciﬁc neck pain of any durati Participants Patients with non-speciﬁc neck pain of any duration (n = 1193 i Intervention Usual care manual therapy. Outcome measures Recovery deﬁned in terms of pain intensity, disability, and global perceived improvement directly post-treatment and at 1-year follow-up. utcome measures Recovery deﬁned in terms of pain intensity, disability, and global perceived improvement d 1-year follow-up. Results All post-treatment models exhibited acceptable discriminative performance after derivation (AUC ≥ 0.7). The developed post- treatment disability model exhibited the best overall performance (R2 = 0.24; IQR, 0.22–0.26), discrimination (AUC = 0.75; 95% CI, 0.63–0.84), and calibration (slope 0.92; IQR, 0.91–0.93). After internal validation and penalization, this model retained acceptable discrimi- native performance (AUC = 0.74). The ﬁve other models, including those predicting 1-year recovery, did not reach acceptable discriminative performance after internal validation. Baseline pain duration, disability, and pain intensity were consistent predictors across models. Conclusion A post-treatment prognostic model for disability was successfully developed and internally validated. This model has potential to inform primary care clinicians about a patient’s individual prognosis after treatment, but external validation is required before clinical use can be recommended Results All post-treatment models exhibited acceptable discriminative performance after derivation (AUC ≥ 0.7). The developed post- treatment disability model exhibited the best overall performance (R2 = 0.24; IQR, 0.22–0.26), discrimination (AUC = 0.75; 95% CI, 0.63–0.84), and calibration (slope 0.92; IQR, 0.91–0.93). After internal validation and penalization, this model retained acceptable discrimi- native performance (AUC = 0.74). The ﬁve other models, including those predicting 1-year recovery, did not reach acceptable discriminative performance after internal validation. Baseline pain duration, disability, and pain intensity were consistent predictors across models. native performance (AUC = 0.74). The ﬁve other models, including those predicting 1-year recovery, did not reach acceptable discriminative performance after internal validation. Baseline pain duration, disability, and pain intensity were consistent predictors across models. Conclusion A post-treatment prognostic model for disability was successfully developed and internally validated. Design Neck pain is a top ﬁve cause of Years Lived with Dis- ability in high and middle income countries and, after low back pain, the second worldwide largest cause of muscu- loskeletal disability [1]. Recovery from non-speciﬁc neck pain mainly takes place in the ﬁrst six weeks with very little further long-term improvement of pain and disabil- ity [2,3]. The prevalence of chronic neck pain, i.e. pain lasting longer than three months, has increased from 2005 to 2015 by 21% up to approximately 358 million people worldwide and it is likely to increase further in West- ern countries due to an ageing population [4]. Noninvasive primary care interventions (e.g. mobilisations and manipula- tions, exercise, psychosocial interventions, or combinations) are reported as effective treatments for non-speciﬁc neck pain [5–7]. For this model derivation study, the authors used data from a prospective cohort study, the ‘Amersfoorts Nekonderzoek of the Master manuele therapie Opleiding’ (ANIMO), con- ducted from 2007 to 2009. In total, 345 manual therapists in the Netherlands recruited 1311 consecutive patients between 18 and 80 years presenting with non-speciﬁc neck pain of any duration. Participants providing baseline data and hav- ing signed informed consent were deemed eligible (n = 1193). Neck pain with or without associated arm pain was classiﬁed as non-speciﬁc if the pain could not be attributed to a speciﬁc underlying pathology (i.e., no red ﬂags were present). Study characteristics (e.g., setting, inclusion criteria, measurement procedures) have been described in detail elsewhere [17]. Participating patients received usual care multimodal manual therapy which may have included speciﬁc joint mobiliza- tions, high velocity thrust techniques, myofascial techniques, giving advice, or speciﬁc exercises. Mean treatment duration was 37.9 days, mean number of treatment sessions was 4.3. The Erasmus Medical Centre Ethics Committee Rotterdam, the Netherlands (MEC-2007-359) approved this study. An accurate individual prognosis at intake can inform clinicians and patients in shared clinical decisions [8]. For example, in patients with a high risk of poor prognosis, sub- sequent effective treatment interventions may improve the patients’ prognosis; at the same time, a wait-and-see approach in patients with a very low risk of poor prognosis can limit exposure to unnecessary treatments and reduce costs [8]. Sep- arate prognostic factors which are consistently reported for outcomes on neck-related pain, physical functioning, and per- ceived recovery are: age, sex, baseline pain intensity, baseline disability, and past history of neck pain [9–11]. Contribution of the paper • Existing prognostic models for patients with non-speciﬁc neck pain present substantial methodological shortcomings, which prevent their clinical use. W d l d d i ll lid d i d l di i i i h k i • Existing prognostic models for patients with non-speciﬁc neck pain present substantial methodological shortcomings, which prevent their clinical use. ernally validated prognostic models to predict recovery in patients with neck pain. developed and internally validated prognostic models to predict recovery in patients with neck pain. • The prognostic model for post-treatment disability exhibited good performance and calibration, showing promise for external validation and clinical use. • The prognostic model for post-treatment disability exhibited good performance and calibration, showing promise for external validation and clinical use. Keywords: Neck pain; Recovery; Prognostic models; Model development; Prognostic factors ∗Corresponding author at: SOMT University of Physiotherapy, P.O. Box 585, 3800 AN Amersfoort, The Netherlands. E-mail addresses: r.wingbermuhle@somtuniversity.nl, roelwingbermuhle@me.com (R.W. Wingbermühle). ∗Corresponding author at: SOMT University of Physiotherapy, P.O. Box 585, 3800 AN Amersfoort, The Netherlands. E-mail addresses: r.wingbermuhle@somtuniversity.nl, roelwingbermuhle@me.com (R.W. Wingbermühle). https://doi.org/10.1016/j.physio.2021.05.011 0031-9406/© 2021 The Authors. Published by Elsevier Ltd on behalf of Chartered Society of Physiotherapy. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1016/j.physio.2021.05.011 0031-9406/© 2021 The Authors. Published by Elsevier Ltd on behalf of Chartered Society of Physiotherapy. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). R.W. Wingbermühle et al. / Physiotherapy 113 (2021) 61–72 62 Design Prognos- tic prediction models (in short: prognostic models) provide probabilities for patients based on their individual combina- tion of predictor values and can support clinicians in their clinical decisions [12]. Prognostic models have been shown to improve prognostic accuracy in various healthcare ﬁelds [13,14]. However, a recent systematic review concluded that the clinical utility of currently available prognostic mod- els in people with neck pain is limited [15]. Overall, the methodological quality of the studies included in this review was low with the large majority of studies lacking sufﬁ- cient sample size and internal validation [15]. Furthermore, from the three promising models as deﬁned in the systematic review, two appeared invalid in a subsequent external valida- tion study and a third model speciﬁcally focusing on patients with whiplash associated disorders could not be tested [16]. Therefore, there is a need to develop a prognostic model for recovery in patients with neck pain that exhibits satisfactory prediction. This model should be developed in a cohort of patients with adequate sample size, and it should be internally validated. This study was conducted following the PROGRESS group recommendations [18] and reported according to the Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD) statement [19]. Candidate model predictors The authors based selection of candidate predictors for the models on the literature and clinical credibility of vari- ables in combination with their reliability, applicability, and costs [20–23], while avoiding univariable pre-selection [8]. The following predictors were considered: age, sex, pre- vious neck pain episode, neck pain duration (acute 0–6 weeks, sub-acute 6–12 weeks and chronic >12 weeks), pain intensity (measured with a Numerical Rating Scale (NRS)), and disability (Neck Disability Index – Dutch version (NDI-DV)) [11,24,25]. Furthermore, the authors included six additional candidate predictors regarded in the litera- ture as clinically credible and relatively easy to collect at intake [9,11,25]: accompanying headache (yes/no), accom- panying low back pain (yes/no), accompanying radiating arm pain (yes/no), smoking status (yes/no), fear-avoidance beliefs (Fear-Avoidance Beliefs Questionnaire – Dutch version (FABQ-DV) physical activity subscale [26,27]), and psycho- logical functioning (Neck Bournemouth Questionnaire-DV (NBQ-DV) anxiety and depression subscale [28–30]). Addi- tionally, the authors considered other potentially relevant predictors from a clinical perspective: general sleeping prob- The aim of this study was to develop and internally vali- date prognostic models that predict at intake post-treatment and 1-year follow-up recovery of neck pain, disability, and global perceived improvement in patients treated with manual therapy in primary care. R.W. Wingbermühle et al. / Physiotherapy 113 (2021) 61–72 63 lems (yes/no), partaking in sporting activities (yes/no), and patients’ expectation to change due to treatment (5-point Likert scale, ranging from ‘much better’ to ‘much worse’) [31]. as Nagelkerke’s R2; calibration was estimated by the calibra- tion slope, calibration curve, and the Hosmer–Lemeshow test; and the Area Under Curve (AUC) of the receiver-operating characteristic Curve (ROC) was calculated for quantifying discriminative performance [8,23]. Perfect discriminative performance has a value of 1 and the authors considered dis- criminative performance acceptable if AUC was ≥0.7 [48]. The calibration plot is obtained across multiply imputed data sets by the following approach that is commonly used to make a calibration plot. In each imputed dataset the predicted prob- abilities are determined and used to make 10 groups by using 10 deciles. Within these groups the observed outcomes were divided by the sample size of each group to obtain the pre- dicted probabilities. The agreement between these 10 groups is plotted on the calibration curve and a natural cubic spline curve is plotted between the black dots. Sample size and candidate model predictors The authors performed a priori sample size calculations for each model to decide on the amount of candidate pre- dictor parameters, using the procedure described by Riley et al. with a shrinkage of 0.8 and R2 of 0.1 [51]. The propor- tion post-treatment non-recovery was 21%, 58%, and 21% for pain intensity, disability, and global perceived improve- ment, respectively, and after 1 year it was 45%, 62%, and 39%, respectively. This resulted in a maximum amount of candidate predictor categories, depending on these outcome proportions, ranging from 14 to 18. Calculations were made with the pmsamplesize package in R. Outcomes In this study, recovery was used as an umbrella term for three different constructs and outcome measures, which were: (1) for pain as an NRS (10-point Likert scale) score dichotomized into >2 for non-recovery and ≤2 for recov- ery as the latter is considered as a satisfactory state by patients [32]; (2) for disability, by dichotomizing the NDI- DV (0–50 scale range), after values were multiplied by two to yield percentages, into <8% for recovery and ≥8% for non- recovery, which is a threshold used before [33,34]; and (3) for global perceived improvement as Global Perceived Effect (GPE) measured on a 7-point Likert scale where recovery was deﬁned by response options “completely recovered” or “much improved”, while non-recovery by responses “slightly improved”, “no change”, “slightly worse”,” much worse”, and “worse than ever” represented non-recovery [35,36] Post- treatment follow-up was measured in ANIMO immediately after a course of treatment and deﬁned as no more than three months after intake, and long-term follow-up was measured after one year from inclusion. Outcome questionnaires were returned by post through provided prepaid envelopes. Results Multivariable logistic regressions were estimated for all the models in the imputed ANIMO datasets as primary anal- ysis. A backward elimination approach with the P-value set at <0.157 was used as this corresponds to the Akaike infor- mation criterion [43,47]. Overall performance was expressed Statistical analysis Regression model assumptions such as linear relationship between predictor variables and the outcome were evaluated using restricted cubic splines and multicollinearity (Tol- erance > 0.2, Variance Inﬂation Factor < 3). Variables were coded before entering the regression models and categorical variables were transformed into dummy variables [44–46]. Sensitivity analyses Missing values were evaluated by comparing patients with and without missing values on relevant predictors and by performing t-tests [37–40]. Missing At Random (MAR) was most plausible based on the data not being MCAR accord- ing to compared patients and the performed t-tests. Multiple imputation on predictors as well as outcomes using all pre- dictor and outcome variables was performed [38–41]. The method of Multivariate Imputation by Chained Equations (MICE) procedure with generation of 50 imputed data sets was applied [41]. Regression coefﬁcient estimates and stan- dard errors were pooled according to Rubin’s Rules, and model performance measures estimated in each of the 50 completed datasets and then combined [39,42,43]. In addition, the authors estimated all models and their per- formance measures on the complete case data as sensitivity analyses to allow comparison of models and performance measures obtained on the imputed data. Candidate model predictors The groups and cali- bration curves of each imputed data set are plotted in the same ﬁgure, distinguished by the multiple blue lines and multiple black dots for the groups. This makes it possible to evaluate agreement across multiply imputed data sets. Internal val- idation of all models was performed with bootstrapping in 250 samples, and repeating all development steps. [49]. The authors corrected the models’ regression coefﬁcients with the optimism-adjusted calibration slope value and updated the intercept using an “offset” procedure by calculating the linear predictor with the new regression coefﬁcients ﬁxed [50]. All analyses were performed in IBM SPSS 24.0 and R version 3.4.3. Outcomes Baseline characteristics and candidate model predictors Patients’ baseline characteristics and candidate factors were comparable for complete cases (Supplement 2) and R.W. Wingbermühle et al. / Physiotherapy 113 (2021) 61–72 64 Table 1 Baseline characteristics and candidate model predictors of patients with non-speciﬁc neck pain (n = 1193). Baseline characteristics Missing n (%) Age (years), mean (SD) 44.7 (13.7) 23 (2) Gender 7 (1) Female sex, n (%) 823 (69) Previous neck pain episode 64 (5) Yes, n (%) 755 (67) Neck pain duration 122 (10) Acute 0 to 6 weeks, n (%) 420 (39) Subacute 6 to 12 weeks, n (%) 138 (13) Chronic >12 weeks, n (%) 513 (48) Pain intensity (NRS, scale 1 to 10)c, mean (SD) 4.8 (2.1) 10 (1) Disability (NDI, scale 0 to 50)d, median [IQR] 12.0 [8.0 to 17.0] 97 (8) Accompanying headache 0 (0) Yes, n (%) 707 (59) Accompanying low back pain 0 (0) Yes, n (%) 538 (45) Accompanying radiating arm pain 0 (0) Yes, n (%) 536 (45) Accompanying general sleeping problems 0 (0) Yes, n (%) 337 (28) Smoking status 3 (0) Yes, n (%) 300 (25) Fear-avoidance believes (FABQ-PA, scale 0 to 24)a, median [IQR] 11.0 [6.0 to 15.0] 85 (7) Emotional functioning (NBQ-AD, scale 0 to 20)b, median [IQR] 7.0 [3.0 to 10.0] 16 (1) Partaking in sporting activities 4 (0) Yes, n (%) 783 (66) Patients’ expectation to change due to treatment 3 (0) Much better, n (%) 517 (43) Better, n (%) 662 (56) No change, n (%) 10 (1) Worse, n (%) 1 (0) Much worse, n (%) 0 (0) % d d l i Table 1 characteristics and candidate model predictors of patients with non-speciﬁc neck pain (n = 1193). Smoking status Yes, n (%) Worse, n (%) % rounded up to closest integer. % rounded up to closest integer. c NRS = numeric rating scale. d NDI = neck disability index. cases with no outcome data (Table 1). Mean age of patients was 44.7 (SD 13.7) years, 69% (n = 823) were female, and 67% (n = 755) experienced a previous episode and 48% (n = 513) was classiﬁed as chronic. Mean baseline pain inten- sity was 4.8 (SD 2.1) and median disability was 12.0 [IQR 8.0–17.0]. The candidate factor for treatment expectations was excluded since it showed an extreme standardised error and coefﬁcient during model estimation. Outcome values Outcome values are presented in Table 2. Pain intensity was 2.0 [IQR 1.0–2.0] and 2.8 [IQR 1.0–4.0] post-treatment and at 1-year, respectively. Disability was 5.0 [IQR 1.0–9.0] and 5.0 [IQR 2.0–8.0] post-treatment and at 1-year, respec- tively. Baseline characteristics and candidate model predictors ues reached about 45% missing values and the post-treatment about 55% (Table 2). Baseline characteristics were com- parable between complete cases (Supplement 2) and those without outcome data, and the means of several variables differed signiﬁcantly depending on the missingness of indi- cator variables, indicating that the MAR assumption is more plausible. Therefore, the authors assumed data were MAR. The authors chose 50 imputed datasets as the rule of thumb is the number of imputations is as large as the percentage of missing data [41]. In fact, the authors had missing data of 46, 43, 43, 53, 54 and 56% in the outcomes. This is on aver- age 42% for all outcomes. The authors applied one run of 50 imputed datasets and developed the different models in the same imputed data to eliminate the inﬂuence of missing data imputation on the development of the models. Multicollinear- ity in the MI model was not checked, but checked between variables before the models were developed. Further, the authors evaluated the convergence plots of the imputed vari- ables and these showed healthy convergence, i.e., no irregular patterns were visible, which is often an indication of that there is no multicollinearity between variables. % rounded up to closest integer. a FABQ-PA = fear-avoidance beliefs questionnaire, physical activity subscale (scale 0–24). b NBQ-AD = Neck Bournemouth Questionnaire, anxiety and depression subscale (scale 0–20), sum score of 11-point numeric subscale of items 4 and 5. c NRS = numeric rating scale. Missing values Several baseline characteristics had more than 5% missing values and a few up to 13% (Table 1). The 1-year outcome val- R.W. Wingbermühle et al. / Physiotherapy 113 (2021) 61–72 65 Table 2 Pain intensity, disability, and perceived recovery post-treatment (n = 1125)a and at 1 year (n = 1193). Outcomes Post-treatmenta Missing, n % 1 year Missing, n % Pain intensity (NRS, 1 to 10 scale)e, median [IQR] 2.0 [1.0 to 2.0] 591 (53) 2.0 [1.0 to 4.0] 552 (46) Not recoveredb, n % 112 (21) 286 (45) Disability (NDI, 0 to 50 scale)f, median [IQR] 5.0 [1.0 to 9.0] 628 (56) 5.0 [2.0 to 8.0] 515 (43) Not recoveredc, n % 290 (58) 423 (62) Global perceived improvement (GPE, 7-point Likert scale)g, n % 605 (54) 508 (43) Completely recovered 127 (24) 149 (23) Much improved 287 (55) 247 (39) Slightly improved 83 (16) 143 (22) No change 24 (5) 81 (13) Slightly worse 0 (0) 11 (2) Much worse 0 (0) 8 (1) Worse than ever 0 (0) 2 (0) Not recoveredd, n % 107 (21) 264 (39) % rounded up to closest integer. a Deﬁned as no more than three months after intake, n = 68 not eligible. a Deﬁned as no more than three months after intake, n = 68 not eligible. c Score multiplied by 2 to yield %, not-recovered ≥8%, recovered <8%. e NRS = numeric rating scale. f NDI = neck disability index. g GPE = global perceived effect. None of the 1-year models reached the level of acceptable discriminative performance after derivation and after adjust- ment for optimism, and showed lower calibration slope values and explained variances. Predictors in the models Neck pain duration was a predictor in all models (Supple- ment 3). Baseline pain was a predictor in all pain models and baseline disability in all disability models. Age was a predic- tor included in all post-treatment models and headache in all 1-year models. Derived models The derived models for post-treatment prediction are described in Table 3 and Supplement 1 and those for 1-year prediction in Table 4. The authors compared spline models’ performance to linear models’ performance for non-linear variable and outcome relations (i.e. Disability model at 1 year and Disability model post-treatment). Spline models’ perfor- mance appeared not superior to linear models’ performance and the authors choose to present these as linear models as they are more straightforward for clinical use. Models’ inter- cept, predictors, and assigned weights (beta’s) are displayed together with their performance and optimism-adjusted per- formance measures as evaluated in imputed data [8]. For all models the Hosmer–Lemeshow test was not-signiﬁcant. a Deﬁned as no more than three months after intake, n = 68 not eligible. Sensitivity analyses g All derived post-treatment models exhibited acceptable discriminative performance. The disability model obtained the highest discriminative performance, and showed a cali- bration slope of 0.92 (IQR, 0.91–0.93), and R2 of 0.24 (IQR, 0.22–0.26). The derived post-treatment pain and perceived improvement models exhibited somewhat lower discrimina- tive performance, with calibration slope values of 0.86 (IQR, 0.91–0.93) and 0.86 (IQR, 0.84–0.87), respectively, and low explained variances. Calibration plots of post-treatment mod- els are presented in Fig. 1. After adjustment for optimism, only the post-treatment disability model retained acceptable discriminative performance of AUC 0.74 (IQR, 0.72–0.75), and R2 of 0.21 (IQR, 0.19–0.23). Sensitivity analyses on complete cases (post-treatment pain, disability, perceived improvement models, n = 532, 495, 518 respectively; 1-year pain, disability, perceived improvement models, n = 476, 508, 511 respectively) showed comparable performance measure values. The post-treatment pain model and the 1-year models derived in complete case data yielded the same or almost the same predictors (Supple- ment 3). The post-treatment disability model in the complete cases contained also sporting and previous episode as predic- tors and the perceived improvement model did not contain the sporting, previous episode, age and baseline disability predictors. nt recovery of neck pain (n = 1193)#. OR ## R2 Optimism-adjusted R2 AUC Optimism-adjusted AUC 57) 0.03 (0.01 to 0.08) 0.13 [0.12 to 0.14] $ 0.09 [0.08 to 0.11] $ 0.70 (0.56 to 0.81) $$ 0.67 [0.66 to 0.69] $ 1.56 (0.78 to 3.10) 2.62 (1.60 to 4.31) 1.21 (1.07 to 1.36) 0) 1.05 (1.00 to 1.10) 1.01 (1.00 to 1.02) 93) 0.06 (0.03 to 0.15) 0.24 [0.22 to 0.26] $ 0.21 [0.19 to 0.23] $ 0.75 (0.63 to 0.84) $$ 0.74 [0.72 to 0.75] $ 1.34 (0.77 to 2.36) 2.62 (1.70 to 4.03) 1.13 (1.08 to 1.17) 1.02 (1.01 to 1.03) 1.36 (0.91 to 2.05 1.02 (0.99 to 1.05) 64) 0.07 (0.02 to 0.19) 0.13 [0.11 to 0.15] $ 0.09 [0.07 to 0.11] $ 0.70 (0.56 to 0.80) $$ 0.67 [0.65 to 0.69] $ 1.18 (0.49 to 2.81) 2.57 (1.60 to 4.17) 1.51 (0.98 to 2.30) 1.04 (1.00 to 1.08) 1.02 (0.10 to 1.03) 1) 0.97 (0.93 to 1.01) 1.04 (1.00 to 1.08) 1.46 (0.95 to 2.25) nterval (CI); $ In logit scale as median with interquartile range [IQR]; $$ In logit scale as mean with 95% CI. ; not-recovered >2. ltiplied by 2 to yield %); not-recovered ≥8%. k pai ## (0.1 (1.0 (0.4 (0.9 (1.0 (1 0 Model comparison with literature One study with six months follow up and a GPE out- come derived a model in a primary care population (n = 468) treated for non-serious neck pain and validated this model in a primary care setting treated with manual therapy and electrotherapy (n = 346) [35]. This model performed less well if compared to the post-treatment model on GPE out- come in our study but similarly to the 1-year model. Its external validation study revealed a possibly helpful dis- criminative ability of AUC 0.65 (95% CI, 0.59 to 0.71), a value slightly better compared to our internal validation [35]. Another study developed models also using the NDI as an outcome in people with acute whiplash associated disorder (WAD) at one-year [55]. Models’ overall performance (R2) was presented but no model calibration and discrimination were calculated, which hampers comparison of model perfor- mance. However, these models performed not well at external validation [16]. Another study developed a prognostic model for WAD, with six months follow-up, in an insurance com- pany subcohort treated with physical therapy physiotherapy and collected self-reported recovery outcome through tele- phone interview [56]. An AUC of 0.67 (95% CI, 0.63 to 0.70) was reported after internal validation. This is comparable to the post-treatment model on GPE outcome after internal val- idation in our study and somewhat better compared to our 1-year model after internal validation. In the current study, the authors recruited patients with non-speciﬁc neck pain of any duration including neck pain with trauma, and, in con- Main result Neck pain duration was a predictor in all models and independent of type of outcome or follow-up time. Baseline disability was a predictor in almost all models except for pain outcome. Baseline pain was a predictor in almost all mod- els except for disability outcome. Age was a predictor that corresponded consistently with post-treatment follow-up and headache with 1-year follow-up. The derived model for post-treatment disability contain- ing baseline pain duration, baseline disability, age, sleeping problem and FABQ-physical activity as predictors exhibited the best overall performance, calibration, and discrimination and it also exceeded the threshold for acceptable discrim- inative ability after adjustment for optimism. The other post-treatment models almost reached acceptable discrimi- native ability after adjustment. None of the derived 1-year models reached acceptable discriminative performance and showed lower calibration slope values and explained vari- ances. Sensitivity analyses kert scale); non-recovered as “slightly improved”, “no change”, “slightly worse”, much worse”, “worse than ever”. depression subscale (scale 0–20), sum score of 11-point numeric subscale of items 4 and 5. al Activity subscale (scale 0–24). ck pain (n = 1193).#. ## R2 Optimism-adjusted R2 AUC Optimism-adjusted AUC 8 (0.17 to 0.45) 0.09 [0.08 to 0.10] $ 0.06 [0.05 to 0.07] $ 0.65 (0.52 to 0.76) $$ 0.62 [0.62 to 0.63] $ 4 (1.06 to 1.24) 2 (0.43 to 0.88) 4 (0.96 to 1.86) 0 (1.03 to 1.89) 5 (1.00 to 1.83) 6 (0.18 to 0.71) 0.09 [0.08 to 0.10] $ 0.06 [0.05 to 0.07] $ 0.65 (0.53 to 0.76) $$ 0.63 [0.62 to 0.64] $ 5 (0.66 to 1.67) 2 (1.13 to 2.32) 5 (1.02 to 1.08) 1 (1.00 to 1.02) 4 (1.01 to 2.05) 5 (0.16 to 0.40) 0.10 [0.09 to 0.11] $ 0.07 [0.06 to 0.08] $ 0.66 (0.53 to 0.77) $$ 0.64 [0.63 to 0.65] $ 5 (0.85 to 2.49) 9 (1.03 to 2.15) 4 (1.01 to 1.06) 8 (1.13 to 2.20) 7 (0.47 to 0.97) 0 (0.48 to 0.99) 2 (1.25 to 2.38) CI); $ In logit scale as median with interquartile range [IQR]; $$ In logit scale as mean with 95% CI. overed >2. y 2 to yield %); not-recovered ≥8%. e); non-recovered as “slightly improved”, “no change”, “slightly worse”, much worse”, “worse than ever”. R.W. Wingbermühle et al. / Physiotherapy 113 (2021) 61–72 68 Fig. 1. Calibration plots. a. Disability model. b. Pain model. c. Perceived improvement model. Fig. 1. Calibration plots. a. Disability model. b. Pain model. c. Perceived improvement model. Conclusions A post-treatment prognostic model for disability was suc- cessfully developed and internally validated. This model has potential to inform primary care clinicians about a patient’s individual prognosis after treatment, but external validation is required before broad clinical use can be recommended. Predictors in the models compared with literature A recent overview of systematic reviews on prognostic factors in neck pain reported that higher baseline NDI and pain at inception were predictors of outcomes after WAD [11]. In our study, in which patients with non-speciﬁc neck pain and WAD were included, all models that predicted dis- ability yielded baseline disability as predictor, and models that predicted pain contained baseline pain as predictor. This is in line with the vast majority of models that predicted dis- ability outcomes and pain outcomes as described in a recent systemic review [15]. Baseline NDI and baseline pain are consistent reported prognostic factors [11,24,25] for predic- tion of disability and neck pain, respectively. This is also the case for neck pain duration as a consistently reported prog- nostic factor [11,24,25] that retains its predictive ability in relation to other prognostic factors for all outcomes as well as age and headache who are consistently reported prognostic factors [11,24,25] that retain their predictive ability in rela- tion to other prognostic factors, for short-term and long-term prognosis, respectively. Sex and previous neck pain episode [11,24,25] appeared less consistent in relation to other prog- nostic factors. The authors reached sample size for the post-treatment disability model and all 1-year models. However, the post- treatment pain and perceived improvement models fell one predictor parameter short to reach effective sample size (the excluded candidate factor for treatment expectations was considered). The authors believe to have corrected for this overﬁtting by penalizing the post-treatment models after internal validation. Important results models The post-treatment models performed better than the 1- year models and exhibited discrimination of 0.70 or upward and calibration slopes more or less around a value of 0.90. Is seems plausible that short-term prediction is more accurate compared to long-term prediction. The post-treatment dis- ability model performed best, possibly because the outcome was measured with the NDI, which is an instrument that cov- ers various health constructs [52]. The NRS is a single-item questionnaire which measures a narrower domain and may also have larger measurement error that can inﬂuence the performance of the models [53]. The same may apply to the GPE which, additionally, is an instrument reﬂecting the cur- rent health status more than change in health status over time [36]. On the whole, our derived models, especially the post- treatment disability model, performed better as compared to existing models that predict recovery in neck pain patients, although few derivation studies allow proper comparison of model performance as both discrimination and calibration performance measures were seldom presented [15,54]. R.W. Wingbermühle et al. / Physiotherapy 113 (2021) 61–72 69 trast with the two aforementioned studies, the authors did not develop a model speciﬁc for WAD. amount of 50 imputed sets. There is some evidence from sim- ulation studies that this high missing data rate can be handled with multiple imputation [57]. To address the potential limi- tation of gain from imputation, complete case analyses were also performed as sensitivity analyses and these showed very similar parameter estimates and this consistency supports our conclusions. Another limitation to be addressed is that the authors used binary outcomes for reason of comparison with previous developed models. The use of other cut-offs may have resulted in other model predictors or model performance and the derived models have to be interpreted in relation to the cut-offs points used at issue. Strengths and limitations In contrast with previously published prognostic models for neck pain [15] the models in our study were developed in a large cohort with sufﬁcient power, and the cohort closely resembles clinical practice in primary care manual therapy in The Netherlands. The authors used the most recent methods in terms of a priori model sample size calculation, develop- ment, and internal validation. After internal validation, the authors presented penalized full models for the models that demonstrated acceptable performance. Implications for practice and further research / Physiotherapy 113 (2021) 61–72 70 suit to avoid key methodological shortcomings and therefore likely obtaining models that are less overﬁtted than the large majority of those developed for neck pain so far [15]. Addi- tionally, the post-treatment pain and perceived improvement models exhibited also precise optimism-adjusted AUCs of 0.67 with small 95% CI widths of 0.03 and 0.04, respectively. The authors strongly believe there is room to expand models’ performance by updating these models with other predictors that were not evaluated in the ANIMO cohort (e.g., clinical examination ﬁndings). eral population: the HUNT study. Pain 2013;154:1237–44, http://dx.doi.org/10.1016/j.pain.2013.03.032. [4] Hurwitz EL, Randhawa K, Yu H, Côté P, Haldeman S. The global spine care initiative: a summary of the global burden of low back and neck pain studies. Eur Spine J 2018:1–6, http://dx.doi.org/10.1007/s00586-017-5432-9. [5] Babatunde OO, Jordan JL, Van der Windt DA, Hill JC, Foster NE, Protheroe J. Effective treatment options for musculoskeletal pain in primary care: a systematic overview of current evidence. PLoS One 2017;12:e0178621, http://dx.doi.org/10.1371/journal.pone.0178621. [6] Coulter ID, Crawford C, Vernon H, Hurwitz EL, Khorsan R, Booth MS, et al. Manipulation and mobilization for treating chronic nonspeciﬁc neck pain: a systematic review and meta-analysis for an appropriateness panel. Pain Physician 2019;22:E55–70. The models’ relatively low explained variances indicate potential for improvement with relevant predictors that are still missing and literature knowledge seems to provide us only limited information. Further research on new predictors that can strengthen the models is needed. Furthermore, the authors suggest research on predictors of treatment effect (e.g. by randomized controlled trials), since they could not be accounted for in this single cohort study design. Speciﬁcally, causally related modiﬁable factors have potential to change patient outcome [8]. [7] Gross A, Langevin P, Burnie SJ, Bédard-Brochu M-S, Empey B, Dugas E, et al. Manipulation and mobilisation for neck pain contrasted against an inactive control or another active treatment. Cochrane Database Syst Rev 2015;(9):CD004249, http://dx.doi.org/10.1002/14651858.CD004249.pub4. [8] Riley RD, van der Windt DA, Croft P, Moons KGM. Prognosis research in health care, concepts, methods, and impact. 1st ed. Oxford: Oxford University Press; 2019. [9] Artus M, Campbell P, Mallen CD, Dunn KM, van der Windt DAW. Generic prognostic factors for musculoskeletal pain in primary care: a systematic review. BMJ Open 2017;7:e012901, http://dx.doi.org/10.1136/bmjopen-2016-012901. Ethics approval: Obtained from Erasmus Medical Centre, Rotterdam, the Netherlands (Ref. No. MEC-2007-359). Ethics approval: Obtained from Erasmus Medical Centre, Rotterdam, the Netherlands (Ref. No. MEC-2007-359). Implications for practice and further research [10] Carroll LJ, Hogg-Johnson S, van der Velde G, Haldeman S, Holm LW, Carragee EJ, et al. Course and prognostic factors for neck pain in the general population. Spine (Phila Pa 1976) 2008;33:S75–82 https://doi.org/10.1097/BRS.0b013e31816445be. Data was deidentiﬁed and the study complies with the Declaration of Helsinki. Funding: This study was partly funded by a program grant of the Dutch Arthritis Foundation. The funder played no role in the design, conduct, or reporting of this study. [11] Walton DM, Carroll LJ, Kasch H, Sterling M, Verhagen AP, Mac- dermid JC, et al. An overview of systematic reviews on prognostic factors in neck pain: results from the International Collaboration on Neck Pain (ICON) project. Open Orthop J 2013;7:494–505 https://doi.org/10.2174/1874325001307010494. Conﬂict of interest: None declared. Conﬂict of interest: None declared. [12] Moons KGM, Royston P, Vergouwe Y, Grobbee DE, Altman DG. Prog- nosis and prognostic research: what, why, and how? BMJ 2009;338, http://dx.doi.org/10.1136/bmj.b375, b375–b375. Availability of data: The dataset used and analyzed during the current study is available through SOMT upon reasonable request. [13] Bordini BJ, Stephany A, Kliegman R. Overcoming diagnos- tic errors in medical practice. J Pediatr 2017;185:19–25.e1, http://dx.doi.org/10.1016/j.jpeds.2017.02.065. [14] Chifﬁ D, Zanotti R. Fear of knowledge: clinical hypotheses in diagnostic and prognostic reasoning. J Eval Clin Pract 2016:1–7, http://dx.doi.org/10.1111/jep.12664. Appendix A. Supplementary data [15] Wingbermühle RW, van Trijffel E, Nelissen PM, Koes B, Verha- gen AP. Few promising multivariable prognostic models exist for recovery of people with non-speciﬁc neck pain in musculoskele- tal primary care: a systematic review. J Physiother 2018;64:16–23, http://dx.doi.org/10.1016/j.jphys.2017.11.013. Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j. physio.2021.05.011. g j j y [16] Wingbermühle RW, Heymans MW, Trijffel E van, Koes B, Arianne P. Verhagen. External validation study of three promising models for prediction of neck pain recovery. Submitted n.d. Implications for practice and further research Recovery is a multidimensional construct and clinical guidelines usually promote the use of several outcome measures simultaneously [58]. For this reason, the authors propose that, if all adequately performing during external validation, the future potential clinical use will be of all the three separate models developed in this study. The post- treatment models for prediction of recovery in patients with non-speciﬁc neck pain, especially the disability model, have good potential for clinical use. The post-treatment disability model can inform clinicians at intake about patient’s indi- vidual prognosis after therapy. To illustrate this for an intake situation where a physiotherapist wants to inform a neck pain patient about his or her speciﬁc prognosis: “based on this model and you being 30 years of age, having 10 weeks neck pain duration, a 7/50 NDI score, sleeping problems and a 4/24 FABQ-PA score, the authors expect there is a 35% chance you will not be recovered post-treatment (or vice-versa a 65% chance that you will be recovered after treatment). However, before clinical use can be promoted, the authors suggest post- treatment models’ further external validation, especially the disability model. The post-treatment disability model derived in our study showed precise optimism-adjusted AUC of 0.74 with small 95% CI width of 0.03. The authors argue this is a promising value for external validation, given our pur- p p The main limitation of this study is the cohort’s missing data, especially for the outcome variables. The high dropout can be explained by the fact that participants returned out- come questionnaire booklets by post that had to be number marked by themselves and when this was missing the book- lets could be labelled at their arrival by the researchers. However, the labels with the patient number on them were frequently lost or separated from the booklets and then the total questionnaire information could not be used anymore. The authors think due to these reasons that the underly- ing missing data mechanism tends towards an MCAR and MAR mechanism but certainly not MNAR. Also, because the majority of predictors are shared by MI and complete case data, especially for one-year follow-up and baseline charac- teristics were comparable between complete cases and those without outcome data. As recommended in the literature [41], missing value analysis was conducted and multivariable mul- tiple imputation on predictors as well as outcomes with an R.W. Wingbermühle et al. References [24] Bruls VEJ, Bastiaenen CHG, de Bie RA. Prognostic factors of complaints of arm, neck, and/or shoulder. Pain 2015;156:765–88, http://dx.doi.org/10.1097/j.pain.0000000000000117. [41] Lee KJ, Roberts G, Doyle LW, Anderson PJ, Carlin JB. Multi- ple imputation for missing data in a longitudinal cohort study: a tutorial based on a detailed case study involving imputation of missing outcome data. Int J Soc Res Methodol 2016;19:575–91, http://dx.doi.org/10.1080/13645579.2015.1126486. [25] Carroll LJ, Hogg-Johnson S, Van Der Velde G, Haldeman S, Holm LW, Carragee EJ, et al. Course and prognostic factors for neck pain in the general population: results of the bone and joint decade 2000–2010 Task Force on Neck Pain and its associated disorders. Spine (Phila Pa 1976) 2008;33:S75–82. [26] Landers MR, Creger RV, Baker CV, Stutelberg KS. The use of fear-avoidance beliefs and nonorganic signs in predicting prolonged disability in patients with neck pain. Man Ther 2008;13:239–48, http://dx.doi.org/10.1016/j.math.2007.01.010. [42] Marshall A, Altman DG, Holder RL, Royston P. Combining estimates of interest in prognostic modelling studies after multiple imputation: current practice and guidelines. BMC Med Res Methodol 2009;9:1–8, http://dx.doi.org/10.1186/1471-2288-9-57. [27] Lundberg M, Grimby-Ekman A, Verbunt J, Simmonds MJ. Pain-related fear: a critical review of the related measures. Pain Res Treat 2011;2011, http://dx.doi.org/10.1155/2011/494196. [43] Heymans MW. R package psfmi: Predictor Selection Functions for Logistic and Cox regression models in multiply imputed datasets; 2019, 0.1.0 2019. [28] Geri T, Piscitelli D, Meroni R, Bonetti F, Giovannico G, Traversi R, et al. Rasch analysis of the Neck Bournemouth Questionnaire to measure disability related to chronic neck pain. J Rehabil Med 2015;47:836–43, http://dx.doi.org/10.2340/16501977-2001. [44] Altman DG, Royston P. The cost of dichotomising continuous variables. BMJ 2006;332:1080, http://dx.doi.org/10.1136/bmj.332.7549.1080. [45] Collins GS, Ogundimu EO, Cook JA, Le Manach Y, Altman DG. Quantifying the impact of different approaches for handling contin- uous predictors on the performance of a prognostic model. Stat Med 2016;35:4124–35, http://dx.doi.org/10.1002/sim.6986. [29] Geri T, Signori A, Gianola S, Rossettini G, Grenat G, Checchia G, et al. Cross-cultural adaptation and validation of the Neck Bournemouth Questionnaire in the Italian population. Qual Life Res 2015;24:735–45, http://dx.doi.org/10.1007/s11136-014-0806-5. [46] Royston P, Altman DG, Sauerbrei W. Dichotomizing continuous pre- dictors in multiple regression: a bad idea. Stat Med 2006;25:127–41, http://dx.doi.org/10.1002/sim.2331. [30] Schmitt MA, de Wijer A, van Genderen FR, van der Graaf Y, Helders PJ, van Meeteren NL. The Neck Bournemouth Question- naire cross-cultural adaptation into Dutch and evaluation of its psychometric properties in a population with subacute and chronic whiplash associated disorders. References Spine (Phila Pa 1976) 2009;34:2551–61, http://dx.doi.org/10.1097/BRS.0b013e3181b318c4. [47] Moons KGM, Altman DG, Reitsma JB, Ioannidis JPA, Macaskill P, Steyerberg EW, et al. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): explanation and elaboration. Ann Intern Med 2015;162:W1–73, http://dx.doi.org/10.7326/M14-0698. [31] Palmlöf L, Holm LW, Alfredsson L, Skillgate E. Expectations of recovery: a prognostic factor in patients with neck pain under- going manual therapy treatment. Eur J Pain 2016;20:1384–91, http://dx.doi.org/10.1002/ejp.861. [48] Hosmer DW, Lemeshow S, Sturdivant RX. Applied logistic regression. 3rd ed. Wiley; 2013. [49] Steyerberg EW, Harrell FE, Borsboom GJJM, Eijkemans MJC, Vergouwe Y, Habbema JDF. Internal validation of predictive models: efﬁciency of some procedures for logis- tic regression analysis. J Clin Epidemiol 2001;54:774–81, http://dx.doi.org/10.1016/S0895-4356(01)00341-9. [32] Wright Aa, Hensley Cp, Gilbertson J, Leland Jm, Jackson S. Deﬁning patient acceptable symptom state thresholds for commonly used patient reported outcomes measures in general orthopedic practice. Man Ther 2015;20:814–9, http://dx.doi.org/10.1016/j.math.2015.03.011. [50] Van Calster B, Nieboer D, Vergouwe Y, De Cock B, Pencina MJ, Steyerberg EW. A calibration hierarchy for risk models was deﬁned: from utopia to empirical data. J Clin Epidemiol 2016;74:167–76, http://dx.doi.org/10.1016/j.jclinepi.2015.12.005. [33] Sterling M, Jull G, Kenardy J. Physical and psychological factors maintain long-term predictive capacity post-whiplash injury. Pain 2006;122:102–8. [34] MacDermid JC, Walton DM, Avery S, Blanchard A, Etruw E, McAlpine C, et al. Measurement properties of the neck disability index: a systematic review. J Orthop Sports Phys Ther 2009;39:400–17, http://dx.doi.org/10.2519/jospt.2009.2930. [51] Riley RD, Snell KIE, Ensor J, Burke DL, Harrell Jr FE, Moons KGM, et al. Minimum sample size for developing a multivariable predic- tion model: PART II—binary and time-to-event outcomes. Stat Med 2019;38:1276–96, http://dx.doi.org/10.1002/sim.7992. [35] Schellingerhout JM, Heymans MW, Verhagen AP, Lewis M, de Vet HCW, Koes BW. Prognosis of patients with non- speciﬁc neck pain. Spine (Phila Pa 1976) 2010;35:E827–835, http://dx.doi.org/10.1097/BRS.0b013e3181d85ad5. [52] Ailliet L, Knol DL, Rubinstein SM, De Vet HCW, Van Tulder MW, Terwee CB. Deﬁnition of the construct to be measured is a prerequisite for the assessment of validity. The neck disabil- ity index as an example. J Clin Epidemiol 2013;66:775–82.e2, http://dx.doi.org/10.1016/j.jclinepi.2013.02.005. [36] Kamper SJ, Ostelo RWJG, Knol DL, Maher CG, de Vet HCW, Han- cock MJ. Global perceived effect scales provided reliable assessments of health transition in people with musculoskeletal disorders, but ratings are strongly inﬂuenced by current status. J Clin Epidemiol 2010;63:760–6.e1, http://dx.doi.org/10.1016/j.jclinepi.2009.09.009. [53] Chiarotto A, Maxwell LJ, Ostelo RW, Boers M, Tugwell P, Terwee CB. References [17] Peters R, Mutsaers B, Verhagen AP, Koes BW, Pool-Goudzwaard AL. Prospective cohort study of patients with neck pain in a manual therapy setting: design and baseline measures. J Manipulative Physiol Ther 2019;42:471–9, http://dx.doi.org/10.1016/j.jmpt.2019.07.001. [1] Abajobir AA, Abate KH, Abbafati C, Abbas KM, Abd-Allah F, Abdulkader RS, et al. Global, regional, and national disability- adjusted life-years (DALYs) for 333 diseases and injuries and healthy life expectancy (HALE) for 195 countries and territories, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 2017;390:1260–344, http://dx.doi.org/10.1016/S0140-6736(17)32130-X. [18] Steyerberg E, Moons KGM, van der Windt D, Hayden J, Perel P, Schroter S, et al. Prognosis research strategy (PROGRESS) series 3: prognostic models. Br Med J 2012;10, http://dx.doi.org/10.1371/jour-. [19] Collins GS, Reitsma JB, Altman DG, Moons KGM. Transparent reporting of a multivariable prediction model for individual progno- sis or diagnosis (TRIPOD): the TRIPOD statement. Ann Intern Med 2015;162:55, http://dx.doi.org/10.7326/M14-0697. [2] Hush JM, Lin CC, Michaleff ZA, Verhagen A, Refshauge KM. Prognosis of acute idiopathic neck pain is poor: a systematic review and meta-analysis. Arch Phys Med Rehabil 2011;92:824–9, http://dx.doi.org/10.1016/j.apmr.2010.12.025. [20] Heinze G, Wallisch C, Dunkler D. Variable selection—a review and recommendations for the practicing statistician. Biom J 2018:1–19, http://dx.doi.org/10.1002/bimj.201700067. [3] Vasseljen O, Woodhouse A, Bjørngaard JH, Leivseth L. Natural course of acute neck and low back pain in the gen- R.W. Wingbermühle et al. / Physiotherapy 113 (2021) 61–72 71 tation of missing values. J Clin Epidemiol 2006;59:1087–91, http://dx.doi.org/10.1016/j.jclinepi.2006.01.014. [21] Moons KGM, Wolff RF, Riley RD, Whiting PF, Westwood M, Collins GS, et al. PROBAST: a tool to assess risk of bias and applicability of prediction model studies: explanation and elaboration. Ann Intern Med 2019;170:W1, http://dx.doi.org/10.7326/M18-1377. [38] Schafer JL, Graham JW. Missing data: our view of the state of the art. Psychol Methods 2002;7:147–77, http://dx.doi.org/10.1037/1082-989X.7.2.147. [22] Royston P, Moons KGM, Altman DG, Vergouwe Y. Progno- sis and prognostic research: developing a prognostic model. BMJ 2009;338:1373–7, http://dx.doi.org/10.1136/bmj.b604. [39] Vergouwe Y, Royston P, Moons KGM, Altman DG. Develop- ment and validation of a prediction model with missing predictor data: a practical approach. J Clin Epidemiol 2010;63:205–14, http://dx.doi.org/10.1016/j.jclinepi.2009.03.017. [23] Steyerberg EW, Vergouwe Y. Towards better clinical prediction models: seven steps for development and an ABCD for validation. Eur Heart J 2014;35:1925–31, http://dx.doi.org/10.1093/eurheartj/ehu207. [40] Sterne JAC, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, et al. Multiple imputation for missing data in epidemiologi- cal and clinical research: potential and pitfalls. BMJ 2009;338:1–10, http://dx.doi.org/10.1136/bmj.b2393. References Measurement properties of visual analogue scale, numeric rating scale, and pain severity subscale of the brief pain inventory in patients with low back pain: a systematic review. J Pain 2019;20:245–63, http://dx.doi.org/10.1016/j.jpain.2018.07.009. [37] Donders ART, van der Heijden GJMG, Stijnen T, Moons KGM, Van Der Heijden G. Review: a gentle introduction to impu- [54] Kelly J, Ritchie C, Sterling M. Clinical prediction rules for prognosis and treatment prescription in neck pain: a sys- R.W. Wingbermühle et al. / Physiotherapy 113 (2021) 61–72 72 tematic review. Musculoskelet Sci Pract 2017;27:155–64, http://dx.doi.org/10.1016/j.math.2016.10.066. model for recovery. BMC Musculoskelet Disord 2012;13:264, http://dx.doi.org/10.1186/1471-2474-13-264. [57] Kontopantelis E, White IR, Sperrin M, Buchan I. Outcome-sensitive multiple imputation: a simulation study. BMC Med Res Methodol 2017;17:1–13, http://dx.doi.org/10.1186/s12874-016-0281-5. [55] Ritchie C, Hendrikz J, Kenardy J, Sterling M. Derivation of a clini- cal prediction rule to identify both chronic moderate/severe disability and full recovery following whiplash injury. Pain 2013;154:2198–206, http://dx.doi.org/10.1016/j.pain.2013.07.001. [58] Hush JM, Refshauge K, Sullivan G, De Souza L, Maher CG, McAuley JH. Recovery: what does this mean to patients with low back pain? Arthritis Care Res (Hoboken) 2008;61:124–31, http://dx.doi.org/10.1002/art.24162. [56] Bohman T, Cote P, Boyle E, Cassidy JD, Carroll LJ, Skill- gate E. Prognosis of patients with whiplash-associated dis- orders consulting physiotherapy: development of a predictive Available online at www.sciencedirect.com
https://openalex.org/W2508040471	https://www.nature.com/articles/srep32783.pdf	English	null	miR-762 promotes porcine immature Sertoli cell growth via the ring finger protein 4 (RNF4) gene	Scientific reports	2,016	cc-by	8,430	miR-762 promotes porcine immature Sertoli cell growth via the ring finger protein 4 (RNF4) gene Changping Ma1, Huibin Song1, Lei Yu1, Kaifeng Guan1, Pandi Hu1, Yang Li1, Xuanyan Xia2, Jialian Li1, Siwen Jiang1,3 & Fenge Li1,3 received: 17 May 2016 accepted: 15 August 2016 Published: 06 September 2016 A growing number of reports have revealed that microRNAs (miRNAs) play critical roles in spermatogenesis. Our previous study showed that miR-762 is differentially expressed in immature and mature testes of Large White boars. Our present data shows that miR-762 directly binds the 3′ untranslated region (3′UTR) of ring finger protein 4 (RNF4) and down-regulates RNF4 expression. A single nucleotide polymorphism (SNP) in the RNF4 3′UTR that is significantly associated with porcine sperm quality traits leads to a change in the miR-762 binding ability. Moreover, miR-762 promotes the proliferation of and inhibits apoptosis in porcine immature Sertoli cells, partly by accelerating DNA damage repair and by reducing androgen receptor (AR) expression. Taken together, these findings suggest that miR-762 may play a role in pig spermatogenesis by regulating immature Sertoli cell growth. Spermatogenesis is a highly integrated process in which spermatogonia undergo mitotic proliferation once to produce primary spermatocytes, and diploid primary spermatocytes undergo meiosis twice to form haploid spermatids. These haploid spermatids then differentiate into spermatozoa1. These events occur in the testicular seminiferous epithelium and require accurate gene expression, which is strictly controlled at the transcriptional and post-transcriptional levels2,3. MicroRNAs (miRNAs) are a class of small endogenous noncoding RNAs (approximately 22 nt long). miRNAs bind to the 3′ untranslated region (3′UTR) or coding sequence region (CDS) of their target messenger RNAs (mRNAs), resulting in the degradation (perfect complementarity) or translational inhibition (incomplete match) of the mRNAs4,5. Accumulating evidence has indicated that miRNAs participate in numerous biological processes, including nervous and cardiovascular system development, cell proliferation and death, and spermatogenesis6–8.if g y p p p g We have identified 129 miRNAs that are differentially expressed in immature and mature testes of Large White boars9. Several of these miRNAs have already been reported to play essential roles in spermatogenesis. For example, miR-17-92 disruption in the testes of adult mice led to severe testicular atrophy, empty seminiferous tubules, and depressed sperm production, these effects were partly due to the activation of the mTOR signalling pathway10. miR-133b is up-regulated in Sertoli cells of Sertoli-cell-only syndrome patients and promotes human Sertoli cell proliferation by targeting the transcription factor GLI family zinc finger 3 (GLI3) and activating Cyclin B1 and Cyclin D111. miR-383 is down-regulated in the testes of infertile men with maturation arrest12. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports received: 17 May 2016 accepted: 15 August 2016 Published: 06 September 2016 miR-762 promotes porcine immature Sertoli cell growth via the ring finger protein 4 (RNF4) gene Changping Ma1, Huibin Song1, Lei Yu1, Kaifeng Guan1, Pandi Hu1, Yang Li1, Xuanyan Xia2, Jialian Li1, Siwen Jiang1,3 & Fenge Li1,3 miR-383 has been proposed to form a feedback loop with fragile X mental retardation protein (FMRP) to suppress cell proliferation and increase DNA damage, consequently regulating spermatogenesis12. The male mice with both miR-449 cluster and miR-34b/c knocked out are sterile and display low sperm counts, low or no motility, and deformed sperm13. miR-762 is differentially expressed in immature and mature testes of Large White boars9. However, the regulatory mechanism by which miR-762 regulates spermatogenesis is still unknown. y g p g In this study, miR-762 inhibited RNF4 expression. A spontaneous RNF4 c.* 566 T > C mutation in the RNF4 3′UTR was significantly associated with sperm quality traits, possibly due to a change in the binding site number between miR-762 and the RNF4 3′UTR. Furthermore, miR-762 promoted immature Sertoli cell proliferation and inhibited apoptosis not via RNF4 gene directly but possibly by accelerating DNA damage repair and weakening 1Key Laboratory of Pig Genetics and Breeding of Ministry of Agriculture & Key Laboratory of Agricultural Animal Genetics, Breeding and Reproduction of Ministry of Education, Huazhong Agricultural University, Wuhan 430070, PR China. 2College of Informatics, Huazhong Agricultural University, Wuhan 430070, PR China. 3The Cooperative Innovation Center for Sustainable Pig Production, Wuhan 430070, PR China. Correspondence and requests for materials should be addressed to F.L. (email: lifener@mail.hzau.edu.cn) Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 1 www.nature.com/scientificreports/ Figure 1. RNF4 is a target gene of miR-762. (A) Two binding sites for miR-762 in the RNF4 3′UTR were predicted using RNA22. (B) pmirGLO-RNF4-3′UTR was co-transfected into ST cells with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC. Whole cellular lysates were obtained 24 h after transfection, and relative luciferase activity was measured. (C) pmirGLO-RNF4-3′UTR-Mut1 and pmirGLO-RNF4-3′UTR- Mut2 were constructed. Black boxes in the seed-matched sites indicated the sequences that were mutated to abolish the interaction between miR-762 and RNF4 3′UTR. (D) pmirGLO-RNF4-3′UTR-Mut1, pmirGLO- RNF4-3′UTR-Mut2 and pmirGLO-RNF4-3′UTR were co-transfected into ST cells with the miR-762 mimic or NC, and the relative luciferase activity was measured. (E) Endogenous RNF4 mRNA levels were detected in ST cells 24 h after transfection with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC. (F) RNF4 protein levels were detected by western blotting 48 h after transfection with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC. Data are expressed as the means ± SEs of three replicates. P < 0.01. miR-762 promotes porcine immature Sertoli cell growth via the ring finger protein 4 (RNF4) gene Changping Ma1, Huibin Song1, Lei Yu1, Kaifeng Guan1, Pandi Hu1, Yang Li1, Xuanyan Xia2, Jialian Li1, Siwen Jiang1,3 & Fenge Li1,3 Figure 1. RNF4 is a target gene of miR-762. (A) Two binding sites for miR-762 in the RNF4 3′UTR were predicted using RNA22. (B) pmirGLO-RNF4-3′UTR was co-transfected into ST cells with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC. Whole cellular lysates were obtained 24 h after transfection, and relative luciferase activity was measured. (C) pmirGLO-RNF4-3′UTR-Mut1 and pmirGLO-RNF4-3′UTR- Mut2 were constructed. Black boxes in the seed-matched sites indicated the sequences that were mutated to abolish the interaction between miR-762 and RNF4 3′UTR. (D) pmirGLO-RNF4-3′UTR-Mut1, pmirGLO- RNF4-3′UTR-Mut2 and pmirGLO-RNF4-3′UTR were co-transfected into ST cells with the miR-762 mimic or NC, and the relative luciferase activity was measured. (E) Endogenous RNF4 mRNA levels were detected in ST cells 24 h after transfection with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC. (F) RNF4 protein levels were detected by western blotting 48 h after transfection with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC. Data are expressed as the means ± SEs of three replicates. P < 0.01. androgen receptor (AR) transcriptional regulatory activity. Our findings suggest that miR-762 may affect sper- matogenesis by promoting immature Sertoli cell growth. Results miR-762 directly targets the RNF4 3′UTR. miR-762 was identified in our pervious study via a microarray approach as a microRNA that is expressed differentially between porcine sexually immature and mature testes9. miR-762 was predicted to target RNF4 according to a RNA22 online prediction and expression pattern analysis9. Our further prediction showed that 2 putative miR-762-binding sites (455–476 bp and 651–677 bp) are present in the RNF4 3′UTR (Fig. 1A). g To ascertain the relationship between miR-762 and RNF4, a fragment of the RNF4 3′UTR containing the 2 putative miR-762-binding sites was inserted into the dual-luciferase reporter vector pmirGLO. A miR-762 mimic and a miR-762 inhibitor were synthesized to overexpress and suppress, respectively, the expression of miR-762 in ST cells (immature porcine Sertoli cells) (Supplementary Fig. S1). Reporter assay results revealed that RNF4 is a genuine miR-762 target gene because the co-transfection of the reporter plasmid with the miR-762 mimic into ST cells reduced luciferase activity (P < 0.01) and co-transfection of the reporter plasmid with the miR-762 inhibitor into ST cells increased luciferase activity (P < 0.01) (Fig. 1B). Next, two mutated RNF4 3′UTR dual-luciferase reporter vectors were constructed (pmirGLO-RNF4-3′UTR-Mut1 contained 4-base substitutions within the first putative miR-762 seed-matched sequence in 455–476 bp, and pmirGLO-RNF4-3′UTR-Mut2 contained 4-base substitutions within the second putative miR-762 seed-matched sequence in 651–677 bp; Fig. 1C) and transfected into ST cells. miR-762 significantly inhibited the activity of pmirGLO-RNF4-3′UTR-Mut1, which indicated that miR-762 bound to the second putative binding site when the first putative binding site was destroyed. miR- 762 had no apparent inhibitory effect on pmirGLO-RNF4-3′UTR-Mut2 (Fig. 1D), which indicated that miR-762 could not bind to the first putative binding site when the second putative binding site was destroyed. These results demonstrate that the second binding site of miR-762 (651–677 bp) in the RNF4 3′UTR exists. Results Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 2 www.nature.com/scientificreports/ Populations Traits Genotype (μ ± SE) Effect (μ ± SE) CC TC TT Additive Dominance Duroc N 135 44 7 VOL (ml) 197.90 ± 2.76a 186.75 ± 4.83b 178.29 ± 12.11 9.80 ± 6.21 0.67 ± 3.93 SCON (108/ml) 2.67 ± 0.05 2.59 ± 0.10 3.00 ± 0.24 −0.16 ± 0.12 0.12 ± 0.08 MOT (%) 86.10 ± 0.37 86.55 ± 0.65 88.14 ± 1.63 −1.02 ± 0.84 0.29 ± 0.53 ASR (%) 8.55 ± 0.20 8.45 ± 0.35 9.71 ± 0.87 −0.58 ± 0.44 0.34 ± 0.28 Large White N 7 36 80 VOL (ml) 197.00 ± 20.01 190.72 ± 8.82 193.78 ± 5.92 32.70 ± 26.16 30.48 ± 15.70 SCON (108/ml) 2.15 ± 0.35b 3.12 ± 0.15a 2.69 ± 0.10b −1.36 ± 0.46 −0.87 ± 0.28 MOT (%) 89.57 ± 1.68 87.56 ± 0.74 86.54 ± 0.50 4.07 ± 2.24 1.52 ± 1.34 ASR (%) 8.66 ± 1.21 7.89 ± 0.53 7.32 ± 0.36 −1.11 ± 1.54 1.12 ± 0.93 Landrace N 26 79 52 VOL (ml) 199.00 ± 11.58a 190.53 ± 6.64 184.81 ± 8.19b 7.10 ± 7.09 0.69 ± 4.86 SCON (108/ml) 2.52 ± 0.22 2.25 ± 0.13 2.47 ± 0.16 0.03 ± 0.14 0.12 ± 0.09 MOT (%) 84.46 ± 1.34 84.77 ± 0.77 84.56 ± 0.95 −0.05 ± 0.82 −0.13 ± 0.56 ASR (%) 5.92 ± 0.66 6.28 ± 0.38 6.21 ± 0.47 −0.15 ± 0.41 −0.11 ± 0.28 Table 1. Association between RNF4 c.566 T > C and boar sperm quality traits in 3 pig populations. Note: VOL: semen volume per ejaculate; SCON: sperm concentration; MOT: sperm motility; ASR: abnormal sperm rate. a and b, P < 0.05; P < 0.05. N: Number of boars. Table 1. Association between RNF4 c.566 T > C and boar sperm quality traits in 3 pig populations. Note: VOL: semen volume per ejaculate; SCON: sperm concentration; MOT: sperm motility; ASR: abnormal sperm rate. a and b, P < 0.05; *P < 0.05. N: Number of boars. Furthermore, endogenous RNF4 mRNA and protein expression in ST cells were also suppressed and increased by the miR-762 mimic and miR-762 inhibitor, respectively (Fig. 1E,F). These results demonstrated that miR-762 directly targeted the RNF4 3′UTR. Furthermore, endogenous RNF4 mRNA and protein expression in ST cells were also suppressed and increased by the miR-762 mimic and miR-762 inhibitor, respectively (Fig. 1E,F). Results These results demonstrated that miR-762 directly targeted the RNF4 3′UTR. A spontaneous T > C mutation in the RNF4 3′UTR is associated with sperm quality traits. A spontaneous T > C mutation in the RNF4 3′UTR was detected by comparative sequencing between Large White and Chinese Meishan pigs and was named RNF4 c. 566 T > C according to standard mutation nomenclature14. RNF4 c.* 566 T > C was genotyped, and the allele frequencies were analysed in 466 unrelated boars from 3 different pig populations (Supplementary Table S1). The association analysis results showed that the CC boars had more semen volume per ejaculate (VOL) than the TC and TT boars in Duroc and Landrace pigs (P < 0.05) and that the TC boars had a higher sperm concentration (SCON) than the CC and TT boars in Large White pigs (P < 0.05) (Table 1). RNF4 c.* 566 C creates a novel miR-762-binding site. To uncover the mechanism by which RNF4 c.* 566 T > C affects sperm quality traits, we predicted that RNF4 c.* 566 C may lead to a novel miR-762-binding site (557–580 bp) in the RNF4 3′UTR through RNA22 (Fig. 2A). Three dual-luciferase reporter vectors, pmirGLO-RNF4-3′UTR-C, pmirGLO-RNF4-3′UTR-T and pmirGLO-RNF4-3′UTR-C-Mut3 (which contained 3-base substitutions within the putative miR-762 seed-matched sequence in 557–580 bp), were constructed and co-transfected with the miR-762 mimic or negative control (NC) into ST cells with the endogenous genotype TT at RNF4 c.* 566 T > C locus (Supplementary Fig. S2). The relative luciferase activities of the three groups were all significantly reduced compared with the activities of the NC groups (Fig. 2B), suggesting that all three vectors contained miR-762-binding sites. The relative luciferase activity of pmirGLO-RNF4-3′UTR-C was sig- nificantly reduced compared with pmirGLO-RNF4-3′UTR-T, whereas the relative luciferase activities between pmirGLO-RNF4-3′UTR-C-Mut3 and pmirGLO-RNF4-3′UTR-T had no remarkable difference (Fig. 2B). These results demonstrated that an additional miR-762-binding site was created by RNF4 c.* 566 C. miR-762 promotes immature Sertoli cell proliferation and inhibits apoptosis. ST cells were iden- tified as a collection of immature porcine Sertoli cells in our previous report15. The number of immature Sertoli cells is important for spermatogenesis because mature Sertoli cells have no proliferative capacity and their final number depends on immature Sertoli cells16. miR-762 promotes immature Sertoli cell proliferation and inhibits apoptosis. ST cells were iden- tified as a collection of immature porcine Sertoli cells in our previous report15. The number of immature Sertoli cells is important for spermatogenesis because mature Sertoli cells have no proliferative capacity and their final number depends on immature Sertoli cells16. Compared with controls, miR-762-transfected cells displayed an increased growth rate, as monitored by the xCELLigence system (Fig. 3A), and showed a significantly higher mRNA expression of proliferating cell nuclear antigen (PCNA), which is a cell proliferation marker gene (Fig. 3B). Cell cycle analysis revealed that fewer miR- 762-transfected ST cells were detected in the G0/G1 phase (quiescent resting phase/pre-DNA- synthetic phase) but more cells were detected in the S phase (DNA synthesis phase) compared with the controls (Fig. 3C and Supplementary Fig. S3), suggesting that miR-762 drives ST cells into the DNA synthesis phase (G1/S transition) and promotes cell cycle progression. The cell apoptosis assay results showed that the miR-762 mimic group had a lower apoptosis rate than the NC group (Fig. 3D and Supplementary Fig. S4). The effect of the miR-762 inhib- itor on ST cell growth and apoptosis was opposite to the effect caused by the miR-762 mimic (Fig. 3A–D). All of these results demonstrated that miR-762 promoted the proliferation of immature Sertoli cells and inhibited cell apoptosis. Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 3 www.nature.com/scientificreports/ Figure 2. RNF4 c.* 566 C in the RNF4 3′UTR creates a novel miR-762-binding site in ST cells. (A) For allele RNF4 c.* 566 C, 3 binding sites for miR-762 were predicted to be present in the RNF4 3′UTR using RNA22. The red box indicates the novel miR-762-binding site, and the black box indicates the SNP site. (B) pmirGLO-RNF4- 3′UTR-C, pmirGLO-RNF4-3′UTR-T and pmirGLO-RNF4-3′UTR-Mut3 were co-transfected with the miR-762 mimic or NC into ST cells, and relative luciferase activity was measured. Black box in the seed-matched sites indicated the sequences that were mutated to abolish the interaction between miR-762 and the novel binding site in the RNF4 3′UTR. P < 0.05. Figure 2. RNF4 c. 566 C in the RNF4 3′UTR creates a novel miR-762-binding site in ST cells. (A) For allele RNF4 c.* 566 C, 3 binding sites for miR-762 were predicted to be present in the RNF4 3′UTR using RNA22. The red box indicates the novel miR-762-binding site, and the black box indicates the SNP site. miR-762 promotes immature Sertoli cell proliferation and inhibits apoptosis. ST cells were iden- tified as a collection of immature porcine Sertoli cells in our previous report15. The number of immature Sertoli cells is important for spermatogenesis because mature Sertoli cells have no proliferative capacity and their final number depends on immature Sertoli cells16. (B) pmirGLO-RNF4- 3′UTR-C, pmirGLO-RNF4-3′UTR-T and pmirGLO-RNF4-3′UTR-Mut3 were co-transfected with the miR-762 mimic or NC into ST cells, and relative luciferase activity was measured. Black box in the seed-matched sites indicated the sequences that were mutated to abolish the interaction between miR-762 and the novel binding site in the RNF4 3′UTR. P < 0.05. RNF4 promotes immature Sertoli cell proliferation and inhibits apoptosis. miR-762 was crucial for determining the porcine Sertoli cell fate by potentially regulating its target gene RNF4. We transfected RNF4 pcDNA3.1 and RNF4 siRNA into ST cells to overexpress and knock down RNF4, respectively (Supplementary Fig. S5). RNF4 pcDNA3.1-transfected cells displayed an increased growth rate (Fig. 4A) and PCNA mRNA expression level (Fig. 4B). Cell cycle analysis showed that fewer RNF4 pcDNA3.1-transfected ST cells stayed in the G0/G1 phase; however, more cells were driven into the S phase compared with the pcDNA3.1 group (Fig. 4C and Supplementary Fig. S6), suggesting that RNF4 promoted the G1/S transition. The cell apoptosis assay results showed that the RNF4 pcDNA3.1 group had a lower apoptosis rate than the control group (Fig. 4D and Supplementary Fig. S7). The effect of RNF4 siRNA on ST cell growth and apoptosis was opposite to that of RNF4 pcDNA3.1 (Fig. 4A–D). All of these results indicated that miR-762 and RNF4 exerted a similar effect on immature Sertoli cell proliferation and apoptosis. miR-762 reduces AR protein and AR transcriptional regulatory activity. RNF4 is a co-activator in AR-dependent transcription17, and AR is a nuclear transcription factor that plays a vital role in Sertoli cell growth and spermatogenesis18. The miR-762 mimic decreased the AR protein level in ST cell nuclei, and the miR-762 inhibitor increased the AR protein level in ST cell nuclei (Fig. 5A). In cell nuclei, AR recognizes and binds to androgen response elements (AREs) in the promoter regions of androgen-regulated genes, and it regu- lates the transcriptional activities of these genes19. The AR transcriptional regulatory activity influenced by miR- 762 was checked by testing the AR-responsive genes cyclin B1 (CCNB1), FK506 binding protein 5 (FKBP5) and transmembrane protease, serine 2 (TMPRSS2)20–22. The miR-762 mimic up-regulated CCNB1 mRNA, which was down-regulated by AR, and down-regulated both FKBP5 and TMPRSS2 mRNA, which were up-regulated by AR (Fig. 5B). The results showed that miR-762 reduced AR transcriptional regulatory activity. miR-762 promotes DNA damage repair. miR-762 promotes immature Sertoli cell proliferation and inhibits apoptosis. ST cells were iden- tified as a collection of immature porcine Sertoli cells in our previous report15. The number of immature Sertoli cells is important for spermatogenesis because mature Sertoli cells have no proliferative capacity and their final number depends on immature Sertoli cells16. DNA damage response was assessed by testing the protein level of γ-H2AX (phosphorylation of H2AX at Ser139), which is an indirect marker of unrepaired DNA breaks23. The miR-762 mimic reduced the γ-H2AX protein level in ST cell nuclei (Fig. 6A), indicating that miR-762 enhanced DNA damage repair or reduced DNA damage. After DNA was damaged by ultraviolet light, the γ-H2AX protein Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 4 www.nature.com/scientificreports/ Figure 3. miR-762 promotes ST cell proliferation and inhibits apoptosis. ST cells were transfected with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC. (A) Cell growth dynamics were continuously monitored by the xCELLigence system. (B) The mRNA levels of PCNA (a marker gene for cell proliferation) were detected in ST cells 24 h after transfection. (C) Cell cycle phase was analysed 48 h after transfection by propidium iodide flow cytometry. (D) Cell apoptosis was measured 48 h after transfection using Annexin-V/PI staining, followed by flow cytometer analysis. UR: late apoptosis, LR: early apoptosis. P < 0.01, P < 0.05. igure 3. miR-762 promotes ST cell proliferation and inhibits apoptosis. ST cells were transfected with the Figure 3. miR-762 promotes ST cell proliferation and inhibits apoptosis. ST cells were transfected with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC. (A) Cell growth dynamics were continuously monitored by the xCELLigence system. (B) The mRNA levels of PCNA (a marker gene for cell proliferation) were detected in ST cells 24 h after transfection. (C) Cell cycle phase was analysed 48 h after transfection by propidium iodide flow cytometry. (D) Cell apoptosis was measured 48 h after transfection using Annexin-V/PI staining, followed by flow cytometer analysis. UR: late apoptosis, LR: early apoptosis. *P < 0.01, P < 0.05. Figure 3. miR-762 promotes ST cell proliferation and inhibits apoptosis. ST cells were transfected with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC. (A) Cell growth dynamics were continuously monitored by the xCELLigence system. (B) The mRNA levels of PCNA (a marker gene for cell proliferation) were detected in ST cells 24 h after transfection. (C) Cell cycle phase was analysed 48 h after transfection by propidium iodide flow cytometry. (D) Cell apoptosis was measured 48 h after transfection using Annexin-V/PI staining, followed by flow cytometer analysis. UR: late apoptosis, LR: early apoptosis. *P < 0.01, P < 0.05. miR-762 promotes immature Sertoli cell proliferation and inhibits apoptosis. ST cells were iden- tified as a collection of immature porcine Sertoli cells in our previous report15. The number of immature Sertoli cells is important for spermatogenesis because mature Sertoli cells have no proliferative capacity and their final number depends on immature Sertoli cells16. level in the miR-762 mimic group decreased faster than the γ-H2AX protein level in the NC group (Fig. 6B), indicating that miR-762 promoted the DNA damage repair response. Furthermore, DNA damage repair-related genes, including tumour protein p53 binding protein 1 (TP53BP1), mediator of DNA-damage checkpoint 1 (MDC1), poly (ADP-ribose) polymerase 1 (PARP1), ring finger protein 8, E3 ubiquitin protein ligase (RNF8), ring finger protein 168, E3 ubiquitin protein ligase (RNF168) and breast cancer 1, early onset (BRCA1)23,24, showed increased mRNA expression in the cells treated with the miR-762 mimic (Fig. 6C). This finding suggested that miR-762 promoted immature Sertoli cell proliferation and inhibited apoptosis partially due to accelerating DNA damage repair. Discussion Spermatogenesis is a complex process that occurs in the seminiferous tubules of the testis. Sertoli cells, which are the only somatic cells in seminiferous tubules, play a critical role in regulating spermatogenesis because they secrete growth factors to facilitate germ cell development25,26 and provide structural supports and facilitate germ cell movement27. Each Sertoli cell can support limited germ cells; therefore, the Sertoli cell number in seminifer- ous tubules is important for germ cell production28. Immature Sertoli cells determine the final number of mature Sertoli cells16,29. Our results indicated that miR-762 promoted cell proliferation, inhibited cell apoptosis, and accelerated DNA damage repair in immature Sertoli cells, thus suggesting a potential function of miR-762 in pig spermatogenesis. In the present study, miR-762 directly targeted RNF4 3′UTR. A SNP (RNF4 c.* 566 T > C) in RNF4 3′UTR altered the number of miR-762-binding sites. Several reports revealed that SNPs within the miRNA binding sites alter the interactions between miRNAs and mRNAs and are associated with specific biological characteristics30–32. The myostatin (MSTN) allele of Texel sheep was characterized by a G to A transition in the 3′UTR that created a target site for miR-1 and miR-206, which caused the translational inhibition of the MSTN gene and contributed Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 5 www.nature.com/scientificreports/ Figure 4 RNF4 promotes ST cell proliferation and inhibits apoptosis ST cells were transfected with Figure 4. RNF4 promotes ST cell proliferation and inhibits apoptosis. ST cells were transfected with RNF4 pcDNA3.1 or pcDNA3.1 and RNF4 siRNA or siRNA NC. (A) Cell growth dynamics were continuously monitored by the xCELLigence system. (B) The PCNA mRNA levels were detected in ST cells 24 h after transfection. (C) Cell cycle phase was analysed 48 h after transfection by propidium iodide flow cytometry. (D) Cell apoptosis was measured 48 h after transfection using Annexin-V/PI staining, followed by flow cytometer analysis. UR: late apoptosis, LR: early apoptosis. *P < 0.01, P < 0.05. Figure 4. RNF4 promotes ST cell proliferation and inhibits apoptosis. ST cells were transfected with RNF4 pcDNA3.1 or pcDNA3.1 and RNF4 siRNA or siRNA NC. (A) Cell growth dynamics were continuously monitored by the xCELLigence system. (B) The PCNA mRNA levels were detected in ST cells 24 h after transfection. (C) Cell cycle phase was analysed 48 h after transfection by propidium iodide flow cytometry. Discussion pf pi p miR-762 accelerated DNA damage repair in ST cells and caused the cells to quickly engage in normal DNA replication and division, and this finding supported both the cell cycle results demonstrating that miR-762 6 Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 www.nature.com/scientificreports/ / p / promoted ST cell transformation from G0/G1 phase to S phase and the cell proliferation results showing that Figure 5. miR-762 reduces AR protein and AR transcriptional regulatory activity. (A) AR protein levels were detected by western blotting 48 h after transfection with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC in ST cells. (B) The mRNA expression levels of CCNB1, FKBP5 and TMPRSS2 were detected 24 h after transfection with the miR-762 mimic or NC. P < 0.01. Figure 5. miR-762 reduces AR protein and AR transcriptional regulatory activity. (A) AR protein levels were detected by western blotting 48 h after transfection with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC in ST cells. (B) The mRNA expression levels of CCNB1, FKBP5 and TMPRSS2 were detected 24 h after transfection with the miR-762 mimic or NC. P < 0.01. Figure 5. miR-762 reduces AR protein and AR transcriptional regulatory activity. (A) AR protein levels were detected by western blotting 48 h after transfection with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC in ST cells. (B) The mRNA expression levels of CCNB1, FKBP5 and TMPRSS2 were detected 24 h after transfection with the miR-762 mimic or NC. P < 0.01. Figure 5. miR-762 reduces AR protein and AR transcriptional regulatory activity. (A) AR protein levels were detected by western blotting 48 h after transfection with the miR-762 mimic or NC and the miR-762 inhibitor or inhibitor NC in ST cells. (B) The mRNA expression levels of CCNB1, FKBP5 and TMPRSS2 were detected 24 h after transfection with the miR-762 mimic or NC. P < 0.01. promoted ST cell transformation from G0/G1 phase to S phase and the cell proliferation results showing that miR-762 promoted ST cell proliferation. In summary, our study showed that miR-762 promotes porcine immature Sertoli cell proliferation and inhibits apoptosis in a comprehensive way by regulating RNF4, AR and DNA damage repair genes (Fig. 8). We speculate that miR-762 functions in pig spermatogenesis by determining the fate of immature Sertoli cells. Discussion (D) Cell apoptosis was measured 48 h after transfection using Annexin-V/PI staining, followed by flow cytometer analysis. UR: late apoptosis, LR: early apoptosis. *P < 0.01, P < 0.05. to the muscular hypertrophy of Texel sheep33. RNF4 c.* 566 T > C was significantly associated with sperm quality traits in 3 pig populations, possibly due to changing the binding status between miR-762 and RNF4 3′UTR. The genetic effect of RNF4 c.* 566 T > C on SCON is different among the breeds, which is possibly caused by different genetic backgrounds of the three pig breeds. g g p g RNF4 is a nuclear transcription factor that acts as a co-activator in AR-dependent transcription17. In our study, RNF4 co-localized with AR in ST cell nuclei, as observed by immunofluorescence (Fig. 7A). In addition, the AR protein levels increased and decreased when RNF4 was overexpressed and inhibited in ST cells, respectively (Fig. 7B). Furthermore, decreased RNF4 protein weakened the AR transcriptional regulatory activity (Fig. 7C), which was consistent with the effect of miR-762 on AR transcriptional regulatory activity (Fig. 5B). These results demonstrated that miR-762 affected immature Sertoli cells by partially weakening the AR transcriptional regula- tory activity caused by decreased RNF4. y y y AR is a nuclear receptor well known for its role in spermatogenesis as AR is required for complete meiosis, adhesion of spermatids, spermiation and blood-testis barrier formation18,34. The overexpression of AR in imma- ture Sertoli cells caused the cells to become premature and decreased the final mature Sertoli cell number, result- ing in a reduced final germ cell number35. AR inactivation by a daily oral dose of an androgen receptor antagonist beginning at one-week of age caused porcine Sertoli cell proliferation36. In this study, miR-762 promoted imma- ture Sertoli cell proliferation and inhibited apoptosis, which was possibly due to lower AR expression and lower transcriptional regulatory activity. Although Sertoli cell-specific AR-knockout male mice were infertile, the Sertoli cell number was normal18. These findings indicated that low AR levels in immature Sertoli cells could maintain cell proliferation and inhibit cell differentiation at a specific phase. Materials and Methods Animals. All animal procedures were performed according to protocols approved by the Biological Studies Animal Care and Use Committee of Hubei Province, PR China. All of the studies involving animals were con- ducted according to the regulation (No. 5 proclamation of the Standing Committee of Hubei People’s Congress) approved by the Standing Committee of Hubei People’s Congress, PR China. The sample collection was approved by the Ethics Committee of Huazhong Agricultural University with the permit number No. 30700571. The animals used for genotyping and association analysis included Duroc (n = 186), Large White (n = 123) and Landrace (n = 157) pigs, with sperm quality records including semen volume per ejaculate (VOL), sperm concentration (SCON), sperm motility (MOT) and abnormal sperm rate (ASR). The animals were allowed access to food and water ad libitum under normal conditions. SNP analysis. Based on the sus scrofa RNF4 mRNA sequence (NM_001044528.2), primers RNF4-SNP-PF and RNF4-SNP-PR (Supplementary Table S3) were designed to amplify the 3′UTR of the RNF4 gene. The ampli- fied sequences of a Large White pig and a Chinese Meishan pig were aligned using Clustalw2 (http://www.ebi. ac.uk/Tools/msa/clustalw2/) to detect the SNP. Boars were genotyped for RNF4 c.* 566 T > C in the 3′UTR by PCR-RFLP. A primer pair (RNF4-SNP-PF and RNF4-SNP-PR, Supplementary Table S3) was used to amplify a specific fragment of RNF4 3′UTR, the restriction enzyme MbiI (Thermo Scientific, Waltham, MA, USA) was used to digest the PCR products, and the digested products were separated by electrophoresis on a 2% agarose gel. Population association analysis was performed as previously described37. Briefly, the relationships between genotypes and sperm quality traits were assessed using the general linear model (GLM) procedure of SAS version 9.2, and the additive and dominance effects were estimated using the REG procedure of SAS version 9.2. Cell culture and transfection. The swine testicular (ST) cell line (ATCC® CRL-1746™) was isolated from swine foetal testes of 80 to 90 day pigs and was identified as a collection of immature porcine Sertoli cells in our previous report15. ST cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium (HyClone, Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 7 www.nature.com/scientificreports/ Figure 6. miR-762 promotes DNA damage repair. ST cells were transfected with the miR-762 mimic or NC. (A) After transfection for 48 h, γ-H2AX protein levels were detected using western blotting. Materials and Methods (B) After transfection for 24 h, ST cells were irradiated with ultraviolet light for 20 s to cause DNA damage, γ-H2AX protein levels were tested using immunofluorescence after irradiation for 2 h, 4 h, 8 h and 16 h. γ-H2AX was positive for secondary antibody (red), and the cell nuclei were positive for DAPI (blue). Scale bar = 200 μm. (C) The mRNA expression levels of TP53BP1, MDC1, PARP1, RNF8, RNF168 and BRCA1 were detected 24 h after transfection. *P < 0.01, P < 0.05. Figure 6. miR-762 promotes DNA damage repair. ST cells were transfected with the miR-762 mimic or NC. (A) After transfection for 48 h, γ-H2AX protein levels were detected using western blotting. (B) After transfection for 24 h, ST cells were irradiated with ultraviolet light for 20 s to cause DNA damage, γ-H2AX protein levels were tested using immunofluorescence after irradiation for 2 h, 4 h, 8 h and 16 h. γ-H2AX was positive for secondary antibody (red), and the cell nuclei were positive for DAPI (blue). Scale bar = 200 μm. (C) The mRNA expression levels of TP53BP1, MDC1, PARP1, RNF8, RNF168 and BRCA1 were detected 24 h after transfection. *P < 0.01, P < 0.05. Figure 7. RNF4 increases AR protein in ST cell nuclei and regulates AR transcriptional regulatory activity. (A) Immunofluorescence was applied to verify the protein-protein interaction between RNF4 and AR, and the ST cells were immunostained by anti-RNF4 (green) and anti-AR (red) antibody. Cell nuclei were positive for DAPI (blue). Scale bar = 10 μm. (B) AR protein levels were detected 48 h after transfection with RNF4 pcDNA3.1 or pcDNA3.1 and RNF4 siRNA or siRNA NC in ST cells. (C) The mRNA expression levels of CCNB1, FKBP5 and TMPRSS2 were detected 24 h after transfection with RNF4 siRNA or siRNA NC. *P < 0.01, P < 0.05. Figure 7. RNF4 increases AR protein in ST cell nuclei and regulates AR transcriptional regulatory activ Figure 7. RNF4 increases AR protein in ST cell nuclei and regulates AR transcriptional regulatory activity. (A) Immunofluorescence was applied to verify the protein-protein interaction between RNF4 and AR, and the ST cells were immunostained by anti-RNF4 (green) and anti-AR (red) antibody. Cell nuclei were positive for DAPI (blue). Scale bar = 10 μm. Materials and Methods (B) AR protein levels were detected 48 h after transfection with RNF4 pcDNA3.1 or pcDNA3.1 and RNF4 siRNA or siRNA NC in ST cells. (C) The mRNA expression levels of CCNB1, FKBP5 and TMPRSS2 were detected 24 h after transfection with RNF4 siRNA or siRNA NC. *P < 0.01, P < 0.05. Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 8 8 www.nature.com/scientificreports/ Figure 8. Model of the main findings of this study. miR-762 promotes immature Sertoli cells proliferation and inhibits apoptosis by inhibiting RNF4 and AR and by enhancing DNA damage repair capacity, and miR-762 is speculated to function in pig spermatogenesis. Figure 8. Model of the main findings of this study. miR-762 promotes immature Sertoli cells proliferation and inhibits apoptosis by inhibiting RNF4 and AR and by enhancing DNA damage repair capacity, and miR-762 is speculated to function in pig spermatogenesis. Logan, Utah, USA) with 10% (v/v) foetal bovine serum (Gibco, Grand Island, NY, USA) at 5% CO2 and 37 °C. The cells were transfected with the miR-762 mimic, negative control (NC), miR-762 inhibitor and inhibitor NC (Supplementary Table S2) (GenePharma, Shanghai, China) using Lipofectamine 2000 transfection reagent (Invitrogen, Carlsbad, CA, USA). Opti-MEM I Reduced Serum Medium (Gibco, Grand Island, NY, USA) was used to dilute Lipofectamine 2000 and nucleic acids. Logan, Utah, USA) with 10% (v/v) foetal bovine serum (Gibco, Grand Island, NY, USA) at 5% CO2 and 37 °C. The cells were transfected with the miR-762 mimic, negative control (NC), miR-762 inhibitor and inhibitor NC (Supplementary Table S2) (GenePharma, Shanghai, China) using Lipofectamine 2000 transfection reagent (Invitrogen, Carlsbad, CA, USA). Opti-MEM I Reduced Serum Medium (Gibco, Grand Island, NY, USA) was used to dilute Lipofectamine 2000 and nucleic acids. Dual-luciferase reporter assays. The 3′UTR of the pig RNF4 gene was amplified by reverse transcription-polymerase chain reaction (RT-PCR) with the primer pair RNF4-3′UTR-PF and RNF4-3′UTR-PR (Supplementary Table S3) using total RNA extracted from untreated ST cells with the HP Total RNA Kit (Omega Bio-tek, Norcross, GA, USA). The PCR products were directionally cloned into the pmirGLO dual-luciferase miRNA target expression vector (Promega, Madison, WI, USA). Site-directed mutagenesis in the seed regions of the putative miR-762-binding sites in the RNF4 3′UTR were generated using overlap-extension PCR with mutagen primer pairs (RNF4-3′UTR-Mut1-PF and RNF4-3′UTR-Mut1-PR, RNF4-3′UTR-Mut2-PF and RNF4- 3′UTR-Mut2-PR, RNF4-3′UTR-Mut3-PF and RNF4-3′UTR-Mut3-PR; Supplementary Table S3). Materials and Methods ST cells were co-transfected with the 3′UTR luciferase reporter vectors and the miR-762 mimic or miR-762 inhibitor. After 24 h of transfection, luciferase activity was measured using the Dual-Glo Luciferase Assay System (Promega, Madison, WI, USA). Firefly luciferase activity was normalized to the corresponding Renilla luciferase activity. RNF4 pcDNA3.1-expression vector and RNF4 siRNA. Porcine RNF4 CDS was amplified by RNF4-pcDNA3.1-PF and RNF4-pcDNA3.1-PR primers (Supplementary Table S3) and inserted into pcDNA3.1 vector to construct an RNF4 gene overexpression vector. RNF4 double-stranded siRNAs (Supplementary Table S2) were obtained from GenePharma (Shanghai, China). Quantitative (Q)-PCR. Total RNA was extracted from ST cells with the HP Total RNA Kit (Omega Bio-tek, Norcross, GA, USA) and was treated with DNase I (Thermo Scientific, Waltham, MA, USA). The concentration and quality of RNA were assessed with the NanoDrop 2000 (Thermo Scientific, Waltham, MA, USA) and agarose gel electrophoresis. Total RNA was reverse transcribed using the Revert Aid First Strand cDNA Synthesis Kit (Thermo Scientific, Waltham, MA, USA). For miR-762 quantification, stem-loop RT-PCR was executed with U6 small nuclear RNA (snRNA) as an internal control. All primers were designed based on miR-762 sequence col- lected from MirBase release 21 (http://www.mirbase.org/). RT-PCR was used to quantify mRNA (RNF4, CCNB1, FKBP5, TMPRSS2, TP53BP1, BRCA1, MDC1, PARP1, RNF8, RNF168, PCNA) with β-actin as an internal con- trol. qPCR was performed in triplicate using iQ SYBR green Supermix (Bio-Rad, Hercules, CA, USA) and the LightCycler®480 (Roche Applied Science, Penzberg, Upper Bavaria, Germany). Western blotting. Whole-cell protein and cell nuclear protein were extracted using RIPA buffer (Beyotime, Haimen, China) and the Nuclear and Cytoplasmic Protein Extraction Kit (Beyotime, Haimen, China), respec- tively. Sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed, followed by protein transfer to polyvinylidene fluoride membrane (Merk Millipore, Billerica, MA, USA). Primary antibodies for RNF4 (Santa Cruz, CA, USA), AR (BIOSS, Beijing, China), Phospho-Histone H2A.X (Ser139) (Cell Signalling Technology, Danvers, MA, USA) and Histone H3 (Wuhan Guge, Wuhan, China) and secondary antibodies for HRP-labelled rabbit anti-goat IgG (Wuhan Guge, Wuhan, China), HRP-labelled goat anti-rabbit IgG (Wuhan Guge, Wuhan, China) and HRP-labelled goat anti-mouse IgG (Wuhan Guge, Wuhan, China) were used to detect RNF4, AR, γ-H2AX and Histone H3 protein expression. Immunofluorescence. Immunofluorescence was performed as previously described15. The primary antibodies included anti-RNF4 antibody (Santa Cruz, CA, USA), anti-AR antibody (BIOSS, Beijing, China) and anti-Phospho-Histone H2A.X (Ser139) antibody (Cell Signalling Technology, Danvers, MA, USA). Materials and Methods The Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 9 www.nature.com/scientificreports/ secondary antibodies included donkey anti-goat FITC-IgG (Wuhan Guge, Wuhan, China), goat anti-rabbit Cy3-IgG (Wuhan Guge, Wuhan, China) and goat anti-rabbit FITC-IgG (Wuhan Guge, Wuhan, China). secondary antibodies included donkey anti-goat FITC-IgG (Wuhan Guge, Wuhan, China), goat anti-rabbit Cy3-IgG (Wuhan Guge, Wuhan, China) and goat anti-rabbit FITC-IgG (Wuhan Guge, Wuhan, China). DNA damage repair analysis. After transfection with the miR-762 mimic and NC for 24 h, ST cells were irradiated with ultraviolet light (253.7 nm wavelength) for 20 s to cause DNA damage. DNA damage repair response was checked by testing γ-H2AX protein levels via immunofluorescence after irradiation for 2 h, 4 h, 8 h and 16 h. Cell proliferation analysis. ST cells were seeded on a 16-well E-Plate with 5000 cells per well and allowed to grow for 12–24 h. The cells were transfected with miR-762 mimic or miR-762 inhibitor when the cell index reached 1.0–2.0 with three wells per treatment. Cell growth and proliferation were monitored by an xCELLigence RTCA DP instrument (Roche Applied Science, Penzberg, Upper Bavaria, Germany).hl pp g pp y The cell cycle was analysed with the Cell Cycle Detection Kit (KeyGEN BioTECH, Nanjing, China). Briefly, 48 h after transfection, the ST cells were fixed in 70% (v/v) ethanol overnight at −20 °C. Following incubation in 50 mg/ml propidium iodide (PI) for 30 min at 4 °C, the cells were analysed using FACSCalibur Flow Cytometry (Becton Dickinson, Franklin Lakes, NJ, USA) and the ModFit software (Verity Software House). The proliferative index was derived by estimating the proportion of mitotic cells from a total of 20,000 cells. Cell apoptosis analysis. Cell apoptosis was analysed using the Annexin V-FITC Apoptosis Detection Kit (KeyGEN BioTECH, Nanjing, China). Briefly, the ST cells were harvested through trypsinization (with- out EDTA) and washed twice with phosphate-buffered saline (HyClone, Logan, Utah, USA). Following cen- trifugation at 2000 rpm for 5 min, the sediment was re-suspended in 500 μl Binding Buffer and incubated with 5 μl FITC-conjugated Annexin V and 5 μl PI for 15 min. The samples then were analysed by FACSCalibur Flow Cytometry (Becton Dickinson, Franklin Lakes, NJ, USA). Bioinformatics method and statistical analysis. The miR-762 targets were predicted by a computer-aided algorithm RNA22 (https://cm.jefferson.edu/rna22/Interactive/). All results are presented as the mean ± standard error (SE), based on at least three replicates for each treatment. Materials and Methods Two-tailed Student’s t-tests were used for P-value calculations. References Targeted disruption of miR-17-92 impairs mouse spermatogenesis by activating mTOR signaling pathway. Medicine 95, e2713 (2016). 1. Yao, C. et al. MiRNA-133b promotes the proliferation of human Sertoli cells through targeting GLI3. Oncotarget 7, 2201–2219 (2016). 12. Tian, H. et al. The targeting and functions of miRNA-383 are mediated by FMRP during spermatogenesis. Cell Death Dis. 4, (2013). 13. Wu, J. et al. Two miRNA clusters, miR-34b/c and miR-449, are essential for normal brain development, motile ciliogenesis, and spermatogenesis. Proc. Natl. Acad. Sci. USA. 111, E2851–E2857 (2014). p g 4. Den Dunnen, J. & Antonarakis, S. Nomenclature for the description of human sequence variations. Hum. Genet. 109, 121–124 (2001). 15. Ma, C. et al. Characterization of swine testicular cell line as immature porcine Sertoli cell line. In Vitro Cell. Dev. Biol. Anim. 52, 427–433 (2016). ( ) 16. Sharpe, R. M., McKinnell, C., Kivlin, C. & Fisher, J. S. Proliferation and functional maturation of Sertoli cells, and their relevance to disorders of testis function in adulthood. Reproduction 125, 769–784 (2003).ii p 7. Moilanen, A. M. et al. Identification of a novel RING finger protein as a coregulator in steroid receptor-mediated gene transcription Mol. Cell. Biol. 18, 5128–5139 (1998). ( ) 18. Walters, K. A., Simanainen, U. & Handelsman, D. J. Molecular insights into androgen actions in male and female reproductive function from androgen receptor knockout models. Hum. Reprod. Update 16, 543–558 (2010).hi g p p p 19. Riegman, P. H., Vlietstra, R. J., Van der Korput, J. A., Brinkmann, A. Q. & Trapman, J. The promoter of the prostate-specific antigen gene contains a functional androgen responsive element. Mol. Endocrinol. 5, 1921–1930 (1991). g g p 0. Li, Y. et al. Regulation of a novel androgen receptor target gene, the cyclin B1 gene, through androgen-dependent E2F family member switching. Mol. Cell. Biol. 32, 2454–2466 (2012). g 21. Magee, J. A., Chang, L. W., Stormo, G. D. & Milbrandt, J. Direct, androgen receptor-mediated regulation of the FKBP5 gene via a distal enhancer element. Endocrinology 147, 590–598 (2006). 22. Lin, B. et al. Prostate-localized and androgen-regulated expression of the membrane-bound serine protease TMPRSS2. Cancer Res. 59, 4180–4184 (1999). ( ) 3. Polkinghorn, W. R. et al. Androgen receptor signaling regulates DNA repair in prostate cancers. Cancer Discov 3, 1245–1253 (2013) 4. Mok, M. T. & Henderson, B. R. Three-dimensional imaging reveals the spatial separation of γH2AX–MDC1–53BP1 and 23. Polkinghorn, W. R. et al. References 1. Braun, R. E. Post-transcriptional control of gene expression during spermatogenesis. Semin. Cell Dev. Biol. 9, 483–489 (1998). 2 P i M D & N f S i RNA i h i b ildi l f ili J A d l 31 26 33 (2010) g g g 2. Papaioannou, M. D. & Nef, S. microRNAs in the testis: building up male fertility. J. Androl. 31, 26–33 (2010). p g p y ( ) 3. Idler, R. K. & Yan, W. Control of messenger RNA fate by RNA‐binding proteins: an emphasis on mammalian spermatogenesis J. Androl. 33, 309–337 (2012). J , ( ) 4. Bartel, D. P. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116, 281–297 (2004 4. Bartel, D. P. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116, 281–297 (2004).h . Bartel, D. P. MicroRNAs: genomics, biogenesis, mechanism, and g g 5. Tay, Y., Zhang, J., Thomson, A. M., Lim, B. & Rigoutsos, I. MicroRNAs to Nanog, Oct4 and Sox2 coding regions modulate embryonic stem cell differentiation. Nature 455, 1124–1128 (2008).h 5. Tay, Y., Zhang, J., Thomson, A. M., Lim, B. & Rigoutsos, I. MicroRNAs to Nanog, Oct4 and Sox2 coding stem cell differentiation. Nature 455, 1124–1128 (2008).h f 6. Ambros, V. The functions of animal microRNAs. Nature 431, 350–355 (2004). f 6. Ambros, V. The functions of animal microRNAs. Nature 431, 350–355 (2004). H hi K l Mi RNA bi i i i d f i di l ll d l d i PL S O 3 f 6. Ambros, V. The functions of animal microRNAs. Nature 431 6. Ambros, V. The functions of animal microRNAs. Nature 431, 350–355 (2004). 7 Hayashi, K et al MicroRNA biogenesis is required for mouse primordial germ cell development and spermatogenesis PLoS One 3, h 7. Hayashi, K. et al. MicroRNA biogenesis is required for mouse primordial germ cell development and spermatogenesis. PLoS One 3 e1738 (2008).i ( ) 8. Sun, K. & Lai, E. C. Adult-specific functions of animal microRNAs. Nat. Rev. Genet. 14, 535–548 (2013).if 8. Sun, K. & Lai, E. C. Adult-specific functions of animal microRN 8. Sun, K. & Lai, E. C. Adult-specific functions of animal microRNAs. Nat. Rev. Genet. 14, 535–548 (2013).if i 9. Luo, L. et al. Microarray-based approach identifies differentially expressed microRNAs in porcine sexually immature and mature testes. PLoS One 5, e11744 (2010). 10. Xie, R. et al. p g , ( ) 5. He, Z. et al. Gdnf upregulates c-Fos transcription via the Ras/Erk1/2 pathway to promote mouse spermatogonial stem cel proliferation. Stem Cells 26, 266–278 (2008). References Discovery of two potential DAZL gene markers for sperm quality in boars by population association studies. Anim Reprod. Sci. 143, 97–101 (2013). Acknowledgements g We thank Yangxiang Group Corporation for collecting the pig semen samples. This work was supported financially by the National Natural Science Foundation of China (31572362), Key Projects in Doctoral Fund of Ministry of Education of China (20120146110018), National Science R&T Program (2014BAD20B01, 2015BAD03B02), Hubei Science R&T Program (2014BBB008), and Fundamental Research Funds for the Central Universities. g We thank Yangxiang Group Corporation for collecting the pig semen samples. This work was supported financially by the National Natural Science Foundation of China (31572362), Key Projects in Doctoral Fund of Ministry of Education of China (20120146110018), National Science R&T Program (2014BAD20B01, 2015BAD03B02), Hubei Science R&T Program (2014BBB008), and Fundamental Research Funds for the Central Universities. Author Contributions C.M. and F.L. conceived and designed the experiments; C.M. and Y.L. performed the experiments; C.M. and F.L. analysed the data; H.S., L.Y., K.G., P.H., X.X., J.L. and S.J. contributed reagents/materials/analysis tools; and C.M. and F.L. wrote the manuscript. C.M. and F.L. conceived and designed the experiments; C.M. and Y.L. performed the experiments; C.M. and F.L. analysed the data; H.S., L.Y., K.G., P.H., X.X., J.L. and S.J. contributed reagents/materials/analysis tools; and C.M. and F.L. wrote the manuscript. References Androgen receptor signaling regulates DNA repair in prostate cancers. Cancer Discov 3, 1245–1253 (2013). 24. Mok, M. T. & Henderson, B. R. Three-dimensional imaging reveals the spatial separation of γH2AX–MDC1–53BP1 and RNF8–RNF168–BRCA1-A complexes at ionizing radiation-induced foci. Radiother. Oncol. 103, 415–420 (2012). Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 10 www.nature.com/scientificreports/ 6. Hai, Y. et al. The roles and regulation of Sertoli cells in fate determinations of spermatogonial stem cells and spermatogenesis. Semin Cell Dev. Biol. 29, 66–75 (2014). 27. Wang, R. S., Yeh, S., Tzeng, C. R. & Chang, C. Androgen receptor roles in spermatogenesis and fertility: lessons from testicular specific androgen receptor knockout mice. Endocr. Rev. 30, 119–132 (2009).h pi g p k k , ( ) 28. Grimaldi, P., Di Giacomo, D. & Geremia, R. The endocannabinoid system and spermatogenesis. Front Endocrinol (Lausanne) 4, 192 (2013). 8. Grimaldi, P., Di Giacomo, D. & Geremia, R. The endocannabinoid system and spermatogenesis. Front Endocrinol (Lausanne) 4, 19 (2013). 9. Sato, Y. et al. Establishment of adult mouse Sertoli cell lines by using the starvation method. Reproduction 145, 505–516 (2013). . Establishment of adult mouse Sertoli cell lines by using the starva Sato, Y. et al. Establishment of adult mouse Sertoli cell lines by using 30. Brest, P. et al. A synonymous variant in IRGM alters a binding site for miR-196 and causes deregulation of IRGM-depen xenophagy in Crohn’s disease. Nat. Genet. 43, 242–245 (2011). p gy 1. Chen, A. X. et al. Germline genetic variants disturbing the Let-7/LIN28 double-negative feedback loop alter breast cance susceptibility. PLoS Genet. 7, e1002259 (2011). p y 32. Wang, X. et al. Single nucleotide polymorphism in the microRNA-199a binding site of HIF1A gene is associated with pancreatic ductal adenocarcinoma risk and worse clinical outcomes. Oncotarget 7, 13717–13729 (2016).f 3. Clop, A. et al. A mutation creating a potential illegitimate microRNA target site in the myostatin gene affects muscularity in sheep Nat. Genet. 38, 813–818 (2006). 4. O’Hara, L. & Smith, L. B. Androgen receptor roles in spermatogenesis and infertility. Best Pract. Res. Clin. Endocrinol. Metab. 29 595–605 (2015). 35. Hazra, R. et al. Temporal role of Sertoli cell androgen receptor expression in spermatogenic development. Mol. Endocrinol. 27, 12–24 (2013).t 36. Legacki, E., Conley, A. J., Nitta-Oda, B. J. & Berger, T. Porcine Sertoli cell proliferation after androgen receptor inactivation. Biol. Reprod. 92, 93 (2015). p ( ) 7. Ma, C. et al. Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 Additional Information upplementary information accompanies this paper at http://www.nature.com/srepi Supplementary information accompanies this paper at http://www.nature.com/srep Competing financial interests: The authors declare no competing financial interests. Competing financial interests: The authors declare no competing financial interests. Competing financial interests: The authors declare no competing financial interests. to cite this article: Ma, C. et al. miR-762 promotes porcine immature Sertoli cell growth via the ring finger ein 4 (RNF4) gene. Sci. Rep. 6, 32783; doi: 10.1038/srep32783 (2016). How to cite this article: Ma, C. et al. miR-762 promotes porcine immature Sertoli cell growth via the ring finger rotein 4 (RNF4) gene. Sci. Rep. 6, 32783; doi: 10.1038/srep32783 (2016). This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ © The Author(s) 2016 Scientific Reports \| 6:32783 \| DOI: 10.1038/srep32783 11
https://openalex.org/W4390651888	https://journals.iaepan.pl/jue/article/download/3086/3319	Polish	null	Uliczny karnawał czy ludowy sąd? Ludyczne elementy współczesnych protestów miejskich	Journal of Urban Ethnology	2,023	cc-by	4,504	Journal of Urban Ethnology 21/2023 PL ISSN 1429-0618 e-ISSN: 2719-6526 DOI : 10.23858/JUE21.2023.007 Journal of Urban Ethnology 21/2023 PL ISSN 1429-0618 e-ISSN: 2719-6526 DOI : 10.23858/JUE21.2023.007 Journal of Urban Ethnology 21/2023 PL ISSN 1429-0618 e-ISSN: 2719-6526 DOI : 10.23858/JUE21.2023.007 Justyna Laskowska-Otwinowska ORCID: https://orcid.org/0000-0001-5566-7193 Państwowe Muzeum Etnograficzne w Warszawie Abstract The article shows the differences between two types of contemporary street demonstrations, starting from their formal differences. Carnival-type protests, as much as the charivari-type ones, are set in different ide ological backgrounds and have their own development stories. The article describes the characteristics and genesis of these phenomena. Key words: Charivari, carnival, contemporary protests, political anthropology * * * W artykule wskazano różnice między dwoma rodzajami współczesnych demonstracji ulicznych, wychodząc od ich odmienności formalnych. Protesty typu karnawałowego, podobnie jak protesty typu charivari, osa dzone są w odmiennych kontekstach ideowych, a także posiadają własne historie rozwoju. W tekście poda no charakterystykę i genezę tych zjawisk. Słowa kluczowe: Charivari, karnawał, współczesne protesty, antropologia polityczna Odebrano / Received: 22.04.2022 Zaakceptowano / Accepted: 6.06.2023 Odebrano / Received: 22.04.2022 Zaakceptowano / Accepted: 6.06.2023 Odebrano / Received: 22.04.2022 Zaakceptowano / Accepted: 6.06.2023 126 Journal of Urban Ethnology 21/2023 1 Wszystkie wymienione zespoły i artyści mają w Chile status „kultowych” wykonawców muzyki protestu. 2 Tłumaczenie własne. Wstęp W 2019 roku miałam odwiedzić Chile na zaproszenie Fundacji Violety Parry, boha terki narodowej tego kraju, której zaginione cztery piosenki, nagrane w Polsce w 1956 roku, odnalazłam. Jednak gdy wylądowałam w Nowym Jorku, skąd miałam lecieć do San tiago, otrzymałam wiadomość o wybuchu gwałtownych zamieszek w docelowym kraju mojej peregrynacji. Protestom, podobnie jak większości wystąpień ulicznych na świecie towarzyszyły przeróżne dźwięki. Oczywiście muzyka, na przykład taka jak ta ze spotkania z legendarnym zespołem Inti Illimani, będącym symbolem walki przeciw dyktaturze Pi nocheta, a także spontaniczne wykonywanie przez tłum kultowych dla narodu pieśni sprzeciwu, takich wykonawców jak: Victora Jary, Los Prisioneros, Violeta Parra czy Ska-P1. Anita Tijoux to autorka piosenki Antipatriarca – śpiewanej m.in. na największym w histo rii Chile marszu kobiet 8 marca 2020 roku. Inna piosenka Anity Tijoux, Cacerolazo, skom ponowana została w pierwszych dniach zamieszek i opowiada zarówno o wydarzeniach w tym kraju, jak i towarzyszących im nastrojach. Cytuję jej fragmenty: Za dwieście metrów skręć w prawo I uciekaj, nadchodzą „chłopcy” La Moneda jest nasza wzdłuż centrum handlowego, A o godzinie policyjnej cacerolazo To nie 30 pesos, to 30 lat Konstytucji! Przed klaunami stańmy z pokrywką i garnkiem Nadszedł bunt i cacerolazo2. Przytoczona piosenka jest przykładem materiału, który będzie w niniejszym arty kule służył przeglądowi historycznych i współczesnych form protestów ulicznych. i W pracy zweryfikowano dotychczasowe tezy zawarte w literaturze przedmiotu, a dotyczące genezy i klasyfikacji tych zjawisk, w celu zakładanej przeze mnie potrzeby uzupełnienia wyłonionych już kategorii. 2 Tłumaczenie własne. Cacerolazo – charivari – kocia muzyka Cytowana we wstępie chilijska piosenka odnosi się do innej niż śpiew formy za chowań z użyciem dźwięków składających się na muzyczny folklor współczesnego pro testu ulicznego. Jest to bardzo stara i stworzona właśnie na potrzeby wyrażania opinii publicznej tzw. kocia muzyka, czyli kakofoniczne koncerty grane głównie na metalo wych naczyniach kuchennych, zwana w Ameryce Łacińskiej cacerolazo. Zwyczaj ten wy wodzi się ze średniowiecznej ludowej formy sądzenia, znanej w Europie pod regionalnymi nazwami: po hiszpańsku cencerrada, Katzenmusik, czyli właśnie kocia 127 Justyna Laskowska-Otwinowska \| Uliczny karnawał czy ludowy sąd? muzyka w Niemczech, scampanate we Włoszech, a najpowszechniej z francuskiego cha rivari. Wspólnoty o mniejszej liczebności i oddalone od ustanowionego przez władców odgórnie sądownictwa, stosowały ten zwyczaj, aby wyrazić swoją dezaprobatę i potępie nie dla różnych rodzajów naruszania norm moralnych, społecznych i obyczajowych, zwłaszcza cudzołóstwa, bicia współmałżonków (zarówno żon, jak i mężów), czy zbyt szybkiego zawierania ponownego małżeństwa przez wdowy i wdowców, co łamało zwy czajowy nakaz odbycia odpowiednio długiej żałoby. Pochodom prowadzonym pod oknami winowajcy, czasem nawet przez 3 do 7 dni z rzędu, towarzyszył parodystyczny i kakofoniczny śpiew oraz hałas wywołany uderza niem w garnki i patelnie lub w cokolwiek, co było pod ręką: brzęczenie kośćmi i dzwon kami, pohukiwania, dęcie w zwierzęce rogi. W przypadku zbyt wczesnego ponownego ślubu wdowy lub wdowca hałas symbolizował krzyk zmarłego męża lub żony. Podczas występu muzycznego przewinienia stawały się przedmiotem pantomimy, przedstawień teatralnych czy recytatywów, wraz z szyderczą litanią wulgaryzmów i obelg oraz dekla macją listy win w formie rymowanej recytacji zwanej „nominacją”. Ich treść wskazuje, że niestosowane zachowania winowajców obciążało nie tylko ich sumienia, lecz było trak towane jako ujma na honorze całej lokalnej społeczności: Pobił swoją żonę! Pobił swoją żonę! to bardzo wielki wstyd i hańba dla wszystkich, którzy mieszkają w tym miejscu3 (Palmer 2005: 49). Pobił swoją żonę! Pobił swoją żonę! to bardzo wielki wstyd i hańba dla wszystkich, którzy mieszkają w tym miejscu3 (Palmer 2005: 49). Pobił swoją żonę! Pobił swoją żonę! to bardzo wielki wstyd i hańba dla wszystkich, którzy mieszkają w tym miejscu3 (Palmer 2005: 49). Charivari był również wykorzystywany jako sankcja wobec tych, którzy popełnili określone przestępstwa gospodarcze, takie jak blokowanie dróg w celu wyłudzania do datkowych opłat za przejazd, czerpanie zysków w okresach słabych zbiorów, zabranianie tradycyjnego angielskiego prawa feudalnego do zbierania przez najuboższych resztek po żniwach na wszystkich polach wsi. 3 Tłumaczenie własne. Cacerolazo – charivari – kocia muzyka Znajdujemy je też w przypadku nieuczciwych prak tyk konkurencyjnych w obrębie tych samych rzemiosł, czyli przy łamaniu solidarności cechowej. Sankcja ludowej woli sądzenia była tak potężna, że odnotowywano samobój stwa ośmieszanych ofiar. Pomimo zakazu stosowania charivari, najpierw przez Kościół katolicki – już w XVII wieku, a później także przez władze świeckie, np. w Anglii – na mocy ustawy o drogach z 1882 roku, zwyczaj przetrwał, zmniejszając jedynie swą czę stotliwość. Ciekawym współczesnym przypadkiem był charivari piętnujący zażyłość uczuciową księdza ze swoją gospodynią. Wzmianka pochodzi z lat 50. XX w. z baskij skiej prowincji Labourd we Francji (Itcaina 2017: 199). Ostatni zanotowany przez folk lorystę Theo Browna przypadek na terenie Anglii miał miejsca w 1973 roku w jednej z wiosek w Devonshire (Banks 2014: 71). 128 Journal of Urban Ethnology 21/2023 Wraz z europejską emigracją do obu Ameryk zwyczaj zakorzenił się też w Stanach Zjednoczonych, Kanadzie i krajach Ameryki Południowej. Pierwsi francuscy koloniza torzy przenieśli go do swoich osad w Quebecu w połowie XVII wieku. Jeden z najbar dziej znaczących rytuałów miał miejsce 28 czerwca 1683 roku. Po tym, jak wdowa po François Vézier dit Laverdure wyszła ponownie za mąż zaledwie trzy tygodnie po śmier ci męża, mieszkańcy Quebec City przeprowadzili głośne i ostre protesty typu charivari przeciwko nowożeńcom (Palmer 2005: 51). Niektórzy badacze uważają, że zwyczaj ten rozprzestrzenił się z francuskojęzycznych obszarów Kanady ku terenom anglojęzycz nym, w tym na tereny USA. Możliwe jednak, że w Stanach Zjednoczonych zwyczaj pojawił się niezależnie – wszak był powszechny również w społeczeństwie angielskim. Jest on dobrze udokumentowany w Dolinie Hudson od najwcześniejszych dni osadnic twa aż do początku XX wieku (Wermuth 2006: 27-53) W niektórych częściach środko wego zachodu Stanów Zjednoczonych, takich jak Kansas, w latach 1960-1970 zjawisko charivari było kontynuowane, jednak już jedynie jako żart weselny polegający na robie niu hałasu, na przykład przez przywiązywanie krowich dzwonków pod łożem weselnym czy mocowaniu puszek do samochodu nowożeńców, co widzimy dziś na niemal każdej amerykańskiej produkcji filmowej pokazującej pożegnanie pary młodych odjeżdżającej w podróż poślubną. Tradycyjne formy zbiorowego działania nie zniknęły z dnia na dzień, a jednocześnie przeniknęły w obszar życia politycznego w XIX w. Pierwsze udokumentowane protesty polityczne w stylu charivari miały miejsce we Francji w latach 30. XIX wieku, na początku monarchii lipcowej, ze strony przeciwników Ludwika Filipa I. Cacerolazo – charivari – kocia muzyka W przeddzień rewolucji 1848-1849 roku w Niemczech ilość wydarzeń tego typu była tak liczna, że zaczęto oficjal nie mówić o politycznym charivari (Brophy 2017: 29). Składały się nań takie elementy, jak: udawane procesy, odgrywane pogrzeby, jazda na osłach, golenie, palenie lub wieszanie ku kły itp. W XX wieku stosowali go francuscy komuniści w okresie międzywojennym, a od lat 70. występowało na terytorium Kraju Basków, gdzie tutejsi nacjonaliści wykorzystywa li parady charivari jako formę teatru ulicznego, która umożliwiała wyrażanie kwestii poli tycznych, w tym opinii na temat ochrony języka i dziedzictwa baskijskiego. W drugiej połowie XX wieku obserwujemy jeszcze częstsze protesty w stylu charivari. Wspomnijmy choćby „noce garnków” w 1961 roku podczas algierskiej wojny o niepodległość lub chari vari na Filipinach w 1978 roku przeciwko autorytarnemu rządowi prezydenta Ferdinanda Marcosa. XXI wiek przynosi protesty z użyciem kociej muzyki w Argentynie, w latach 2001-2002, 2008 oraz 2012, w Meksyku w 2006, na Islandii w 2008, w Maroku w 2017- 2018, w Ekwadorze i Libanie w 2019, w Brazylii i w Indiach w 2020. Przykłady można by mnożyć, gdyż częstotliwość ich występowania narasta. Narodziny współczesnego protestu Skąd współcześnie taka popularność kociej muzyki towarzyszącej protestom? Forma ta wydaje się odradzać na szeroką skalę po czasie, gdy w wyniku uprzemysłowienia 129 Justyna Laskowska-Otwinowska \| Uliczny karnawał czy ludowy sąd? i rozprzestrzenienia się polityki partyjnej, przedindustrialne wspólnotowe formy protestu wydawały się stopniowo zanikać na rzecz nowoczesnego repertuaru obejmującego bojkoty, petycje, demonstracje i strajki. O takiej przemianie form sprzeciwu piszą szczegółowo au torzy pracy: Protest, popular culture and tradition in modern and contemporary western Euro pe (Favretto, Itcaina 2017). Sugestywnie o genezie nowych form nacisku i o ludowym oporze wobec nich napisał w zawartym tam artykule Stefan Nyzell (Nyzell 2017). W pierwszy weekend października 1937 roku, w miejscowości Eskilstuna w Szwecji odbywał się coroczny targ jesienny, z czym związane było przyzwolenie na upijanie się mężczyzn. Pijaństwo należało wówczas wśród szwedzkich warstw ludowych do elemen tów kultury męskiej. W Eskilstuna od czasu gospodarczego boomu w 1930 roku rozwijał się przemysł metalurgiczny, co wywoływało napływ dużej liczby chętnych do pracy z oko licznych wsi. Zatrudnieni chłopi stawali się inną klasą społeczną – proletariatem, który wzbudzał zainteresowanie szwedzkiej Partii Socjaldemokratycznej (SAP), związków za wodowych, a także kościoła. Organizacje te postrzegały robotników jako chaotyczną zbie raninę ludową, którą należało przekształcić w zdyscyplinowaną klasą robotniczą: „Istniał ścisły związek między ruchem robotniczym, ruchem wstrzemięźliwości [od alkoholu – dop. J.L.-O.], ruchem edukacyjnym i ruchem niezależnego Kościoła, które przyczyniły się do powstania «kultury poważania»”4 (Ambjornsson 1988, za: Stefan Nyzell 2017: 93). Dlatego władze miasta postanowiły tym razem wysłać policjantów do pilnowania trzeź wości świętujących, co spotkało się z gwałtowną i wrogą reakcją tłumu i momentalnie przerodziło w kilkudniowe zamieszki z elementami charivari. l Sytuacja ta jest podwójnym przykładem konfliktu pomiędzy ludową władzą oby czaju i opinii publicznej, zarówno w formie przyzwolenia na świąteczne pijaństwo, jak i w zastosowaniu protestu – charivari, a kształtującą się wówczas koncepcją nowoczesnej, zdyscyplinowanej klasy robotniczej stosującej nowy arsenał broni politycznej. Warto podkreślić, że tę drugą strategię oraz samą koncepcję klasy robotniczej kształtowała w większości… klasa średnia, spośród której wywodzili się XIX-wieczni teoretycy myśli socjalistycznej i komunistycznej. Nowe formy protestów robotniczych odpowiadały ich kryteriom i wyobrażeniom nowej klasy jako zunifikowanej, zdyscyplinowanej i godnej zaufania (Martini 2007: 22, za: Favretto, Fincardi 2017: 159). Obserwatorzy tych zja wisk widzieli w nich połączenie stylu wojskowego ze stylem procesji kościelnych (Perrot 1987: 158, za: Favretto, Fincardi 2017: 159). 4 „There were close link between the labour movement, the temperance movement, the educational move ment and the independent church movement, all of which contributed to this «culture of respectability»” (Ambjornsson 1988, za: Stefan Nyzell 2017: 93). Narodziny współczesnego protestu Historycy polityki sugerują ponadto, że wybrana forma miała być wyraźną sugestią ze strony partii socjalistycznych i związków zawodowych skierowaną do władz państwowych, że organizacje te panują nad ludem, czyli masą, co miało legitymizować ich istnienie (Favretto, Fincardi 2017: 159). Formy protestu ulicznego były zatem przystosowywane do zmieniającej się sytuacji politycznej, 130 Journal of Urban Ethnology 21/2023 którą było pojawienie się struktur władzy nowoczesnego państwa wraz z jego organiza cjami reprezentacyjnymi: partiami i związkami zawodowymi przy jednoczesnym wypie raniu społecznych funkcji tradycyjnych wspólnot. Obok różnic organizacyjnych badacze dopatrują się w tych dwóch formach protestu również poważnej odrębności ideologicz nej; podczas gdy charivari odwołuje się do pojęć moralności i niemoralności, to formy protestu społecznego ruchów robotniczych podnosiły kwestie łamania umowy społecz nej przez pracodawców czy agendy państwowe (della Porta 2017: 259). Jednak pośród tych ostatnich występował ruch, który odwoływał się do starszych, ludowych form protestu. Był nim anarchizm. Jak pisze Marco Manfredi: „Włoski anar chizm, w porównaniu ze wszystkimi innymi ruchami politycznymi związanymi z klasa mi niższymi, był szczególnie biegły w wykorzystywaniu przedindustrialnych form i przejawów kultury popularnej i tradycyjnych metod protestu społecznego”5(Manfredi 2017: 103). W przeciwieństwie do ruchu socjalistycznego liderzy anarchizmu często wywodzili się spośród warstw niższych przez co bardziej utożsamiali się z kulturą ludo wą. „Przepaść społeczna i kulturowa między profesorami, którzy byli publicznymi twa rzami włoskiej Partii Socjalistycznej, a tymi przywódcami «podporządkowanych» była bardzo duża”6 (Manfredi 2017: 113). Agitacja włoskich anarchistów odbywała się pod czas zorganizowanych pikników, amatorskich przedstawień teatralnych i wspólnego śpiewania; trafiała wprost do ludzkich serc, a siłą przyciągającą był luźny, niemal niefor malny charakter tych zgromadzeń. Anarchiści wypracowali styl agitacji oparty głównie na przemówieniach, rzadziej na druku, korzystali z figur retorycznych zaczerpniętych wprost z języka ludowego – dosadnych epitetów i gróźb „pocięcia na kawałki” czy „zar żnięcia jak kurczaka” (Manfredi 2017: 110). Ten ekspresyjny język funkcjonujący w krę gu propagandy anarchistycznej był natomiast nie do przyjęcia dla reprezentantów ruchu socjalistycznego, krytykujących jego inklinacje do sentymentalizmu, chaotycznej narra cji, rebelii i przemocy oraz nadużycie pieśni w miejsce rzeczowej, dyskusji doktrynerskiej (Zoccoli 1907: 365, za Manfredi 2017: 114). Rzeczywiście redagowanie i dystrybucja śpiewników były jedną z głównych aktywności tego nurtu ideologicznego, a wspólne śpiewanie najważniejszą częścią anarchistycznych zgromadzeń ludowych. Dostrzegano bowiem, że śpiew jest najbardziej efektywną formą komunikacji zawierającą nie tylko treści, ale też pożądane emocje. 5 „Italian anarchism, compared to all other political movements connected to the lower classes, was particu larly adept at exploitation pre-industrial forms and expressions of popular culture and traditional meth ods of social protest” (Manfredi 2017: 103). 6 „The social and cultural gap between the professors, who were the public face of the Italian Socialist Party, and these «subnersive» leaders was a very wide one” (Manfredi 2017: 113). 5 „Italian anarchism, compared to all other political movements connected to the lower classes, was particu larly adept at exploitation pre-industrial forms and expressions of popular culture and traditional meth ods of social protest” (Manfredi 2017: 103). 6 „The social and cultural gap between the professors, who were the public face of the Italian Socialist Party, d h b i l d id ” (M f di 2017 113) 7 O procesie transformacji karnawału w polityczne charivari pisali m.in.: David I. Kertzer, Edward D. Thompson, Charles Tilly. 8 Margaret Canovan wypełnia za Claudem Lefortem pojęcie „lud” określeniem „puste miejsce władzy” (Conavan 2008: 152za: Lefort 1986). Narodziny współczesnego protestu Uniwersalność tego przekazu powodowała, że wiele z tych utworów, wbrew złej o nich opinii środowisk socjalizujących, przenikało do folk loru tego ruchu we Włoszech. Przyczynę popularności anarchizmu we Włoszech końca XIX wieku Manfredi upatruje właśnie w wykorzystaniu bogactwa folkloru, dobrze za korzenionego w kulturze ludowej (Manfredi 2017: 106). 131 Justyna Laskowska-Otwinowska \| Uliczny karnawał czy ludowy sąd? 10 Kryzys nie spowodował ich zanikania, ale modyfikację. Przykładem takiej przemiany podstawowych war tości jest według Castellsa odstąpienie przez związki zawodowe od walki klasowej na rzecz domagania się sprawiedliwości społecznej (Castells 2003: 160). Odejście od form protestu charakterystycznego dla starych ruchów społecznych dostrzegł również generał Marcos, legendarny przywódca meksykańskich zapatystów: „Rewolucji generalnie nie postrzega się już zgodnie z socjalistycznymi wzorami realizmu, czyli jako grupę mężczyzn i kobiet stoicko maszerujących za czerwoną, powiewającą flagą w kierunku świetlanej przyszłości. Stała się ona raczej rodzajem karnawału” (Antoniewicz 2010: 44). Wieczne charivari Wydaje się, że niektóre momenty historyczne są szczególnie podatne na pojawianie się ludowego protestu. Pośród czynników sprzyjających wykorzystywaniu tej formy na leżą: niepełna demokracja, brak prawa lub możliwości legalnego głosowania dla wszyst kich, momenty kryzysów czy politycznej nierównowagi i utraty autorytetu władzy państwowej (Favretto 2017: 11), rewolucje czy wspomniany ruch anarchistyczny. Poli tyczne protesty typu charivari, które z terenów wiejskich przeniosły się w XIX wieku na ulice miast nie osądzały już krnąbrnych współmieszkańców czy bezpośrednich zwierzch ników, lecz ogarniały swym buntem ośrodki władzy państwowej, a współcześnie ponad państwowej i globalnej7. Elementem niezmiennym ich miejskich form pozostaje odniesienie protestów do kontekstu moralnego (Thompson 1991: 338) oraz utożsamie nie ich z oporem „podporządkowanych”, który to termin zastępuje od jakiegoś czasu określenie „lud”. Zmiana ta wynika z rozszerzenia określenia na wszystkie warstwy nie mające dostępu do władzy (Rauszer 2021: 26)8. Od ostatnich dziesięcioleci XX wieku, wobec toczącego się procesu globalizacji i przenoszenia władzy w ręce wielkich, międzynarodowych korporacji, lokalne środowi ska polityczne zaczęły coraz wyraźniej tracić zaufanie swoich wyborców. Podczas gdy rządy państwowe stają się już tylko częścią różnych „węzłów” decyzyjnych, rozproszo nych w świecie globalnych instytucji, takich jak korporacje czy Bank Światowy, warstwy na dole starają się koncentrować władzę na poziomie lokalnym jako pozarządowe grupy wspólnotowe czy organizacje etniczne (Lewellen 2010: 206). Pośród metod, którymi posługują się „podporządkowani” są metody sprzeciwu indywidualnego, na przykład od mowa odbycia służby wojskowej lub dezercja z armii oraz działania zorganizowane gru powo. Sięgając po środki, by wyrazić władzę ludu, organizatorzy sprzeciwów aktywują też dawną moc charivari. Wydaje się, że pośród opisanych form współczesnego protestu charivari zasługuje na szczególną uwagę z racji swej wartości historycznej. Jego wyjątkowa trwałość jest efektem uniwersalnej formy, która pozwala na wykorzystywanie w różnych kontekstach, nawet przez aktorów społecznych o odmiennych inklinacjach ideologicznych. Charivari nigdy nie był identyfikowany z konkretnymi przekonaniami politycznymi. Służył za równo chłopom sprzeciwiającym się łamaniu norm wspólnoty w całej Europie Zachod niej aż po wiek XIX, a nawet później, republikanom za króla Ludwika Filipa I czy komunistom w okresie międzywojennym we Francji – jednym słowem zarówno środo wiskom konserwatywnym, jak i lewicowym, nacjonalistom czy zwolennikom internacjo nalizmu. Pierwszy polityczny charivari w Ameryce Południowej wszak odbył się 132 Journal of Urban Ethnology 21/2023 w Chile, gdy demonstranci zastosowali go jako protest wobec pogarszającej się sytuacji gospodarczej za czasów Salvadore Allende! Ten fakt chyba najlepiej ilustruje uniwersal ne znaczenie protestu w stylu kociej muzyki. 9 “The carnival theatre of inversion and mockery provided Italian workers with a highly expressive and com fortably familiar language in which they could forcefully and graphically communicate their desire for change (…)” (Favretto, Fincardi 2017: 156). Wieczne charivari Uniwersalność wyrazu jest też czynnikiem przemawiającym za swobodnym rozpowszechnieniem się charivari w odmiennych kul turowo częściach świata. Powszechne jego zrozumienie wynika z odwołania się do argu mentów emocjonalnych: śmiechu, gniewu, skargi. Część badaczy upatruje przyczyn trwałości i „wiecznych powrotów” charivari w jego karnawałowym charakterze, zaś związek tych dwóch form rytuału oceniany jest jako duża bliskość czy nawet pokrewieństwo (Brophy 2017: 27). Wskazuje się na zasto sowanie charivari, podobnie jak karnawału, w okresach przejścia od przestarzałych ukła dów społecznych ku nowym, co w przypadku protestu politycznego przekłada się na żądanie porzucenia uprzednich, niekorzystnych i zastosowania nowych, pożądanych form rządzenia (Favretto, Fincardi 2017: 155). „(…) teatr inwersji i kpiny zapewnił wło skim robotnikom bardzo ekspresyjny i komfortowo znajomy język, w którym mogli silnie i obrazowo komunikować swoje pragnienie zmiany (…)”9 (Favretto, Fincardi 2017: 156). Zadaniem karnawału jest odtworzenie harmonii w systemie społecznych hierar chii poprzez budowę symbolicznego „świata na opak” (Turner 2010: 177). Proces ten odbywa się w ramach „folkloru zawierającego wizję świata” (Rauszer 2021: 21).i Sądzę jednak, że definiowanie najnowszych form protestów ulicznych jedynie w kontekście ich karnawalizacji jest błędem – chodzi raczej o ich „charivarizację”, co nie jest ze sobą tożsame. „Charivarizacja”, co prawda, wykorzystuje formy karnawałowe, jed nak zarazem wprowadza bezpośrednio skierowany wobec winowajców przekaz sprzeci wu i osądzenia w oparciu o niepisany kodeks moralny, przekształcając „form(y) kolektywnego działania w formy kolektywnego oporu” (Rauszer 2021: 46). Zawiera też mechanizm konstruowania kary, czym różni się zarówno od niegdysiejszej, jak i współ czesnej formy karnawału. Różnice między dwoma wspomnianymi formami protestów odnajdujemy w anali zach nowych ruchów społecznych. Badacze upatrują ich pochodzenie w kontrkulturze, która wyposażyła te ruchy w postawę zdecentralizowaną, zindywidualizowaną i sieciową. Manuel Castells połączył odmienność nowych ruchów także z kryzysem „pionowo zo rientowanych organizacji odziedziczonych po erze przemysłowej10 (Castells 2003: 160). 9 “The carnival theatre of inversion and mockery provided Italian workers with a highly expressive and com fortably familiar language in which they could forcefully and graphically communicate their desire for change (…)” (Favretto, Fincardi 2017: 156). i 133 Justyna Laskowska-Otwinowska \| Uliczny karnawał czy ludowy sąd? Również polscy badacze ruchu alterglobalistycznego podzielają pogląd o kontrkulturo wym pochodzeniu nowych ruchów społecznych, co ma wpływ na formę organizowanych przez nie protestów, umiejscawianych między zabawą a protestem. Idea ta wprowadza do polityki charakterystyczną dla kontrkultury koncepcję przyjemności. „Pojęcia karnawału i karnawalizacji są podstawowym tropem w analizach dotyczących współczesnych ruchów antykapitalistycznych” (Antoniewicz 2010: 45). Wieczne charivari Koncepcja karnawału z jego tworzeniem światów „na opak” odpowiada też postrzeganiu buntu jako rodzaju „uniwersalnej postawy kulturowej towarzyszącej władzy” (Kuligowski, Pomieciński 2012: 8), będącej stałym me chanizmem przywracania wartości zmarginalizowanych (Burszta 2012: 64). Należy więc zgodzić się z Antoniewiczem, że karnawalizacja protestu, charakterystyczna dla końca XX i początku XXI wieku, miała w myśl jeszcze kontrkulturowych zasad służyć „natychmia stowej próbie urzeczywistnienia projektu odmiennej rzeczywistości” (Antoniewicz 2010: 43) poprzez wprowadzenie elementu zabawy. Klasyfikację rodzajów protestu, sporządzoną w początkach XXI wieku, po dostrze żeniu coraz większej roli Internetu dla ruchów społecznych, najczęściej ograniczano do jednej z trzech kategorii: a) opartych na centralnej tożsamości partyjneji b) opartych na afirmacji jednej grupy – separatystycznych i c) opartych na sieciowym modelu wielości – karnawale (Antoniewicz 2010: 46 za Hardt, Negri 2004: 217). g Podsumowując dotychczasowe rozważania, stwierdzić trzeba, że w typologii tej brakuje kategorii protestu zawierającego w sobie element osądzenia, a takim jest właśnie protest charivari. Ta forma odróżnia się wyraźnie na tle pozostałych. Prześledzenie jej historii ujawnia, że rozwijała się równolegle do pozostałych, zaś jej korzenie sięgają do kultury ludowej, w której odnajdujemy coś, co nie mieści się w pochodzącej z kontrkul tury wolnościowej orientacji indywidualistycznej. Jest to potrzeba ekspresji wspólnoto wej. Pomieciński, badając polski ruch alterglobalistyczny, dostrzega ten problem i odnosi potrzebę budowania wspólnej tożsamości w nowych ruchach społecznych do idei wspólnego wroga, jakim jest kapitalizm (Pomiecińki 2012: 138-139). Jednak pożą dane elementy tożsamości wspólnotowej, czyli „zagęszczania sieci międzyjednostko wych i międzygrupowych” oraz dezindywidualizacja jako świadomy wybór (Drozdowski 2012: 125-127) są znacznie wyraźniejsze w charivari. Odmienność tych dwóch form protestu najlepiej zobrazuje porównanie karnawałowego zasypywania policji pluszowy mi miśkami przez uczestników demonstracji przebranych w clownowskie stroje (Anto niewicz 2010: 43) do grup wybijających jeden rytm na pateleniach i pokrywkach, wykrzykujących obelgi podczas protesów typu charivarii, z czym spotkaliśmy się w cza sie wspomnianych na początku protestów w Chile. Także w Polsce, gdzie protesty kobiece sięgnęły pierwszy raz po broń z ludowego arsenału – wulgarny język. Słynne już „Jebać PIS” nie jest anomalią, która dotknęła do brze wychowane Polki, lecz dobitnym dowodem ludowego rodowodu formy ich działań. 134 Journal of Urban Ethnology 21/2023 Co ciekawe – charivari nie występowało w polskiej kulturze ludowej. Dopiero ostatnie wystąpienia kobiet przeciw zaostrzeniu zakazu aborcji wprowadzają ten zwyczaj na na szą scenę polityczną. Wieczne charivari I to właśnie on – cały ten zgiełk charivari, który zatrzymał mnie u progu mojej wielkiej życiowej przygody w Chile w 2019 roku – dosięgnął mnie w Pol sce w 2021 roku. Co ciekawe – charivari nie występowało w polskiej kulturze ludowej. Dopiero ostatnie wystąpienia kobiet przeciw zaostrzeniu zakazu aborcji wprowadzają ten zwyczaj na na szą scenę polityczną. I to właśnie on – cały ten zgiełk charivari, który zatrzymał mnie u progu mojej wielkiej życiowej przygody w Chile w 2019 roku – dosięgnął mnie w Pol sce w 2021 roku. Bibliografia Ambjornsson R. 1988. Den skotsamme arbetaren: ideer oh ideal i ett norrlandskt sagverkssamballe 1880-1930. Stockholm: Carlsson. Antoniewicz P. 2010. Karnawał i codzienność ruchu alterglobalistycznego. [W:] Czas Kultury 4, 42–49. Banks S. 2014. Informal Justice in England and Wales 1760-1914. Woodbridge-Suffolk: Boydell Press. Brophy J. M. 2017. The Modernity of tradition: Popular culture and protest in Ninetheenth- Century Germany. [W:] I. Favretto. i X. Itcaina (red.), Protest, popular culture and tradition in modern and contemporary western. London: Palgrave Macmillan, 21-43. Burszta W. J. 2012. U źródeł buntu. Lata sześćdziesiąte w cyklu śmierci i zmartwychwstania. [W:] W. Kuligowski i A. Pomieciński (red.), Oblicza buntu. Praktyki i teorie sprzeciwu w kul turze współczesnej. Poznań: Wydawnictwo Poznańskie, 45-68.l Castells M. 2003. Galaktyka Internetu. Refleksje nad Internetem, biznesem i społeczeństwem, (przeł. T. Hornowski). Poznań: Rebis. Canovan M. 2008. Lud (przeł. S. Szymański). Warszawa: Sic!l Drozdowski R. 2012. Strategie refleksyjnej dezindywiduacji jako nowa forma postaw kultur opo ru. [W:] W. Kuligowski i A. Pomieciński (red.), Oblicza buntu. Praktyki i teorie sprzeciwu w kulturze współczesnej. Poznań: Wydawnictwo Poznańskie, 119-132. Favretto I. Introduction. 2017. [W:] I. Favretto i X. Itcaina (red.), Protest, popular culture and tra dition in modern and contemporary western. London: Palgrave Macmillan, 1-20. Favretto I. i Fincardi M. 2017. Carnavalesque and charivari Repertoires in 1960s and 1970s Italian Protest. [W:] I. Favretto i X. Itcaina (red.), Protest, popular culture and tradition in modern and contemporary western. London: Palgrave Macmillan, 149-183. Itcaina X. 2017 Popular Justice and Informal Politics: the Charivari In Nineteenth and Twentieth Century France. [W:] I. Favretto i X. Itcaina (red.), Protest, popular culture and tradition in modern and contemporary western. London: Palgrave Macmillan, 185-207. Kerzner, D. 2010. Rytuał, polityka, władza. Warszawa: Oficyna Wydawnicza Volumen.Kuligowski W. i Pomieciński A. 2012. Antropologia wobec buntu. [W:] W. Kuligowski i A. Pomieciński (red.), Oblicza buntu. Praktyki i teorie sprzeciwu w kulturze współczesnej, Poznań: Wydawnictwo Poznańskie, 7-15. Lefort C. 1986. The political forms of modern society: bureaucracy, democracy, totalitarianism, Cambridge: Polity. 135 Justyna Laskowska-Otwinowska \| Uliczny karnawał czy ludowy sąd? Lewellen T. C. 2010. Antropologia polityczna. Wprowadzenie (przeł. A. Dąbrowska, T. Sieczkowski). Kraków: Wydawnictwo UJ. Lewellen T. C. 2010. Antropologia polityczna. Wprowadzenie (przeł. A. Dąbrowska, T. Sieczkowski). Kraków: Wydawnictwo UJ. Manfredi M. 2017. Italian Anarchism and Popular Culture: History of a Close Relationship. [W:] I. Favretto i X. Itcaina (red.), Protest, popular culture and tradition in modern and con temporary western. London: Palgrave Macmillan, 103-123. Martini L. Bibliografia (red.) 2007. L’irruzione delle classi lavoratrici nella storia. [W:] Martini L. Rossa: immagine e communicazione 1848-2006, Milano: Skira, 17-32. Nyzell S. 2017. A fight for right to get drunk: the autumn fair riot in Eskilstuna 1937. [W:] I. Favretto i X. Itcaina (red.), Protest, popular culture and tradition in modern and contemporary western. London: Palgrave Macmillan, 89-102. Palmer B. D. 2005. Discordant Music: Charivaris and Whitecapping in Nineteenth-Century North America. Toronto: Canadian Scholars Press. Perrot M. 1987. Workers on strike: France 1871-1890. New Haven: Yale University Pre Pomieciński A. 2012. Ruch alterglobalny: między buntem a happeningiem. [W:] W. Kuligowski i A. Pomieciński (red.), Oblicza buntu. Praktyki i teorie sprzeciwu w kulturze współczesnej, Poznań: Wydawnictwo Poznańskie, 135-149. (della) Porta D. 2017. Afterwords: Old and New Repertoires of Contention. [W:] I. Favretto i X. Itcaina (red.), Protest, popular culture and tradition in modern and contemporary western. London: Palgrave Macmillan, 249-260. Rauszer M. 2021. Siła podporządkowanych. Warszawa: Wydawnictwo UW. Tilly Ch. 1995. Popular contention in Great Britain, 1758-1834, Cambridge: Harvard University Press. Thompson E. 1991. Customs in Common: Studies in Traditional Popular Culture. Londyn: Merlin Press. Turner V. 2010. Proces rytualny (przeł. E. Dżurak). Warszawa: PIW. Zoccoli E. 1907. L’Anarchia. Gili agitatori, le idee, i fatti. Turin: Bocca. Źródła internetowe Wermuth T. S. 2016. Charivari on the Hudson: Misrule, Disorder, and Festive Play in the Countryside, 1750-1900. The Hudson River Valley Review. Hudson River Valley Institute at Marist College 32 (2), 27-52. https://web.archive.org/web/20180304113242/ http://www.hudsonrivervalley.org/review/pdfs/hrvr_32pt2_online.pdf 5.06.2023. Hardt M, Negri A. 2004. Multitude: War and Democracy in the Age of Empire. New York: Penguin Press, https://selforganizedseminar.files.wordpress.com/2011/07/hardt_negri_multitude.pdf 5.06.2023 Autorka: Dr Justyna Laskowska-Otwinowska e-mail: laskowskaotwinowska@gmail.com Autorka:
https://openalex.org/W2895808928	https://ogsh.abvpress.ru/jour/article/download/350/335	Russian	null	Management of the total glossectomy defect with latissimus dorsi free flap	Opuholi golovy i šei	2,018	cc-by	3,160	Опухоли ГОЛОВЫ и ШЕИ HEAD and NECK tumors 2’2018 Клинический случай Опухоли ГОЛОВЫ и ШЕИ HEAD and NECK tumors 2’2018 Клинический случай Том 8 Vol. 8 Опухоли ГОЛОВЫ и ШЕИ HEAD and NECK tumors 2’2018 Клинический случай For citation: Davudov M.M., Ragimov Ch.R., Akhundov A.A. et al. Management of the total glossectomy defect with latissimus dorsi free flap. Opukholi golovy i shei = Head and Neck Tumors 2018;8(2):77–82 Реконструкция языка свободным лоскутом широчайшей мышцы спины после тотальной глоссэктомии М.М. Давудов1, Ч.Р. Рагимов1, А.А. Ахундов2, Ф.Г. Ираннежат1, Ш.А. Магомедов1, М.Ч. Рагимли1, Д.А. Сафаров2 Д уд , , у д , р , Ш.А. Магомедов1, М.Ч. Рагимли1, Д.А. Сафаров2 1Азербайджанский медицинский университет; Азербайджанская Республика, AZ 1022 Баку, ул. Бакиханова, 23; 2ФГБУ «Национальный медицинский исследовательский центр онкологии им. Н.Н. Блохина» Минздрава России; Россия, 1154784 Москва, Каширское шоссе, 24 1Азербайджанский медицинский университет; Азербайджанская Республика, AZ 1022 Баку, ул. Бакиханова, 23; 2ФГБУ «Национальный медицинский исследовательский центр онкологии им. Н.Н. Блохина» Минздрава России; Россия, 1154784 Москва, Каширское шоссе, 24 Контакты: Магомед Мирабдуллаевич Давудов mahammad_davud@mail.ru Контакты: Магомед Мирабдуллаевич Давудов mahammad_davud@mail.r Удаление языка при распространенных опухолях ведет к нарушению функций глотания, жевания и артикуляции, а также неза- щищенными остаются дыхательные пути. Для их восстановления требуется адекватное восполнение дефицита мышечной мас- сы языка и резецированных тканей дна полости рта, а в некоторых случаях – и боковых отделов нижней челюсти. С различным успехом используют для этого лоскуты большой грудной мышцы, переднебокового отдела бедра, прямой брюшной мышцы, пред- плечья и широчайшей мышцы спины. В статье описан клинический случай реконструкции языка лоскутом широчайшей мышцы спины на сосудистой ножке после тотальной глоссэктомии, а результаты сопоставлены с данными литературы. Состояние трансплантата в послеоперационном периоде оценивалось как по клиническим критериям, так и по данным лазерной доплеровской флоуметрии. Начиная с 5-го дня после операции показатель микроциркуляции повышался, на 14-й день приблизившись к значениям донорского участка. Достаточно большая масса трансплантата позволила создать утолщение над надгортанником, а также закрыть дефект в полости рта. Лоскут широчайшей мышцы спины при реконструкции языка обладает высоким по- тенциалом: его использование обеспечивает относительно хорошее качество артикуляции, восстановление глотания. Ключевые слова: рак языка, глоссэктомия, реконструкция свободным лоскутом широчайшей мышцы спины Для цитирования: Давудов М.М., Рагимов Ч.Р., Ахундов А.А. и др. Реконструкция языка свободным лоскутом широчайшей мыш- цы спины после тотальной глоссэктомии. Опухоли головы и шеи 2018;8(2):77–82. DOI: 10.17650/2222‑1468‑2018‑8‑2‑77‑82 DOI: 10.17650/2222‑1468‑2018‑8‑2‑77‑82 Введение Из анамнеза: в 2014 г. больная была подвергнута ко- лоректальной резекции по поводу карциномы толстой кишки. В послеоперационном периоде прошла курс химио лучевой терапии по стандартному протоколу в онколо- гическом учреждении. Через 9 мес в связи с рецидивом опухоли дополнительно получила курс химиотерапии препаратом авастин. Рак слизистой оболочки полости рта часто локали- зуется в области языка, нередко распространяясь на оро- и гипофарингеальную области, что обусловли- вает большой объем резекции при оперативных вме- шательствах [1]. Язык является наиболее важной структурой, уча- ствующей в оральной фазе глотания, а также способ- ствует продвижению пищевого комка в фарингеальной фазе глотания. Он обеспечивает артикуляцию и защи- щает дыхательные пути. Такая комбинация нескольких функций в одном органе возможна благодаря сложному строению мускулатуры языка и его стратегическому расположению во рту – над надгортанником. Кроме того, язык имеет собственную двигательную, сенсорную и специализированную висцеральную иннервацию. Все это делает его органом, отвечающим за жизненно важ- ные функции, серьезно влияющие на качество жизни. В 2016 г. больная обратилась в клинику с жалобами на припухлость в области языка и в области шеи слева. Исследование биоптатов подтвердило диагноз рака язы- ка. Выполнена левосторонняя гемиглоссэктомия и лево- сторонняя функциональная шейная диссекция. Гистоло- гическое заключение: высокодифференцированная плоскоклеточная карцинома языка, метастазы в 3 лим- фатических узлах, реактивная гиперплазия 6 лимфати- ческих узлов. После операции пациентка прошла курс лучевой терапии на область первичной опухоли и регио- нарных лимфатических узлов. Через 1 год выявлен рецидив опухоли языка, в связи чем больная была госпитализирована. При осмотре по- лости рта на остаточной части языка обнаружено бо- лезненное, кровоточащее при контакте образование инфильтративно-язвенного характера с распростране- нием на слизистую оболочку дна полости рта (рис. 1). Глотание затруднено, с правой стороны шеи выявлена лимфаденопатия. При компьютерной томографии визу- ализировано образование размерами 3,7 × 2,1 см в левой ретромандибулярной области, возможно с распадом, увеличение лимфатических узлов правой стороны шеи и правой поднижнечелюстной области. При тотальной резекции или ампутации языка больные могут принимать жидкую и мягкую пищу, но способность к жеванию, раздавливанию пищи и формированию пищевого комка утрачивается. На- ряду с нарушением функции жевания, существенно страдает акт глотания, остаются незащищенными дыхательные пути, возникает риск попадания в них пищи. Таким образом, пациенту после глоссэктомии жизненно необходима реабилитация, а именно раз- личные реконструктивные операции [2–8] и ортопе- дическое лечение – протезирование языка. Опухоли ГОЛОВЫ и ШЕИ HEAD and NECK tumors 2’2018 Клинический случай Том 8 Vol. 8 Введение При хирургическом вмешательстве требуется адек- ватное восполнение дефицита мышечной массы языка и тканей дна полости рта после резекции для сохране- ния функций артикуляции и проталкивания пищевого комка, глотания [8]; при этом не должно быть создано препятствий для прохождения воздуха в дыхательные пути. Эту задачу можно решить с различным успехом при использовании свободного лоскута большой груд- ной мышцы, переднебокового бедренного трансплан- тата [9], прямой брюшной мышцы [10] и широчайшей мышцы спины. Сообщения об их применении носят единичный характер, и для обобщения результатов еще не накоплен достаточный объем клинических данных. В настоящем исследовании мы описываем клинический случай реконструкции языка после глос- сэктомии с использованием свободного лоскута ши- рочайшей мышцы спины. После консультации с химиотерапевтом и радиоло- гом было принято решение о проведении радикальной операции в объеме тотальной глоссэктомии, шейной фасциально-футлярной лимфодиссекции справа и ра дикальной шейной лимфодиссекции слева, реконструкции пострезекционного дефекта кожно-мышечным лоскутом широчайшей мышцы спины на микрососудистых анасто- мозах. Рис. 1. Кровоточащая опухоль инфильтративно-язвенного характера с распространением на слизистую оболочку дна полости рта Fig. 1. Bleeding infiltrating ulcerating tumor extending into the mucosa of the floor of the mouth Management of the total glossectomy defect with latissimus dorsi free flap Management of the total glossectomy defect with latissimus dorsi free flap M.M. Davudov1, Ch.R. Rahimov1, A.A. Akhundov2, F.G. Irannezhat2, Sh.A. Magomedov2, M.Ch. Rahimli1, D.A. Safarov2 1Azerbaijan Medical University; 23 Bakikhanov St., Baku AZ 1022, Azerbaijan; 2N.N. Blokhin National Medical Research Center of Oncology, Ministry of Health of Russia; 24 Kashirskoe Shosse, Moscow 115478, Russia M.M. Davudov1, Ch.R. Rahimov1, A.A. Akhundov2, F.G. Irannezhat2, Radical surgical management of tongue cancer results in sever speech and swallowing disruption, impaired airway protection and life-threat- ening aspiration. Surgical objective of total glossectomy defect management is an adequate restoration of lingual mass and affected tissues of the floor of the mouth. Range of flaps are known to provide the adequate outcome, i. e. pediculed pectoralis major flap, anterolateral free thigh flap, rectus abdominis flap, radial free forearm flap and latissimus dorsi flap. g f p, f p, f f f p f p In the current report, we present a case of glossectomy defect reconstruction with pediculed latissimus dorsi flap. Postoperative assessment of the transplant was made according to clinical criteria and viability of the flap was assessed via laser Doppler flowmetry. Signs of microcirculation improved starting from day 5 postoperatively, and were almost equal with donor site microcirculation signs on day 14. A sufficiently large mass of the transplant allowed to create a thickening over the epiglottis, as well as to close the defect in the oral cavity. The latissimus dorsi flap in the tongue reconstruction has a high potential: its use provides a relatively good quality of ar- ticulation, recovery of deglutition. Key words: tongue cancer, glossectomy, reconstruction with latissimus dorsi free flap For citation: Davudov M.M., Ragimov Ch.R., Akhundov A.A. et al. Management of the total glossectomy defect with latissimus dorsi free flap. Opukholi golovy i shei = Head and Neck Tumors 2018;8(2):77–82 77 77 Клиническое наблюдение Пациентка М., 62 лет, обратилась в клинику че- люстно-лицевой хирургии Азербайджанского медицин- ского университета с жалобами на болезненное язвенное образование в области проведения предыдущей операции (удаления части языка). Пациентка отмечала периоди- ческие кровотечения из образования, а также затрудне- ние глотания. Рис. 1. Кровоточащая опухоль инфильтративно-язвенного характера с распространением на слизистую оболочку дна полости рта Fig. 1. Bleeding infiltrating ulcerating tumor extending into the mucosa of the floor of the mouth 78 78 Опухоли ГОЛОВЫ и ШЕИ HEAD and NECK tumors 2’2018 Клинический случай Том 8 Vol. 8 В ходе операции, выполняемой под общим обезболива- нием вначале произвели левостороннюю радикальную Рис. 2. Удаленный из полости рта макропрепарат Fig. 2. Gross specimen removed from the oral cavity Рис. 3. Отмечены границы кожно-мышечного лоскута для транс плантации Fig. 3. Margins of the musculocutaneous flap for transplantation are high lighted Рис. 4. Кожно-мышечный лоскут на сосудистой ножке подготовлен к трансплантации Fig. 4. Musculocutaneous flap on a vascular pedicle is ready for transplantation Рис. 5. Свободный аутотрансплантат Fig. 5. Free autologous transplant Рис. 4. Кожно-мышечный лоскут на сосудистой ножке подготовлен Рис. 2. Удаленный из полости рта макропрепарат Fig. 2. Gross specimen removed from the oral cavity Рис. 2. Удаленный из полости рта макропрепарат Fig. 2. Gross specimen removed from the oral cavity Рис. 4. Кожно-мышечный лоскут на сосудистой ножке подготовлен к трансплантации Рис. 3. Отмечены границы кожно-мышечного лоскута для транс плантации Fig. 3. Margins of the musculocutaneous flap for transplantation are high lighted Fig. 4. Musculocutaneous flap on a vascular pedicle is ready for transplantation Рис. 5. Свободный аутотрансплантат Fig. 5. Free autologous transplant Рис. 5. Свободный аутотрансплантат Fig. 5. Free autologous transplant Рис. 3. Отмечены границы кожно-мышечного лоскута для транс плантации Fig. 3. Margins of the musculocutaneous flap for transplantation are high lighted Рис. 5. Свободный аутотрансплантат Fig. 5. Free autologous transplant В ходе операции, выполняемой под общим обезболива- нием, вначале произвели левостороннюю радикальную лимфодессекцию с резекцией грудино-ключично-сосцевид- ной мышцы и внутренней яремной вены, а также право- стороннюю модифицированную радикальную шейную диссекцию (с сохранением внутренней яремной вены). Далее осуществили тотальную глоссэктомию с удалени- ем тканей дна полости рта (рис. 2). Затем начали под- готовку лоскута широчайшей мышцы спины. Больную перевернули на бок, отметили границы забираемого ло- скута, достаточные для закрытия пострезекционного дефекта и восполнения объема тканей (рис. 3). Выполни- ли разрез кожи и подкожной клетчатки, обнажили и от- сепарировали широчайшую мышцу спины, по периферии разреза кожную часть лоскута пришили к мышце (рис. 4). ния таким трансплантатом жизненно важных функций языка. ния таким трансплантатом жизненно важных функций языка. Использование в реконструкции языка лоскутов из паховой области (m. transversus abdominis и m. rec- tus abdominis – TRAM flap) или области предплечья (radial forearm free flap), в том числе с костным ком- понентом, приносит удовлетворительные функцио- нальные результаты [11, 12], однако со временем наблюдается потеря мышечной массы. Кроме того, не учитывается возможность активных произвольных движений, обеспечиваемых подъязычной реиннерва- цией. Трансплантат из широчайшей мышцы спины име- ет достаточный объем, позволяющий не только вос- полнить дефицит тканей дна полости рта, но и создать утолщение над надгортанником, чтобы лучше направ- лять пищевой комок вниз по рото- и гортаноглотке. Это также значительно облегчает артикуляцию. Транс- плантат достаточно гибок и подвижен при произволь- ных попытках глотания и речи. Рис. 7. Трансплантат в зоне дефекта спустя 2 нед после операции Fig. 7. Transplant in the defect area 2 weeks after the surgery Первые 48 ч после операции тщательно контролиро- вали жизнеспособность трансплантата. Его окраска оставалась розовой, заполнение капилляров кровью при повторных нажатиях на трансплантат происходи- ло через 1–3 с, а при прокалывании иглой на поверхность выступали капли алой крови, что свидетельствовало об отсутствии признаков тромбоза анастомозов. При использовании этого трансплантата в случае необходимости можно осуществить одномоментный забор васкуляризированной кости (ребра) для рекон- струкции нижней челюсти. Микроциркуляцию в трансплантате в зоне дефекта изучали методом лазерной доплеровской флоуметрии. Мониторинг в послеоперационном периоде проводили ежедневно в течение 14 дней (рис. 7). Показатель микро- циркуляции увеличивался начиная с 5‑го дня после хирур- гического вмешательства, достигнув на 14‑й день наи- высшего уровня (15,38 перф. ед.) и приблизившись, таким образом, к значениям донорского участка. Печальная статистика, свидетельствующая о том, что при таких распространенных процессах, когда требуется глоссэктомия, в лучшем случае могут вы- жить лишь 35 % больных [6], не должна стать камнем преткновения для хирургов. Опубликованы и другие данные, согласно которым 50 % пациентов, подверг- шихся полной глоссэктомии при раке языка, могут прожить до 1,5 года [11]. И задача хирурга – обеспе- чить этим больным как можно лучшее качество остав- шейся жизни. Кроме того, не ослабевает тенденция к дальнейшему увеличению показателей излечения больных, поэтому актуальность реконструктивных операций не уменьшается, а, напротив, растет, что диктует необходимость их совершенствования. Через 1 мес после операции состояние транспланта- та оценивали как удовлетворительное. Больная прини- мала пищу через рот, ее речь была внятной, проблем с дыханием не наблюдалось. Обсуждение Реконструкция языка у больных после тотальной или субтотальной глоссэктомии остается сложной за- дачей. Отсутствие донорской ткани, которая хотя бы отдаленно напоминала ткань этого специализирован- ного органа и соответствовала бы его функциональным задачам, вынуждает хирургов выбирать лучшее из до- ступного. Дальнейшие исследования функциональных ре- зультатов использования трансплантата из широчай- шей мышцы спины при полной реконструкции языка позволят дать более объективную оценку данному методу. Клиническое наблюдение Произвели диссекцию a. toracodorsalis и v. toracodorsalis, формирующих сосудистую ножку трансплантата соот- ветствующей длины, после чего сосуды были пересечены и перевязаны (рис. 5). Далее на резецированные концы сосудистой ножки трансплантата наложили микрохи- рургические анастомозы соответственно с a. thyroidea superior и v. thyroidea superior, проверили их функциониро- вание (рис. 6). Затем трансплантат провели через под- Рис. 6. Наложение сосудистого анастомоза на резецированные концы сосудистой ножки трансплантата Fig. 6. Imposition of vascular anastomosis on the resectied ends of the transplant’s vascular pedicle челюстную область в область дефекта и подшили узло- выми швами к краям дефекта. После завершения операции больную перевели в реанимационное отделение. 79 79 Опухоли ГОЛОВЫ и ШЕИ HEAD and NECK tumors 2’2018 Клинический случай Том 8 Vol. 8 Рис. 7. Трансплантат в зоне дефекта спустя 2 нед после операции Fig. 7. Transplant in the defect area 2 weeks after the surgery Л И Т Е Р А Т У Р А / R E F E R E N C E S 1. Матякин Е.Г., Голдобенко Г.В., Воро- бьев Ю.И. и др. К проблеме элективно- го воздействия на путях лимфооттока при раке подвижной части языка (ре- зультаты кооперированного рандоми- зированного исследования). Вестник РОНЦ им. Н.Н. Блохина РАМН 1995;6(1):23–6. [Matyakin E.G., Goldobenko G.V., Vorobiev Yu.I. et al. To the problem of elective treatment of lymph channels in cancer of the tongue movable part (results of a cooperative randomized study). Vestnik RONC im. N.N. Blokhina RAMN = Journal of N.N. Blokhin Russian Cancer Research Center 1995;6(1):23–6. (In Russ.)]. 1. Матякин Е.Г., Голдобенко Г.В., Воро- бьев Ю.И. и др. К проблеме элективно- го воздействия на путях лимфооттока при раке подвижной части языка (ре- зультаты кооперированного рандоми- зированного исследования). Вестник РОНЦ им. Н.Н. Блохина РАМН 1995;6(1):23–6. [Matyakin E.G., 8. Miyamoto S., Sakuraba M., Nagamatsu S. et al. Risk factors for gastric-tube depen- dence following tongue reconstruction. Ann Surg Oncol 2012;19(7):2320–6. DOI: 10.1245/s10434-012-2298-2. PMID: 22396003. functional outcomes after long-term fol- low-up of total glossectomy. J Craniofac Surg 2010;21(1):111–6. DOI: 10.1097/ SCS.0b013e3181c46692. PMID: 20061966. functional outcomes after long-term fol- low-up of total glossectomy. J Craniofac Surg 2010;21(1):111–6. DOI: 10.1097/ SCS.0b013e3181c46692. PMID: 20061966. 5. Kimata Y., Sakuraba M., Hishinuma S. et al. Analysis of the relations between the shape of the reconstructed tongue and postoperative functions after subtotal or total glossectomy. Laryngoscope 2003;113(5):905–9. DOI: 10.1097/00005537-200305000-00024. PMID: 12792331. 9. Ng R.W., Chan J.Y., Mok V. et al. Clinical implications of anterolateral thigh flap shrinkage. Laryngoscope 2008;118(4):585– 8. DOI: 10.1097/MLG.0b013e31816067a3. PMID: 18176347. 10. Lyos A.T., Evans G.R., Perez D., Schus- terman M.A. Tongue reconstruction: out- comes with the rectus abdominis flap. Plast Reconstr Surg 1999;103(2):442–7. PMID: 9950529. 6. Sinclair C.F., Carroll W.R., Desmond R.A., Rosenthal E.L. Functional and survival outcomes in patients undergoing total glossectomy compared with total laryngo- glossectomy. Otolaryngol Head Neck Surg 2011;145(5):755–8. DOI: 10.1177/0194599811412724. PMID: 21670476. 11. Chien C.Y., Su C.Y., Hwang C.F. et al. Ablation of advanced tongue or base of tongue cancer and reconstruction with free flap: functional outcomes. Eur J Surg Oncol 2006;32(3);353–7. DOI: 10.1016/j. ejso.2005.12.010. PMID: 16455224. 2. Engel H., Huang J.J., Lin C.Y. et al. A strategic approach for tongue recon- struction to achieve predictable and im- proved functional and aesthetic outcomes. Plast Reconstr Surg 2010;126(6):1967–77. DOI: 10.1097/PRS.0b013e3181f44742. PMID: 21124135. 7. Hanasono M.M., Skoracki R.J., Yu P. A prospective study of donor-site morbid- ity after anterolateral thigh fasciocutane- ous and myocutaneous free flap harvest in 220 patients. Plast Reconstr Surg 2010;125(1):209–14. Вклад авторов М.М. Давудов: проведение операций; М.М. Давудов: проведение операций; Ч.Р. Рагимов: разработка дизайна исследования; А.А. Ахундов: анализ полученных данных, обзор публикаций по теме статьи; Ф.Г. Ираннежат: анализ полученных данных; Ш.А. Магомедов: анализ полученных данных; М.Ч. Рагимли: написание текста статьи; Д.А. Сафаров: получение данных для анализа. Authors’ contributions M.M. Davudov: surgical treatment; M.M. Davudov: surgical treatment; Ch.R. Rahimov: developing the research design; A. A. Akhundov: analysis of the obtained data, reviewing of publications of the article's theme; F. G. Irannezhat: analysis of the obtained data; Sh.A. Magomedov: analysis of the obtained data; M.Ch. Rahimli: article writing; D.A. Safarov: obtaining data for analysis. Л И Т Е Р А Т У Р А / R E F E R E N C E S DOI: 10.1097/ PRS.0b013e3181c495ed. PMID: 19910852. 12. Dziegielewski P.T., Ho M.L., Rieger J. et al. Total glossectomy with laryngeal preservation and free flap reconstruction: objective functional outcomes and system- atic review of the literature. Laryngoscope 2013;123(1):140–5. DOI: 10.1002/ lary.23505. PMID: 22952109. 3. Sultan M.R., Coleman J.J. 3rd. Oncologic and functional considerations of total glossectomy. Am J Surg 1989;158(4):297– 302. PMID: 2552853. and functional considerations of total glossectomy. Am J Surg 1989;158(4):297– 302. PMID: 2552853. 4. Yun I.S., Lee D.W., Lee W.J. et al. Corre- lation of neotongue volume changes with Заключение При анализе литературы об использовании транс- плантатов из различных областей следует отметить работу M. R. Sultan и J. J. Coleman [3], которые сооб- щили об 11 случаях применения лоскута большой грудной мышцы на сосудистой ножке для полной реконструкции языка после глоссэктомии. Авторы отметили, что через некоторое время подвижность трансплантата уменьшалась из‑за его провисания, а также указали на малые возможности восстановле- ния иннервации трансплантата, что не позволяет счи- тать этот метод приоритетным вследствие невыполне- Лоскут широчайшей мышцы спины для рекон- струкция языка у больных после тотальной или субто- тальной глоссэктомии обладает высоким потенциалом для реабилитации: его использование обеспечивает относительно хорошее качество артикуляции, восста- новление глотания. Достаточная мышечная масса ло- скута компенсирует потерю большого объема тканей после расширенных хирургических вмешательств, а также восстанавливает функции органа, включая питание через рот. 80 80 Опухоли ГОЛОВЫ и ШЕИ HEAD and NECK tumors 2’2018 Клинический случай Том 8 Vol. 8 Статья поступила: 13.05.2018. Принята к публикации: 19.06.2018. Article received: 13.05.2018. Accepted for publication: 19.06.2018. ORCID авторов ORCID авторов А. А. Ахундов: https://orcid.org / 0000‑0002‑9543‑990X Д. А. Сафаров: https://orcid.org / 0000‑0003‑2793‑5597 ORCID of authors A. A. Akhundov: https://orcid.org / 0000‑0002‑9543‑990X D. A. Safarov: https://orcid.org / 0000‑0003‑2793‑5597 Конфликт интересов. Авторы заявляют об отсутствии конфликта интересов. Conflict of interest. The authors declare no conflict of interest. 81 81 81 81 Том 8 Vol. 8 Финансирование. Исследование проведено без спонсорской поддержки. Financing. The study was performed without external funding. Финансирование. Исследование проведено без спонсорской поддержки. Информированное согласие. Пациент подписал информированное согласие на публикацию своих данных. Informed consent. The patient gave written informed consent to the publication of his data. Информированное согласие. Пациент подписал информированное согласие на публикацию своих данных. Informed consent. The patient gave written informed consent to the publication of his data. 82 82
https://openalex.org/W3158660694	https://ojs-gr.zrc-sazu.si/primerjalna_knjizevnost/article/download/8101/7565	Slovene	null	Svetovna primerjalna književnost	Primerjalna književnost	2,021	cc-by	1,821	Primerjalna književnost (Ljubljana) 44.1 (2021) Svetovna primerjalna književnost David Damrosch: Comparing the Literatures. Literary Studies in a Global Age. Princeton, NJ: Princeton University Press, 2020. 386 str. Blaž Zabel Filozofska fakulteta, Univerza v Ljubljani, Oddelek za filozofijo, Aškerčeva 2, 1000 Ljubljana https://orcid.org/0000-0001-6512-1565 blaz.zabel@ff.uni-lj.si Blaž Zabel Blaž Zabel Filozofska fakulteta, Univerza v Ljubljani, Oddelek za filozofijo, Aškerčeva 2, 1000 Ljubljana https://orcid.org/0000-0001-6512-1565 blaz.zabel@ff.uni-lj.si Blaž Zabel Filozofska fakulteta, Univerza v Ljubljani, Oddelek za filozofijo, Aškerčeva 2, 1000 Ljubljana https://orcid.org/0000-0001-6512-1565 blaz.zabel@ff.uni-lj.si Naslov te recenzije, »Svetovna primerjalna književnost« se morda zdi nenavaden. Vemo, kaj je to »primerjalna književnost«, literarnovedna disciplina, ki jo včasih imenujemo tudi komparativistika. Prav tako si lahko vsaj nekako predstavljamo, kaj je to »svetovna književnost«, pa čeprav je bilo o tem v zadnjih dvajsetih letih prelito že mnogo črnila. A »svetovna primerjalna književnost«, kaj je to? S tem vprašanjem se v svoji najnovejši monografiji z naslovom Comparing the Literatures: Literary Studies in a Global Age ukvarja David Damrosch, profesor primerjalne književnosti na Univerzi Harvard. Povedano drugače, v študiji, ki je nedavno izšla pri Princeton University Press, se eden najslavnejših komparativistov enaindvajsetega stoletja sprašuje o vlogi primerjalne književnost v času globalizacije, to pa na način, da razpra- vlja o zgodovini komparativistike ter o zgodovini primerjalnih metod v literarni vedi. Pri tem se posebej osredotoča na tiste zgodovinske vidike, ki so jih raziskovalci v zadnji dveh desetletjih izpostavili kot bistvene za razumevanje transnacionalizacije ter globalizacije, kot so na primer medkulturni vplivi, vznik postkolonialnih študij, razvoj prevodoslovja, svetovni obtok idej itd. j Da se je Damrosch odločil preučiti »razvoj literarnih ved v dobi globalizacije«, ni povsem nepričakovano. Damrosch je namreč tako pri nas kot po svetu najbolj znan po svoji odmevni študiji What is World Literature? (2003), v kateri se je ukvarjal predvsem s književnostjo v dobi globalizacije, medkulturno in transhistorično recepcijo literar- nih del ter s prevodoslovjem. Če se je torej v monografiji s preloma tisočletja, ki je v komparativistiki sprožila pravo revolucijo, ukvarjal predvsem z vprašanjem svetovne književnosti, pa lahko njegovo najno- vejše delo razumem kot nadaljevanje, le da se tokrat namesto na samo književnost osredotoči na vedo, ki svetovno književnost preučuje, torej na »svetovno primerjalno književnost«. A glavna teza njegove najnovejše raziskave je podobna: tako kot je v globaliziranem svetu potrebno razmišljati o svetovni književnosti (torej o literarnih delih, ki 223 Primerjalna književnost (Ljubljana) 44.1 (2021) PKn, letnik 44, št 1, Ljubljana, maj 2021 predvsem v obliki prevodov potujejo preko svojih lokalnih okvirov in bralce seznanjajo z drugačnimi svetovi), tako je tudi o komparativistiki nujno razmišljati z vidika transnacionalizacije. Blaž Zabel Tudi akademske ideje namreč potujejo med različnimi kulturnimi konteksti, raziskovalci raz- mišljajo o prevodih, o perifernih ali ne-kanoniziranih literarnih delih, se soočajo z vprašanji in pritiski globalizacije, vse to pa vpliva na pojav in razvoj komparativistike kot literarnovedne discipline. Rezultat te raziskave, Comparing the Literatures je odličen pregled razvoja kompa- rativistike, ki ponuja povsem svež pogled na preteklost te vede, preko raziskovanja njene preteklosti pa avtor oblikuje tudi svoj odgovor na sodobne raziskovalne izzive. V prvem poglavju se David Damrosch posveti zgodnjim začetkom primerjalne književnosti. Za razliko od What is World Literature?, ki v uvodu obravnava idejo Weltliteratur Johanna Wolfganga von Goetheja, se Comparing the Literatures začne z njegovim antagonistom Johannom Gottfriedom Herderjem. Avtor razpravlja o različnih vidikih nacio- nalne in primerjalne književnosti, ki jih je zagovarjal romantični mislec, hkrati pa ponuja svežo tezo o produktivnosti vpliva sočasnih francoskih akademikov (ki predstavlja zanimiv antipod ideji Isaiah Berlina o anti- razsvetljenstvu). Prav s tem namenom Damrosch v poglavju obravnava še Madame de Staël, ki prav tako predstavlja eno izmed utemeljiteljic komparativistike, posveti pa se njenemu razmišljanju o odnosu med literaturo in družbenimi institucijami. Ker je prvo poglavje namenjeno začetkom primerjalne književnosti, avtor monografije seveda ne more mimo Huga Meltzla in Hutchesona Maucalaya Posnetta, pri čemer prvemu pripiše začetek zavezanosti naše vede poliglotizmu, drugemu pa idejo estetske relativnosti. Prvo poglavje je tako odličen uvod v nekaj najpomembnejših in najprepoznavnejših začetkov komparativistike, a posebna dodana vrednost razprave je sinteza najnovejših ugotovitev in tez o delu teh zgodnjih komparativistov. g j p Če se prvo poglavje ukvarja predvsem z izvori primerjalne knji ževnosti (torej 18. in 19. stoletjem), so preostala poglavja monogra- fije posvečena različnim tematskim vprašanjem. Drugo poglavje tako obravnava akademske migrante, ki veljajo za ustanovitelje ameriške komparativistike. Poseben doprinos tega dela razprave je Damroschev ne-evropocentričen pogled na to tematiko: čeprav v svojem delu obrav- nava tudi najbolj znana povojna imigranta v primerjalni književnosti, Lea Spitzerja in Ericha Auerbacha (kjer ponuja izvrstno sintezo najno- vejših in preštevilnih razprav), pa enakovredno pozornost nameni tudi delu dveh kitajskih (i)migrantov, Hu Shih (znan jezikovni revolucio- nar, ki se je šolal na univerzi Cornell) in Lin Yutang (kitajski pisatelj 224 RECENZIJA ter odmeven prevajalec), pa tudi delu komparativistke Lilian Furst, ki jo je zaznamovala izkušnja pobega pred nacizmom. Blaž Zabel Prav marginalnost Auerbacha in Spitzerja, dveh »velikanov« komparativistike, ponuja svež in inovativen pogled na zgodovino discipline, ki so jo v mnogih pogle- dih še bolj kot »velikani« zaznamovale tudi druge kulture in pozabljene komparativistke. V tretjem poglavju se Damrosch posveti vprašanju ameriške kompa- rativistike po drugi svetovni vojni, ki se ga loti s politično-historičnega vidika, ter vprašanju političnosti literarnih ved. Najprej razpravlja o komparativistiki v petdesetih in šestdesetih letih, ki jo je močno zazna- movala hladna vojna (Balakian, Welek, Frye, Greene), a zanimivejši del razprave predstavlja drugi del poglavja, ki obravnava delo Edwarda Saida ter Gayatri Spivak. V obeh primerih avtor postreže z zanimivo interpretacijo njune postkolonialne teorije v kontekstu sočasnih poli- tičnih dogajanj (predvsem v Palestini) in kulturnih vojn v akadem- skem prostoru. Četrto poglavje obravnava obdobje prevlade literarne teorije v ameriški komparativistiki, predvsem dekonstrukcije Paula de Mana in Barbare Johnson, pri čemer Damroscha še posebej zanimajo nedavno odkriti medvojni spisi, ki so pokazali na de Manov antisemi- tizem. Podobno kot v drugih poglavjih avtor bolj znane dekonstrukti- viste zoperstavi še kitajski in indijski literarni teoriji (v tem kontekstu pa ponuja zanimivo medkulturno branje del indijskega pesnika in dra- matika Kālidāsa). V nadaljevanju Damrosch obravnava vprašanje poliglotizma, ki predstavlja enega bistvenih vidikov naše vede. Znan je poziv in zahteva številnih oddelkov za primerjalno književnost po svetu, da mora vsak komparativist tekoče znati vsaj še dva ali tri tuje jezike. Da je za primer- janje literarnih tradicij potrebno znanje več jezikov, je teza, ki spremlja komparativistiko praktično od njenih začetkov. Damrosch se v sklopu razmisleka o poliglotizmu posveti razvoju tega koncepta, in sicer od poudarka na aktivnem znanju francoščine in nemščine v ameriški komparativistiki šestdestih let do utemeljitelja koncepta poliglotizma v komparativistiki Renéja Étiembla, pa do sodobnih prevodoslovnih teorij in vprašanja, ali lahko komparativist pravzaprav primerja več lite- rarnih del v prevodu. Primerjanju preko prevodov je Damrosch, kot bodo dobro vedeli vsi, ki poznajo njegova ostala dela, precej naklonjen. Š Še posebej bo slovenske bralce zanimalo šesto poglavje, v kate- rem avtor raziskuje različne koncepte književnosti in se sprašuje, kaj sploh je literatura. Razpravo otvori podpoglavje z naslovom »A Tale of Two Knjižnici« in se začne takole: »Dobra točka, kjer lahko zač- nemo, se nahaja v Ljubljani. Slovensko literarno izročilo je staro več 225 PKn, letnik 44, št 1, Ljubljana, maj 2021 stoletij in prestolnica ga hrani v dveh glavnih knjižnicah (knjižnici, iz slovanskega pojma za 'knjige')…« (209). Blaž Zabel Avtor razpravo nadaljuje z opisom Semeniške knjižnice ter NUK-a, nato pa se posveti poeziji Franceta Prešerna in vprašanju nacionalnega pomena tega pesnika. V prihodnosti bo zanimivo spremljati, kako močno bo Comparing the Literatures popularizirala slovensko književnost in Prešerna po svetu, s čimer bo mogoče aktivno opazovali vlogo globalne akademske mreže, ki za kanonizacijo perifernih literarnih tradicij potrebuje prav takšne spodbude akademskih centrov. Nato avtor razpravlja še o kanonizaciji, majhnih ter perifernih literarnih tradicijah, globalnem literarnem trgu in podobno. Posebej je zanimiva razprava o odnosih med literaturo in drugimi mediji (predvsem videoigrami, sodobno umetnostjo in perfor- mansom), pri čemer avtor zagovarja tezo, da mora komparativistika te intermedialne fenomene vzeti resno in jih začeti natančno preučevati. j p Zadnji dve poglavji sta posvečeni razpravi o svetovi književnost in o literarnih primerjavah. Glede svetovne književnosti Damrosch ponuja precej dolgo refleksijo najnovejših tematik in razprav, ki jih na zanimiv način preplete z različnimi literarnimi primeri iz svetovnega kanona (Calvino, Yourcenar, Marquéz, Kadare) ter popularne književnosti (J. R. R. Tolkien, Conan Doyle). Nato se posveti bistvenemu vprašanju komparativistike, torej vprašanju, kaj so literarne primerjave in kako primerjati literature. Pri tem se spopade z nekaterimi novejšimi raz- pravami v primerjalni književnosti (Pheng Cheah, Emily Apter, Haun Saussy, Natalie Melas, Rajagopalan Radhakrishnan) in v drugih vedah, ki se posvečajo vprašanju medkulturnega primerjanja (Sheldon Pollock, Marcel Detienne), odlično pa v razpravo vključuje tudi ne-Evropske akademske kontekste, na primer, delo kitajskih komparativistov Cao Shunqing in Zhang Longxi. Bralca, ki pozna Damroschevo raziskova- nje, ne bo presenetilo, da se tudi v svoji najnovejši študiji zavzema za čim širše literarno primerjanje tako preko prostora (primerjave naj zaje- majo čim več različnih kultur) kot časa (zajamejo naj čim daljše časovno obdobje), to tezo pa utemelji še z zanimivim poskusom primerjave lite- rarnih modernizmov po svetu in v različnih časovnih obdobjih. V zaključku, ki je naslovljen s pomenljivim naslovom »Rebirth of a Discipline«, oblikovan kot odgovor na polemično študijo Gayatri Spivak The Death of the Discipline (2003), David Damrosch ponudi svoj pogled na prihodnost primerjalne književnosti. Njegova vizija je optimistična, prihodnost komparativistike pa vidi v pluralnosti sodobnih literarnovednih pristopov. V prvi vrsti te prepozna v bogati interdisciplinarni zasedenosti letnih konferenc Ameriške zveze za pri- merjalno književnost (ACLA), ki se odražajo tudi v nedavnem poro- 226 RECENZIJA RECENZIJA čilu o stanju vede (»State of the discipline report«). Blaž Zabel Po drugi strani pa izpostavi interdisciplinarnost ter implementacijo novih literarnove- dnih pristopov v seriji najnovejših komparativističnih študij, ki po nje- govem mnenju kažejo na revitalizacijo in svetlo prihodnost naše vede (npr. Edmon, Levy, Ungureanu, Venkat Mani, pa tudi Apter, Cheah, Friedman, Hitchcock, Gupta, Moser, Jay, Casanova, Ganguly, Feng, Bhattacharya in drugi). y g Comparing the Literatures je po What is World Literature? najpo- membnejša knjiga Davida Damroscha in bo v naslednjih letih gotovo močno zaznamovala sodobno komparativistiko. Harvardski kompara- tivist predstavi optimističen odgovor na krizo, v kateri se je primerjalna književnost nahajala nekje od osemdesetih let naprej. Odgovor išče in prepričljivo najde v sodobnejših komparativistični pristopih (postko- lonialna teorija, vprašanja globalizacije, prevodoslovje, kulturna hibri- dnost, intermedialnost, svetovni literarni modernizmi…). Povedano drugače, prihodnost primerjalne književnosti je »svetovna primerjalna književnost«, veda, ki bo upoštevala in preučevala kulturno, histo- rično, intermedialno in socialno večplastnost književnost in literarnih tradicij. A še bolj kot sama teza glede prihodnosti discipline je zani- miva Damrosceva argumentacija, ki ni usmerjena naprej v prihodnost, ampak nazaj v preteklost. Kot odlično pokaže avtor, je namreč zgo- dovina naše discipline še vedno precejšnja neznanka, a je mnogo bolj interkulturna, transnacionalna, interdisciplinarna ter politična, kot se zdi na prvi pogled. Morda za prepoznanje tega dejstva potrebujemo nekaj brskanja po preteklosti in je potrebno včasih usmeriti pogled stran od kanoniziranih komparativistov, a preteklost primerjalne knji- ževnost je vedno bila »svetovna«, torej takšna kot bo nedvomno tudi njena prihodnost. 1.19 Recenzija / Review UDK 82.0 DOI: https://doi.org/10.3986/pkn.v44.i1.12 227
https://openalex.org/W1978335090	https://bmchealthservres.biomedcentral.com/track/pdf/10.1186/1472-6963-10-S2-A22	English	null	Patient classification in rehabilitation as a basis for performance-related payments	BMC health services research	2,010	cc-by	504	Patient classification in rehabilitation as a basis for performance-related payments G Neubauer1*, J Ranneberg2, A Wick3 From 26th Patient Classification Systems International (PCSI) Working Conference Munich, Germany. 15-18 September 2010 From 26th Patient Classification Systems International (PCSI) Working Conference Munich, Germany. 15-18 September 2010 compensation system for our RTGs. Moreover, we would like to think about how RTGs and DRGs can be assembled to determine complex lump-sum rates (“Komplexpauschalen”). Another question to be dis- cussed is how health insurance funds can be included in this process. Results Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit In cooperation with 10 rehab clinics, we calculated the average costs for each RTG. By grouping, we found out a coefficient of determination (R^2) of 0.4. By defining the various subgroups via the “Barthel-Index”, we were able to improve the coefficient of determination (R^2) by a further 5 percentage points. Neubauer et al. BMC Health Services Research 2010, 10(Suppl 2):A22 http://www.biomedcentral.com/1472-6963/10/S2/A22 Neubauer et al. BMC Health Services Research 2010, 10(Suppl 2):A22 http://www.biomedcentral.com/1472-6963/10/S2/A22 Open Access Author details 1 1IfG München, München, Germany. 2BKK Landesverband Bayern, München, Germany. 3Universität der BW München, Neubiberg, Germany. Methods h There are several proposals to solve the problem described in Germany. One approach is the “patient classification system for rehab patients” which was developed by the Institute of Health-Economics (IfG). It will be described in the following section. Published: 6 October 2010 Published: 6 October 2010 doi:10.1186/1472-6963-10-S2-A22 Cite this article as: Neubauer et al.: Patient classification in rehabilitation as a basis for performance-related payments. BMC Health Services Research 2010 10(Suppl 2):A22. doi:10.1186/1472-6963-10-S2-A22 Cite this article as: Neubauer et al.: Patient classification in rehabilitation as a basis for performance-related payments. BMC Health Services Research 2010 10(Suppl 2):A22. Following the logic of DRG, the system starts with major indication fields like musculoskeletal disorder, neurology, etc. In the next step, with the assistance of rehabilitation physicians, we defined basic rehabilitation treatment groups (RTGs/RBGs) for each indication field. As a result, we found 7 basic groups and 12 severity groups for orthopaedics, and 10 basic groups and 13 severity groups for cardiology. Introduction Since Germany introduced a comprehensive system for its acute-care sector, post-acute rehabilitation is under economic pressure. Patients are referred earlier and less healed to rehabilitation clinics. In addition, “payers” are not willing to pay more for post-acute rehabilitation. This situation requires an adequate classification system for rehab patients. 1IfG München, München, Germany Full list of author information is available at the end of the article Conclusions At the conference we would like to discuss our results, methods and how to establish a performance-based 1IfG München, München, Germany Full list of author information is available at the end of the article © 2010 Neubauer et al; licensee BioMed Central Ltd.
https://openalex.org/W2174792007	https://europepmc.org/articles/pmc4645477?pdf=render	English	null	Variable responses of human and non-human primate gut microbiomes to a Western diet	Microbiome	2,015	cc-by	8,513	© 2015 Amato et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Variable responses of human and non-human primate gut microbiomes to a Western diet Katherine R. Amato1,2,3, Carl J. Yeoman4, Gabriela Cerda5, Christopher A. Schmitt6,16, Jennifer Danzy Cramer7, Margret E. Berg Miller8, Andres Gomez8,9, Trudy R. Turner10,11, Brenda A. Wilson8,12, Rebecca M. Stumpf5,8, Karen E. Nelson13, Bryan A. White8,14, Rob Knight15 and Steven R. Leigh2,8 Abstract Background: The human gut microbiota interacts closely with human diet and physiology. To better understand the mechanisms behind this relationship, gut microbiome research relies on complementing human studies with manipulations of animal models, including non-human primates. However, due to unique aspects of human diet and physiology, it is likely that host-gut microbe interactions operate differently in humans and non-human primates Results: Here, we show that the human microbiome reacts differently to a high-protein, high-fat Western diet than that of a model primate, the African green monkey, or vervet (Chlorocebus aethiops sabaeus). Specifically, humans exhibit increased relative abundance of Firmicutes and reduced relative abundance of Prevotella on a Western diet while vervets show the opposite pattern. Predictive metagenomics demonstrate an increased relative abundance of genes associated with carbohydrate metabolism in the microbiome of only humans consuming a Western diet. Conclusions: These results suggest that the human gut microbiota has unique properties that are a result of changes in human diet and physiology across evolution or that may have contributed to the evolution of human physiology. Therefore, the role of animal models for understanding the relationship between the human gut microbiota and host metabolism must be re-focused. Keywords: Gut microbiome, Vervet, Chlorocebus aethiops, Western diet, Human evolution Correspondence: katherine.amato@northwestern.edu 1Department of Anthropology, Northwestern University, Evanston, USA 2Department of Anthropology, University of Colorado Boulder, Boulder, USA Full list of author information is available at the end of the article RESEARCH Open Access Variable responses of human and non-human primate gut microbiomes to a Western diet Katherine R. Amato1,2,3, Carl J. Yeoman4, Gabriela Cerda5, Christopher A. Schmitt6,16, Jennifer Danzy Cramer7, Margret E. Berg Miller8, Andres Gomez8,9, Trudy R. Turner10,11, Brenda A. Wilson8,12, Rebecca M. Stumpf5,8, Karen E. Nelson13, Bryan A. White8,14, Rob Knight15 and Steven R. Leigh2,8 Amato et al. Microbiome (2015) 3:53 DOI 10.1186/s40168-015-0120-7 RESEARCH Open Access Variable responses of human and non-human primate gut microbiomes to a Western diet Katherine R. Amato1,2,3, Carl J. Yeoman4, Gabriela Cerda5, Christopher A. Schmitt6,16, Jennifer Danzy Cramer7, Margret E. Berg Miller8, Andres Gomez8,9, Trudy R. Turner10,11, Brenda A. Wilson8,12, Rebecca M. Stumpf5,8, Karen E. Nelson13, Bryan A. White8,14, Rob Knight15 and Steven R. Leigh2,8 Amato et al. Microbiome (2015) 3:53 DOI 10.1186/s40168-015-0120-7 Amato et al. Microbiome (2015) 3:53 DOI 10.1186/s40168-015-0120-7 Background in many of these conditions [8–10], indicating a link be- tween the human gut microbiota, diet, and physiology. A common theme in human gut microbiome research currently is the effect of diet on gut microbiota compos- ition, and ultimately, host physiology [1]. In particular, a number of studies focus on the impact of a “Western” diet (high in animal fat and protein and low in fiber) compared to a “non-Western” diet (low in animal fat and protein and high in fiber; [2–6]). Western and non- Western human diets are consistently associated with distinct gut microbial communities [2–6], and humans consuming a high-protein, high-fat Western diet also have higher rates of obesity and other metabolic syndromes such as diabetes [7]. Gut microbes appear to be implicated As data describing patterns in the human gut micro- biota accumulate, researchers are beginning to turn to- ward animal models to test the mechanisms driving host-gut microbe interactions and their physiological consequences. While the use of rodent models is particu- larly popular due to ease of manipulation, non-human primate models are being integrated as well as a result of their close phylogenetic relationships with humans and presumably similar physiology [1, 11–13]. For example, a recent study examined the impact of a high-fat diet on the gut microbiota of adult female macaques and their off- spring [12]. However, human evolution appears to reflect a major dietary shift from a predominantly plant-based diet to an increasingly carnivorous diet, with relatively recent in- creases in food digestibility due to tool use, cooking, and Amato et al. Microbiome (2015) 3:53 Page 2 of 9 other processing techniques [14–16]. Additionally, com- pared to non-human primates, humans are characterized by a number of unique physiological adaptations, includ- ing increased brain size, reduced gut size, increased fat deposition, and decreased muscle mass [17–19]. Differ- ences also exist between human and non-human primate Toll-like receptors in the gut as well as expression of enzymes for the production of hormones such as prosta- glandin [20–22]. Because diet can alter the taxonomic composition of the human gut microbiota during an indi- vidual’s lifetime [23, 24], and human metabolism and physiology is closely linked to the gut microbiota [25], it is likely that evolutionary shifts in human diet and physi- ology were accompanied by shifts in the gut microbiota. Background As a result, the human gut microbiota should exhibit unique properties when compared to the non-human pri- mate gut microbiota. If this is the case, non-human primate models of host-gut microbe relationships may be less ideal than assumed for addressing questions regarding human diet and physiology in the context of the gut microbiota. Physiological similarities to humans make vervets, an at- tractive model for biomedical trials [28], particularly with regards to human metabolic disorders and obesity [29, 30]. Vervets may therefore provide an excellent reference system for understanding how unique the responses of the human gut microbiota to diet are compared to other primates. Despite the utility of vervets as models for human physiology in many contexts, we hypothesized that the vervet gut microbiota would exhibit distinct responses to a Western diet when compared with the human gut microbiota. Specifically, we predicted that the relative abundances of fewer microbial taxa and genes would differ in response to diet in the vervet gut microbiota since vervets have not evolved to include large amounts of animal fat and protein in their natural diets [31]. Consequently, the vervet gut microbiota should be less able to adapt to this type of Western diet and may not be an appropriate model for understanding human gut microbe interactions in the context of diet. Recent data provide evidence that the human gut microbiota is less diverse and better adapted to a meat-rich diet compared to closely related extant pri- mates [26], but no direct comparison of host-gut mi- crobe dynamics in humans and non-human primates currently exists. Therefore, it is unclear if the relation- ship between the gut microbiota and host diet and physiology differs in humans compared to other pri- mates. To improve our understanding of how host-gut microbe interactions are similar or different in human and non-human primates, here, we use 454 titanium pyrose- quencing and predictive metagenomics (PICRUSt) [27] to compare the gut microbiota of humans and African green monkeys, or vervets (Chlorocebus aethiops sabaeus), consuming both a Western and a non-Western diet. Results Vervets Results show that although vervets fed a typical West- ern diet (TWD, see Methods for composition) had similar gut microbial richness compared to wild vervets consuming a non-Western diet (Fig. 1), the compos- ition of their gut microbiota was distinct (unweighted UniFrac distance, F1,26 = 4.8, p < 0.001, r2 = 0.16, Fig. 2; weighted UniFrac distance, F1,26 = 9.6, p < 0.001, r2 = 0.28, Additional file 1: Figure S1, Table S1-S5). Specifically, at the phylum level, TWD-fed vervets had significantly lower relative abundances of Firmicutes, Lentisphaerae, Proteo- bacteria, Tenericutes, and Verrucomicrobia than wild vervets (Additional file 1: Table S1). In contrast, TWD-fed Fig. 1 A Western diet reduces gut microbial diversity in humans but not vervets. Chao1 estimates (average ± SD) of microbial community richness at 1000 sequence reads per sample for vervets and humans consuming a non-Western vs. a Western diet. Star indicates significant differences (FDR-corrected p < 0.05) in microbial richness between diets. Western humans are from Italy (Schnorr et al. [3]) and the USA (Yatsunenko et al. [2]). Non-western humans are from Tanzania (Hadza, Schnorr et al. [3]), Venezuela (Guahibo, Yatsunenko et al. [2]), and Malawi (Yatsunenko et al. [2]). (N = 13 non-Western vervets; N = 14 Western vervets; N = 17 non-Western (Schnorr et al.) humans; N = 11 Western (Schnorr et al.) humans; N = 76 non-Western (Yatsunenko et al.) humans; N = 118 Western (Yatsunenko et al.) humans) Fig. 1 A Western diet reduces gut microbial diversity in humans but not vervets. Chao1 estimates (average ± SD) of microbial community richness at 1000 sequence reads per sample for vervets and humans consuming a non-Western vs. a Western diet. Star indicates significant differences (FDR-corrected p < 0.05) in microbial richness between diets. Western humans are from Italy (Schnorr et al. [3]) and the USA (Yatsunenko et al. [2]). Non-western humans are from Tanzania (Hadza, Schnorr et al. [3]), Venezuela (Guahibo, Yatsunenko et al. [2]), and Malawi (Yatsunenko et al. [2]). (N = 13 non-Western vervets; N = 14 Western vervets; N = 17 non-Western (Schnorr et al.) humans; N = 11 Western (Schnorr et al.) humans; N = 76 non-Western (Yatsunenko et al.) humans; N = 118 Western (Yatsunenko et al.) humans) Fig. 1 A Western diet reduces gut microbial diversity in humans but not vervets. Humans vs. vervets Fig. 2 The vervet gut microbiome resembles a non-Western human gut microbiome regardless of the vervet diet. Non-metric multi-dimensional scaling (NMDS) plot based on unweighted UniFrac distances illustrating clustering patterns in gut microbiomes across sampling groups at the OTU level Both human studies revealed differences in the gut micro- biota associated with diet both in terms of microbial taxo- nomic composition (unweighted UniFrac distance, Hadza/ Italy, F1,27 = 4.9, p < 0.001, r2 = 0.16, Malawi/Venezuela/ U.S., F1,193 = 31.6, p < 0.001, r2 = 0.14, Fig. 2; weighted Uni- Frac distance, Hadza/Italy, F1,27 = 5.2, p = 0.001, r2 = 0.17; Malawi/Venezuela/USA, F1,193 = 72.7, p < 0.001, r2 = 0.27 Additional file 1: Figure S1, Additional file 1: Table S1-S5) and gene relative abundances (Hadza/Italy, F1,42 = 6.1, p = 0.003, r2 = 0.13; Malawi/Venezuela/USA, F1,193 = 61.3, p < 0.001, r2 = 0.24), but we detected more differences in the relative abundances of microbial taxa between Western and non-Western gut microbiomes in the Malawi/ Venezuela/USA dataset, presumably as a result of larger sample sizes (Additional file 1: Tables S1-S7). In addition, several microbial taxa reacted similarly to a Western diet across both human and vervet datasets despite different storage, DNA extraction, and sequencing technologies as well as variation in Western and non-Western diet across countries and species. Like the vervets, Western diet humans from the USA and Italy exhibited lower relative abundances of Proteobacteria, Lentisphaerae, and Teneri- cutes (p = 0.06 for Italians) compared to those on a non- Western diet. Relative abundances of Spirochaetes were also lower in the humans from Malawi and Venezuela com- pared to the USA, and we detected lower relative abun- dances of Clostridium and higher relative abundances of Desulfovibrionaceae and Collinsella in humans from the USA compared to Malawi and Venezuela. Fig. 2 The vervet gut microbiome resembles a non-Western human gut microbiome regardless of the vervet diet. Non-metric multi-dimensional scaling (NMDS) plot based on unweighted UniFrac distances illustrating clustering patterns in gut microbiomes across sampling groups at the OTU level vervets had significantly higher abundances of Bacteroi- detes and TM7, with a trend for higher relative abun- dances of Spirochaetes (Additional file 1: Table S1). Results Vervets Chao1 estimates (average ± SD) of microbial community richness at 1000 sequence reads per sample for vervets and humans consuming a non-Western vs. a Western diet. Star indicates significant differences (FDR-corrected p < 0.05) in microbial richness between diets. Western humans are from Italy (Schnorr et al. [3]) and the USA (Yatsunenko et al. [2]). Non-western humans are from Tanzania (Hadza, Schnorr et al. [3]), Venezuela (Guahibo, Yatsunenko et al. [2]), and Malawi (Yatsunenko et al. [2]). (N = 13 non-Western vervets; N = 14 Western vervets; N = 17 non-Western (Schnorr et al.) humans; N = 11 Western (Schnorr et al.) humans; N = 76 non-Western (Yatsunenko et al.) humans; N = 118 Western (Yatsunenko et al.) humans) Page 3 of 9 Page 3 of 9 Page 3 of 9 Amato et al. Microbiome (2015) 3:53 Humans vs. vervets While people in the USA exhibited increased relative abun- dances of Firmicutes and reduced relative abundances of Page 4 of 9 Page 4 of 9 Amato et al. Microbiome (2015) 3:53 Bacteroidetes compared to people in Venezuela and Malawi, vervets exhibited the following opposite pattern: a reduction in Firmicutes relative abundances and an in- crease in Bacteroidetes relative abundances in response to a Western diet (Fig. 3a). Furthermore, Western human populations exhibited lower relative abundances of Prevo- tella and higher relative abundances of Bacteroides and Bifidobacterium, which contrasts sharply with the patterns observed in the vervets (Fig. 3b). Vervets also exhibited increased levels of Cantenibacterium on a Western diet, but both human populations we examined showed the opposite pattern. Finally, in both human populations, rela- tive abundances of Succinovibrio and Treponema were lower for people consuming a Western diet, but diet did not signficantly affect these taxa in vervets. in humans on a Western diet while in vervets its relative abundance increased (Additional file 1: Table S6). The same was true of K01858 (myo-inositol-1-phosphate synthase), K02548 (general secretion pathway protein I), and K02619 (4-amino-4-deoxychorismate lyase), associated with carbohydrate metabolism, cofactor and vitamin metabolism/biosynthesis, and folate biosynthesis, respect- ively (Additional file 1: Table S6). Additionally, vervets showed shifts in the relative abundances of eight genes that did not shift in response to diet in humans in either study. At the pathway level, humans also showed an increased rela- tive abundance of genes associated with carbohydrate me- tabolism when consuming a Western diet, a difference that was not observed in vervets (Additional file 1: Table S7). Humans vs. vervets At the genus level, we also detected lower relative abundances of Clostridium and a tendency for lower relative abundances of Bifidobacterium in TWD-fed vervets as well as higher relative abundances of Desulfovibrio, Prevotella, Canteni- bacterium, and Collinsella (Additional file 1: Table S5). Despite the similarities noted above, we also observed key differences between the gut microbiota of humans and vervets associated with each type of diet. First, exam- ination of the presence or absence of microbial taxa revealed that the vervets clustered more strongly with non-Western humans, especially the Hadza, regardless of whether the vervets were consuming a wild or Western diet (Fig. 2, Additional file 1: Figure S2). Although recent bottlenecks have resulted in low genetic diversity in both vervet populations sampled [28, 32], gut microbiome data from another wild, non-human primate species (Cercoce- bus agilis, Central African Republic) also cluster with the vervets (Additional file 1: Figure S3), indicating that in- creased genetic diversity among human populations com- pared to the vervet populations is not driving the observed patterns. ( ) Predictive metagenomics also revealed differences in the vervet gut microbiome in response to diet (F1,57 = 4.94, p = 0.01, r2 = 0.08). Specifically, 22 genes differed in relative abundance between wild vervets and TWD-fed vervets. For example, K00432, a glutathione peroxidase, which is associ- ated with lipid and amino acid metabolism, was present at higher relative abundances in vervets consuming a TWD than wild vervets (Additional file 1: Table S6). The same pattern was observed in the relative abundance of K02619, a 4-amino-4-deoxychorismate lyase, which is associated with folate biosynthesis. In contrast, K00068 (sorbitol-6- phosphate 2-dehydrogenase), K00844 (hexokinase), and K05884 (L-2-hydroxycarboxylate dehydrogenase), with roles in fructose and mannose metabolism, carbohy- drate and lipid metabolism, and coenzyme M biosyn- thesis, respectively, were all predicted in lower relative abundances in TWD-fed vervets (Additional file 1: Table S6). At the pathway level, TWD-fed vervets exhib- ited increased relative abundances of genes associated with amino acid metabolism (Additional file 1: Table S7). In addition, Chao1 species diversity estimates tended to be lower in humans consuming a Western diet vs. a non-Western diet (Fig. 1), while diversity in vervets was unaffected by diet. Compared to humans, vervets also exhibited distinct changes in the relative abundances of microbial taxa in response to a Western diet. Discussion Predictive metagenomic data also showed strong differ- ences in the reactions of the vervet and human gut micro- biota to diet. Although the relative abundance of K00432 (glutathione peroxidase) changed in response to diet in both humans and vervets, its relative abundance decreased As hypothesized, our data indicate that the composition of the human and vervet gut microbiota is distinct on a Western diet. However, we did not observe differences in the relative abundance of fewer microbial taxa and Fig. 3 Human and vervet gut microbiomes react differently to a Western diet. Relative abundances of key a phyla and b genera in humans and vervets consuming non-Western and Western diets. Western humans are from Italy (Schnorr et al. [3]) and the USA (Yatsunenko et al. [2]). Non-Western humans are from Tanzania (Hadza, Schnorr et al. [3]), Venezuela (Guahibo, Yatsunenko et al. [2]), and Malawi (Yatsunenko et al. [2]). Stars indicate significant differences (FDR-corrected p < 0.05) in relative abundances between diets for both humans and vervets. However, Bacteroides relative abundances were not significantly different between diets for vervets Fig. 3 Human and vervet gut microbiomes react differently to a Western diet. Relative abundances of key a phyla and b genera in humans and vervets consuming non-Western and Western diets. Western humans are from Italy (Schnorr et al. [3]) and the USA (Yatsunenko et al. [2]). Non-Western humans are from Tanzania (Hadza, Schnorr et al. [3]), Venezuela (Guahibo, Yatsunenko et al. [2]), and Malawi (Yatsunenko et al. [2]). Stars indicate significant differences (FDR-corrected p < 0.05) in relative abundances between diets for both humans and vervets. However, Bacteroides Fig. 3 Human and vervet gut microbiomes react differently to a Western diet. Relative abundances of key a phyla and b genera in humans and vervets consuming non-Western and Western diets. Western humans are from Italy (Schnorr et al. [3]) and the USA (Yatsunenko et al. [2]). Non-Western humans are from Tanzania (Hadza, Schnorr et al. [3]), Venezuela (Guahibo, Yatsunenko et al. [2]), and Malawi (Yatsunenko et al. [2]). Stars indicate significant differences (FDR-corrected p < 0.05) in relative abundances between diets for both humans and vervets. However, Bacteroides relative abundances were not significantly different between diets for vervets Fig. 3 Human and vervet gut microbiomes react differently to a Western diet. Relative abundances of key a phyla and b genera in humans and vervets consuming non-Western and Western diets. Discussion Western humans are from Italy (Schnorr et al. [3]) and the USA (Yatsunenko et al. [2]). Non-Western humans are from Tanzania (Hadza, Schnorr et al. [3]), Venezuela (Guahibo, Yatsunenko et al. [2]), and Malawi (Yatsunenko et al. [2]). Stars indicate significant differences (FDR-corrected p < 0.05) in relative abundances between diets for both humans and vervets. However, Bacteroides relative abundances were not significantly different between diets for vervets Amato et al. Microbiome (2015) 3:53 Page 5 of 9 exposure to local microbial pools, and antibiotic use. Like- wise, future studies must control for the potential effects of non-human primate captivity such as host social contact networks and early life influences on the gut microbiota. However, it is unlikely that the results we present here are primarily driven by these potential confounds. Patterns in confounding factors such as antibiotic use across populations of the same host are likely to be similar for both humans and vervets, with Western humans and captive vervets generally having more exposure to antibiotics than non-Western humans and wild vervets. Therefore, while diet may not be the only factor causing differences between populations of the same host species, the comparisons of patterns between the two host species remains valid. Additionally, the captive population of vervets was taken from St. Kitts between 1975 and 1980. As a result, the two popula- tions are only separated by 3–8 generations, reducing the impact of host genetics as a potential confound for the vervets. genes in the vervets compared to the humans. Instead, the effect of a Western diet on the vervet gut microbiota was driven by distinct patterns in microbial taxa and genes compared to humans. Many of the patterns we detected in both vervets and humans have been reported in other studies. For ex- ample, we observed elevated microbial richness and higher relative abundances of Prevotella in non-Western humans and elevated relative abundances of Bacteroides in Western humans. These results are concordant with a study of the human gut microbiota that associates diets high in protein and animal fat with high levels of Bacter- oides and diets high in plant carbohydrates with high levels of Prevotella [23, 33]. Similarly, all published stud- ies of Western and non-Western humans to-date report higher microbial richness and higher relative abundances of Prevotella in non-Western populations [2–6, 34]. Discussion In contrast, vervets on a Western diet showed similar micro- bial richness and higher relative abundances of Prevotella, a change that is mirrored in the macaque model [12]. In addition, we detected an elevated relative abundance of Collinsella in humans on a Western diet, and Collinsella has been associated with obesity in other studies of humans [35]. Similarly, while decreased relative abundances of Bac- teroidetes coupled with increased relative abundances of Firmicutes, like those we observed in Western humans from the USA, are not reported in many studies comparing Western and non-Western humans [4, 5, 34]; they have been associated with obesity in human and mouse studies [36]. In our TWD-fed vervets, we measured the opposite patterns, and these findings are analogous to previous work contrasting a high-fat diet to a captive control (chow) diet in a macaque model (Macaca fuscata) where a shift to a high-fat diet also resulted in significant increases in Bacter- oidetes, especially Prevotella and no significant variations in Firmicutes [12]. Despite the comparison of multiple datasets in this analysis, studies generally indicate that host diet has the strongest effect on the gut microbiota compared to other factors [37], reducing the potential for confounding fac- tors associated with different host populations to drive the observed patterns. For example, when human studies broadly control for host ethnic backgrounds, they continue to illustrate strong impacts of Western and non- Western diets, suggesting that differences in gut micro- biota composition between populations of the same host species with distinct diets are unlikely to be genetically- driven [6]. Furthermore, although the vervet Western diet is distinct from a human Western diet, it is important to note that Western and non-Western human diets vary markedly across populations (e.g., USA vs. Italy and Malawi vs. Venezuela vs. Tanzania; [2, 3]). Despite differ- ences within these diet categories, Western and non- Western human populations still cluster together in terms of gut microbiota composition, and other studies compar- ing Western and non-Western humans consistently report similar patterns [2–6, 34], suggesting that differences in the human and vervet gut microbiota are not driven by subtle variation in the composition of either a Western or non- Western diet or by idiosyncrasies in the selected datasets. Discussion Likewise, while the data we present were not all generated using the same methodology, the potential effects of dis- tinct DNA extraction protocols, PCR primers, or sequen- cing platforms on gut microbiota data have been shown to be small [38, 39] compared to the effects of diet. Addition- ally, all DNA extractions utilized bead-beating step, which reduces extraction bias [40], and our use of the same closed-reference OTU picking pipeline (see methods) on rarefied data also reduces the potential effect of se- quencing error and read depth on the results [41]. These commonalities suggest that our results are not unique to the human and vervet populations that we ex- amined in this study. The human gut microbiota and its response to an easy-to-digest Western diet that is low in fiber differ fundamentally from the non-human primate gut microbiota. These findings provide the first evidence implying a specialization of the human gut microbiota. Therefore, while non-human primates may serve as comparative biomedical models for other aspects of hu- man physiology [28], we suggest that the non-human primate gut microbiota may not provide an ideal direct model for understanding the effect of the human gut microbiota on host metabolism and nutrition in the context of a Western diet. It is important to note that both the human and vervet data used in this study compare distinct populations, and therefore it is impossible to control for potential non-diet influences on the gut microbiota such as host genetics, Page 6 of 9 Amato et al. Microbiome (2015) 3:53 protein, high-fat diet can impact the gut microbiota almost immediately in humans, and changes are similar to those observed across populations with distinct diets [24], suggesting that a 6-month interval was sufficient. Although the captive vervets weighed more than the wild vervets both before and after the diet challenge, indicating an effect of captivity on body weight, this effect is likely a result of increased food availability and decreased activity, both of which would be expected to exacerbate weight gain and negative health outcomes on a Western diet, not mitigate them. While additional research is necessary to confirm and examine this pat- tern in more detail, we suggest that a side effect of the proposed unique human gut microbiota may be an increased susceptibility to obesity and metabolic disor- ders, particularly when hosts are consuming a high- protein, high-fat diet. Discussion If the non-human primate gut microbiota does, in fact, possess properties that make it resistant to obesity when subjected to a Western diet, it may open new avenues of exploration for translational metabolic therapies. What might confer this resistance and whether it can be maintained over generations of Western diet consumption remains to be investigated, but understanding these factors could help develop treatments to improve human resistance to obesity via the gut microbiota. Therefore, while we cannot completely eliminate the biases of study-specific methodology on this dataset, we are confident that the patterns presented are biologically-driven. Although data for other non-human primate taxa must be collected, we propose two explanations for the distinct responses of the human and non-human primate gut microbiota to diet. First, it is possible that the unique properties observed in the human gut microbiota are simply the result of unique human diet and physiology (namely an easy-to-digest, meat-heavy diet and increased brain size, reduced gut size, increased fat deposition, and decreased muscle mass [17–19]). Since both host diet and physiology drive gut microbiota composition [42], evolu- tionary changes in human diet and physiology could have easily led to a distinct gut microbiota. Another possible explanation is that, in addition to other factors such as diet, unique properties of the human gut microbiota con- tributed to the evolution of human physiology. For ex- ample, if over time the human gut microbiota shifted in a way that confers an increased capacity for energy produc- tion and storage, it could have promoted increased brain size during human evolution. Similarly, because the hu- man gut microbiota plays a role in regulating host energy intake and fat production [43], and an increased capacity to store energy as fat has been hypothesized to have en- abled humans to develop larger brains [18, 19], changes in the gut microbiota that affected host metabolic pathways could have contributed to the evolution of the human brain as well. To distinguish between these two alterna- tives, further studies are necessary that measure the meta- bolic potential of the human and non-human gut microbiota and more directly compare the physiological consequences of consuming a Western diet. However, a recent study suggests that large-brained primates endure seasonal periods of food limitation more successfully than small-brained primates [44]. Conclusions Although animal models such as non-human primates are commonly used to test mechanisms of human gut microbe interactions, direct comparisons of the influ- ence of commonly investigated factors on the gut micro- biota of humans and non-human primates are lacking. This study demonstrates that similar host diets differen- tially affect the human and non-human primate gut microbiota. These results indicate that non-human pri- mates are not appropriate models for directly testing the relationships between human diet, physiology and the gut microbiota. However, the observed patterns have implica- tions for human evolution since they suggest an association between the gut microbiota and unique shifts in human diet and physiology across evolutionary time. A more de- tailed examination of this association has the potential to transform our understanding of the role of gut microbes in human biology. Finally, our data indicated that, in addition to the gut microbiota, the physiological responses of humans and non-human primates to a Western diet may be distinct. Although vervets have an adverse metabolic reaction to a Western diet and can become obese in captivity [46, 47], most of the TWD-fed vervets did not gain weight during the study (Additional file 1: Table S8). It is possible that two factors restricted vervet weight gain: (1) finite food availability and (2) study duration (6 months). We regard these factors as unlikely to be limiting because the ver- vets were provided with enough food to result in a 10 % daily surplus, and no obvious reduction in food intake was observed [47]. Moreover, a sufficiently high- Discussion Together with evidence that gut microbes compensate for periodically reduced energy intake in some wild primates [45], these data could indicate a role for the gut microbiota in buffering hosts nutritionally against energetically expensive physiological adaptations. Sample collection and processing l b d l DNA was extracted from rectal swabs using a MoBio UltraClean Soil DNA Isolation kit (MoBio, Carlsbad, CA, USA) per manufacturer instruc- tions. The V1–V3 region of the bacterial 16S rRNA gene was amplified using pyrotagged primers (MID 1-15) 27f and 534r. Amplicons were checked for specificity with gel electrophoresis and cleaned using the AMPure XP system (Beckman Coulter, Danvers, MA, USA). Amplicons were sequenced using 454 pyrosequencing technology at the J. Craig Venter Institute and/or the University of Illinois KECK Center. No differences were seen in the microbial taxonomic distributions of samples sequenced at both sequencing centers. There- fore, sequencing center was eliminated as a possible effect. Chao1 diversity estimates were produced for each sample using QIIME. We also predicted the metagen- ome associated with each sample using Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt 1.0.0) [27] after normal- izing for 16S copy number. The average Nearest Se- quenced Taxon Index was 0.11 ± 0.03, with Western humans falling at the lower end of the range and non- Western humans and vervets falling at the higher end. We tested for significant differences between wild and TWD vervets as well as Western and non-Western humans from each human study. Although inter-study effects on patterns of gut microbial community com- position have been demonstrated to be low [38, 39], we tested for the effect of diet within each individual data- set independently to eliminate the potential for con- founding factors. Principle coordinate analyses based on unweighted and weighted UniFrac distances were used to visualize differences in gut microbial commu- nity composition among populations. PERMANOVA (R software, version 3.0.2, adonis package) was used to test for the effect of diet on microbial community compos- ition and predicted metagenomes for all three datasets. Similarly, for each dataset, we tested for differences in the relative abundances of microbial taxa and genes be- tween wild and TWD vervets or between Western and non-Western humans using a series of Kruskal-Wallis tests at each taxonomic level. p values were adjusted using either a family-wide detection rate (microbial taxonomy) or a Bonferroni correction (predicted meta- genomes). All statistics were performed using QIIME’s implementation of R. Raw sequence data from non-Western and Western human populations published by Yatsunenko et al. [2] and Schnorr et al. [3] were also obtained for analysis. Non- Western human populations were sampled in Malawi, Venezuela (Guahibo Amerindians), and Tanzania (Hadza). Sample collection and processing l b d l Rectal swabs were used to sample the gut microbiota of 27 wild and captive vervets. In January 2010, 15 wild vervet monkeys were captured, sampled, and released on the island of St. Kitts. Wild vervets on St. Kitts are not provisioned and consume mainly ripe fruits, flowers, and seeds [48]. We considered this a non-Western diet. In Page 7 of 9 Page 7 of 9 Amato et al. Microbiome (2015) 3:53 database with OTUs defined as sharing ≥97 % identity. This strategy made it possible to compare patterns across studies despite the use of different primers and/ or sequencing platforms and reduced the impact of se- quencing error on the data. Due to the use of 454 tech- nology in two of the datasets, we rarefied the number of sequences from each sample to 1000 for all studies. Previ- ous studies suggest that this level allows for meaningful comparison [50]. Additionally, nine TWD vervet samples were removed from analysis as a result of unusually high relative abundances of Brachyspira. Several Brachyspira species are known gut pathogens, and loose stools were observed in some of the enclosures during the diet trial. Removing these samples did not affect the overall patterns observed in beta diversity within and across studies (Additional file 1: Figure S4, S5) but did lead to Brachy- spira relative abundances in a more typical range. September 2009, 23 vervets at the Wake Forest University Primate Center were sampled in a similar manner. These vervets are directly descended from the St. Kitts popula- tion (with an initial group obtained in 1975) and were fed a typical Western human diet (TWD; LabDiet 5L0P, Purina, St. Louis, MO; 18 % protein, 37 % fat, 45 % carbohydrates, 9 % fiber) for 6 months before sampling. Sampling of and care for the captive TWD vervets was approved by the Institutional Animal Care and Use Committee at Wake Forest University (IACUC #A10-091). Sampling of the wild monkeys was approved by the University of Illinois Institu- tional Animal Care and Use Committee (IACUCs #08044, #11046) as well as the University of California Los Angeles IACUC (#2009-053-13). All sampling was carried out in ac- cordance with the approved guidelines at each institution. All rectal swabs were immediately placed in RNAlater (Qiagen Inc., Valencia, CA, USA) and stored at −80 °C until processing. Data analysis All All sequence data were quality filtered so that se- quences shorter than 200 nt, longer than 1000 nt, con- taining incorrect primer sequences, more than six ambiguous base calls and/or homopolymers longer than 7 nt were discarded, resulting in 387,785,923 sequences across all studies. All subsequent data ana- lyses were performed using QIIME (version 1.8.0) [49]. Operational taxonomic units (OTUs) were picked closed-reference against the Green Genes 13_8 Sample collection and processing l b d l Western human populations were sampled in the USA and Italy. Only individuals between the ages of 18 and 50 were included from the Yatsunenko et al. [2] dataset (Malawi/Venezuela N = 76 people, USA N = 118 people), while all samples from Schnorr et al. [3] were included (Hadza N = 17 people, Italy N = 11 people). References References 1. Turnbaugh PJ, Ridaura VK, Faith JJ, Rey FE, Knight R, Gordon HA. The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice. Sci Transl Med. 2009;1:6ra14. 2. Yatsunenko T, Rey FE, Manary MJ, Trehan I, Dominguez-Bello MG, Contreras M, et al. Human gut microbiome viewed across age and geography. Nature. 2012;486(7402):222–7. 3. Schnorr SL, Candela M, Rampelli S, Centanni M, Consolandi C, Basaglia G, et al. Gut microbiome of Hadza hunter-gatherers. Nat Comm. 2014;5:3654. doi:10.1038/ncomms4654. 4. Obregon-Tito AJ, Tito RY, Metcalf JL, Sankaranarayanan K, Clemente JC, Ursell LK, et al. Subsistence strategies in traditional societies distinguish gut microbiomes. Nat Comm. 2015;6:6505. 5. Clemente JC, Pehrsson EC, Blaser MJ, Sandhu K, Gao Z, Wang B, et al. The microbiome of uncontacted Amerindians. Sci Adv. 2015;1(3):e1500183. 6. Ou J, Carbonero F, Zoetendal EG, DeLaney JP, Wang M, Newton K, et al. Diet, microbiota, and microbial metabolitesin colon cancer risk in rural Africans and African-Americans. Am J Clin Nutr. 2013;98(1):111–20. 7. Hu F, van Dam R, Liu S. Diet and risk of type II diabetes: the role of types of fat and carbohydrate. Diabetol. 2001;44:805–17. 8. Ley RE. Obesity and the human microbiome. Curr Opin Gastroenterol. 2010;26(1):5–11. 1. Turnbaugh PJ, Ridaura VK, Faith JJ, Rey FE, Knight R, Gordon HA. The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice. Sci Transl Med. 2009;1:6ra14. 1. Turnbaugh PJ, Ridaura VK, Faith JJ, Rey FE, Knight R, Gordon HA. The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice. Sci Transl Med. 2009;1:6ra14. 1. Turnbaugh PJ, Ridaura VK, Faith JJ, Rey FE, Knight R, Gordon HA. The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice. Sci Transl Med. 2009;1:6ra14. 29. Gray SB, Langefeld CD, Ziegler JT, Hawkins GA, Wagner JD, Howard TD. Single-nucleotide polymorphisms in the TBF gene are associated with obesity-related phenotypes in vervet monkeys. Obesity. 2011;15:1427–32. 2. Yatsunenko T, Rey FE, Manary MJ, Trehan I, Dominguez-Bello MG, Contreras M, et al. Human gut microbiome viewed across age and geography. Nature. 2012;486(7402):222–7. 2. Yatsunenko T, Rey FE, Manary MJ, Trehan I, Dominguez-Bello MG, Contreras M, et al. Human gut microbiome viewed across age and geography. Nature. 2012;486(7402):222–7. 30. Kavanagh K, Fairbanks LA, Bailey JN, Jorgensen MJ, Wilson M, Zhang L, et al. Availability of supporting data The raw sequence files supporting the results of this art- icle are available in the EBI repository, [deposited upon acceptance]. Page 8 of 9 Page 8 of 9 Page 8 of 9 Amato et al. Microbiome (2015) 3:53 Authors’ contributions KRA wrote the manuscript. KRA and CJY analyzed the data. CJY and GC processed the vervet samples. CAS, JDC, MEBM, AG, TRT, BAW, RMS, KEN, BAW, RK, and SRL provided feedback on the manuscript. SRL, RMS, KEN, BAW, and BAW provided funding. All authors read and approved the final manuscript. 13. Kisidayova S, Varadyova Z, Pristas P, Piknova M, Nigutova K, Petrzelkova KJ, et al. Effects of high- and low-fiber diets on fecal fermentation and fecal microbial populations of captive chimpanzees. Am J Primatol. 2009;71:548–57. Received: 20 May 2015 Accepted: 29 September 2015 Received: 20 May 2015 Accepted: 29 September 2015 28. Jasinska AJ, Schmitt CA, Service SK, Cantor RM, Dewar K, Jentsch JD, et al. Systems biology of the vervet monkey. ILAR J. 2013;54(2):122–43. Acknowledgements 14. Leonard WR, Snodgrass JJ, Robertson ML. Effects of brain evolution on human nutrition and metabolism. Annu Rev Nutr. 2007;27:311–27. This work was supported by the NSF grant #0935347 (Human Origins Moving in New Directions – HOMINID, PI: Rebecca M. Stumpf), NSF grant BCS-0820709 (PI: Rebecca M. Stumpf), NCRR P40 grant #RR019963 and VA contract #VA247-P-0447 (PI: Jay Kaplan) and the NIH grant #5R01RR016300 (Integrated genetic and genomic resources for a model system, PI: Nelson Freimer). The authors are especially grateful to Jay Kaplan, Matt Jorgensen, and Nelson Freimer for providing samples from the Wake Forest University-UCLA-VA Vervet Research Colony and to Klara Petrzelkova and WWF, Bayanga CAR for facilitating access to mangabey samples. They also wish to thank Marcus Gillis and Manolito Torralba for assistance in sequencing at JCVI. Nelson Freimer and Ania Jasinska provided helpful commentary. 15. Teaford MF, Ungar PS. Diet and the evolution of the earliest human ancestors. PNAS. 2000;97(25):13506–11. 16. Carmody RN, Wrangham RW. The energetic significance of cooking. J Human Evol. 2009;57(4):379–91. 17. Aiello LC, Wheeler P. The expensive-tissue hypothesis: the brain and digestive system in human and primate evolution. Curr Anthr. 1995;32(2):199–221. 18. Kuzawa CW. Adipose tissue in human infancy and childhood: an evolutionary perspective Yearb Phys Anthropol 1998 41 177 209 8. Kuzawa CW. Adipose tissue in human infancy and childhood: an 18. Kuzawa CW. Adipose tissue in human infancy and childhood: an evolutionary perspective. Yearb Phys Anthropol. 1998;41:177–209. 18. Kuzawa CW. Adipose tissue in human infancy and childhood: an evolutionary perspective. Yearb Phys Anthropol. 1998;41:177–209. 19. Leonard WR, Robertson ML, Snodgrass JJ, Kuzawa CW. Metabolic correlates of hominid brain evolution. Comp Biochem Physiol A. 2003;135:5–15. 20. Wlasiuk G, Nachman MW. Adaption and constraint at toll-like receptors in primates. Mol Biol Evol. 2010;27(9):2172–86. Additional file 9. Larsen N, Vogensen FK, van den Berg FWJ, Nielsen DS, Andreasen AS, Pedersen BK, et al. Gut microbiota in human adults with type 2 diabetes differs from non-diabetic adults. PLoS One. 2010;5(2):e9085. Competing interests Competing interests 11. Faith JJ, McNulty NP, Rey FE, Gordon JI. Predicting a human gut microbiota’s response to diet in gnotobiotic mice. Science. 2011;333:101–4. The authors declare that they have no competing interests. 12. Ma J, Prince AL, Bader D, Hu M, Ganu R, Baquero K, et al. High-fat maternal diet during pregnancy persistently alters the offspring microbiome in a primate model. Nat Comm. 2014;5:3889. doi:10.1038/ncomms4889. 34. Martinez I, Stegen JC, Maldonado-Gomez MX, Eren AM, Siba PM, Greenhill AR, et al. The gut microbiota of rural Papua New Guineans: Composition, diversity patterns, and ecological processes. Cell Rep. 2015;11:527–38. Additional file 1: Supplementary tables and figures. (DOC 2591 KB) 10. Hosseini E, Grootaert C, Verstraete W, Van de Wiele T. Propionate as a health-promoting microbial metabolite in the human gut. Nutr Rev. 2011;69(5):245–58. doi:10.1111/j.1753-4887.2011.00388.x. Author details 1D f 1Department of Anthropology, Northwestern University, Evanston, USA. 2Department of Anthropology, University of Colorado Boulder, Boulder, USA. 3BioFrontiers Institute, University of Colorado Boulder, Boulder, USA. 4Department of Range Sciences, Montana State University, Bozeman, USA. 5Department of Anthropology, University of Illinois, Urbana, USA. 6Department of Anthropology, Boston University, Boston, USA. 7Department of Sociology, Anthropology, and Women’s Studies, American Military University and American Public University, Charles Town, USA. 8The Institute for Genomic Biology, University of Illinois, Urbana, IL 61801, USA. 9Department of Ecology, Evolution, and Behavior, University of Minnesota, Minneapolis, USA. 10Department of Anthropology, University of Wisconsin, Milwaukee, USA. 11Department of Genetics, University of the Free State, Bloemfontein, South Africa. 12Department of Microbiology, University of Illinois, Urbana, USA. 13The J. Craig Venter Institute, Rockville, USA. 1Department of Anthropology, Northwestern University, Evanston, USA. 2Department of Anthropology, University of Colorado Boulder, Boulder, USA. 3BioFrontiers Institute, University of Colorado Boulder, Boulder, USA. 4Department of Range Sciences, Montana State University, Bozeman, USA. 5Department of Anthropology, University of Illinois, Urbana, USA. 6Department of Anthropology, Boston University, Boston, USA. 7Department of Sociology, Anthropology, and Women’s Studies, American Military University and American Public University, Charles Town, USA. 8The Institute for Genomic Biology, University of Illinois, Urbana, IL 61801, USA. 9Department of Ecology, Evolution, and Behavior, University of Minnesota, Minneapolis, USA. 10Department of Anthropology, University of Wisconsin, Milwaukee, USA. 11Department of Genetics, University of the Free State, Bloemfontein, South Africa. 12Department of Microbiology, University of Illinois, Urbana, USA. 13The J. Craig Venter Institute, Rockville, USA. 14 15 21. Casanova JL, Abel L, Quintana-Murci L. Human TLRs and IL-1Rs in host defense: Natural insights from evolutionary, epidemiological, and clinical genetics. Immunology. 2011;29:447–91. 22. Kargman S, Charleson S, Cartwright M, Frank J, Riendeau D, Mancini J, et al. Characterization of prostaglandin G/H synthase 1 and 2 in rat, dog, monkey, and human gastrointestinal tracts. Gastroenterol. 1996;111(2):445–54. 23. Wu GD, Chen J, Hoffmann C, Bittinger K, Chen YY, Keilbaugh SA, et al. Linking long-term dietary patterns with gut microbial enterotypes. Science. 2011;334:105–8. 24. David LA, Maurice CF, Carmody RN, Gootenberg DB, Button JE, Wolfe BE, et al. Diet rapidly and reproducibly alters the human gut microbiome. Nature. 2014;505:559–66. doi:10.1038/nature12820. 25. Tremaroli V, Backhed F. Functional interactions between the gut microbiota and host metabolism. Nature. 2012;489:242–9. 14Department of Animal Sciences, University of Illinois, Urbana, USA. 15School of Medicine, University of California San Diego, La Jolla, USA. Author details 1D f 16Center for Neurobehavioral Genetics, University of California, Los Angeles, CA, USA. 14Department of Animal Sciences, University of Illinois, Urbana, USA. 15School of Medicine, University of California San Diego, La Jolla, USA. 16Center for Neurobehavioral Genetics, University of California, Los Angeles, CA, USA. 26. Moeller AH, Li Y, Ngole EM, Ahuka-Mendeke S, Lonsdorf EV, Pusey AE, et al. Rapid changes in the gut microbiome during human evolution. PNAS. 2014;111(46):16431–5. doi:10.1073/pnas.1419136111. 27. Langille MGI, Zaneveld J, Caporaso JG, McDonald D, Knights D, Reyes JA, et al. Predictive functional profiling of microbial communitiesusing 16S rRNA marker gene sequences. Nat Biotech. 2013;8:1–10. Received: 20 May 2015 Accepted: 29 September 2015 References Characterization and heritability of obesity and associated risk factors in vervet monkeys. Obesity. 2007;15:1666–74. 3. Schnorr SL, Candela M, Rampelli S, Centanni M, Consolandi C, Basaglia G, et al. Gut microbiome of Hadza hunter-gatherers. Nat Comm. 2014;5:3654. doi:10.1038/ncomms4654. 3. Schnorr SL, Candela M, Rampelli S, Centanni M, Consolandi C, Basaglia G, et al. Gut microbiome of Hadza hunter-gatherers. Nat Comm. 2014;5:3654. doi:10.1038/ncomms4654. 31. Jaffe KE, Isbell LA. The guenons: polyspecific associations in socioecological perspective. In: Campbell C, Fuentes A, MacKinnon KC, Bearder SK, Stumpf RM, editors. Primates in perspective. Second Editionth ed. New York: Oxford University Press; 2011. p. 277–99. 4. Obregon-Tito AJ, Tito RY, Metcalf JL, Sankaranarayanan K, Clemente JC, Ursell LK, et al. Subsistence strategies in traditional societies distinguish gut microbiomes. Nat Comm. 2015;6:6505. 4. Obregon-Tito AJ, Tito RY, Metcalf JL, Sankaranarayanan K, Clemente JC, Ursell LK, et al. Subsistence strategies in traditional societies distinguish gut microbiomes. Nat Comm. 2015;6:6505. 32. Aarnink A, Jacquelin B, Dauba A, Hebrard S, Moureaux E, Muller-Trutwin M, et al. MHC polymorphism in Caribbean African green monkeys. Immunogen. 2014;66(6):353–60. 5. Clemente JC, Pehrsson EC, Blaser MJ, Sandhu K, Gao Z, Wang B, et al. The microbiome of uncontacted Amerindians. Sci Adv. 2015;1(3):e1500183. 5. Clemente JC, Pehrsson EC, Blaser MJ, Sandhu K, Gao Z, Wang B, et al. The microbiome of uncontacted Amerindians. Sci Adv. 2015;1(3):e1500183. 6. Ou J, Carbonero F, Zoetendal EG, DeLaney JP, Wang M, Newton K, et al. Diet, microbiota, and microbial metabolitesin colon cancer risk in rural Africans and African-Americans. Am J Clin Nutr. 2013;98(1):111–20. 33. De Filippo C, Cavalieri D, Di Paola M, Ramazzotti M, Poullet JB, Massart S, et al. Impact of diet in shaping gut microbiota revealed by a comparative study in children from Europe and rural Africa. PNAS. 2010;107(33):14691–6. 7. Hu F, van Dam R, Liu S. Diet and risk of type II diabetes: the role of types of fat and carbohydrate. Diabetol. 2001;44:805–17. 34. Martinez I, Stegen JC, Maldonado-Gomez MX, Eren AM, Siba PM, Greenhill AR, et al. The gut microbiota of rural Papua New Guineans: Composition, diversity patterns, and ecological processes. Cell Rep. 2015;11:527–38. 8. Ley RE. Obesity and the human microbiome. Curr Opin Gastroenterol. 2010;26(1):5–11. Page 9 of 9 Page 9 of 9 Amato et al. Microbiome (2015) 3:53 Amato et al. Microbiome (2015) 3:53 35. Turnbaugh PJ, Hamady M, Yatsunenko T, Cantarel BL, Duncan A, Ley RE, et al. References A core gut microbiome in obese and lean twins. Nature. 2009;457:480–4. 36. Turnbaugh PJ, Ley RE, Mahowald MA, Magrini V, Mardis ER, Gordon JI. An obesity-associated gut microbiome with increased capacity for energy harvest. Nature. 2006;444(21):1027–31. harvest. Nature. 2006;444(21):1027–31. 37. Xu ZZ, Knight R. Dietary effects on human gut microbiome diversity. Br J Nutr. 2015;113(S1):S1–5. 38. Lozupone C, Stombaugh J, Gonzalez A, Ackermann G, Wendel D, Vasquez- Baeza Y, et al. Meta-analysis studies of the human microbiota. Genome Res. 2013;23:1704–14. doi:10.1101/gr.151803.112. 39. Wesolowski-Andersen A, Bahl MI, Carvalho V, Kristiansen K, Sicheritz-Ponten T, Gupta R, et al. Choice of bacterial DNA extraction method from fecal material influences community structure as evaluated by metagenomic analysis. Microbiome. 2014;5(2):19. 40. Walker AW, Martin J, Scott P, Parkhill J, Flint HJ, Scott KP. 16s rRNA gene- based profiling of the human infant gut microbiota is strongly influenced by sample processing and PCR primer choice. Microbiome. 2015;3:26. 41. Schloss PD, Gevers D, Westcott SL. Reducing the effects of PCR amplification and sequencing artifacts on rRNA-based studies. PLoS One. 2011;6(12):e27310. 42. Dethlefsen L, Eckburg PB, Bik EM, Relman DA. Assembly of the human intestinal microbiota. Trends Ecol Evol. 2006;21(9):517–23. 43. Backhed F, Ding H, Wang T, Hooper LV, Koh GY, Nagy A, et al. The gut microbiota as an environmental factor that regulates fat storage. PNAS. 2004;101:15718–23. 44. van Woerden JT, Willems EP, Van Schaik CP, Isler K. Large brains buffer energetic effects of seasonal habitats in catarrhine primates. Evol. 2012;66(1):191–9. 45. Amato KR, Leigh SR, Kent A, Mackie RI, Yeoman CJ, Stumpf RM, et al. The gut microbiota appears to compensate for seasonal diet variation in the wild black howler monkey (Alouatta pigra). Microb Ecol. 2014;69(2):434–43. doi:10.1007/s00248-014-0554-7. 46. Voruganti VS, Jorgensen MJ, Kaplan JR, Kavanagh K, Rudel LL, Temel R, et al. Significant genotype by diet (GxD) interaction effects on cardiometabolic responses to a pedigree-wide, dietary challenge in vervet monkeys (Chlorocebus aethiops sabaeus). Am J Primatol. 2013;75(5):491–9. 47. Jorgensen MJ, Aycock ST, Clarkson TB, Kaplan JR. Effects of a Western-type diet on plasma lipids and other cardiometabolic risk factors in African green monkeys (Chlorocebus aethips sabaeus). J Am Assoc Lab Anim Sci. 2013;52(4):448–53. 48. Chapman C, Fedigan LM, Fedigan L. Ecological and demographic influences on the pattern of association in St. Kitts vervets. Primates. 1988;29:417–21. 49. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. Amato et al. Microbiome (2015) 3:53 References QIIME allows analysis of high-throughputcommunity sequencing data. Nature Methods. 2010;7:335-336. 50. Hamady M, Knight R. Microbial community profiling for human microbiome projects: Tools, techniques and challenges. Genome Res. 2009;19:1141–52. 50. Hamady M, Knight R. Microbial community profiling for human microbiome projects: Tools, techniques and challenges. Genome Res. 2009;19:1141–52. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission
https://openalex.org/W2461436928	https://www.nature.com/articles/srep28567.pdf	English	null	Inhomogeneous Point-Processes to Instantaneously Assess Affective Haptic Perception through Heartbeat Dynamics Information	Scientific reports	2,016	cc-by	11,797	Inhomogeneous Point-Processes to Instantaneously Assess Affective Haptic Perception through Heartbeat Dynamics Information G. Valenza1,2, A. Greco2, L. Citi3, M. Bianchi2,4, R. Barbieri1 & E. P. Scilingo2 received: 05 November 2015 accepted: 07 June 2016 Published: 30 June 2016 This study proposes the application of a comprehensive signal processing framework, based on inhomogeneous point-process models of heartbeat dynamics, to instantaneously assess affective haptic perception using electrocardiogram-derived information exclusively. The framework relies on inverse-Gaussian point-processes with Laguerre expansion of the nonlinear Wiener-Volterra kernels, accounting for the long-term information given by the past heartbeat events. Up to cubic-order nonlinearities allow for an instantaneous estimation of the dynamic spectrum and bispectrum of the considered cardiovascular dynamics, as well as for instantaneous measures of complexity, through Lyapunov exponents and entropy. Short-term caress-like stimuli were administered for 4.3–25 seconds on the forearms of 32 healthy volunteers (16 females) through a wearable haptic device, by selectively superimposing two levels of force, 2 N and 6 N, and two levels of velocity, 9.4 mm/s and 65 mm/s. Results demonstrated that our instantaneous linear and nonlinear features were able to finely characterize the affective haptic perception, with a recognition accuracy of 69.79% along the force dimension, and 81.25% along the velocity dimension. Emotions such as anger, fear, and love can be communicated through touch to the forearm1,2. Indeed, the human tactile system is very sensible to affective stimuli like caressing, stroking or patting, thanks to the activity of low-threshold mechanoreceptors called CT fibers1,2. Through this knowledge, the relationship between the phys- ical characteristics of a haptic stimulus and its pleasantness was previously studied3–5. Most pleasant stimuli were associated to the lowest applied force3, with movements velocity in the range 1–10 cm/s (3 cm/s considered as the most pleasant)4. Of note, these findings are consistent with stimuli administered in different parts of the body (forehead, arm, palm, thigh)6.h p g The dynamics of hearth contractions, which is modulated by the Autonomic Nervous System (ANS) activity, is significantly affected by emotion elicitation and by passive touch7–15. Specifically, linear and nonlinear analysis of Heart Rate Variability (HRV)16 revealed parasympatetic changes after massage13,17,18. Of note, the use of nonlinear quantifiers is justified by the fact that heartbeat dynamics shows exemplary nonlinear behavior, mainly generated through integration of multiple neural signaling at the level of the sinoatrial node19.f Major shortcomings of current signal processing methodologies used to objectively assess affective haptic stimuli are related to the stimulus duration. Standard linear and nonlinear HRV measures, in fact, require rela- tively long-term recordings to accurately characterize the emotional state of a subject. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 www.nature.com/scientificreports/ estimates are biased by noise statistics (e.g., white or 1/f noise) underlying physiological dynamics, and interpo- lation techniques20,21. q To overcome these limitations, we here propose the application of inhomogeneous point-process models of heartbeat dynamics to assess short-term affective haptic perception in an instantaneous fashion, such that caress-like stimuli lasting for 4.3–25 seconds can be therefore automatically recognized. As an input, the proposed methodology takes unevenly sampled heartbeat interval information exclusively, derived from the electrocardi- ogram. To our knowledge, the use instantaneous measures of linear and nonlinear autonomic control derived from cardiovascular dynamics in order to characterize and automatically discern physical properties of affective haptic perception is novel in the current literature. Motivations of this study are related to scientific knowledge, and possible potential application in the field of experimental psychology and assessment of patients with mental disorders.h The model underlying these estimates comprises probability density functions (pdf) characterizing and pre- dicting the time until the next event occurs, as a function of the past history. Of note, the derived instantaneous measures can be estimated without applying any interpolation techniques to the original series, and are associ- ated to effective goodness-of-fit measures20–23. The framework relies on inverse-Gaussian point-processes with Laguerre expansion of the nonlinear Wiener-Volterra kernels, accounting for the long-term information given by the past heartbeat events20,21,23. This methodology has been successfully applied to the study of brain-heart interaction24, assessment of Parkinson’s disease25 and depression in bipolar disorder26,27. p p In this study, the use of instantaneous nonlinear estimates is emphasized. We consider up to cubic-order nonlinearities, allowing for a comprehensive assessment through the dynamic spectrum and bispectrum of the considered cardiovascular dynamics23, as well as for instantaneous measures of complexity, through Lyapunov exponents and entropy20,21. As mentioned above, although the underlying physiological dynamics linked to these measures is still unknown, it has been widely accepted that the quantification of ANS nonlinear and complex dynamics provides meaningful information on psychophysiological and pathological states28–32. Furthermore, measures of time-varying complexity have enhanced discriminant power with respect to standard complexity measures20,21. Short-term caress-like stimuli were administered on the forearms of 32 healthy volunteers (16 females) through an ad-hoc wearable haptic device able to mimic caresses33. In order to develop a processing chain able to actually discern affective perception using electrocardiogram-derived information exclusively, we moved from looking for simple statistical differences between stimulus physical characteristics to building an automatic clas- sification algorithm. www.nature.com/scientificreports/ i Next, we describe the mathematical bases of the inhomogeneous point-process modeling, extending the pre- viously proposed approach reported in23. Then, a brief description of related instantaneous estimates is reported, allowing with experimental set-up and results which are expressed in terms of confusion matrix of classification and complementary statistics. Inhomogeneous Point-Processes to Instantaneously Assess Affective Haptic Perception through Heartbeat Dynamics Information G. Valenza1,2, A. Greco2, L. Citi3, M. Bianchi2,4, R. Barbieri1 & E. P. Scilingo2 For example, observations from 30 seconds to 5 minutes are the least needed for simple HRV spectral analysis16. Nevertheless, actual affec- tive haptic stimuli, such as caresses, usually last for few seconds. Therefore, standard signal processing meth- odologies are not suitable to perform such an assessment. Even considering affective haptic stimuli lasting for more than 30 seconds, the actual emotional perception will be compromised by the saturation of the CT fibers activity3,5. Furthermore, from a methodological point of view, standard methods are generally not suitable to provide accurate nonlinear fitting in the absence of information regarding the system phase space, and their 1Neuroscience Statistics Research Laboratory, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA. 2Department of Information Engineering and Research Centre E Piaggio, University of Pisa, Pisa, Italy. 3School of Computer Science and Electronic Engineering, University of Essex, Colchester, CO43SQ, UK. 4Department of Advanced Robotics, Istituto Italiano di Tecnologia, via Morego 30, 16163 Genova, Italy. Correspondence and requests for materials should be addressed to G.V. (email: g.valenza@ieee.org) Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 1 www.nature.com/scientificreports/ Materials and Methods Point-Process Model of the Heartbeat. A random point process is a stochastic process characterizing the occurrence in time of discrete events. Here, this methodology aims to deriving estimates from the intrinsically discrete, unevenly sampled heartbeat events at each moment in time. To this extent, starting from the electrocar- diogram (ECG), it is possible to define the pdf predicting the next ventricular contraction (R-peaks), through a parametric formulation of the past heartbeat events. Materials and Methods = − = ∞ = = = − ∼ t t g t g i t l t g i i t l t ( , , ( )) RR ( ) ( , ) ( ) ( , , , ) ( ) (2) t N t i p i n i M i M n n j n i RR ( ) 0 0 1 2 1 1 1 1 n n j 1 1 ∑∑ ∑ ∏ + ⋅⋅ .. = ∞ = = = − g i i t l t ( , , , ) ( ) (2) n i M i M n n j n i 2 1 1 1 1 n n j 1 1 (2) where ∑φ = − − = − −− ∼ ∼ ∼ l t n ( ) ( )(RR RR ) (3) i n N t i N t n N t n 1 ( ) ( ) ( ) 1 ∑φ = − − = − −− ∼ ∼ ∼ l t n ( ) ( )(RR RR ) i n N t i N t n N t n 1 ( ) ( ) ( ) 1 (3) is the output of the Laguerre filters just before time t, and is the output of the Laguerre filters just before time t, and ∑ φ α α α α = − − − − = − ( ) ( ) n n k i k ( ) (1 ) ( 1) (1 ) (4) i n i k i k i k k 2 1 2 0 (4) is the ith-order discrete time orthonormal Laguerre function, with n ≥ 0 and α is the discrete-time Laguerre parameter (0 < α < 1) which determines the rate of exponential asymptotic decline of these functions. The coef- ficients g0, {g1(i)}, and {g2(i, j)} correspond to the time-varying zero-, first-, second-order NARL coefficients, respectively. Considering the derivative R-R series improves the achievement of stationarity within the sliding time window W (in this work we have chosen W = 90 sec.)23. ( ) Of note, the corresponding nonlinear autoregressive (NAR) Wiener-Volterra model with degree of nonline- arity 2 and long-term memory34 becomes: ∑ ∑∑ ∑ ∏ µ ξ γ γ γ = + + − + ⋅⋅ .. Materials and Methods − = ∞ − −− = ∞ = ∞ = ∞ = − −− ∼ ∼ ∼ ∼ ∼ t t t i t i i t ( , , ( )) RR ( ) ( , ) (RR RR ) ( , , , ) (RR RR ) (5) t N t i N t i N t i n i i n n j n N t i N t i RR ( ) 0 1 1 ( ) ( ) 1 2 1 1 1 1 ( ) ( ) 1 n j j 1  (5) As mentioned above, as µ ξ t t ( , , ( )) t RR  is defined in continuous time, it is possible to obtain an instantaneous R–R mean estimate at a very fine timescale, which requires no interpolation between the arrival times of two beats. In this work, we consider nonlinearities associated to eq. 5 up to the third-order. Cubic terms, in fact, allow for the estimation of the dominant Lyapunov exponent, whereas quadratic terms account for the high-order spectral estimation (see sections below). Parameter Estimation, Model Selection, Goodness-of-Fit. The optimal time-varying parameter vector ξ(t) is defined as the set maximizing the following local log-likelihood23:   ∫ ∑ ξ ξ τ ξ τ \| = − \| + = + − − + + ∞ + L t U w t u f u t f t d ( ( ) ) ( ) log [ ( , ( ))] log ( , ( )) m m k m i m k k u t t : 1 1 1 1 max k 1 2 1 2 1 (6) where: ( where: • (t − l, t] is a local observation interval of duration l; • (t − l, t] is a local observation interval of duration l; • Um m :1 2 is a subset of R-wave events; 1 2 • m1 = N(t − l) + 1; • m2 = N(t); • imax = max{i1, i2, …, in}; • w(τ) = e−ϖτ is an exponential weighting function. • w(τ) = e−ϖτ is an exponential weighting function. We use a Newton-Raphson procedure to maximize the local log-likelihood in eq. 6. Materials and Methods As such a pdf is a continuous function defined in the time domain, it is possible to obtain instantaneous cardiovascular estimates at any desired time resolution.i p y Mathematically, motivated by both goodness-of-fit and physiological reasons, the pdf characterizing cardio- vascular control dynamics follows an inverse-Gaussian model20,21,23:    ξ ξ π ξ µ ξ µ ξ = − ×      − − −  −      f t t t t u t t u t t t t t u ( , ( )) ( ) 2 ( ) exp 1 2 ( ) ( , , ( )) ( , , ( )) ( ) (1) t j j t t j 0 3 0 RR 2 RR 2 (1) with: • t ∈ (0, T], the observation interval; J ∼ • t ∈ (0, T], the observation interval; J • t ∈ (0, T], the observation interval; J • = u { } j j J 1 the R-wave events, in this study, detected from the ECG, and = ∼ j N t( ) the index of the previous R-wave event before time t • ≤ < < < < < ≤ +   u u u u T 0 k k K 1 1 the times of the events; • RRj = uj − uj−1 > 0 the jth RR interval; • ≤ < < < < < ≤ +   u u u u T 0 k k K 1 1 the times of the events; • RRj = uj − uj−1 > 0 the jth RR interval; j j j 1 j ; N(t) = max{k : uk ≤ t} the sample path of the counting process of the RR interval seris; ∼ j j j 1 j • N(t) = max{k : uk ≤ t} the sample path of the counting process of the RR interval seris; ∼ • τ = = = < ∼ τ − →− N t N t N k u t ( ) ( ) lim ( ) max{ : } t k • = … − − + u ( , RR , RR , , RR ) t j j j j M 1 1 ; • τ = = = < ∼ τ − →− N t N t N k u t ( ) ( ) lim ( ) max{ : } t k • = … − − + u ( , RR , RR , , RR ) t j j j j M 1 1 ; + ( ) t j j j j M 1 1 • ξ(t) the vector of the model time-varying parameters;i • µ ξ t t ( , , ( )) t RR  the first-moment statistic (mean) of the distribution; • µ ξ t t ( , , ( )) t RR  the first-moment statistic (mean) of the distribution; • ξ0(t) > 0 the shape parameter of the inverse Gaussian distribution; Accounting for history dependence, ξ f t t ( , ( )) t  is thus able to predict the next heartbeat event, being para- metrized in its first-order moment as a Nonlinear Autoregressive model with Laguerre expansions (NARL) of the Volterra terms: Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 2 www.nature.com/scientificreports/  ∑ ∑∑ ∑ ∏ µ ξ = + + + ⋅⋅ .. This transformation can be performed in the frequency domain by using the following relationships23 This transformation can be performed in the frequency domain by using the following relationships23 ∑ ∑ ω … + … + × Γ′ Γ′ × Γ′ Γ′ = σ σ σ σ σ σ σ σ σ σ σ σ σ σ = ∈ + + − + + −   H f f f f f f f f f f f ( , , , , , ) ( ) ( ) ( , ) ( , ) 0 p M M p r r r M M r r r M M mid( ) (1) ( ) ( 1) ( 2) ( 1) ( ) 1 (1) 1 ( ) 2 ( 1) ( 2) 2 ( 1) ( ) M ∑ ∑ ω … + … + × Γ′ Γ′ × Γ′ Γ′ = σ σ σ σ σ σ σ σ σ σ σ σ σ σ = ∈ + + − + + −   H f f f f f f f f f f f ( , , , , , ) ( ) ( ) ( , ) ( , ) 0 (7) p M M p r r r M M r r r M M mid( ) (1) ( ) ( 1) ( 2) ( 1) ( ) 1 (1) 1 ( ) 2 ( 1) ( 2) 2 ( 1) ( ) M (7) where where p(f1, , fn) p • Γ′ f( ) 1 1 and Γ′ f f ( , ) 2 1 2 are the Fourier transforms of the extended terms of γ1(i) and γ2(i, j)23, respectively • M is a given integer representing the kernel order ⌈ ⌉ M M id( ) /2 p f f p • Γ′ f( ) 1 1 and Γ′ f f ( , ) 2 1 2 are the Fourier transforms of the extended terms of γ1(i) and γ2(i, j)23, respectively • M is a given integer representing the kernel order p • Γ′ f( ) 1 1 and Γ′ f f ( , ) 2 1 2 are the Fourier transforms of the e • r = 2p − M and σM is the permutation set of q • r = 2p − M and σM is the permutation set of q Since the Volterra kernels induced by the NAR model are nested, the Mth-order kernel can be deduced recursively23. This transformation can be performed in the frequency domain by using the following relationships23 where σ = S f t ( , ) xx RR 2 . where σ = S f t ( , ) xx RR. By integrating the (11) in each frequency band, we can compute the index within the very low frequency (VLF = 0.01–0.05 Hz), low frequency (LF = 0.05–0.15 Hz), and high frequency (HF = 0.15–0.5 Hz) ranges22. Instantaneous Bispectral Analysis. The higher-order spectral representation allows for statistics beyond the sec- ond order, and phase relations between frequency components otherwise suppressed.h Instantaneous Bispectral Analysis. The higher-order spectral representation allows for statistics beyond the sec- ond order, and phase relations between frequency components otherwise suppressed.h p q y p pp The analytical solution for the bispectrum of a nonlinear system response subject to stationary, zero-mean Gaussian input can be found in36. Further details on the instantaneous bispectrum derivation from point-process nonlinear models can be found in23.h Through this powerful computational tool, we here evaluate the instantaneous presence of nonlinearity in heartbeat series by calculating the nonlinerar sympatho-vagal interactions. Specifically, by integrating \|Bis(f1, f2, t)\| in the appropriate frequency bands, it is possible to obtain: ∫ ∫ = = . = . + + t f f t df df LL( ) Bis( , , ) (12) f f 0 0 15 0 0 15 1 2 1 2 1 2 ∫ ∫ = = . = . . + + t f f t df df LH( ) Bis( , , ) (13) f f 0 0 15 0 15 0 4 1 2 1 2 1 2 ∫ ∫ = = . . = . . + + t f f t df df HH( ) Bis( , , ) (14) f f 0 15 0 4 0 15 0 4 1 2 1 2 1 2 ∫ ∫ = = . = . + + t f f t df df LL( ) Bis( , , ) (12) f f 0 0 15 0 0 15 1 2 1 2 1 2 (12) ∫ ∫ = = . = . . + + t f f t df df LH( ) Bis( , , ) (13) f f 0 0 15 0 15 0 4 1 2 1 2 1 2 ∫ ∫ = = . . = . . Materials and Methods This formulation is also used to preprocess all the actual heartbeat data with a previously developed algorithm35, performing a real-time R-R interval error (e.g., peak detection errors and ectopic beats) detection and correction. h fi l b d l l l k l h d l h h g p p Because there is significant overlap between adjacent local likelihood intervals, we start the Newton-Raphson procedure at t with the previous local maximum-likelihood estimate at time t − Δ, where Δ defines the time interval shift to compute the next parameter update. t p p p We determine the optimal model orders based on the Kolmogorov-Smirnov (KS) test and associated KS sta- tistics23. Autocorrelation plots are also considered to test the independence of the model-transformed intervals23. Instantaneous Time, Frequency, and Higher-Order Spectral Analysis. Estimation of the Input-Output Volterra Kernels. In order to provide quantitative tools related to representations defined in the time, frequency, and higher order spectral domains, considering quadratic nonlinearities (with n = 2), estimated parameters of the fully autoregressive eq. 5 have to be linked to classical input-output Wiener-Volterra model23. Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 3 3 www.nature.com/scientificreports/ This transformation can be performed in the frequency domain by using the following relationships23 Instantaneous Analysis in the Time and Frequency Domain. The time-domain characterization is based on the first and the second order moments of the underlying probability structure22. Namely, given the time-varying parameter set ξ(t), the instantaneous estimates of mean µ ξ t t ( , , ( )) t RR  , R-R interval standard deviation σ ξ t t ( , , ( )) RR t 2  , mean heart rate µ ξ t t ( , , ( )) HR t  , and heart rate standard deviation σ ξ t t ( , , ( )) HR t  can be derived at each moment in time as follows22:  σ ξ µ ξ = . t t t t ( , , ( )) ( )/ ( ) (8) RR t RR 2 3 0 µ ξ µ ξ = + − −  t t t ( , , ( )) ( ) (9) HR t RR 1 0 1   σ ξ µ ξ µ ξ =    +      t t t t ( , , ( )) 2 ( ) ( ) (10) HR t RR RR 0 0 2 1/2  σ ξ µ ξ = . t t t t ( , , ( )) ( )/ ( ) (8) RR t RR 2 3 0 (8) (9)  σ ξ µ ξ µ ξ =    +      t t t t ( , , ( )) 2 ( ) ( ) (10) HR t RR RR 0 0 2 1/2 (10) The linear power spectrum can be estimated as: The linear power spectrum can be estimated as: Q H ∫ ξ ω π = − − − − f t t cos S f t H f t H f t H f f f t S f t df ( , , , ( )) 2(1 ( )) ( , ) ( , ) ( , ) 3 2 ( , , , ) ( , ) (11) t xx xx 1 1 3 2 2 2 2 (11) where σ = S f t ( , ) xx RR 2 . where σ = S f t ( , ) xx RR 2 . This transformation can be performed in the frequency domain by using the following relationships23 + + t f f t df df HH( ) Bis( , , ) (14) f f 0 15 0 4 0 15 0 4 1 2 1 2 1 2 (13) (14) Instantaneous Measures of Complexity. Instantaneous Lyapunov Exponents. Considering cubic non- linearities (with n = 3) in the fully autoregressive eq. 5, and using a Fast Orthogonal Search algorithm, it is pos- sible to estimate the complete Lyapunov Exponents (LE) spectrum at each moment in time21. In this study, we Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 4 www.nature.com/scientificreports/ tificreports/ Figure 1. (Top) Experimental protocol timeline and (bottom) the haptic system worn by a subject. Figure 1. (Top) Experimental protocol timeline and (bottom) the haptic system worn by a subject. use the Instantaneous Dominant Lyapunov Exponent (IDLE, λ1), which is the first exponent of the LE spectrum, along with λ2: use the Instantaneous Dominant Lyapunov Exponent (IDLE, λ1), which is the first exponent of the LE spectrum, along with λ2: ∑ λ = ∆ = − N R 1 ln (15) i j N j ii 0 1 ( ) (15) where Δ is the sampling time step, N the length of data samples, and R(j)ii is part of the QR decomposition of the Jacobian of the time series21. where Δ is the sampling time step, N the length of data samples, and R(j)ii is part of the QR decomposition of the Jacobian of the time series21. Instantaneous Entropy Measures. In this study, the estimated instantaneous entropy measures refer to the inho- mogeneous point-process approximate and sample entropy, AI and SI, respectively20. These measures have their foundation in the instantaneous phase space estimation, in which the distance between two points is calculated through KS distance (i.e. the maximum value of the absolute difference between two cumulative distribution functions) between the two pdfs associated to these points. The time-varying radius r(t) is instantaneously expressed as σ = . µ r t( ) 0 2 ( ) 20. µ t e( ) Moreover, AI and SI dynamical values are not seriously affected by the kind of noise underlying the complex system, thus ensuring truly instantaneous tracking of the dynamic system complexity20. Experimental Setup. Thirty-two participants aged 27 ± 2 (16 females) gave their informed consent to take part in the study. This transformation can be performed in the frequency domain by using the following relationships23 No participants reported physical limitations and any experience of mental or personality disorder in their life that would affect the experimental outcomes. All participants gave written informed con- sent to participating in the study, which was approved by the Ethical Committee of the University of Pisa-Pisa University Hospital, Pisa (Italy). All experimental procedures and analyses were carried out in accordance with such approved guidelines. During the experimental protocol the participants were comfortably seated, the right forearm was horizontal and placed on the forearm support, hand palm down. For all trials, participants wore earplugs in order to prevent any auditory cues.hh The ad-hoc robotic device uses a layer of elastic fabric to convey haptic like-caressing stimuli33 (see Fig. 1). The system is endowed with a load cell placed at the basis of the forearm support, prior to the experiment, the load cell was auto calibrated with respect to the forearm weight. When the device was active, two different operating phases were distinguishable: a calibration phase, when forces exerted by the fabric on the user were calibrated, and a movement phase. In this phase, the motors started to coherently rotate and the fabric moves forward and backward over the user forearm, simulating a caress. For further technical details on the device, the reader is invited to refer to ref. 33.f In this study, we considered 4 different stimuli among 2 levels of force (F1 = 2 N, F2 = 6 N) and 2 levels of velocity (V1 = 9.4 mm/s, V2 = 65 mm/s), obtained by feeding the motors with two different sinusoidal input tra- jectories, at the frequencies of 0.1 Hz and 0.4 Hz. These values were chosen according to previous studies4,33. Exemplary timeline is shown in Fig. 1. Between two stimuli, the motors were stopped and the force was set to 0 N, in this case the fabric was only lightly in contact with the forearm. As the force increases, the fabric is more closely wrapped around the forearm and there is no more pure sliding (as with light forces) but also skin torsion. This behavior was coherent with the goal to reproduce as exhaustively as possible the behavior of the human caress. This transformation can be performed in the frequency domain by using the following relationships23 Then, to average among multiple subjects, we consider group values expressed as median and its respective absolute deviation (i.e., for a feature ∼X, = ± ∼ ∼ ∼ X X X Median( ) MAD( ) where = \| − \| ˜ ˜ ˜ X X X MAD( ) Median( Median( ) )). l l f h f l d h l d ff b imum value X, the median ∼X and its respective absolute deviation ∼ ∼ X, all calculated along the time of last second of caressing. Accordingly, since we derived 13 instantaneous features, the total number of parameter used for the force/velocity classification was 39. A summary of all considered features is reported in Table 1. The median ∼X is considered as a measure of central tendency, which can be related to classical estimation methods that provide one finite value when considering a given time window, whereas the median absolute deviation ∼ ∼ X and the maxi- mum value X depend on the feature variability along the time. Then, to average among multiple subjects, we consider group values expressed as median and its respective absolute deviation (i.e., for a feature ∼X, = ± ∼ ∼ ∼ X X X Median( ) MAD( ) where = \| − \| ˜ ˜ ˜ X X X MAD( ) Median( Median( ) )). A l li i f h f l d h i i l diff b imum value X, the median ∼X and its respective absolute deviation ∼ ∼ X, all calculated along the time of last second of caressing. Accordingly, since we derived 13 instantaneous features, the total number of parameter used for the force/velocity classification was 39. A summary of all considered features is reported in Table 1. The median ∼X is considered as a measure of central tendency, which can be related to classical estimation methods that provide one finite value when considering a given time window, whereas the median absolute deviation ∼ ∼ X and the maxi- l X d d h f b l l h Th l l b one finite value when considering a given time window, whereas the median absolute deviation ∼X and the maxi- mum value X depend on the feature variability along the time. This transformation can be performed in the frequency domain by using the following relationships23 Throughout the experiment, there were two phases of resting sessions with a duration of two minutes: the first at the beginning of the protocol, the second at the end of the stimulation.h The four kinds of stimulation were suitably randomized, with a pre-stimulus and a post-stimulus interval of 35 seconds each. During the elicitation, the ECG was continuously acquired, following the Einthoven triangle configuration, by means of a dedicate hardware module, i.e. the ECG100C Electrocardiogram Amplifier from BIOPAC inc. with a sampling rate of 500 Hz. To obtain the HRV series from the ECG, a QRS complex detection Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 5 www.nature.com/scientificreports/ Feature Symbol Description Meaning Reference µ∼ RR, µ∼ ∼ RR, µ RR Mean of the Inverse-Gaussian pdf Instantaneous Mean of the RR Interval Series 22,23 σ∼ RR, σ∼ ∼ RR, σ RR Standard Deviation of the Inverse- Gaussian pdf Instantaneous Standard Deviation of the RR Interval Series 22,23 σ∼ HR, σ∼ ∼ HR, σ HR Heart rate Standard Deviation Instantaneous Std of the Heartbeat Series 22,23 ∼ LF, ∼ ∼ LF,  LF Low-Frequency Power of the RR interval series spectrum Instantaneous Sympathetic and Parasympathetic Activity 22,23 ∼ HF, ∼ ∼ HF,  HF High-Frequency Power of the RR interval series spectrum Instantaneous Parasympathetic Activity 22,23 ∼ LF HF / , ∼ ∼ LF HF / ,  LF HF / Ratio between Low- and High- Frequency Power of the RR interval series spectrum Instantaneous Sympatho-Vagal Balance 22,23 ∼ LL, ∼ ∼ LL,  LL Low-Freq./Low-Freq. integration of the RR interval series bispectrum Instantaneous Nonlinear Sympatho-Vagal Interactions 23 ∼ LH, ∼ ∼ LH,  LH Low-Freq./High-Freq. integration of the RR interval series bispectrum Instantaneous Nonlinear Sympatho Vagal Interactions 23 ∼ HH, ∼ ∼ HH,  HH High-Freq./High-Freq. integration of the RR interval series bispectrum Instantaneous Nonlinear Sympatho-Vagal Interactions 23 ∼ AI, ∼ ∼ AI,  AI Inhomogeneous Point-Process Approximate Entropy of the RR interval series Measure of Instantaneous Complexity 20 ∼ SI, ∼ ∼ SI,  SI Inhomogeneous Point-Process Sample Entropy of the RR interval series Measure of Instantaneous Complexity 20 λ∼ 1, λ∼ ∼ 1, λ1 First Lyapunov Exponent of the RR interval series Measure of Instantaneous Complexity 21 λ∼ 2, .., λ2 Second Lyapunov Exponent of the RR interval series Measure of Instantaneous Complexity 21 Table 1. A summary of all features used in this study. Table 1. This transformation can be performed in the frequency domain by using the following relationships23 A summary of all features used in this study. Table 1. A summary of all features used in this study. algorithm was used, i.e. the automatic algorithm developed by Pan-Tompkins37. Peak detection and consequent correction of ectopic beats was performed with an automatic technique previously described in ref. 35. Statistical Analysis and Pattern Recognition. All features were instantaneously calculated with a Δ = 5 ms tem- poral resolution. We considered the time-varying dynamics of a given feature X estimated throughout the short-time window when the affective stimulus occurs. Although the length of such a short-time window depends on the caressing velocity ranging from 4.3 to 25 seconds, in order to avoid possible biases related to signal length, we considered feature values derived from the last second of each caressing stimulus. In order to reduce possible intra-, and inter-subject variability, this value was then normalized by the feature value gathered from the fist 10 s of the initial resting state. Quantification of the feature dynamical information was performed through the max- Statistical Analysis and Pattern Recognition. All features were instantaneously calculated with a Δ = 5 ms tem- poral resolution. We considered the time-varying dynamics of a given feature X estimated throughout the short-time window when the affective stimulus occurs. Although the length of such a short-time window depends on the caressing velocity ranging from 4.3 to 25 seconds, in order to avoid possible biases related to signal length, we considered feature values derived from the last second of each caressing stimulus. In order to reduce possible intra-, and inter-subject variability, this value was then normalized by the feature value gathered from the fist 10 s of the initial resting state. Quantification of the feature dynamical information was performed through the max- imum value X, the median ∼X and its respective absolute deviation ∼ ∼ X, all calculated along the time of last second of caressing. Accordingly, since we derived 13 instantaneous features, the total number of parameter used for the force/velocity classification was 39. A summary of all considered features is reported in Table 1. The median ∼X is considered as a measure of central tendency, which can be related to classical estimation methods that provide one finite value when considering a given time window, whereas the median absolute deviation ∼ ∼ X and the maxi- mum value X depend on the feature variability along the time. Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 This transformation can be performed in the frequency domain by using the following relationships23 Then, to average among multiple subjects, we consider group values expressed as median and its respective absolute deviation (i.e., for a feature ∼X, = ± ∼ ∼ ∼ X X X Median( ) MAD( ) where = \| − \| ˜ ˜ ˜ X X X MAD( ) Median( Median( ) )).f \| \| As an exploratory preliminary step, for each feature, we evaluated the statistical differences between groups of caresses, identified by levels of force/velocity. The difference was expressed in terms of p-values from a non-parametric Wilcoxon test for paired data, under the null hypothesis that the medians of the two sample groups are equal. g p q In order to actually discern between force and velocity of the caresses using heartbeat dynamics exclusively, an automatic classification algorithm is needed. To this extent, each feature constituted a single dimension of the feature space. A multidimensional point was considered an outlier if z-scores associated to its dimensions were greater than 4. The obtained feature set is taken as an input of the Leave-One-Out (LOO) procedure applied on a Support Vector Machine (SVM)-based pattern recognition38 (see Fig. 2). Within the LOO scheme, the training set was normalized by subtracting the median value and dividing by the MAD over each dimension. More spe- cifically, we used a nu-SVM (nu = 0.5) having a radial basis kernel function with γ = n−1, with n = 39 equal to the number of features.i Additionally, in order to explore the relative importance of all features in the classification problem, we employed a support vector machine recursive feature elimination (SVM-RFE) procedure in a wrapper approach Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 6 www.nature.com/scientificreports/ Figure 2. Overview of the signal processing and classification chain. ANS dynamics on cardiovascular control modulates the heartbeat dynamics. Starting from data acquisition, RR series are extracted by using automatic peak detection algorithms applied on artifact-free signals. The nonlinear point-process model is fitted to the HRV series, and all features are estimated in an instantaneous fashion. Successively, for each subject, a feature set is defined and fed into support vector machine-based classification using leave-one our procedures. Figure 2. Overview of the signal processing and classification chain. ANS dynamics on cardiovascular control modulates the heartbeat dynamics. Starting from data acquisition, RR series are extracted by using automatic peak detection algorithms applied on artifact-free signals. Results I Instantaneous series from a representative subject are shown in Figs 3 and 4. The model order selection anal- ysis revealed optimal NARL order as p = 3 and M1 = 1, M2 = 0. Over all the considered subjects, 31 out of a total of 32 recordings showed KS plots and more than 98% of the autocorrelation samples within the 95% of the confidence interval. Of note, KS distances were as low as 0.0313 ± 0.0062. Figures 5 and 6 show an exem- plary Kolmogorow-Smirnov and autocorrelation plots demonstrating how our model well predicted all heartbeat events of a given RR interval series.i g As a first preliminary investigation, we performed a statistical analysis of all the features between force and velocity levels. Significant results (p < 0.05) from such an analysis are shown in Table 2.iff i Significant differences were found in discerning caresses performed at different velocities only. No relevant changes, in fact, were found in instantaneous heartbeat linear and nonlinear dynamics between caresses per- formed at different force levels. On the other hand, caresses performed at higher velocity affected the variability of all of the instantaneous measures, as well as median value of four features, one of which was related to heartbeat complex dynamics, and maximum value of 8 features, six of which were related to heartbeat complex/nonlinear dynamics (see Table 2). Caressing Force Classification. Table 3 shows the confusion matrix as well as the total average accuracy whilst discerning caresses-like stimuli performed at different force levels, considering first 26 features selected by the nonlinear SVM-RFE algorithm. This specific feature set, giving the best total average accuracy, which was 69.79%, is partially listed in Table 4 ordered by median rank over every fold computed during the LOO procedure. Figure 7 shows the recognition accuracy while discerning caressing force levels, as a function of the feature rank estimated through the SVM-RFE procedure.i g Furthermore, we performed the classification of caressing at low force vs. high force levels using features only related to the central tendency (median) of the time-varying estimates, such as µ∼ RR, σ∼ RR, σ∼ HR, ∼ LF, ∼ HF, and ∼ LF HF / , ∼ LL, ∼ LH, ∼ HH, λ∼ 1, λ∼ 2, ∼ AI, ∼ SI, avoiding any measure of variability (i.e., MAD and maximum values). Results of this classification are shown in Table 5. This transformation can be performed in the frequency domain by using the following relationships23 The nonlinear point-process model is fitted to the HRV series, and all features are estimated in an instantaneous fashion. Successively, for each subject, a feature set is defined and fed into support vector machine-based classification using leave-one our procedures. (RFE was performed on the training set of each fold and we computed the median rank for each feature over all folds). We specifically chose a recently developed nonlinear SVM-RFE which employes a radial basis function kernel and includes a correlation bias reduction strategy into the feature elimination procedure39.i Classification results are here expressed in terms of recognition accuracy, and in form of confusion matrix. The generic element rij of the confusion matrix indicated a percentage as to how many times a pattern belonging to class i was classified as belonging to class j. A more diagonal confusion matrix corresponded to a higher degree of classification. All of the algorithms were implemented by using Matlab© v8.4 endowed with an additional toolbox for pattern recognition, i.e., LIBSVM40. p g In the next section, we report experimental results in classifying force and velocity level of the administered caress-like stimuli. Specifically, we show best classification results given the ad-hoc number of feature identified through the SVM-RFE procedure. Results I In this case, the total average accuracy was 57.66%. Caressing Velocity Classification. Table 6 shows the confusion matrix as well as the total average accu- racy whilst discerning caresses performed at different velocity levels, considering first 35 features selected by the Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 7 nonlinear SVM-RFE algorithm. This specific feature set, giving the best total average accuracy, which was 81.25%, is partially listed in Table 7 ordered by median rank over every fold computed during the LOO procedure. www.nature.com/scientificreports/ www.nature.com/scientificreports/ re.com/scientificreports/ nonlinear SVM-RFE algorithm. This specific feature set, giving the best total average accuracy, which was 81.25%, Figure 3. Instantaneous heartbeat estimates computed from a representative subject, obtained through a point-process NARL model. Black lines and gray areas indicate median and MAD values among subjects. Vertical red lines mark the short-time window of caressing. From the top panel, the estimated mean, μRR(t), the RR standard deviation, σRR(t), the high frequency (HF), the low frequency (LF), the (LF/HF) ratio, and LL are reported. Figure 3. Instantaneous heartbeat estimates computed from a representative subject, obtained through a point-process NARL model. Black lines and gray areas indicate median and MAD values among subjects. Vertical red lines mark the short-time window of caressing. From the top panel, the estimated mean, μRR(t), the RR standard deviation, σRR(t), the high frequency (HF), the low frequency (LF), the (LF/HF) ratio, and LL are reported. nonlinear SVM-RFE algorithm. This specific feature set, giving the best total average accuracy, which was 81.25%, is partially listed in Table 7 ordered by median rank over every fold computed during the LOO procedure. Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 8 www.nature.com/scientificreports/ Figure 8 shows the recognition accuracy, while discerning caressing velocity levels, as a function of the feature rank estimated through the SVM RFE procedure Figure 4. Instantaneous heartbeat estimates computed from a representative subject, obtained through a point-process NARL model. Black lines and gray areas indicate median and MAD values among subjects. Vertical red lines mark the short-time window of caressing. From the top panel, LL and HH bispectral statistics, the instantaneous first and second Lyapunov exponent, λ1 and λ2, respectively, and the instantaneous point- process entropy measures AI and SI are reported. Figure 4. Instantaneous heartbeat estimates computed from a representative subject, obtained through a point-process NARL model. Black lines and gray areas indicate median and MAD values among subjects. Vertical red lines mark the short-time window of caressing. From the top panel, LL and HH bispectral statistics, the instantaneous first and second Lyapunov exponent, λ1 and λ2, respectively, and the instantaneous point- process entropy measures AI and SI are reported. Figure 4. Instantaneous heartbeat estimates computed from a representative subject, obtained through a point-process NARL model. Black lines and gray areas indicate median and MAD values among subjects. Vertical red lines mark the short-time window of caressing. Figure 8 shows the recognition accuracy, while discerning caressing velocity levels, as a function of the feature rank estimated through the SVM-RFE procedure. www.nature.com/scientificreports/ From the top panel, LL and HH bispectral statistics, the instantaneous first and second Lyapunov exponent, λ1 and λ2, respectively, and the instantaneous point- process entropy measures AI and SI are reported. Figure 8 shows the recognition accuracy, while discerning caressing velocity levels, as a function of the feature rank estimated through the SVM-RFE procedure. Figure 8 shows the recognition accuracy, while discerning caressing velocity levels, as a function of the feature rank estimated through the SVM-RFE procedure. Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 9 www.nature.com/scientificreports/ Quantiles 0 0.2 0.4 0.6 0.8 1 CDF 0 0.2 0.4 0.6 0.8 1 KS Plot Figure 5. KS plot using the proposed inhomogeneous point process nonlinear models for a representative subject (N. 1) undergoing caressing-like stimuli. Blue diagonal and red lines indicate the 95% confidence bounds. Quantiles 0 0.2 0.4 0.6 0.8 1 CDF 0 0.2 0.4 0.6 0.8 1 KS Plot Quantiles Figure 5. KS plot using the proposed inhomogeneous point process nonlinear models for a representative subject (N. 1) undergoing caressing-like stimuli. Blue diagonal and red lines indicate the 95% confidence bounds. Lag 0 10 20 30 40 50 60 Autocorrelation -0.6 -0.4 -0.2 0 0.2 0.4 Figure 6. Autocorrelation plot using the proposed inhomogeneous point process nonlinear models for a representative subject (N. 1) undergoing caressing-like stimuli. Dashed lines indicate the 95% confidence bounds. Figure 5. KS plot using the proposed inhomogeneous point process nonlinear models for a representative subject (N. 1) undergoing caressing-like stimuli. Blue diagonal and red lines indicate the 95% confidence bounds. Figure 6. Autocorrelation plot using the proposed inhomogeneous point process nonlinear models for a representative subject (N. 1) undergoing caressing-like stimuli. Dashed lines indicate the 95% confidence bounds. Likewise performed for the caressing force classification, we performed the classification of caressing at low velocity vs. high velocity levels using features only related to the central tendency (median) of the time-varying estimates, avoiding any measure of variability. Results of this classification are shown in Table 8. In this case, the total average accuracy was 57.14%. Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 Discussion and Conclusion l d As a matter of fact, no relevant changes were found between caresses Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 10 www.nature.com/scientificreports/ Feature Velocity 9.4 mm/s Velocity 65 mm/s p-values µ RR 0.88 ± 0.11 0.84 ± 0.11 0.0075  HF 591.10 ± 357.82 547.99 ± 321.54 0.046  LH 8.43 108 ± 4.22 108 5.09 108 ± 2.5 108 0.040  HH 1.6 109 ± 9.37 108 1.29 109 ± 8.3 108 0.0076 λ1 0.080 ± 0.074 0.056 ± 0.066 0.047 λ2 −0.064 ± 0.041 −0.11 ± 0.035 <10−6  AI 0.39 ± 0.051 0.33 ± 0.051 <7 * 10−5  SI 0.30 ± 0.059 0.24 ± 0.064 <4 * 10−5 σ∼ HR 4.48 ± 1.89 5.47 ± 1.99 0.049 ∼ LF 538.25 ± 354.08 706.40 ± 496.25 0.015 ∼ HF 384.12 ± 231.03 451.63 ± 269.10 0.013 λ∼ 1 −0.0059 ± 0.057 0.016 ± 0.058 0.01 µ∼ ∼ RR 0.020 ± 0.0065 0.015 ± 0.0061 0.016 σ∼ ∼ RR 29.83 ± 18.92 9.24 ± 5.57 <10−6 σ∼ ∼ HR 0.341 ± 0.17 0.138 ± 0.085 <10−4 ∼ ∼ LF 79.17 ± 48.63 29.71 ± 26.19 <10−3 ∼ ∼ HF 41.44 ± 31.21 16.21 ± 11.30 <10−3 ∼ ∼ LF HF / 0.30 ± 0.23 0.12 ± 0.089 <10−3 ∼ ∼ LL 5.08 107 ± 3.8 107 2.27 107 ± 2.02 107 0.041 ∼ ∼ LH 6.88 107 ± 3.76 107 2.29 107 ± 1.09 107 <10−4 ∼ ∼ HH 1.18 108 ± 8.05 107 5.52 107 ± 4.57 107 <10−5 λ∼ ∼ 1 0.026 ± 0.0079 0.011 ± 0.0044 <10−6 λ∼ ∼ 2 0.020 ± 0.0035 0.011 ± 0.0040 <10−6 ∼ ∼ AI 0.021 ± 0.0077 0.0060 ± 0.0034 <10−6 ∼ ∼ SI 0.019 ± 0.0080 0.0057 ± 0.0033 <10−6 Table 2. Significant features and p-values. p-values are gathered from the Wilcoxon non-parametric test. Discussion and Conclusion l d In conclusion, we proposed a comprehensive signal processing framework to assess short-term affective haptic stimuli through the analysis of heartbeat dynamics exclusively. We built on a previously proposed set of equations defining inhomogeneous point-process nonlinear models23 to obtain instantaneous linear, nonlinear, and com- plex estimates of cardiovascular dynamics. Key concepts relay on the definition of a continuous pdf predicting the next heartbeat event, identified through R-waves from the ECG, parametrized using a nonlinear autoregressive model with Laguerre expansion of the Wiener-Volterra terms23. Such a parametrization is defined up to the cubic terms, allowing for the instantaneous estimation of Lyapunov exponents21 and approximate and sample entropy20, as well as instantaneous nonlinear bispectral measures, and linear estimates defined in the time and frequency domain23. To our knowledge, this is the only signal processing method proposed so far able to objectively assess short-term (4.3 to 25 seconds) affective haptic stimuli such as caresses. Standard linear and nonlinear HRV meas- ures16, in fact, would be unable to perform such an assessment as they require relatively long-term recordings to accurately characterize the emotional state of a subject. On the other hand, recognition of long-term caressing (i.e., >30 seconds) would lead to the saturation of the CT fibers activity3,5, strongly biasing the affective part of the emotional perception. Of note, motivation of this study are strictly related to objective assessments performed in experimental psychology and, more in general, in the evaluation of patients with mental disorders. l l d h d l d fi l ff d b h l f Our analysis revealed that instantaneous cardiovascular dynamics is significantly affected by the velocity of the affective haptic stimuli (see Table 2). Discussion and Conclusion l d Feature Velocity 9.4 mm/s Velocity 65 mm/s p-values µ RR 0.88 ± 0.11 0.84 ± 0.11 0.0075  HF 591.10 ± 357.82 547.99 ± 321.54 0.046  LH 8.43 108 ± 4.22 108 5.09 108 ± 2.5 108 0.040  HH 1.6 109 ± 9.37 108 1.29 109 ± 8.3 108 0.0076 λ1 0.080 ± 0.074 0.056 ± 0.066 0.047 λ2 −0.064 ± 0.041 −0.11 ± 0.035 <10−6  AI 0.39 ± 0.051 0.33 ± 0.051 <7 * 10−5  SI 0.30 ± 0.059 0.24 ± 0.064 <4 * 10−5 σ∼ HR 4.48 ± 1.89 5.47 ± 1.99 0.049 ∼ LF 538.25 ± 354.08 706.40 ± 496.25 0.015 ∼ HF 384.12 ± 231.03 451.63 ± 269.10 0.013 λ∼ 1 −0.0059 ± 0.057 0.016 ± 0.058 0.01 µ∼ ∼ RR 0.020 ± 0.0065 0.015 ± 0.0061 0.016 σ∼ ∼ RR 29.83 ± 18.92 9.24 ± 5.57 <10−6 σ∼ ∼ HR 0.341 ± 0.17 0.138 ± 0.085 <10−4 ∼ ∼ LF 79.17 ± 48.63 29.71 ± 26.19 <10−3 ∼ ∼ HF 41.44 ± 31.21 16.21 ± 11.30 <10−3 ∼ ∼ LF HF / 0.30 ± 0.23 0.12 ± 0.089 <10−3 ∼ ∼ LL 5.08 107 ± 3.8 107 2.27 107 ± 2.02 107 0.041 ∼ ∼ LH 6.88 107 ± 3.76 107 2.29 107 ± 1.09 107 <10−4 ∼ ∼ HH 1.18 108 ± 8.05 107 5.52 107 ± 4.57 107 <10−5 λ∼ ∼ 1 0.026 ± 0.0079 0.011 ± 0.0044 <10−6 λ∼ ∼ 2 0.020 ± 0.0035 0.011 ± 0.0040 <10−6 ∼ ∼ AI 0.021 ± 0.0077 0.0060 ± 0.0034 <10−6 ∼ ∼ SI 0.019 ± 0.0080 0.0057 ± 0.0033 <10−6 Table 2 Significant features and p-values p-values are Table 2. Significant features and p-values. p-values are gathered from the Wilcoxon non-parametric test. Table 2. Significant features and p-values. p-values are gathered from the Wilcoxon non-parametric test. Table 2. Significant features and p-values. p-values are gathered from the Wilcoxon non-parametric test. SVM Force 2 N Force 6 N Force 2 N 76.09 23.91 Force 6 N 36.00 64.00 Table 3. Confusion matrix for caressing administered at different force levels. Values are expressed as percentages. Total Accuracy: 69.79%. Table 3. Confusion matrix for caressing administered at different force levels. Values are expressed as percentages. Total Accuracy: 69.79%. Table 3. Confusion matrix for caressing administered at different force levels. Values are expressed as percentages. Total Accuracy: 69.79%. Discussion and Conclusion l d Rank Feature 1  LL 2 σ∼ ∼ HR 3 ∼ HH 4 σ HR 5  LF HF / 6 λ∼ ∼ 1 Table 4. Selected features ordered by their median rank over every fold computed during the LOO procedure for force classification. Table 4. Selected features ordered by their median rank over every fold computed during the LOO procedure for force classification. Table 4. Selected features ordered by their median rank over every fold computed during the LOO procedure for force classification. performed at different force levels. Of note, 14 out of the 25 features which were sensitive to caressing velocity level are derived in a nonlinear fashion, revealing a decrease of complexity variability and maximum values dur- ing the haptic stimulus. This result further justify the use of real-time estimates of heartbeat dynamics through point-process modeling, also highlighting the great role of nonlinear dynamics in cardiovascular control, fer- vently pointed out in the current literature16. In order to develop a processing chain able to actually discern Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 11 www.nature.com/scientificreports/ Feature Rank 0 10 20 30 40 Accuracy [%] 55 60 65 70 Caressing Force Classification Figure 7. Recognition accuracy in discerning caressing force levels as a function of the feature rank estimated through the SVM-RFE procedure. Figure 7. Recognition accuracy in discerning caressing force levels as a function of the feature rank estimated through the SVM-RFE procedure. SVM Force 2 N Force 6 N Force 2 N 53.70 46.30 Force 6 N 38.60 61.40 Table 5 Confusion matrix for caressing at different force levels without features related to instantaneous SVM Force 2 N Force 6 N Force 2 N 53.70 46.30 Force 6 N 38.60 61.40 Table 5. Confusion matrix for caressing at different force levels, without features related to instantaneo variability. Values are expressed as percentages. Total Accuracy: 57.66%. Table 5. Confusion matrix for caressing at different force levels, without features related to instantaneous variability. Values are expressed as percentages. Total Accuracy: 57.66%. Table 6. Confusion matrix for caressing administered at different velocity levels. Values are expressed as percentages. Total Accuracy: 81.25%. Table 6. Confusion matrix for caressing administered at different velocity levels. Values are expressed as percentages. Total Accuracy: 81.25%. Rank Feature 1 σ∼ RR 2 ∼ LF HF / 3  LF LF / 4 ∼ LL 5 σ RR 6  SI Table 7. Selected features ordered by their median rank over every fold computed during the LOSO procedure for velocity classification. Table 7. Selected features ordered by their median rank over every fold computed during the LOSO procedure for velocity classification. Feature Rank 0 10 20 30 40 Accuracy [%] 20 40 60 80 100 Caressing Velocity Classification Figure 8. References 1. Liljencrantz, J. & Olausson, H. Tactile c fibers and their contributions to pleasant sensations and to tactile allodynia. Frontiers in behavioral neuroscience 8, 1–6 (2014).ii ( ) 2. Macefield, V. G. Tactile c fibers. In Encyclopedia of Neuroscience, Springer, 3995–3998 (2009).hl ii y p f p g 3. Klöcker, A., Oddo, C. M., Camboni, D., Penta, M. & Thonnard, J.-L. Physical factors influencing pleasant touch during passive fingertip stimulation. Plos One 9, e101361 (2014).ff i g p 4. Löken, L. S., Evert, M. & Wessberg, J. Pleasantness of touch in human glabrous and hairy skin: order effects on affective ratings Brain research 1417, 9–15 (2011). 5. McGlone, F., Vallbo, A. B., Olausson, H., Loken, L. & Wessberg, J. Discriminative touch and emotional touch. Canadian Journal of Experimental Psychology/Revue canadienne de psychologie expérimentale 61, 173–183 (2007).f 6. Ackerley, R., Carlsson, I., Wester, H., Olausson, H. & Wasling, H. B. Touch perceptions across skin sites: differences between sensitivity, direction discrimination and pleasantness. Frontiers in behavioral neuroscience 8, 1–10 (2014).if y p 7. Agrafioti, F., Hatzinakos, D. & Anderson, A. K. ECG pattern analysis for emotion detection. Affective Computing, IEEE Transaction on 3, 102–115 (2012). 8. Calvo, R. & D’Mello, S. Affect detection: An interdisciplinary review of models, methods, and their applications. Affective Computing, IEEE Transactions on 1, 18–37 (2010).f 9. Eyben, F., Wollmer, M. & Schuller, B. OpenEAR-introducing the Munich open-source emotion and affect recognition toolkit. InAffective Computing and Intelligent Interaction and Workshops, ACII 3rd International Conference on, 1–6 (2009). ff p g g p f 0. Haag, A., Goronzy, S., Schaich, P. & Williams, J. Emotion recognition using bio-sensors: First steps towards an automatic system InADS, 36–48 (2004).h 11. Hatayama, T., Kitamura, S., Tamura, C., Nagano, M. & Ohnuki, K. The facial massage reduced anxiety and negative mood status, and increased sympathetic nervous activity. Biomedical research (Tokyo, Japan) 29, 317–320 (2008). 12. Kim, K. H., Bang, S. W. & Kim, S. R. Emotion recognition system using short-term monitoring of physiological signals. Medical and biological engineering and computing. 42, 419–427 (2004). g g g p g 3. Lindgren, L., Rundgren, S., Winsö, O., Lehtipalo, S. & Wiklund, U. et al. Physiological responses to touch massage in healthy volunteers. Autonomic Neuroscience 158, 105–110 (2010).f 4. Olausson, H., Cole, J., Rylander, K., McGlone, F. & Lamarre, Y. et al. www.nature.com/scientificreports/ This result demonstrates that information on central tendency only are not sufficient to achieve an automatic discrimination of affective haptic perception using heartbeat dynamics exclu- sively. This is also demonstrated by the results of the SVM-RFE feature selection procedure, pointing out how complexity variability measures are the most needed for the classification (see Tables 4 and 7). Indeed, as shown in Fig. 3, the time-varying dynamics of the heartbeat features is highly non-stationary, with most significant changes in nonlinear and complex dynamics (i.e., LL, LH, HH, λ1, λ2, AI, SI) during caressing at slower velocity. Importantly, our results are not biased by possible stochastic components underlying physiological dynamics, as our instantaneous complexity measures are not significantly affected by the statistical properties of the physiolog- ical noise behind the observed dynamics20,21. y Finally, we remark that in this study caressing-like stimuli were administered through an ad-hoc robotic device29. Further research is needed to investigate possible differences/similarities with actual caresses by a human hand, possibly investigating more than two levels of caressing force and velocity. www.nature.com/scientificreports/ www.nature.com/scientificreports/ SVM Velocity 9.4 mm/s Velocity 65 mm/s Velocity 9.4 mm/s 58.33 41.67 Velocity 65 mm/s 43.90 56.10 Table 8. Confusion matrix for caressing administered at different velocity levels, without features related to instantaneous variability. Values are expressed as percentages. Total Accuracy: 57.14%. SVM Velocity 9.4 mm/s Velocity 65 mm/s Velocity 9.4 mm/s 58.33 41.67 Velocity 65 mm/s 43.90 56.10 Table 8. Confusion matrix for caressing administered at different velocity levels, without features related to instantaneous variability. Values are expressed as percentages. Total Accuracy: 57.14%. Table 8. Confusion matrix for caressing administered at different velocity levels, without features related o instantaneous variability. Values are expressed as percentages. Total Accuracy: 57.14%. results from the statistical analysis, cardiovascular dynamics is strongly affected by the caressing velocity, rather than the caressing force. In order to further emphasize the role of our processing framework, which allows for the calculation of features’ variability measures through the definition of instantaneous estimates, we performed the caressing force and velocity classification using features only related to the central tendency. In these cases, recog- nition accuracy were as low as <58%. This result demonstrates that information on central tendency only are not sufficient to achieve an automatic discrimination of affective haptic perception using heartbeat dynamics exclu- sively. This is also demonstrated by the results of the SVM-RFE feature selection procedure, pointing out how complexity variability measures are the most needed for the classification (see Tables 4 and 7). Indeed, as shown in Fig. 3, the time-varying dynamics of the heartbeat features is highly non-stationary, with most significant changes in nonlinear and complex dynamics (i.e., LL, LH, HH, λ1, λ2, AI, SI) during caressing at slower velocity. Importantly, our results are not biased by possible stochastic components underlying physiological dynamics, as our instantaneous complexity measures are not significantly affected by the statistical properties of the physiolog- ical noise behind the observed dynamics20,21. results from the statistical analysis, cardiovascular dynamics is strongly affected by the caressing velocity, rather than the caressing force. In order to further emphasize the role of our processing framework, which allows for the calculation of features’ variability measures through the definition of instantaneous estimates, we performed the caressing force and velocity classification using features only related to the central tendency. In these cases, recog- nition accuracy were as low as <58%. Table 4. Selected features ordered by their median rank over every fold computed during the LOO procedure for force classification. Recognition accuracy in discerning caressing velocity levels as a function of the feature rank estimated through the SVM-RFE procedure. Feature Rank 0 10 20 30 40 Accuracy [%] 20 40 60 80 100 Caressing Velocity Classification Caressing Velocity Classification Figure 8. Recognition accuracy in discerning caressing velocity levels as a function of the feature rank estimated through the SVM-RFE procedure. affective perception, we moved from looking for simple statistical differences between stimulus physical charac- teristics to building an automatic classification algorithm. The automatic classification, embedded in the on-line processing chain, was performed by means of nu-SVM, and following a LOO and nonlinear SVM-RFE proce- dures for cross-validation and feature selection, respectively. Recognition accuracies were very satisfactory, being of 69.79% along the force dimension, and 81.25% along the velocity dimension. This suggest that, in line with Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 12 ficreports/ SVM Velocity 9.4 mm/s Velocity 65 mm/s Velocity 9.4 mm/s 58.33 41.67 Velocity 65 mm/s 43.90 56.10 Table 8. Confusion matrix for caressing administered at different velocity levels, without features related to instantaneous variability. Values are expressed as percentages. Total Accuracy: 57.14%. www.nature.com/scientificreports/ ACM Transactions on Intelligent Systems and Technology 2, 1–27 (2011). www.nature.com/scientificreports/ www.nature.com/scientificreports/ 25. Valenza, G., Orsolini, S., Diciotti, S., Citi, L., Scilingo, E. P., Guerrisi, M., Danti, S., Lucetti, C., Tessa, C., Barbieri, R. & Toschi, N. Assessment of spontaneous cardiovascular oscillations in Parkinson’s disease. Biomedical Signal Processing and Control 26, 80–89 (2016). ( ) 26. Valenza, G., Citi, L., Gentili, C., Lanatá, A., Scilingo, E. P. & Barbieri, R. Point-process nonlinear autonomic assessment of depre states in bipolar patients. Methods of information in medicine, 53(4), 296–302 (2013). states in bipolar patients. Methods of information in medicine, 53(4), 296–302 (2013). 27. Valenza, G., Citi, L., Gentili, C., Lanata, A., Scilingo, E. P. & Barbieri, R. Characterization of depressive states in bipolar patients using bl t til t h l d i t t h t t i bilit t Bi di l d H lth I f ti IEEE J l f 7. Valenza, G., Citi, L., Gentili, C., Lanata, A., Scilingo, E. P. & Barbieri, R. Characterization of depressive states in bipolar patients using wearable textile technology and instantaneous heart rate variability assessment. Biomedical and Health Informatics, IEEE Journal o 19(1), 263–274 (2015). 28. Chang, H. A., Chang, C.-C., Kuo, T. B. & Huang, S.-Y. Distinguishing bipolar ii depression from unipolar major depressive disorder: Differences in heart rate variability. The World Journal of Biological Psychiatry 0, 1–10 (2015). fh 9. Costa, M. D. & Goldberger, A. L. Generalized multiscale entropy analysis: Application to quantifying the complex volatility o human heartbeat time series. Entropy 17, 1197–1203 (2015). human heartbeat time series. Entropy 17, 1197–1203 (2015). py 0. Glass, L. Dynamical disease: Challenges for nonlinear dynamics and medicine. Chaos: An Interdisciplinary Journal of Nonlinea Science 25, 097603 (2015). 31. Lo, M. T., Chang, Y.-C., Lin, C., Young, H.-W. V. & Lin, Y.-H. et al. Outlier-resilient complexity analysis of heartbeat dynamics. Scientific reports 5, 1–7 (2015). fi p 32. Valenza, G., Nardelli, M., Bertschy, G., Lanata, A. & Scilingo, E. Mood states modulate complexity in heartbeat dynamics: A multiscale entropy analysis. EPL (Europhysics Letters) 107, 18003 (2014).f y y p y 3. Bianchi, M., Valenza, G., Serio, A., Lanata, A. & Greco, A. et al. Design and preliminary affective characterization of a novel fabric based tactile display. In Haptics Symposium (HAPTICS), IEEE, 591–596 (2014).i y p y p 34. Marmarelis, V. Identification of nonlinear biological system using Laguerre expansions of kernels. Ann. Biomed. Eng. 21, 573–589 (1993). 35. Citi, L., Brown, E. N. Acknowledgementsh g The research leading to these results has received partial funding from the Department of Anesthesia, Critical Care & Pain Medicine, Massachusetts General Hospital, and Harvard Medical School, Boston, MA, USA, and European Union Seventh Framework Programme FP7/2007–2013 under Grant No. 601165 of the project “WEARHAP”, and, ERC Advanced Grant No. 291166 “Soft- Hands: A Theory of Soft Synergies for a New Generation of Artificial Hands”. European H2020 project “SoftPro: Synergy-based Open-source Foundations and Technologies for Prosthetics and RehabilitatiOn” (No. 688857). European Commission—Horizon 2020 Program under Grant 689691 “NEVERMIND” (NEurobehavioural predictiVE and peRsonalised Modelling of depressIve symptoms duriNg primary somatic Diseases with ICT-enabled self-management procedures). References Functional role of unmyelinated tactile afferents in human hairy skin: sympathetic response and perceptual localization. Experimental brain research 184, 135–140 (2008). 5. Wiens, S., Mezzacappa, E. S. & Katkin, E. S. Heartbeat detection and the experience of emotions. Cognition & Emotion 14, 417–427 (2000). 6. Acharya, U. R., Paul Joseph, K., Kannathal, N., Lim, C. & Suri, J. Heart rate variability: a review. Medical and Biological Engineering and Computing 44, 1031–1051 (2006).hf p g 7. Delaney, J., Leong, K. S., Watkins, A. & Brodie, D. The short-term effects of myofascial trigger point massage therapy on cardiac autonomic tone in healthy subjects. Journal of advanced nursing 37, 364–371 (2002). 8. Lindgren, L., Lehtipalo, S., Winsö, O., Karlsson, M. & Wiklund, U. et al. Touch massage: a pilot study of a complex intervention Nursing in critical care 18 269–277 (2013). 9. Sunagawa, K., Kawada, T. & Nakahara, T. Dynamic nonlinear vago-sympathetic interaction in regulating heart rate. Heart and Vessels 13, 157–174 (1998). 0. Valenza, G., Citi, L., Scilingo, E. P. & Barbieri, R. Inhomogeneous point-process entropy: An instantaneous measure of complexity in discrete systems. Physical Review E 89, 052803 (2014). y y 1. Valenza, G., Citi, L. & Barbieri, R. Estimation of instantaneous complex dynamics through Lyapunov exponents: a study on heartbeat dynamics. PloS one 9, e105622 (2014).i y 22. Barbieri, R., Matten, E., Alabi, A. & Brown, E. A point-process model of human heartbeat intervals: new definitions of heart rate and heart rate variability. American Journal of Physiology-Heart and Circulatory Physiology 288, H424 (2005). y f y gy y y gy 23. Valenza, G., Citi, L., Scilingo, E. P. & Barbieri, R. Point-process nonlinear models with Laguerre and Volterra expans Instantaneous assessment of heartbeat dynamics. Signal Processing, IEEE Transactions On 61, 2914–2926 (2013). 24. Valenza, G., Greco, A., Gentili, C., Lanata, A., Sebastiani, L., Menicucci, D., Gemignani, A. & Scilingo, E. P. Combining electroencephalographic activity and instantaneous heart rate for assessing brain–heart dynamics during visual emotional elicitation in healthy subjects. Phil. Trans. R. Soc. A 374(2067), 20150176 (2016). Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 13 Author Contributions G.V., A.G., M.B. and E.P.S. conceived the experiment; M.B. realized the haptic device; G.V., A.G., L.C. and R.B. analyzed data; All authors discussed the experimental results; All authors wrote the manuscript. Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 www.nature.com/scientificreports/ & Barbieri, R. A real-time automated point-process method for the detection and correction of erroneous and ectopic heartbeats. Biomedical Engineering, IEEE Transactions on 59, 2828–2837 (2012).h p g g 36. Nichols, J., Olson, C., Michalowicz, J. & Bucholtz, F. The bispectrum and bicoherence for non-Gaussian inputs. Signal Processing, IEEE Transactions on 57, 3879–3890 (2009). 36. Nichols, J., Olson, C., Michalowicz, J. & Bucholtz, F. The bispectrum and bicoherence for quadratically nonlinear systems subje non Gaussian inputs Signal Processing IEEE Transactions on 57 3879 3890 (2009) 36. N c o s, J., O so , C., c a ow c , J. & uc o t , .h e b spect u a d b co e e ce o quad at ca y o ea syste s su non-Gaussian inputs. Signal Processing, IEEE Transactions on 57, 3879–3890 (2009). g g 37. Pan, J. & Tompkins, W. J. A real-time qrs detection algorithm. Biomedical Engineering, IEEE Transactions on 32, 230–236 (1985). 38 S hölk f B S l A J Willi R C & B tl tt P L N t t l ith N l t ti 12 1207 1245 . Pan, J. & Tompkins, W. J. A real-time qrs detection algorithm. Bio 37. Pan, J. & Tompkins, W. J. A real-time qrs detection algorithm. Biomedical Engineering, IEEE Transactions on 32, 230–236 (1985). 38. Schölkopf, B., Smola, A. J., Williamson, R. C. & Bartlett, P. L. New support vector algorithms. Neural computation 12, 1207–1245 (2000). 37. Pan, J. & Tompkins, W. J. A real-time qrs detection algorithm. Biomedical Engineering, IEEE Transactions on 32, 230–236 (1985). 38. Schölkopf, B., Smola, A. J., Williamson, R. C. & Bartlett, P. L. New support vector algorithms. Neural computation 12, 1207–1245 7. Pan, J. & Tompkins, W. J. A real-time qrs detection algorithm. Biomedical Engineering, IEEE Transactions on 32, 230–236 (1985). 8. Schölkopf, B., Smola, A. J., Williamson, R. C. & Bartlett, P. L. New support vector algorithms. Neural computation 12, 1207–1245 (2000). 9. Yan, K. & Zhang, D. Feature selection and analysis on correlated gas sensor data with recursive feature elimination. Sensors and Actuators B: Chemical 212, 353–363 (2015). 0. Chang, C. C. & Lin, C.-J. LIBSVM: A library for support vector machines. ACM Transactions on Intelligent Systems and Technology 2, 1–27 (2011). 40. Chang, C. C. & Lin, C.-J. LIBSVM: A library for support vector machines. Additional Informationih Competing financial interests: The authors declare no competing financial interests. How to cite this article: Valenza, G. et al. Inhomogeneous Point-Processes to Instantaneously Assess Affective Haptic Perception through Heartbeat Dynamics Information. Sci. Rep. 6, 28567; doi: 10.1038/srep28567 (2016). This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ Scientific Reports \| 6:28567 \| DOI: 10.1038/srep28567 14
https://openalex.org/W4283691634	https://link.springer.com/content/pdf/10.1007/s00418-022-02119-8.pdf	English	null	Landmark-based retrieval of inflamed skin vessels enabled by 3D correlative intravital light and volume electron microscopy	Histochemistry and cell biology	2,022	cc-by	7,548	* Dagmar Zeuschner D.Zeuschner@mpi-muenster.mpg.de Abstract The nanometer spatial resolution of electron microscopy imaging remains an advantage over light microscopy, but the restricted field of view that can be inspected and the inability to visualize dynamic cellular events are definitely drawbacks of standard transmission electron microscopy (TEM). Several methods have been developed to overcome these limitations, mainly by correlating the light microscopical image to the electron microscope with correlative light and electron micros- copy (CLEM) techniques. Since there is more than one method to obtain the region of interest (ROI), the workflow must be adjusted according to the research question and biological material addressed. Here, we describe in detail the develop- ment of a three-dimensional CLEM workflow for mouse skin tissue exposed to an inflammation stimulus and imaged by intravital microscopy (IVM) before fixation. Our aim is to relocate a distinct vessel in the electron microscope, addressing a complex biological question: how do cells interact with each other and the surrounding environment at the ultrastructural level? Retracing the area over several preparation steps did not involve any specific automated instruments but was entirely led by anatomical and artificially introduced landmarks, including blood vessel architecture and carbon-coated grids. Suc- cessful retrieval of the ROI by electron microscopy depended on particularly high precision during sample manipulation and extensive documentation. Further modification of the TEM sample preparation protocol for mouse skin tissue even rendered the specimen suitable for serial block-face scanning electron microscopy (SBF-SEM). Keywords Intravital microscopy · Dorsal skinfold chamber · Live cell imaging · Correlative light and electron microscopy · Transmission electron microscopy · Serial block-face scanning electron microscopy Landmark‑based retrieval of inflamed skin vessels enabled by 3D correlative intravital light and volume electron microscopy Accepted: 16 May 2022 / Published online: 29 June 2022 © The Author(s) 2022 Histochemistry and Cell Biology (2022) 158:127–136 https://doi.org/10.1007/s00418-022-02119-8 Histochemistry and Cell Biology (2022) 158:127–136 https://doi.org/10.1007/s00418-022-02119-8 ORIGINAL PAPER ORIGINAL PAPER 1 Electron Microscopy Unit, Max Planck Institute for Molecular Biomedicine, Röntgenstr. 20, 48149 Münster, Germany * Dagmar Zeuschner D.Zeuschner@mpi-muenster.mpg.de 1 Electron Microscopy Unit, Max Planck Institute for Molecular Biomedicine, Röntgenstr. 20, 48149 Münster, Germany 2 iBiOs–integrated Bioimaging Facility Osnabrück, CellNanOs- Center for Cellular Nanoanalytics, University of Osnabrück, Osnabrück, Germany 3 Department of Vascular Cell Biology, Max Planck Institute for Molecular Biomedicine, Münster, Germany 4 BioOptic Service Unit, Max Planck Institute for Molecular Biomedicine, Münster, Germany Introduction permeability of blood vessels, stimulating numerous res- cue reactions involving neutrophils and platelets (Claes- son-Welsh et al. 2021). Neutrophils leave the vessel in a controlled manner described as a transmigration process (Vestweber 2015) while platelets are sealing the endothelial barrier to prevent vascular leaks (Braun et al. 2020). Inter- actions of platelets and neutrophils have been observed at sites of inflammation and require further investigation (Lis- man 2018) to improve understanding of the dynamics of inflammation. To this end, a mouse dorsal skinfold chamber was established, allowing high-temporal-resolution imag- ing with an intravital microscopy (IVM) setup (Secklehner et al. 2017). While screening the entire tissue, selective ves- sels with distinct behavior of transmigrating neutrophils and platelets could be observed. As the resolution of intravital microscopy is insufficient to reveal the detailed relation on a single-cell level in the region of interest (ROI), a correla- tive light and electron microscopy (CLEM) approach for Tissue homeostasis is important to maintain the functional- ity of each organ in the body. Inflammatory events chal- lenge this balance and often lead to drastic changes in the 2 iBiOs–integrated Bioimaging Facility Osnabrück, CellNanOs- Center for Cellular Nanoanalytics, University of Osnabrück, Osnabrück, Germany 3 Department of Vascular Cell Biology, Max Planck Institute for Molecular Biomedicine, Münster, Germany 4 BioOptic Service Unit, Max Planck Institute for Molecular Biomedicine, Münster, Germany (0121 3456789) 3 3456789) 3 Histochemistry and Cell Biology (2022) 158:127–136 128 the skin, combining IVM with subsequent ultrastructural investigation, remained to be established. This goal can only be achieved in a correlative manner, resulting in a complex workflow in which almost every step is optimized (Karre- man et al. 2016). As guidance throughout the protocol, arti- ficial and anatomical landmarks are indispensable to relocate the target vessel in the sample volume in the x-, y-, and z dimensions (Goudarzi et al. 2015; Luckner et al. 2018). The composition of skin tissue, a functional multilayered barrier, additionally hinders the infiltration of reagents used for fixa- tion and embedding (Randall Wickett 2006), posing another challenge that required further optimization. The question at hand could only be answered sufficiently by three-dimen- sional volume electron microscopy, here using serial block- face scanning electron microscopy (SBF-SEM) (Denk and Horstmann 2004). As a consequence, extra adjustments for the sample preparation were incorporated into the workflow (Currie et al. Introduction 2022).l dorsal kit, APJ trading Co Inc) was surgically attached to the disinfected skin, and a circular window area (~ 12 mm diameter) was prepared for microscopy. For further details, see Currie et al. (2022). ( ) After analgesia and 24 h of recovery, the epidermis was marked with a stamp (handcrafted), subdivided into 20 fields, thereby introducing the first artificial landmark (Fig. 2a, b). Inflammation was induced by initiating the reverse passive Arthus reaction, as described by Currie et al. (2022). Prior to anesthesia, mice were i.v. injected with 100 µL cell culture graded bovine serum albumin (75 µg BSA/g body weight) in phosphate buffered saline (PBS). Anesthe- tized mice subsequently received an intradermal injection of 20 µL rabbit anti-BSA (30 µg/mouse, MP Biomedicals) anti- body into the stamp-labeled area of the back skin. Live cell imaging was performed using a Zeiss LSM 880 microscope with Airyscan fast. For this procedure, the anesthetized mouse carrying a dorsal skinfold chamber was mounted onto a custom-designed stage. Visualization of blood vessels and platelets was achieved by systemically injecting LysMeGFP mice (fluorescently tagged myelomonocytic cells) with Dylight antibodies against CD31 (30 μg/mouse, eBiosci- ence #16–0311-85) and GPIbβ (2 μg/mouse, #X649, Emfret Analytics). To study the interaction of these immune cells, intravital time-lapse videos were recorded for 90–120 min (Fig. 1 a, Supplementary Movie). The confocal analysis focused on one distinctive postcapillary venule displaying strong neutrophil diapedesis and subsequent single-platelet binding to the vessel wall. At the end of the experiment, the anatomy of the monitored vessel segment (ROI) and the surrounding vascular network was well documented in microscopic z-stacks, using different magnification levels (z-stacks at 20× and 10×). To create an additional reference However, the workflow remains simple and valuable for standard transmission electron microscopy, as well as for SBF-SEM, and could be reproduced in almost any standard electron microscopy laboratory. Dorsal skinfold chamber IVM Illustrated workflow describing all steps introducing landmarks, which are essential to reposition ROI at the electron microscope exactly at the light microscopical location recorded by IVM in a dorsal skin cham- ber (a). The printed stamp on the skin supports to locate the laser point, which illuminates the last position of the IVM settings (ROI) b. After fixation, punching out, and resizing the skin to a minimum, the sample is glued to a carbon grid and marked on one corner c. The endogenous, visible vessels in the tissue serve as the next land- marks while repositioning the sample on the confocal stage f, where Fig. 2 Introduction of landmarks for CLEM orientation. Illustrated workflow describing all steps introducing landmarks, which are essential to reposition ROI at the electron microscope exactly at the light microscopical location recorded by IVM in a dorsal skin cham- ber (a). The printed stamp on the skin supports to locate the laser point, which illuminates the last position of the IVM settings (ROI) b. After fixation, punching out, and resizing the skin to a minimum, the sample is glued to a carbon grid and marked on one corner c. The endogenous, visible vessels in the tissue serve as the next land- marks while repositioning the sample on the confocal stage f, where the fluorescently labeled vessel network in the fixed tissue serve as landmarks to relocate the ROI at the confocal microscope g, h. With the help of the carbon-gridded cover slip and its imprinted letters, the position of the ROI can be documented with respect to these very let- ters d, e. Together with the distance in the z-direction between the ROI and the coverslip, this information about the x–y location defines the starting point for the microtome sectioning j further illustrated in Fig. 3. The measured z height in the sample thus guides this approach to successfully find the correlative view between light and electron microscope (CLEM) k point, a photograph of the laser position within the frame of the introduced stamp was taken (Fig. 2b). the specimen. The flattened tissue was glued onto a carbon gridded coverslip with letters for orientation (coating mask for finder grid; Leica, Austria) using a thin layer of 2% low- melting-point agarose, with the vascularized side of the skin sample directly facing the carbon layer (Fig. 2e, f). Dorsal skinfold chamber IVM The aga- rose was solidified on ice, and the specimen was transferred to a 3-cm glass-bottomed dish (ibidi, Germany). Using the position of the stamp and the edge with the cutoff corner as landmarks, the ROI was relocated in the confocal micro- scope. The localization of the ROI with respect to the carbon letters on the gridded coverslip was documented in z-stacks at different magnification levels. To record special anatomic features of the surrounding tissue serving as additional land- marks, the anatomy of the retrieved ROI and the surround- ing vascular network was captured in microscopic z-stacks, using different magnification levels (z-stacks at 20× and 10×). Finally, the distance in the z-direction between the vessel of interest and the coverslip was measured. The sam- ple was then further fixed in 2% glutaraldehyde (GA), 2% formaldehyde, in 0.1 M cacodylate buffer (CB), pH 7.2, pre- warmed to 37 °C. Dorsal skinfold chamber IVM All animal experiments were approved by the local authori- ties (LANUV, State Agency for Nature, Environment and Consumer Protection). Prior to surgery, mice were anes- thetized, and the dorsal skin was shaved and chemically depilated. A titanium dorsal skinfold chamber frame (small Fig. 1 Vessel segment undergoing neutrophil extravasation during cutaneous inflammation. a Representative time-lapse still image of an inflamed vessel from intravital fluorescence microscopy showing the endothelial cell layer (red), transmigrating neutrophils (green), and platelets (white). Original movie in Supplementary Movie 1. b, c Electron micrographs of inflamed mouse skin sample, convention- ally fixated and processed with CLEM techniques, exhibit poor pres- ervation of the ultrastructure (asterisk marking bulging membranes of platelets) and distortion of the sections because of weak epon infiltra- tion (arrowhead). # vessel lumen, end endothelial cell, neu neutrophil, pl platelet, bulging membranes. Scale bar in a = 10 µm, scale bar in b, c = 5 µm ally fixated and processed with CLEM techniques, exhibit poor pres- ervation of the ultrastructure (asterisk marking bulging membranes of platelets) and distortion of the sections because of weak epon infiltra- tion (arrowhead). # vessel lumen, end endothelial cell, neu neutrophil, pl platelet, bulging membranes. Scale bar in a = 10 µm, scale bar in b, c = 5 µm Fig. 1 Vessel segment undergoing neutrophil extravasation during cutaneous inflammation. a Representative time-lapse still image of an inflamed vessel from intravital fluorescence microscopy showing the endothelial cell layer (red), transmigrating neutrophils (green), and platelets (white). Original movie in Supplementary Movie 1. b, c Electron micrographs of inflamed mouse skin sample, convention- Fig. 1 Vessel segment undergoing neutrophil extravasation during cutaneous inflammation. a Representative time-lapse still image of an inflamed vessel from intravital fluorescence microscopy showing the endothelial cell layer (red), transmigrating neutrophils (green), and platelets (white). Original movie in Supplementary Movie 1. b, c Electron micrographs of inflamed mouse skin sample, convention- ally fixated and processed with CLEM techniques, exhibit poor pres- ervation of the ultrastructure (asterisk marking bulging membranes of platelets) and distortion of the sections because of weak epon infiltra- tion (arrowhead). # vessel lumen, end endothelial cell, neu neutrophil, pl platelet, bulging membranes. Scale bar in a = 10 µm, scale bar in b, c = 5 µm 1 3 3 Histochemistry and Cell Biology (2022) 158:127–136 129 Fig. 2 Introduction of landmarks for CLEM orientation. Target sectioning The carbon imprint of letters and crosses on the block sur- face, now preserved as a mirror picture, is indispensable for retracing the ROI during ultrathin sectioning (Fig. 3). There- fore, all steps modifying the block need to be documented carefully (see also Fig. 3b). Optimizing imaging parameters for SBF-SEM is highly dependent on the type of sample. To achieve the best contrast and resolution possible, both high accelerating voltages and high vacuum are desirable. However, this approach could increase the possible charge accumula- tion in regions of unstained and nonconductive biological tissue or empty resin areas. Whereas backscattered elec- trons are detected for the final image acquisition, second- ary electrons are more prone to charging artifacts (Kim et al. 2008). Utilizing the secondary-electron detector unit within the microscope, charge accumulation manifested as intense brightness can therefore be detected more easily and precautions be taken. The sample investigated here proved to be stable under imaging parameters of 1.5 kV accelerating voltage, utilizing a 30 nm condenser aperture, HV conditions of 10 Pa, and a positive bias of 400 V. The pixel resolution of each ROI was set to 5 nm, in between ablation of 50-nm slices with a dwell time of 2 µs. The multi-ROI acquisition was controlled by Gatan Digital Micrograph software (version 3.32.2403.0). The same was used to convert the raw data (.dm4) into 16-bit TIFF files using a custom script command. Further processing of the datasets, including alignment, rendition, and extrac- tion of subvolumes and segmentation, was handled using Microscopy Image Browser (version 2.7) (Belevich et al. 2016). To keep the considerable data size to a minimum, First, unlike standard protocols, the block was trimmed to a cubic and not pyramidal shape, so that the dimension of the sections will not increase with further cutting into the speci- men. Using a diamond knife (Ultratrim 90, Diatome, Swit- zerland), the size of the specimen was reduced to maximal 1 × 1 mm2 in the x–y direction. One corner of the block was removed, creating a new reference point, as the carbon mask disappears with the first sections. The forward trimming and sectioning steps were all documented. At levels of interest, sections of 200 nm thickness were collected and stained with toluidine blue. Sample fixation and introduction of additional landmarks for CLEM orientation After confocal IVM, mice were terminally anesthetized. The glass cover of the dorsal skinfold chamber was removed, and the skin was submerged into fixative for 2 h in a dark room (4% formaldehyde (FA) in PBS, prewarmed to 37 °C). All further steps were performed in an unlit room to prevent bleaching of fluorescence signals in the tissue. The dorsal skinfold chamber was separated from the body of the mouse, and the skin surface (epidermis) surrounding the stamp was carefully perforated with blood lancets and acupuncture nee- dles to allow the penetration of fluids during further fixation and embedding. Next, the stamped area was punched out (12-mm dermatology skin punch), and trimmed to minimal rectangular size with a scalpel, while keeping the ROI cen- tered. One corner of the four-sided sample was removed, serving as an additional landmark for better orientation of 1 3 3 Histochemistry and Cell Biology (2022) 158:127–136 130 in the electron microscope (Tecnai 12-biotwin, Thermo Fisher Scientific Inc.). No counterstaining was needed if the samples were processed according to the OTOTO protocol. Additionally, consecutive serial sections of 60 nm or 200 nm thickness were collected and analyzed to provide more pre- liminary “volume” information in 3D (Fig. 4a, b). Repre- sentative images were taken with a 2k charge-coupled device (CCD) camera (Veleta, EMSIS, Münster, Germany) and arranged with Adobe Photoshop without further processing. Epon embedding in beam capsule For electron microscopy analysis, the sample was further processed with an adapted protocol that is also suitable for SBF-SEM (Deerinck et al. 2010) (Goudarzi et al. 2017), as summarized in Table 1. Contrast enhancement was spe- cifically achieved by keeping the specimen on a warm heat- ing plate at 40 °C during all incubation times starting with postfixation reagents (osmium tetroxide OsO4, thiocarbo- hydrazide TCH, lead aspartate; OTOTO) and followed by several washing steps. Uranyl (UA) en bloc staining was per- formed at 4 °C overnight, followed by dehydration in ethanol and acetone at room temperature and subsequent incubation with increasing concentrations of epon. The last epon infil- tration solutions were mixed with 0.07% (w/w) Ketjenblack (KB; TAAB Laboratories Equipment Ltd) to improve the conductivity of the resin, as required for SBF-SEM (Nguyen et al. 2016). Finally, a beam capsule was attached above the ROI. After resin polymerization, the gridded glass coverslip was blown off by sequential plunges in liquid nitrogen and hot water. Ideally, the carbon imprint remained on the sam- ple surface. See Table 1 for a more detailed description of the individual steps for the protocol. Target sectioning Samples were inspected with a light micro- scope to identify structures within the vascular network that match the anatomical landmarks of the vasculature recorded by confocal fluorescence microscopy (z stacks). In the ROI, the overlay of all images (carbon grid, fluorescence image, light microscopy and electron microscopy image) verified that the target region was approached successfully. Serial block‑face scanning electron microscopy Preselected correlative samples were trimmed to a size of approximately 500 × 500 × 500 µm3 and mounted on an alu- minum specimen holder using a two-component silver con- ductive epoxy adhesive in a ratio of 1.25:1 (Wanner et al. 2016). The samples were coated with a 30-nm thin layer of gold in a sputter coater (ACE EM 600, Leica, Germany). The trimmed sample was mounted in the sample holder of the SEM (JSM-7200F, JEOL, Japan) integrated ultrami- crotome stage (3View2XP, Gatan, USA). To minimize the imaging time, the trimmed sample block was aligned such that the course of the blood vessel was as parallel to the knife edge as possible, based on the previous TEM analyses (Fig. 2). An additional helpful tool in aligning the sample block was the reference landmark (cut-off edge on one side of the square sample block). Transmission electron microscopy Thin sections of 60 nm of the ROI were collected on 1 slot- filmed copper grids, counterstained with lead, and imaged 1 3 Histochemistry and Cell Biology (2022) 158:127–136 131 Table 1 Step-by-step protocol for optimized mouse skin fixation and SBF embedding The table summarizes all the steps for fixation and embedding of murine skin. The steps were modified to counteract the loss of contrast after conventional treatment of the CLEM sample (Fig. 1b, c). Major changes include the raise of temperature of the reagents and inclusion of OTOTO counterstaining (step 5–15 and 17–19), rendering the sample suitable additionally for SBF-SEM aThis needs to be done carefully because raising the temperature risks detachment of the sample as the low melting point agarose has a ten Step no. The table summarizes all the steps for fixation and embedding of murine skin. The steps were modified to counteract the loss of contrast after conventional treatment of the CLEM sample (Fig. 1b, c). Major changes include the raise of temperature of the reagents and inclusion of OTOTO counterstaining (step 5–15 and 17–19), rendering the sample suitable additionally for SBF-SEM Transmission electron microscopy Solution Time Temperature Comment 1 4% FA in PBS, 37 °C 2 h RT Prefixation in the dark to keep fluorescence 2 Cut out part with ROI, glue on coverslip with agarose, and find back ROI in fluorescence microscope with respect to the carbon grid 3 2% GA + 2% FA in 0.1 M CB, 37 °C 1 h + overnight RT + 4 °C Final fixation 4 1% FA in 0.1 M CB Until embedding 4 °C Storage 5 Demineralized water 3 × 5 min 40 °C on hot platea In four-well plate to reduce chemicals 6 2% OsO4 + 1,5% KFeCN (potassium ferricyanide(III)) in demineralized water) 30 min 7 Demineralized water 3 × 10 min 8 1% TCH in demineralized water 15 min Prepare the total volume you need for the day at the beginning of the embedding and let it cool down to 40 °C 9 Demineralized water 3 × 10 min 10 1% OsO4 in demineralized water 30 min Transfer sample to fresh four-well plate 11 Demineralized water 3 × 10 min 12 1% TCH in demineralized water 20 min 13 Demineralized water 3 × 10 min 14 1% OsO4 in demineralized water 30 min Transfer sample to fresh four-well plate 15 Demineralized water 4 × 10 min 16 0.5% UA in 25% methanol Overnight 4 °C Transfer sample to fresh four-well plate 17 Demineralized water 3 × 10 min 40 °C on hot platea 18 Lead aspartate 60 min Incubation at 60 °C is not possible because the agarose would melt 19 Demineralized water 2 × 5 min 20 70% ethanol Overnight 4 °C 21 90% ethanol 2 × 15 min RT Without agitation, to prevent detachment of the sample 22 96% ethanol 2 × 15 min 23 100% ethanol 3 × 30 min 24 100% acetone 3 × 10 min Transfer sample to glass dishes (40 mm diam- eter) 25 Epon–acetone mixture 1 + 3 1 h 26 Epon–acetone mixture 1 + 1 1 h 27 Epon + 0.7% Ketjen Black (KB)–acetone mixture 3 + 1 1 h Mix the total volume of epon you need with KB (volatile powder!) 28 Epon, pure + 0.7% KB 1.5 h Plus 15 min under vacuum to facilitate infiltra- tion of resin 29 Epon, pure + 0.7% KB Overnight Plus 15 min under vacuum to facilitate infiltra- tion of resin 30 Epon, pure + 0.7% KB 2 h or longer Plus 15 min under vacuum to facilitate infiltra- tion of resin 31 Place on top of a beam capsule (without lid and bottom) Overnight 60 °C 32 Fill the capsule with epon–KB to 3/4 6 h 33 Insert a label with sample name and fill up with pure epon without KB 3 days The label is not readable in KB Table 1 Step-by-step protocol for optimized mouse skin fixation and SBF embedding The table summarizes all the steps for fixation and embedding of murine skin. Transmission electron microscopy The resin material is trimmed away from the ROI, identified by the let- ter of the carbon imprint, resulting in a small square block (second Fig. 3 Target trimming and sectioning. a Outline of the CLEM chal- lenge: retrieve a specific area from skin tissue, recorded by IVM (left panel), repositioned several times through the entire sample prepara- tion (middle panel) and reimaging the area exactly at the ROI () in the electron microscope (right panel). The trimming and sectioning of the embedded sample block needs to be done in a controlled man- ner as further illustrated in b. After detachment from the coverslip, the block surface is covered with traces of the carbon grid (left). The resin material is trimmed away from the ROI, identified by the let- ter of the carbon imprint, resulting in a small square block (second image). One corner is cut off, defining a new reference point. Next, the ROI is approached by careful sectioning with µm/nm step size, until the z-height of the vessel of interest is reached. The approach is documented with 200-nm-thick sections, stained with toluidine blue, which are continuously compared with the pattern of the vessel net- work as imaged in the confocal microscope on the prefixed sample. The overlay of all corresponding layers (fluorescence image, carbon grid, and semithin and thin section) prove that the vessel has been retrieved in a correlative mode and allows further investigation at higher resolution in the electron microscope Transmission electron microscopy The steps were modified to counteract the loss of contrast after conventional treatment of the CLEM sample (Fig. 1b, c). Major changes include the raise of temperature of the reagents and inclusion of OTOTO counterstaining (step 5–15 and 17–19), rendering the sample suitable additionally for SBF-SEM a This needs to be done carefully, because raising the temperature risks detachment of the sample, as the low-melting-point agarose has a ten- dency to melt at 40 °C 1 3 3 Histochemistry and Cell Biology (2022) 158:127–136 132 Fig. 3 Target trimming and sectioning. a Outline of the CLEM chal- lenge: retrieve a specific area from skin tissue, recorded by IVM (left panel), repositioned several times through the entire sample prepara- tion (middle panel) and reimaging the area exactly at the ROI () in the electron microscope (right panel). The trimming and sectioning of the embedded sample block needs to be done in a controlled man- ner as further illustrated in b. After detachment from the coverslip, the block surface is covered with traces of the carbon grid (left). The resin material is trimmed away from the ROI, identified by the let- ter of the carbon imprint, resulting in a small square block (second image). One corner is cut off, defining a new reference the ROI is approached by careful sectioning with µm/nm until the z-height of the vessel of interest is reached. The documented with 200-nm-thick sections, stained with tol which are continuously compared with the pattern of the work as imaged in the confocal microscope on the pref The overlay of all corresponding layers (fluorescence im grid, and semithin and thin section) prove that the vess retrieved in a correlative mode and allows further inve higher resolution in the electron microscope Fig. 3 Target trimming and sectioning. a Outline of the CLEM chal- lenge: retrieve a specific area from skin tissue, recorded by IVM (left panel), repositioned several times through the entire sample prepara- tion (middle panel) and reimaging the area exactly at the ROI () in the electron microscope (right panel). The trimming and sectioning of the embedded sample block needs to be done in a controlled man- ner as further illustrated in b. After detachment from the coverslip, the block surface is covered with traces of the carbon grid (left). Results and discussion all datasets were converted to 8-bit TIFFs and binned to a factor of 2 or 4. In case high resolution was required, an unbinned subvolume of the ROI was isolated from the raw dataset. To keep imaging times short, the dwell times dur- ing acquisition can be reduced at the higher noise ratios. The noise can be removed via edge-preserving filters, e.g., anisotropic diffusion. Subsequent graylevel image histo- gram normalization can restore sufficient contrast for sem- iautomatic segmentation tools (e.g., watershed or global threshold) based on specific gray values. The final models and image stacks were exported for volumetric visualiza- tion in Amira 3D (version 2021.1, Thermo Fisher Scientific Inc.). IVM of inflamed skin tissue allows live cell imaging over a long time period (Fig. 1a, Supplementary Movie 1). In response to the inflammation stimulus, neutrophils leave vessels, mainly through partial openings of endothelial junctions (paracellular transmigration route). Only in dis- tinct vessels could increased numbers of platelets, obviously attracted by the transmigration process of neutrophils, be observed. As this co-interaction is not a general pheno- type, imaging of the same position seen in IVM by electron microscopy at higher resolution was required. This obser- vation cannot be captured in an unsystematic sectioning 1 3 133 Histochemistry and Cell Biology (2022) 158:127–136 process, so a CLEM approach is the only way to retrace the i viz. the epidermis and the glued side on the cover slip, Fig. 4 Analyzing the ROI by thin sectioning, thick sectioning, or serial block-face sectioning. Electron micrographs of successfully recovered CLEM vessels of inflamed mouse skin. Throughout the vessel, several cellular interactions can be observed and studied in more detail with respect to the surrounding tissue components such as the endothelial cell (end) and transmigrating neutrophil (neu) leaving the lumen of the vessel (#). To provide more “volume” information, several techniques can be applied. Serial consecutive sections (60 nm thin or 200 nm thick) can be used to follow structures in 3D on a nor- mal transmission electron microscope a, b. Physical distortions occur regularly, marked by arrowheads. An alternative approach is auto- mated sectioning by SBF-SEM. The sample block is continuously cut via an ultramicrotome, accommodated within the instrument. The scanning detector scans the surface after every section, resulting in a large 3D volume of the selected ROI that can be recorded and analyzed as single sections c. Results and discussion Here, the contour of a transmigrating neutrophil is highlighted in green, and additionally annotated in Sup- plementary Movie 3. Scale bar = 5 µm Fig. 4 Analyzing the ROI by thin sectioning, thick sectioning, or serial block-face sectioning. Electron micrographs of successfully recovered CLEM vessels of inflamed mouse skin. Throughout the vessel, several cellular interactions can be observed and studied in more detail with respect to the surrounding tissue components such as the endothelial cell (end) and transmigrating neutrophil (neu) leaving the lumen of the vessel (#). To provide more “volume” information, several techniques can be applied. Serial consecutive sections (60 nm thin or 200 nm thick) can be used to follow structures in 3D on a nor- mal transmission electron microscope a, b. Physical distortions occur regularly, marked by arrowheads. An alternative approach is auto- mated sectioning by SBF-SEM. The sample block is continuously cut via an ultramicrotome, accommodated within the instrument. The scanning detector scans the surface after every section, resulting in a large 3D volume of the selected ROI that can be recorded and analyzed as single sections c. Here, the contour of a transmigrating neutrophil is highlighted in green, and additionally annotated in Sup- plementary Movie 3. Scale bar = 5 µm mal transmission electron microscope a, b. Physical distortions occur regularly, marked by arrowheads. An alternative approach is auto- mated sectioning by SBF-SEM. The sample block is continuously cut via an ultramicrotome, accommodated within the instrument. The scanning detector scans the surface after every section, resulting in a large 3D volume of the selected ROI that can be recorded and analyzed as single sections c. Here, the contour of a transmigrating neutrophil is highlighted in green, and additionally annotated in Sup- plementary Movie 3. Scale bar = 5 µm viz. the epidermis and the glued side on the cover slip, where fluids can hardly penetrate (Fig. 2 c, Supplementary Fig. 2). At these sides, infiltration of fixatives and contrast- ing reagents were abolished, resulting in a dramatic lack of conservation and contrast compared with previous samples (Frye et al. 2015). It was not possible to further minimize the sample dimension, for example, by vibratome sectioning, as process, so a CLEM approach is the only way to retrace the same event (Kopek et al. 2017). Results and discussion The first attempts allowed us to relocate the vessel of interest, similarly to what we described earlier for zebrafish tissue (Goudarzi et al. 2015). However, we encountered unsatisfactory preservation, con- trast, and physical distortions of the sample (Fig. 1b, c). Two major entry sides of the CLEM tissue block were masked, 1 3 Histochemistry and Cell Biology (2022) 158:127–136 134 the specimen define the adjustments required to optimize the protocol. We performed ten experiments to establish the presented protocol and successfully reproduced the approach eight times. the course of the vessel could not be accurately located into a certain volume (Sousa et al. 2021). Therefore, we started to optimize the protocol step-by-step (Table 1), mainly inspired by the staining methods used for SBF-SEM (Denk and Hor- stmann 2004). Depending on the question, ultrathin serial sections might already answer the original research question. To generate volume data (in 3D), serial sectioning can be quite unpleas- ant when the section quality is disturbed by folds and holes (arrowheads in Fig. 4), sections get lost during retrieval, or the selected area does not develop throughout the series of images. As a consequence, the z-dimension that can be examined is quite limited. For more complex questions, the sample can be further analyzed by SBF-SEM, resulting in a 3D volume data stack that can be inspected and anno- tated similar to the confocal volume (Fig. 4, Supplementary Movie 3). In particular, samples with many extracellular spaces, e.g., blood vessels, or less heavily stained subcellular regions are increasingly prone to higher electron doses. To mitigate possible charging artifacts during image acquisi- tion, it is therefore necessary to introduce conductive fillers in the epon resin and keep the electron dose per unit area to a minimum while maximizing resolution. In the absence of variable pressure modes (Griffin 2007) and focal charge compensation (Deerinck et al. 2018), it is possible to bias the sample stage with a positive charge to account for the nega- tively charged electron accumulation (Reimer 2019). Prior to the final SBF-SEM acquisition, the imaging parameters must be optimized as beam-induced artifacts can only be fully evaluated after a few imaging cycles. To overcome the infiltration barrier of the fixed skin, we opened the epidermis by introducing small lesions with acupuncture needles and blood lancets. Results and discussion A major improve- ment was achieved by increasing the temperature during sample preparation, which led to better infiltration of rea- gents, including osmium tetroxide, as exemplified in Sup- plementary Fig. 2. The risk that raising the temperature could alter the reactivity of the chemicals was known, but a compromise between altered reactivity and the need for the fixatives to reach the restricted regions had to be found. The morphological appearance, obtained with the adjusted pro- tocol, showed no sign of proteolysis compared with smaller sized, conventionally processed control skin samples (not shown here). It should be mentioned that raising the tem- perature additionally risks detachment of the sample, as the low-melting-point agarose has a tendency to liquify at 40 °C. Artificially introduced landmarks, such as the stamp on the skin, removal of corners from the sample, and carbon gridded coverslip, were indispensable for retracing the fluo- rescently labeled ROI from the IVM sample in the electron microscope sample. The imprint of the carbon cover slip was not always retained perfectly on the block surface, but most of the times the remnants were sufficient to decide on the trimming area (Figs. 2j, 3b). Trimming with a 90° diamond knife was essential, resulting in a minimized square size of the block, which does not change shape while sectioning deeper into the tissue. Every step was photographed and contributed to rebuilding missing information.l The image size was highly dependent on the course of the blood vessel, thus fluctuating throughout the overall data acquisition. The natural irregular course of the vessel, due to the cutting plane of the diamond knife, occasionally led to multiple regions in which the vessel surface was exposed before ultimately fusing. To prevent loss of information due to the inherently destructive slicing of SBF-SEM, several ROIs along the vessel were imaged and adjusted frequently to the morphology of the vessel. This process kept the imag- ing time to a minimum. The overlap between the different selected tiles was kept as small as possible to prevent large areas from being exposed twice to the electron dose. Overall, 1200 consecutive slices (60 µm) were acquired, on numer- ous occasions with up to three different ROIs being imaged. Very useful in this context was remote control of the SBF run, which allows readjustment of the actual scanning area with respect to vessel progression into the sectioned block volume. 1 3 Results and discussion The fluorescently labeled vascular network of the sample turned out to be a valuable guide as well (Fig. 2h), especially for retracing the ROI under the confocal microscope in the mildly fixed and immobilized skin sample. Moreover, the anatomy of the network facilitated localization of the vessel of interest during targeted sectioning while comparing 200- nm sections stained for light microscopy with fluorescent images in the confocal stack (Fig. 3). A stepwise sectioning approach is recommended as the z-dimension of the fluorescence data cannot be transferred 1:1 to the EM specimen, since the applied treatments (fixa- tion, heating, dehydration, embedding, etc.) alter the tissue proportions. The sample dimension changes to an unpredict- able extent, especially as skin is a multilayered tissue with varying tissue composition in each layer, so each layer reacts differently to the treatment. Even swelling effects beyond those normally expected shrinkage were observed. The dataset shown in Fig. 4 and Supplementary Movie 3 represents a subset of 153 slices and covers a volume of 68 × 53 × 7.6 µm3 (the corresponding original IVM data are published in Currie et al. 2022, Fig. 1, and Supplementary Movie 1). Taken together, there is no (and probably never will be a) standard CLEM protocol. The unique characteristics of 1 3 Histochemistry and Cell Biology (2022) 158:127–136 135 Outlook and Conclusions SBF-SEM results. D.Z. wrote the first draft of the paper, and all authors read and approved the manuscript. Will CLEM become a routine approach comparable to con- focal microscopy, accessible to all scientists? This was antic- ipated by Karreman (2016). Small, service electron micros- copy units are still not routinely preparing CLEM samples, and when this is requested by customers, new challenges with changing model systems, changing research ques- tions, and collaborating partners arise. Therefore, we tried to establish a simple workflow based on our own experience and with regard to the actual research question, trying to implement published tools (Muller-Reichert and Verkade, 2021).l Funding Open Access funding enabled and organized by Projekt DEAL. Ethical approval All animal experiments were approved by the local authorities (LANUV). Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The workflow had be optimized carefully using a step-by- step approach and by trial and error. Despite all the problems encountered during the establishment of this workflow, the end result was reproducible and of high quality. Recent progress in the implementation of X-ray micro-CT pre-measurements of the sample block allowed the develop- ment of a topological 3D volume well correlated to the light microscopical dataset. From this dataset, even the sectioning angle and depth can be derived and programmed to a motor- ized microtome (https://github.com/K-Meech/crosshair). This is probably an improvement that will make CLEM techniques more popular, as it facilitates retracing of the ROI and omits the stepwise approach. However, this requires financial investment, as not every laboratory has access to X-ray micro-CT. References Belevich I, Joensuu M, Kumar D, Vihinen H, Jokitalo E (2016) Micros- copy image browser: a platform for segmentation and analysis of multidimensional datasets. PLoS Biol 14(1):e1002340. https:// doi.org/10.1371/journal.pbio.1002340 SBF-SEM allows the visualization of almost unlimited 3D volumes, capturing complex cell–cell interactions. This turned out to be the fundamental instrument to solve the original biological question, as platelets could be detected in full size transferring the basement membrane barrier of the activated vessel (as documented in detail in Currie et al. 2022). Braun LJ, Stegmeyer RI, Schafer K, Volkery S, Currie SM, Kempe B, Nottebaum AF, Vestweber D (2020) Platelets docking to VWF prevent leaks during leukocyte extravasation by stimulating Tie- 2. Blood 136(5):627–639. https://doi.org/10.1182/blood.20190 03442 Claesson-Welsh L, Dejana E, McDonald DM (2021) Permeability of the endothelial barrier: identifying and reconciling controversies. Trends Mol Med 27(4):314–331. https://doi.org/10.1016/j.mol- med.2020.11.006 It can be expected that electron microscopy and CLEM techniques will become indispensable to study aspects of vascular dynamics in the future (Tomaiuolo et al. 2020). Currie SM, Stegmeyer RI, Mildner K, Breitsprecher L, Zeuschner D, Psathaki OE, Schafer K, Wilkens M, Volkery S, Nieswandt B, Vestweber D (2022) Confocal real-time analysis of cutaneous platelet recruitment during immune-complex-mediated inflamma- tion. J Invest Dermatol. https://doi.org/10.1016/j.jid.2022.03.011 Supplementary Information The online version contains supplemen- tary material available at https://doi.org/10.1007/s00418-022-02119-8. Deerinck TJ, Bushong EA, Thor A, Ellisman MH (2010) NCMIR methods for 3D EM: a new protocol for preparation of biologi- cal specimens for serial blockface scanning electron microscopy. https://ncmir.ucsd.edu/sbemprotocol Acknowledgements We are grateful for funding by the Max Planck Society. This study was funded by CRC1009, B01 (SC, DV) and KFO342, P2 (RS, DV) with further funding by DFG iBiOs (No. PI 405/14-1) and SFB 944 Z-Project (LB, OP). Acknowledgements We are grateful for funding by the Max Planck Society. This study was funded by CRC1009, B01 (SC, DV) and KFO342, P2 (RS, DV) with further funding by DFG iBiOs (No. PI 405/14-1) and SFB 944 Z-Project (LB, OP). Deerinck TJ, Shone TM, Bushong EA, Ramachandra R, Peltier ST, Ellisman MH (2018) High-performance serial block-face SEM of nonconductive biological samples enabled by focal gas injection- based charge compensation. J Microsc 270(2):142–149. https:// doi.org/10.1111/jmi.12667 Author contributions K.M. and L.B. contributed equally to the study. S.C., D.V., and D.Z. initiated and supervised the project. Material preparation and data collection were performed by K.M., L.B., S.C., and R.S. M.S. and S.V. Competing interest The authors declare no competing interest. Ethical approval All animal experiments were approved by the local authorities (LANUV). Ethical approval All animal experiments were approved by the local authorities (LANUV). Declarations Competing interest The authors declare no competing interest. References assisted with the technical setup of the dorsal skinfold chamber. K.M. and D.Z. developed the CLEM workflow. L.B. and O.P. operated the serial block-face scanning electron microscope and generated resultant 3D volume data. K.M. prepared the figures. R.S. prepared the IVM movie. L.B. prepared the animated 3D movie of Denk W, Horstmann H (2004) Serial block-face scanning electron microscopy to reconstruct three-dimensional tissue nanostructure. PLoS Biol 2(11):e329. https://doi.org/10.1371/journal.pbio.00203 29 Frye M, Dierkes M, Kuppers V, Vockel M, Tomm J, Zeuschner D, Rossaint J, Zarbock A, Koh GY, Peters K, Nottebaum AF, Vest- weber D (2015) Interfering with VE-PTP stabilizes endothelial 1 3 Histochemistry and Cell Biology (2022) 158:127–136 136 junctions in vivo via Tie-2 in the absence of VE-cadherin. J Exp Med 212(13):2267–2287. https://doi.org/10.1084/jem.20150718 Muller-Reichert T, Verkade P (2021) Preface to CLEM IV: broaden the horizon. Methods Cell Biol 162:xix. https://doi.org/10.1016/ S0091-679X(21)00029-7 p g j Goudarzi M, Mildner K, Babatz F, Riedel D, Klambt C, Zeuschner D, Raz E (2015) Correlative light and electron microscopy of rare cell populations in zebrafish embryos using laser marks. Zebrafish 12(6):470–473. https://doi.org/10.1089/zeb.2015.1148 Nguyen HB, Thai TQ, Saitoh S, Wu B, Saitoh Y, Shimo S, Fujitani H, Otobe H, Ohno N (2016) Conductive resins improve charging and resolution of acquired images in electron microscopic volume imaging. Sci Rep 6:23721. https://doi.org/10.1038/srep23721 Goudarzi M, Tarbashevich K, Mildner K, Begemann I, Garcia J, Paksa A, Reichman-Fried M, Mahabaleshwar H, Blaser H, Hartwig J, Zeuschner D, Galic M, Bagnat M, Betz T, Raz E (2017) Bleb expansion in migrating cells depends on supply of membrane from cell surface invaginations. Dev Cell 43(5):577-587 e575. https:// doi.org/10.1016/j.devcel.2017.10.030fi Randall Wickett R, Visscher MO (2006) Structure and function of the epidermal barrier. Am Jo Infect Control 34(10):98–110. https:// doi.org/10.1016/j.ajic.2006.05.295 Reimer R (2019) Image quality in SBF-SEM: obstacles and approaches. In: MC 2019 Berlin microscopy conference. LS7.001, P 785. https://doi.org/10.5283/epub.40685 Griffin BJ (2007) Variable pressure and environmental scanning elec- tron microscopy: imaging of biological samples. Methods Mol Biol 369:467–495. https://doi.org/10.1007/978-1-59745-294-6_23 Secklehner J, Lo Celso C, Carlin LM (2017) Intravital microscopy in historic and contemporary immunology. Immunol Cell Biol 95(6):506–513. https://doi.org/10.1038/icb.2017.25i Karreman MA, Hyenne V, Schwab Y, Goetz JG (2016) Intravital cor- relative microscopy: imaging life at the nanoscale. Trends Cell Biol 26(11):848–863. https://doi.org/10.1016/j.tcb.2016.07.003 Sousa AL, Rodrigues Loios J, Faisca P, Tranfield EM (2021) The Histo-CLEM Workflow for tissues of model organisms. Methods Cell Biol 162:13–37. References https://doi.org/10.1016/bs.mcb.2020.12.001 Kim KW, Choi SJ, Moon TW (2008) Backscattered electron imaging for reduced charging of moisturized corn starch granules: impli- cations for versatile imagery of hygroscopic powder specimens. Micron 39(8):1160–1165. https://doi.org/10.1016/j.micron.2008. 05.007 Tomaiuolo M, Litvinov RI, Weisel JW, Stalker TJ (2020) Use of electron microscopy to study platelets and thrombi. Platelets 31(5):580–588. https://doi.org/10.1080/09537104.2020.1763939 Vestweber D (2015) How leukocytes cross the vascular endothelium. Nat Rev Immunol 15(11):692–704. https://doi.org/10.1038/nri39 08 Kopek BG, Paez-Segala MG, Shtengel G, Sochacki KA, Sun MG, Wang Y, Xu CS, van Engelenburg SB, Taraska JW, Looger LL, Hess HF (2017) Diverse protocols for correlative super-resolution fluorescence imaging and electron microscopy of chemically fixed samples. Nat Protoc 12(5):916–946. https://doi.org/10.1038/nprot. 2017.017l Wanner AA, Genoud C, Friedrich RW (2016) 3-dimensional electron microscopic imaging of the zebrafish olfactory bulb and dense reconstruction of neurons. Sci Data 3:160100. https://doi.org/10. 1038/sdata.2016.100 Lisman T (2018) Platelet-neutrophil interactions as drivers of inflam- matory and thrombotic disease. Cell Tissue Res 371(3):567–576. https://doi.org/10.1007/s00441-017-2727-4 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Luckner M, Burgold S, Filser S, Scheungrab M, Niyaz Y, Hummel E, Wanner G, Herms J (2018) Label-free 3D-CLEM using endog- enous tissue landmarks. iScience 6:92–101. https://doi.org/10. 1016/j.isci.2018.07.012 1 3 1 3 3
https://openalex.org/W4311376658	https://www.biarjournal.com/index.php/matondang/article/download/804/777	English	null	Investigation Of The Psychological Effects Of Social Media Use Among Students In Minnesota, United State America	Matondang Journal/Màtondàng Journal	2,022	cc-by-sa	5,246	Matondang Journal ISSN: 2828-1942 (Online) Vol 1 No 1 January 2022 Page: 1-9 Investigation Of The Psychological Effects Of Social Media Use Among Students In Minnesota, United State America Toyosi Motilola Olola1, Akpan Ubong-Abasi Asukwo2, Funmilayo Odufuwa3 1Department of Communication School, University of North Dakota, Grand Forks, USA 2African and African American Studies, Department of Religion, Rice University, Houston Texas 3Department of Communication School, North Dakota State University toyosi.olola@und.edu, uaa2@rice.edu, odufuwa@ndsu.edu Abstract: The study looked into the psychological effects of social media use among students in Minnesota, USA. Specifically, the study assessed the various psychological effects of social media on students in Minnesota and the United States of America. It investigated the impact of social media use on the psychological well-being of students in Minnesota, United States of America. The study adopted a descriptive survey research design. The sample for the study consisted of two hundred (200) students that were randomly selected using a simple random sampling technique. One research instrument tagged "Psychological Effects of Social Media Questionnaire" (PESMQ) was used to collect data for the study. The data collected were analysed using frequency counts and simple percentage analyses of covariance (ANCOVA). According to the study's findings, the psychological effects of social media use among students in Minnesota, United States of America include depression, stress, anxiety, emotional isolation, low self-esteem, memory loss, and self-harm. ca. The study's findings also revealed that the use of social media had a significant impact on the psychological well-being of students in Minnesota, United States of America. The study concluded that depression, stress, anxiety, emotional isolation, low self-esteem, memory loss, and self-harm are the psychological effects of social media use among students and that there is a significant effect of the use of social media on the psychological well-being of the students in Minnesota, United States of America. Keywords: social media; students; use; online; psychological; effects; communication Matondang Journal ISSN: 2828-1942 (Online) Vol. 1, No. 1, January 2022, Page: 1-9 Matondang Journal ISSN: 2828-1942 (Online) Vol. 1, No. 1, January 2022, Page: 1-9 I. Introduction Over the years, social media has flooded most areas of our society. For youth, social media plays a particularly large part in their lives, as indicated by their extensive use of several social media platforms (Ofcom, 2018). Individuals use social media for many reasons, including entertainment, communication, and searching for information. Remarkably, youth and adults spend more time using online networking sites, e-games, texting, and other social media (Twenge and Campbell, 2019). Some authors (Dhir et al., 2018; Tateno et al., 2019) have suggested that social media has altered the forms of group interaction and its users’ individual and collective behaviour worldwide. In addition, the use of social media and the types of platforms offered have increased at such a rapid pace that there is reason to believe that scientific knowledge about social media in relation to youth's health and well-being is scattered and incomplete (Orben, 2020). Nevertheless, research findings have shown the potential negative effects of social media on mental health and well-being are frequently reported in traditional media (newspapers, radio, and TV) (Bell et al., 2015). Twenge and Campbell (2019) submitted that the use of social media has a negative impact on well-being, while Orben and Przybylski (2019) argued that the association between digital technology use and adolescent well-being is so small that it is more or less inconsequential. According to the World Health Organization (2017), people who have been using the internet for a long time have been shielded from _______________________________________________________________ DOI: https://doi.org/10.33258/matondang.v2i1.804 -11- Matondang Journal ISSN: 2828-1942 (Online), 2828-979X (Print) Vol. 2, No. 1, January 2023, Page: 11-19 dangers, and they are also concerned about a country's future issues.Pater and Mynatt (2017) depicted that social media causes hostile and unsafe conduct in this manner, which is harmful to the emotional well-being of a person, particularly for youngsters. The impact of internet- based life depends upon its use by an individual who utilises too much social media as well as whatever can be harmful to him. Parmar (2017) expressed that in the 21st century, youth have been investing their energy in various social contraptions, spending a normal 12–15 hours per day on stimulation media, including telephones, PCs, workstations, and TVs, among other gadgets. With the assistance of these gadgets, they use WhatsApp, YouTube, Facebook, Instagram, Twitter, and some others. I. Introduction (Olola, 2022) g , , ( , ) Similarly, the number of online life clients has been steadily increasing, posing a genuine health risk to the psychological well-being of youth.Dangerous online life use was unequivocally and openly linked to increased depression side effects (Shensa et al. 2017).Thornicroft et al. (2016) proposed that online networking is important for children's lives because it has an incredible connection to people groups' emotional well-being, however, some speculative portrayal thought it as an extraordinary danger and perilous for human psychological wellness, also youth use social networking sites for digital harassing, similar to ridicule others, mimicry, and so on, as a result others lose their trust.It has small positive effects on online networking among the younger generation. g g y g g Miller et al. (2016) demonstrated that internet-based life alters the method for correspondence and socialisation while also preserving companionship.While becoming a development group has its advantages, on the other hand, investing their energy in society is hazardous to their emotional health.According to Len-Ros et al. (2016), social media is frequently used for long-term interpersonal communication bookkeeping, which is a serious risk for youth or adolescents because regular use of person-to-person communication bookkeeping can impact emotional wellness. Patel et al. (2016) pointed out that online networking is creating problems like uneasiness, sorrow, and dietary issues. Furthermore, the use of social media has become an increasingly difficult issue for young people, as more of them invest their energy looking over social media applications and receive nothing in return, posing a significant risk to young people's mental health. It has been discovered that excessive use of social networking systems can cause increased misery, anxiety, pain, depression, and dissatisfaction with life, thus deteriorating psychological well-being (Woods and Scott, 2016). Social media use has been linked to increased depression (Lin et al., 2016). Unlike previous explanations, Townsend et al. (2016) considered how social media becomes a barrier to meeting people and that people typically live alone when they converse with others; as a result, their proclivity to live alone is an indication of stress, sadness, and self-deduction. As a result, research into the psychological effects of social media use among students in Minnesota, United States of America, is required. 1.2 Research Question What are the psychological effects of social media use among students in Minnesota, United State of America? 2.1 Social Media Engagement Theory g g y Social media engagement (SME) theory accounts for the role of technology as the underlying platform needed to facilitate social interactions among users that are globally and temporally distributed. Clearly, the rise of social media comes in large part from the evolution of technology to provide a unique user experience that enables users to connect in new ways that were never before possible. The central premise of SME theory is that higher user engagement leads to greater usage of the social media platform. Usage is defined as the frequency of a user’s contribution, retrieval, and/or exploration of content within a social media site (Kankanhalli, Tan, & Wei, 2005; Li & Bernoff, 2008). The more frequently users take part in a variety of activities, the more valuable the social media platform becomes to the organisation and fellow users, resulting in the co-creation of value (Kankanhalli et al., 2005; Li & Bernoff, 2008). As such, the Social Media Engagement (SME) theory is considered good for studying the psychological effects of social media use among students in the study area. 2.2 Concept of social media Since the inception of this technology, social media has been defined in a different context.It is difficult to provide a single definition incorporating all the technology and practises related to social media, partly because the technology is not limited by any specific scope, format, topic, audience, or source. (Treem et al., 2016). Appel et al. (2020) described social media as a technology-based system that is not entirely technological and in which a variety of interconnected actors, including individuals, businesses, organizations, and institutions, can engage in a variety of complex and varied behaviors, interactions, and exchanges. In addition, Kaplan and Haenlein (2010) defined social media as "a group of Internet-based applications that builds on the ideological and technological foundations of Web 2.0 and allows the creation and exchange of user-generated content." In terms of marketing, Cheung et al. (2020) described social media as a web tool used to create content tailored for each social media platform to increase interaction and advance business. As a result, social media can be defined in terms other than online communication, such as marketing, politics, education, health, and essential human interaction (Bashir & Bhat, 2016). Building on these preexisting definitions and the concept to which this literature is tilted, we define social media as software-based digital technologies that typically take the shape of applications and websites and provide users the ability to communicate and receive digital content, including images, texts, videos, and other types of media content. A few examples of these apps are Facebook, Instagram, Twitter, Snapchat, and Tiktok. Snapchat's social media app is one of the most popular online activities, with approximately 4.6 billion people using the various platforms in 2022 and a worldwide consumption time of 147 minutes per day. Also, 2 million Snapchat messages are sent per minute (Statista, 2022). 1.3 Hypothesis There is no significant effect of the use of social media on the psychological being of the students in Minnesota, United State of America. yp There is no significant effect of the use of social media on the psychological being of the students in Minnesota, United State of America. II. Review of Literature II. Review of Literature 1.1 Statement of Problem Social media plays an important role in everyone’s life because it’s a source of entertainment, information, and so forth. Despite social media's critical role in society, students abuse it in one way or another, resulting in physical and mental health consequences. Hence, the need to investigate the psychological effects of social media use among students in Minnesota, United States of America, becomes imperative. 1.2 Research Question What are the psychological effects of social media use among students in Minnesota, United State of America? www.biarjournal.com/index.php/matondang 2.3 Psychological Effects and Usage of social media Psychosocial behaviour is seen as the response or reaction of how an individual behaves and develops his or her relationship in the school environment or anywhere the person finds himself or herself as a result of the use of mobile phones, iPads, tablets, laptops, and other internet services. Reasonable attention has been given to the negative side of social networking sites, such as cyberbullying, parent-child interactions, internet addiction, and the -13- influence of social media use on sleep disturbance and academic performance. Suhail and Bargees (2006) conducted a study on the effect of social networking in Pakistan using the Internet Effect Scale (IES) to determine the effect of some factors, specifically those that include behavioural issues, educational problems, internet abuse, interpersonal problems, and psychological issues, as well as their consequences. The result of the studies showed that social media networking has both positive and negative advantages for its users. The connectivism between spare times spent on the internet showed tremendous impact; the time spent on the internet has affected students’ interpersonal, academic, psychological, and physical factors. Also, it is recommended that the use of the internet may perhaps be used to enhance academic performance among students. p g Most adolescents come across different levels of negative experiences through SNSs, and the ferocity of online bullying is exacerbated. Some adolescents may be more exposed to develop at-risk indicators of internet addiction than others, while others may be experiencing other psychological symptoms and syndromes such as depression, ADHD symptoms, or aggression (Shapiro & Margolin, 2014). Akhtar (2013) revealed that immoderate usage of the internet can result in addiction among students, which could lead to poor academic attainment. The author opined that internet addiction can cause a reduction in time spent on studies, a fundamental drop in grades, low interest in extracurricular activities, and a loss of interest in lessons. Furthermore, Internet addiction among adolescents has been related to psychological and intellectual problems, which include anxiety, despair, and loneliness. As a result, other studies on social media networking have shown excellent academic interaction and supported net usage among students. pp g g However, the negative impact of social media has outweighed its tremendous aspects. Thus, excessive utilisation of social media has shown a negative effect on students' physical, psychological, as well as family fitness. III. Research Method Section A contained items on the socio- demographic information of the respondents; Section B contained seven items on the various psychological effects of social media used by the respondents based on a Likert scale of "strongly agree," "agree," "disagree," and "strongly disagree"; and Section C contained seven items on the usage of social media by the respondents based on a projective rating scale of "always," "sometimes," and "never." The questionnaires were administered to the respondents and later retrieved from them for data analysis. The data collected were analysed using frequency counts, simple percentages, and analysis of covariance (ANCOVA). 2.3 Psychological Effects and Usage of social media Furthermore, excessive use has been linked to sleep deprivation, insomnia, and chronic illnesses (O'keeffe and Pearson, 2011).Social media platforms provide students with new opportunities and challenges to express their individual worth to the world (their likes and dislikes) and receive immediate feedback. Most of the students express their views to both known and unknown recipients. They widely share their views with varying degrees of accuracy, honesty, and openness that were previously private or reserved for selected individuals. Students can also join internet or social groups reflecting aspects of their personalities they wish to discover or develop (Kolek & Saunders, 2008). With respect to psychosocial development, the use of social media has influenced matters such as individuality formation and social skill advancement. There are positive aspects of social media use among students, and there are also many negative effects concerning the use of such gadgets. Since their introduction, social networking sites have drawn the attention of many adolescents and university students, as observed by Lenhart and Madden (2007). The manner in which these sites are paving their way through the educational setting cannot be overemphasised, hence the belief that these sites have even more potential for the improvement of teaching, learning, and sharing of information among learners and educators (Ferdig, 2007). ( g ) Generally, students are sent to school to learn but are exposed to numerous experiences that impact their behavior. One of the goals of education is to influence behaviour in a desirable direction, so that one behaves in accordance with society's accepted values and norms. Thus, learning is a behavioural change that manifests in cognitive reasoning, physical development, manipulative skills, and growth in values and interests. All these changes depend on individual differences, the home background, and the school setting. According to Oluwatoyin (2011), educators view social media as encouraging negative traditional literacy abilities (writing, grammar, vocabulary), which will be largely disruptive to learners' reading skills required in the school environment. -14- III. Research Method III. Research Method The study adopted a descriptive survey research design. A descriptive survey aims at describing the distribution of a phenomenon in a population and thereby establishing the facts. The study population comprised students in Minnesota and the United States of America. The sample for the study consisted of two hundred (200) students that were randomly selected using a simple random sampling technique. One research instrument tagged "Psychological Effects of Social Media Questionnaire" (PESMQ) was used to collect data for the study. The instrument has three sections. Section A contained items on the socio- demographic information of the respondents; Section B contained seven items on the various psychological effects of social media used by the respondents based on a Likert scale of "strongly agree," "agree," "disagree," and "strongly disagree"; and Section C contained seven items on the usage of social media by the respondents based on a projective rating scale of "always," "sometimes," and "never." The questionnaires were administered to the respondents and later retrieved from them for data analysis. The data collected were analysed using frequency counts, simple percentages, and analysis of covariance (ANCOVA). The study adopted a descriptive survey research design. A descriptive survey aims describing the distribution of a phenomenon in a population and thereby establishing th The study adopted a descriptive survey research design. A descriptive survey aims at describing the distribution of a phenomenon in a population and thereby establishing the facts. The study population comprised students in Minnesota and the United States of America. The sample for the study consisted of two hundred (200) students that were randomly selected using a simple random sampling technique. One research instrument tagged "Psychological Effects of Social Media Questionnaire" (PESMQ) was used to collect data for the study. The instrument has three sections. Section A contained items on the socio- demographic information of the respondents; Section B contained seven items on the various psychological effects of social media used by the respondents based on a Likert scale of "strongly agree " "agree " "disagree " and "strongly disagree"; and Section C contained seven randomly selected using a simple random sampling technique. One research instrument tagged "Psychological Effects of Social Media Questionnaire" (PESMQ) was used to collect data for the study. The instrument has three sections. 3.1 Purpose of the Study The primary goal of the study is to look into the psychological effects of social media use among students in Minnesota, United States of America. Specifically, the study’s objective is to assess the various psychological effects of social media on students in Minnesota, United States of America, and examine the effect of the use of social media on the psychological well- being of the students in Minnesota, United States of America. 3.2 Significance of the Study 3.2 Significance of the Study g y The study would provide practical information on the various psychological effects of social media and their effects on the psychological well-being of students in Minnesota, USA. The study will help parents and guardians ensure that their wards are appropriately monitored as far as the use of social media is concerned. The study's findings would help to expand the frontiers of knowledge and serve as a catalyst for future research. IV. Result and Discussion IV. Result and Discussion IV. Result and Discussion 4.1 Analysis Table 1. Descriptive analysis of socio-demographic information of the respondents in Minnesota, United State of America S/N Variables Frequency (f) Percentage (%) 1. Gender Male 91 45.5 Female 109 54.5 2. Age Bracket 10-12 years 48 24.0 13-15 years 72 36.0 16-19 years 80 40.0 N = 200 y Table 1. Descriptive analysis of socio-demographic information of the respondents in Minnesota, United State of America N Variables Frequency (f) Percentage (% -15- Results in Table 1 showed the descriptive analysis of socio-demographic information of the respondents in Minnesota, United State of America. Considering gender, it can be gathered from the Table that 45.5% of the respondents are males while 54.5% are females. In the same light, 24.0%, 36.0%, and 40.0% are in the age brackets of 10-12 years, 13-15 years, and above 16-19 years respectively in the study area. Research Question: What are the psychological effects of social media use among students in Minnesota, United State of America? Data collected on the various psychological effects of the use of social media were subjected to descriptive analysis of frequency and percentage to answer this research question. The results are presented in Table 2. Research Question: What are the psychological effects of social media use amon Research Question: What are the psychological effects of social media use among students in Minnesota, United State of America? Data collected on the various psychological effects of the use of social media were subjected to descriptive analysis of frequency and percentage to answer this research question. The results are presented in Table 2. Table 2. Descriptive analysis of the various psychological effects of social media use among students in Minnesota, United State of America S/N Items Frequency (f) Percentage (%) 1. Depression 174 87.0 2. Stress 169 84.5 3. Anxiety 155 77.5 4. Emotional isolation 170 85.0 5. Low self-esteem 145 72.5 6. Memory loss 140 70.0 7. 3.2 Significance of the Study Self-harm 142 71.0 N = 200 ble 2. Descriptive analysis of the various psychological effects of social media use among students in Minnesota, United State of America Results in Table 2 showed the descriptive analysis of psychological effects of social media use among students in Minnesota, United State of America. It can be deduced from the Table that 87.0%, 84.5%, 77.5%, 85.0%, 72.5%, 70.0% and 71.0% of the students submitted that depression, stress, anxiety, emotional isolation, low self-esteem, memory loss and self- harm respectively are the psychological effects of social media use among students in Minnesota, United State of America. Results in Table 2 showed the descriptive analysis of psychological effects of social media use among students in Minnesota, United State of America. It can be deduced from the Table that 87.0%, 84.5%, 77.5%, 85.0%, 72.5%, 70.0% and 71.0% of the students submitted that depression, stress, anxiety, emotional isolation, low self-esteem, memory loss and self- harm respectively are the psychological effects of social media use among students in Minnesota, United State of America. 4.3 Discussion According to the study's findings, the psychological effects of social media use among students in Minnesota, United States of America include depression, stress, anxiety, emotional isolation, low self-esteem, memory loss, and self-harm. The results corroborate the findings of Deepa and Priya (2020), who carried out a study on the impact of social media on the mental health of students. They found out that depression, stress, emotional isolation, and anxiety are the psychological effects of the use of social media among learners in the study area. The study's findings also revealed that the use of social media had a significant impact on the psychological well-being of students in Minnesota, United States of America. The findings are consistent with those of Talatu and Murja (2018), who investigated the impact of social media on secondary school students' psychosocial behaviour and academic performance. It was conducted in Batagarawa Local Government, Katsina State, Nigeria. According to the study's findings, social media use has a negative impact on psychosocial behaviour and academic performance. performance. V. Conclusion The study concluded that depression, stress, anxiety, emotional isolation, low self- esteem, memory loss, and self-harm are the psychological effects of social media use among students and that there is a significant effect of the use of social media on the psychological well-being of the students in Minnesota, United States of America. 4.2 Hypothesis 4.2 Hypothesis There is no significant effect of the use of social media on the psychological being of the students in Minnesota, United State of America. In order to test for this hypothesis, data collected on the psychological effects and the use of social media among students were subjected to analysis of covariance (ANCOVA). The result is presented in Table 3. Table 3: Analysis of covariance (ANCOVA) of the effect of the use of social media on the psychological being of the students in Minnesota, United State of America Table 3: Analysis of covariance (ANCOVA) of the effect of the use of social media on the psychological being of the students in Minnesota, United State of America Tests of Between-Subjects Effects Dependent Variable: PE Source Type III Sum of Squares df Mean Square F Sig. Partial Eta Squared Corrected Model 31.757a 16 1.985 .720 .027 .590 Intercept 10676.328 1 10676.328 3870.969 .000 .955 SM 31.757 16 1.985 .720 .027 .590 -16- Error 504.723 183 2.758 Total 23606.000 200 Corrected Total 536.480 199 R Squared = .059 (Adjusted R Squared = -.023) (F = 0.720; p < 0.05) Results inn Table 3 showed that there was significant effect of the use of social media on the psychological being of the students in Minnesota, United State of America (F = 0.720; p < 0.05). Hence, the null hypothesis that states that there is no significant effect of the use of social media on the psychological being of the students in Minnesota, United State of America is hereby rejected. The partial eta squared value of 0.590 actually accounted for 59.0% effect size of the use of social media on the psychological being of the students in Minnesota, United State of America. References Online disclosure: An empirical examination of undergraduate facebook profiles. NASPA Journal 45(1) 1-5. Lenhart, A., & Madden, M. (2007). Social networking websites and teens: An overwiew. Pew Internet and American Life Project Report. Retrieved from [Online]: http://www pewinternet org/PPF/r/198/report display asp Lenhart, A., & Madden, M. (2007). Social networking websites and teens: An overwiew. Pew Internet and American Life Project Report Retrieved from [Online]: Len-Ríos, M.E., Hughes, H.E., McKee, L.G. and Young, H.N., (2016). Early adolescents as publics: A national survey of teens with social media accounts, their media use preferences, parental mediation, and perceived Internet literacy.Public Relations Review, 42(1), pp. 101-108. pp Miller. et.al. (2016).How the world changed social media. UCL Press (p. 286). O’Keeffe, G. S. & Pearson, K. C. (2011). The impact of social media on children, adolescents and families. American Academy of Paediatrics, 127(4) 45-52. Ofcom (2018). Children and Parents: Media Use and Attitudes Report. Warrington:Ofcom. Olola, T. M. (2022), Social Media Use among Adolescents in North Dakota, USA. Budapest International Research and Critics Instiute Journal. Vol. 5(4), pp. 30654-30660 Oluwatoyin, A. E. (2011). ICT, CGPA: Consequences of social networks in an internet driven learning society. Retrieved from International Journal of Computer Trends & Ofcom (2018). Children and Parents: Media Use and Attitudes Report. Warrington:Ofcom. Olola, T. M. (2022), Social Media Use among Adolescents in North Dakota, USA. Budapest International Research and Critics Instiute Journal. Vol. 5(4), pp. 30654-30660 Oluwatoyin, A. E. (2011). ICT, CGPA: Consequences of social networks in an internet driven learning society. Retrieved from International Journal of Computer Trends & Technology: 2(2) 9-14: http://www.internationaljournals/srg.org Orben, A. (2020). Teenagers, screens and social media: a narrative review ofreviews and key studies. J. Soc. Psychiatry Psychiatr. Epidemiol. 55, 407–414.doi: 10.1007/s00127-019- 01825-4 Orben, A., and Przybylski, A. K. (2019). The association between adolescent wellbeingand digital technology use. Nat. Hum. Behav. 3, 173–182. doi: 10.1038/s41562-018-0506- 1 Pakistan. Cyber Psychology and Behaviour, 9 (3). 139-152 Pater, J. and Mynatt, E.D., (2017). Defining Digital Self-Harm. InCSCW, ISBN: 978-1-4503- 4335-0(pp. 1501-1513). y y gy , ( ) Pater, J. and Mynatt, E.D., (2017). Defining Digital Self-Harm. InCSCW, ISBN: 978-1-4503- 4335-0(pp. 1501-1513). (pp Shapiro, A. L. & Margolin, G. (2014). Growing up wired: Social networking sites and Shapiro, A. L. & Margolin, G. (2014). Growing up wired: Social networking sites and adolescent psychosocial development. Retrieved from Clin Child Psychol Rev. Shapiro, A. L. & Margolin, G. (2014). References Akhtar, N. (2013). Relationship between internet addiction and academic performance among university undergraduates. Educational Research and Reviews. 8(19), 1793-1796. Appel, H., Gerlach, A. L., and Crusius, J. (2020). The interplay between Facebookuse, social comparison,envy, and depression. Curr. Opin. Psychol. 9, 44–49.doi: 10.1016/j.copsyc.2015.10.006 Bell, V., Bishop, D. V., and Przybylski, A. K. (2015). The debate over digitaltechnology and young people. BMJ 351:h3064. doi: 10.1136/bmj.h3064 Deepa, M. & Priya, V. K. (2020). Impact of social media on mental health of students. International Journal of Scientific and Technology Research. Vol. 9 (3). Dhir, A., Yossatorn, Y., Kaur, P., and Chen, S. (2018). Online social mediafatigue and psychological wellbeing—a study of compulsive use, fear ofmissing out, fatigue, Akhtar, N. (2013). Relationship between internet addiction and academic performance among university undergraduates. Educational Research and Reviews. 8(19), 1793-1796. Appel, H., Gerlach, A. L., and Crusius, J. (2020). The interplay between Facebookuse, social comparison,envy, and depression. Curr. Opin. Psychol. 9, 44–49.doi: 10.1016/j.copsyc.2015.10.006 Akhtar, N. (2013). Relationship between internet addiction and academic performance among university undergraduates. Educational Research and Reviews. 8(19), 1793-1796. Appel, H., Gerlach, A. L., and Crusius, J. (2020). The interplay between Facebookuse, social comparison,envy, and depression. Curr. Opin. Psychol. 9, 44–49.doi: 10.1016/j.copsyc.2015.10.006 10.1016/j.copsyc.2015.10.006 Bell, V., Bishop, D. V., and Przybylski, A. K. (2015). The debate over digitaltechnology and young people. BMJ 351:h3064. doi: 10.1136/bmj.h3064 Deepa, M. & Priya, V. K. (2020). Impact of social media on mental health of students. International Journal of Scientific and Technology Research. Vol. 9 (3). Dhir, A., Yossatorn, Y., Kaur, P., and Chen, S. (2018). Online social mediafatigue and psychological wellbeing—a study of compulsive use, fear ofmissing out, fatigue, y g p p J j Deepa, M. & Priya, V. K. (2020). Impact of social media on mental health of students. International Journal of Scientific and Technology Research. Vol. 9 (3). g Dhir, A., Yossatorn, Y., Kaur, P., and Chen, S. (2018). Online social mediafatigue and psychological wellbeing—a study of compulsive use, fear ofmissing out, fatigue, -17- anxiety and depression. Int. J. Inf. Manag. 40, 141–152.doi: 10.1016/j.ijinfomgt.2018.01.012 j j g Ferdig, R. (2007). Editorial: examining social software in teacher education. Journal of Technology and Teacher Education, 15(1), 5-10. gy ( ) Kaplan, A. M., and Haenlein, M. (2010). Users of the world, unite! The challengesand opportunities of social media. Bus. Horiz. 53, 59–68. doi: 10 1016/j b h 2009 09 003 Kolek, E. A., & Saunders, D. (2008). Woods, H. C., and Scott, H. (2016). Sleepyteens: social media use inadolescence is associated with poor sleep quality, anxiety, depression andlow self-esteem. J. Adolesc. 51, 41–49. doi: 10.1016/j.adolescence.2016.05.008 Townsend, L. et.al.(2016). Building virtual bridges: How rural Micro‐Enterprises develop social capital in online and Face‐to‐Face settings. 56(1), pp.29-47. Twenge, J. M., and Campbell, W. K. (2019). Media use islinked to lower psychological well- being: evidence from threedatasets. Psychiatr. Q. 90, 311–331. doi: 10.1007/s11126- 019- 09630-7 References Growing up wired: Social networking sites and adolescent psychosocial development. Retrieved from Clin Child Psychol Rev. 17(1)_18: doi: 10.1007/s10567-013-0135-1 adolescent psychosocial development. Retrieved from Clin Child Psychol Rev. 17(1)_18: doi: 10.1007/s10567-013-0135 Shensa, A., Escobar-Viera, C.G., Sidani, J.E., Bowman, N.D., Marshal, M.P. and Primack, B. A. (2017), “Problematic social media use and depressivesymptoms among US young adults: a nationally representative study”, Social Science & Medicine, 182, 150-157. adults: a nationally representative study , Social Science & Medicine, 182, 150 157. Suhail, K. &Bargees, Z. (2006). Effects of Excessive Internet Use on Undergraduate Students in Talatu, I. U. &Murja, I. (2018). Influence of social media on psychological behaviour and academic performance of secondary school students. Journal of Education and Entrepreneurship. Vol 5 (2), 36 -46. Tateno, M., Teo, A. R., Ukai, W., Kanazawa, J., Katsuki, R., Kubo, H., et al. (2019).Internet addiction, smartphone addiction, and hikikomori trait in Japaneseyoung adult: social isolation and social network. Front. Psychiatry 10:455.doi: 10.3389/fpsyt.2019.00455 Thornicroft. et al., (2016). Evidence for effective interventions to reduce mental-health-related stigma and discrimination.The Lancet,387(10023), pp.1123-1132. -18- Townsend, L. et.al.(2016). Building virtual bridges: How rural Micro‐Enterprises develop social capital in online and Face‐to‐Face settings. 56(1), pp.29-47. Twenge, J. M., and Campbell, W. K. (2019). Media use islinked to lower psychological well- being: evidence from threedatasets. Psychiatr. Q. 90, 311–331. doi: 10.1007/s11126- 019- 09630-7 Woods, H. C., and Scott, H. (2016). Sleepyteens: social media use inadolescence is associated with poor sleep quality, anxiety, depression andlow self-esteem. J. Adolesc. 51, 41–49. doi: 10.1016/j.adolescence.2016.05.008 -19- -19-
https://openalex.org/W4285183730	https://www.degruyter.com/document/doi/10.1515/med-2022-0495/pdf	English	null	<i>In vitro</i> antimicrobial activity of ozonated oil in liposome eyedrop against multidrug-resistant bacteria	Open Medicine	2,022	cc-by	4,883	Open Medicine 2022; 17: 1057–1063 Giuseppe Grandi, Rossana Cavallo, Elisa Zanotto, Raﬀaella Cipriani, Claudio Panico, Romolo Protti, Giovanni Scapagnini, Sergio Davinelli, Ciro Costagliola the ocular surface with a potential role in antimicrobial prophylaxis. the ocular surface with a potential role in antimicrobial prophylaxis. https://doi.org/10.1515/med-2022-0495 received February 13, 2022; accepted May 9, 2022 Abstract: Bacteria are the most common causative agents of ocular infections. Treatment with topical broad-spectrum antibiotics is recommended in severe cases. However, anti- biotic resistance has become a major concern in recent years, although antibiotics are generally eﬀective in treating ocular infections. Antibacterial compound screening is per- formed to identify alternative therapeutic options to anti- biotics. The aim of this study was to assess the in vitro antimicrobial activity of an ophthalmic solution containing ozonated oil. Strains of bacterial species with a multidrug resistance proﬁle, which are responsible for a large propor- tion of ocular infections, were isolated and selected from diﬀerent biological samples. The bacterial isolates were cul- tured, and ozonated oil was used to evaluate the inhibition zones at diﬀerent time points. The treatment exhibited anti- bacterial activity against all the tested species. The eﬀect was lower against the strains of Pseudomonas aeruginosa and more evident against Staphylococcus and Streptococcus spp. Our results suggest that the administration of ozonated oil may be a candidate agent to treat some infections of Keywords: in vitro antimicrobial activity, ozonated oil, bacteria, ocular infections ppe Grandi et al., published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 This work is licensed under the Creative Commons Attribution 4.0 Giuseppe Grandi, Rossana Cavallo, Elisa Zanotto: Microbiology and Virology Unit, University Hospital Città della Salute e della Scienza di Torino, Turin, 10126, Italy Raﬀaella Cipriani: Microbiology and Virology Unit, Amedeo di Savoia Hospital, Turin, 10149, Italy Claudio Panico: Ophthalmology Unit, Humanitas Gradenigo Hospital, Turin, 10153, Italy Romolo Protti: Ophthalmic Unit, Hospital San Biagio, Domodossola, 28845, Italy Giovanni Scapagnini, Ciro Costagliola: Department of Medicine and Health Sciences “V. Tiberio”, University of Molise, Via V. De Sanctis, s.n.c., Campobasso 86100, Italy Corresponding author: Sergio Davinelli, Department of Medicine and Health Sciences “V. Tiberio”, University of Molise, Via V. De Sanctis, s.n.c., Campobasso 86100, Italy, e-mail: sergio.davinelli@unimol.it, tel: +39 0874 404771, fax: +39 0874 404 778 Open Access. © 2022 Giuseppe Grandi et al., published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. 2 Methods From the bacterial strain library of the Comprehensive Structure of Microbiology and Virology Unit, University “Città della Salute e della Scienza,” Turin, Italy, 60 microorganisms with MDR proﬁles (bacteria resistant to at least three classes of antibiotics, including ﬂuoroqui- nolones), previously isolated from diﬀerent biological materials, were selected. These included (1) 20 strains of P. aeruginosa (PA MDR); (2) 20 strains of methicillin- resistant S. aureus (MRSA); and (3) 20 strains of methi- cillin-resistant S. epidermidis (MRSE). In addition, ten strains of Streptococcus spp. (ﬁve S. pneumoniae strains, three S. agalactiae strains, and two S. pyogenes strains) were analyzed even if they showed no signiﬁcant resis- tance to antibiotics because they are widely represented in ocular infections. The antimicrobial susceptibility was determined by Kirby-Bauer antibiogram method. After reviving the microorganisms, which were frozen at −20°C, a suspension of the bacterial cultures with a turbidity of 0.5 McFarland was prepared. McFarland standards were used to prepare bacterial suspensions to a speciﬁed turbidity. In the Kirby-Bauer susceptibility test protocol, the bacterial suspension of the organism to be tested is equivalent to the 0.5 McFarland standard. MRSE, MRSA, and PA MDR cultures were transferred onto Mueller-Hinton agar (BD Diagnostics, Sparks, MD) plates using a spatula, while other Streptococcus spp. strains were transferred onto Mueller Hinton fastidious (BD Diagnostics, Sparks, MD) agar plates. Liposomal ozonated oil eye drops (Ozodrop, FB Vision, Italy), composed of ozonated liposomal sunﬂower oil (LipozonEye®, 10.5%) plus hypromellose (0.2%) and polyhexamethylene biguanide; sunﬂower ozonated oil, and hypromellose (hydroxypropyl methylcellulose), were deposited at the center of the plate at a ﬁnal volume of 75 μL. Zone of inhibition was observed after incubation at 37°C for 2, 4, 6, 8, and 24 h. Streptococcus spp. strains were incubated at 37°C in a 5% CO2-enriched atmosphere. Descriptive statistical analysis, including percentages to characterize data, was performed using Microsoft Excel To prevent postoperative infections, it is important to reduce the bacterial load on the ocular surface through the use of antibiotics. However, if this use is not selective and well managed, there is a risk for multidrug resistance development [23–25]. The latest Antibiotic Resistance Monitoring in Ocular Microorganisms (ARMOR) study, published in 2018, covers 4829 bacterial isolates from ocular infections, collected from 87 centers across 40 states, in the United States, from January 2009 to December 2016, including S. aureus, CoNS, S. pneumoniae, H. inﬂuenzae, and P. aeruginosa [26]. three diﬀerent forms: gaseous ozone, ozonated water, and ozonated oil [29–33]. Several studies have evaluated the eﬃcacy of ozonated oil formulations against microorgan- isms that are normally responsible for ocular infections [34–36]. The aim of this study was to evaluate the anti- bacterial activity of a liposomal ozonated oil solution for ophthalmic use against multidrug-resistant (MDR) bac- terial strains. condition in children [11–13]. S. aureus, CoNS, and Strep- tococcus spp. are the most important pathogens responsible for keratitis. Meanwhile, P. aeruginosa is predominant in contact lens wearers and in patients who have experienced ocular trauma [14–18]. Endophthalmitis, occurring after cataract surgery or after intravitreal injection, is most com- monly caused by CoNS, S. aureus, Bacillus spp., or P. aer- uginosa [19–21]. Although antibiotics are generally eﬀective in treating ocular infections, antibiotic resistance has become a wide- spread problem. This is largely due to the widespread and unwarranted use of broad-spectrum antibiotics in both systemic and topical infections, coupled with an inap- propriate duration of treatment [22]. 1 Introduction Like most parts of the body, the eyelids, conjunctiva, and corneal surface are colonized by resident bacteria. Imbalances in these microorganisms can trigger the onset of ocular infec- tions, such as conjunctivitis, keratitis, and endophthalmitis [1,2]. This can be often attributed to external factors, such as the use of contact lenses, trauma, or surgical operations, or to patient-related factors, such as age, ocular dryness, or chronic obstruction of the nasolacrimal duct [3,4]. The eye is protected by the body’s innate and adap- tive immunity; surface IgA and large amounts of antibac- terial compounds, such as lysozyme and lactoferrin, are present in tears. However, the presence of these antibac- terial compounds alone is not enough to prevent the development of infection, and subsequent inﬂammation and scarring do not heal very easily [5–8]. Severe eye infections require immediate management because they can damage the ocular tissues, leading to visual impair- ment and blindness. Although these conditions are mostly managed empirically, optimal treatment requires knowl- edge of the speciﬁc etiology of the condition. In western countries, the main causes of ocular infections are Gram- positive bacteria, including Staphylococcus aureus, coagu- lase-negative Staphylococci (CoNS), Streptococcus spp., and gram-negative bacteria, particularly Pseudomonas aeruginosa [9,10]. Conjunctivitis in adults is frequently caused by Staphylococcus spp., Streptococcus spp., P. aeruginosa, Klebsiella pneumoniae, and Pasteurellae spp., whereas Haemophilus inﬂuenzae, Streptococcus pneumoniae, and Moraxella spp. are more frequently responsible for the Romolo Protti: Ophthalmic Unit, Hospital San Biagio, Domodossola, 28845, Italy Giovanni Scapagnini, Ciro Costagliola: Department of Medicine and Health Sciences “V. Tiberio”, University of Molise, Via V. De Sanctis, s.n.c., Campobasso 86100, Italy Open Access. © 2022 Giuseppe Grandi et al., published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. 1058 1058  Giuseppe Grandi et al. 2 Methods This study demonstrated geographic variation in resis- tance rates among ocular isolates. Moreover, methicillin- resistant strains of both S. aureus and CoNS were highly resistant to ﬂuoroquinolones, aminoglycosides, and macro- lides. Strains of S. pneumoniae were found to be sensitive to all the antibiotics tested, and strains of P. aeruginosa exhib- ited low levels of resistance, particularly toward ciproﬂox- acin. These data highlight the need to re-evaluate and implement the guidelines for antibiotic treatment not only to prevent the emergence of drug resistance in pathogens but also to highlight the need to develop new drugs with high eﬃcacy, low toxicity, and low resistance potential [27]. g y, y, p [ ] Therefore, research has begun to focus on molecules with antimicrobial and antiseptic potentials that are struc- turally diﬀerent from conventional antibiotics. One such molecule is Ozone, which is a powerful oxidizing agent known for its antiseptic and anti-inﬂammatory properties. This molecule releases free oxygen radicals that facilitate the formation of hydrogen peroxide and lipid peroxidation products that are responsible for bacterial lysis and cell death. The introduction of ozonated agents has been advo- cated on the basis of eﬃcacy against all microorganisms, as well as the lack of induction of antibiotic resistance [28]. Over the last few years, several therapeutic protocols using ozone have been developed to treat dental and skin infec- tions. In these protocols, ozone is used in the following Antimicrobial activity of ozonated oil eye drops  1059 Figure 1: Culture of multidrug resistant Pseudomonas aeruginosa. (a) Bacterial culture after 6 h. An inhibition zone was observed at this time point. (b) Bacterial culture after 8 h. A recolonization in the area where the ozonated solution had been deposited was observed after 8 h. Figure 1: Culture of multidrug resistant Pseudomonas aeruginosa. (a) Bacterial culture after 6 h. An inhibition zone was observed at this time point. (b) Bacterial culture after 8 h. A recolonization in the area where the ozonated solution had been deposited was observed after 8 h. 2013 (Microsoft Corp.; Redmond, WA, USA). All experi- ments were performed in triplicate. ozonated oil droplet after 6 and 8 h. However, after 24 h, the bacterial regrowth was inhibited in 70% of the Streptococcus spp. strains. In the remaining 30% of strains (all S. agalactiae strains), the growth was inhib- ited at 6 and 8 h, but not at 24 h (Figure 2a and b). 2 Methods Table 1 presents the changes observed in the bacterial cultures on the agar plates after the addition of the ozonated oil. At 8 h, 94% of all the strains tested showed an inhibition zone. After 24 h, the growth of more than 10% of the isolates was inhibited in the area of contact with the ozonated oil. 3 Results The inhibition zones of all isolates of the PA MDR strain were observed after 6 h (Figure 1a). At 8 h, 20% of the samples exhibited recolonization in the region where the ozonated solution had been deposited (Figure 1b). The PA MDR strains completely recovered the growth at 24 h. At 6 and 8 h, inhibition zones of the MRSA and MRSE strains were observed in the area around the ozo- nated oil, but not at 24 h (Table 1), suggesting that the potency of the oil to inhibit may decrease over time. Inhi- bition zones were observed in 100% of the Streptococcus spp. strains in the area that was in contact with the 2013 (Microsoft Corp.; Redmond, WA, USA). All experi- ments were performed in triplicate. 4 Discussion Furthermore, ozone is a powerful immunomodulator that can boost immune activity in the body. It can also oxidize lipids to produce hydrogen peroxide, which is also a strong antimicrobial agent. Oxygenated agents, such as ozonated water, vege- table oil, and animal fats, have been used to improve the short half-life of ozone. There are several parameters that characterize the quality of an ozonated oil: the type of oil, ozonization process applied, iodine value, acid value, per- oxide value, viscosity, and absorption [39–41]. antibacterial, antiviral, and antifungal agents. This activity is a result of its high oxidation potential, which allows it to destroy the bacterial cell wall and cytoplasmic membrane, resulting in increased cell permeability, allowing the entry of ozone into the bacterial cell [37,38]. Furthermore, ozone is a powerful immunomodulator that can boost immune activity in the body. It can also oxidize lipids to produce hydrogen peroxide, which is also a strong antimicrobial agent. Oxygenated agents, such as ozonated water, vege- table oil, and animal fats, have been used to improve the short half-life of ozone. There are several parameters that characterize the quality of an ozonated oil: the type of oil, ozonization process applied, iodine value, acid value, per- oxide value, viscosity, and absorption [39–41]. Ozonated olive oil has been used in the treatment of bacterial skin infections since the 1930s. Even today, ozo- nated oils are widely used in dermatology and dentistry to treat infected lesions [42]. This study aimed to verify the action of ozonated oil for topical ophthalmic use against MDR bacteria. Another interesting aspect is the duration of bacterial growth inhibition by the ozonated oil, which was found to be 6 h for all species and strains, with regrowth in 20% of the PA MDR strains after 8 h and 30% of Streptococcus spp. after 24 h. These observations allow us to deﬁne an in vivo drug administration approach, both empiri- cally and when the pathogen is unknown. Therefore, the administration of ozonated oil every 6–8 h, as in the clinical practice, should completely inhibit even the most resistant bacterial species, whether endogenous to the ocular ﬂora like S. aureus and CoNS, or exogenous like P. aeruginosa. However, this hypothesis should be conﬁrmed in vivo. Song et al. evaluated the action of ozonated water and ozonated oil for topical dermatological use on methi- cillin-sensitive and methicillin-resistant strains of S. aureus. 4 Discussion Ozone is a gaseous, triatomic allotrope of oxygen that has long been recognized as one of the most potent Table 1: Growth changes in the bacterial cultures after treatment with the liposomal ozonated oil Microorganism 2–4 h incubation 6 h incubation 8 h incubation 24 h incubation Pseudomonas aeruginosa No visible inhibition Inhibition zones Inhibition zones but regrowth in 20% of strains Regrowth Staphylococcus aureus No visible inhibition Inhibition zones Inhibition zones Regrowth Staphylococcus epidermidis No visible inhibition Inhibition zones Inhibition zones Regrowth Streptococcus spp. No visible inhibition Inhibition zones Inhibition zones Regrowth in 30% of strains (all S. agalactiae) Table 1: Growth changes in the bacterial cultures after treatment with the liposomal ozonated oil 1060  Giuseppe Grandi et al. Figure 2: Inhibition zone of Streptococcus agalactiae strains after contact with the liposomal ozonated oil. (a) Inhibition zone after 6 h of contact with the liposomal ozonated oil. (b) Inhibition zone after 8 h of contact with the liposomal ozonated oil. Figure 2: Inhibition zone of Streptococcus agalactiae strains after contact with the liposomal ozonated oil. (a) Inhibition zone after 6 h of contact with the liposomal ozonated oil. (b) Inhibition zone after 8 h of contact with the liposomal ozonated oil. The ozonated solution exhibited antibacterial activity against all the tested species. The activity was lower against the PA MDR strains than against the other species and was more evident in Staphylococcus and Streptococcus spp. We believe that it was important to analyze multiple strains of each species, since it allowed us to highlight diﬀerences in the behavior of the solution, not only toward the diﬀerent genera and species but also toward individual strains, given that after 8 h, not all strains of PA MDR recolonized the zone of inhibition, and after 24 h, not all Streptococcus species were inhibited (only S. agalactiae, a rare pathogen in eye infection). This ﬁnding is particularly signiﬁcant because the mechanism of antimicrobial action of ozonated oil appears to be nonspeciﬁc and nonselective. Therefore, in theory, we would not expect any diﬀerences in activity against the bacterial strains. Further studies are necessary to investigate and explain these observations. antibacterial, antiviral, and antifungal agents. This activity is a result of its high oxidation potential, which allows it to destroy the bacterial cell wall and cytoplasmic membrane, resulting in increased cell permeability, allowing the entry of ozone into the bacterial cell [37,38]. Antimicrobial activity of ozonated oil eye drops 1061 liposomes has already been demonstrated in an animal model. However, this study was limited to coagulase-posi- tive and coagulase-negative staphylococci (34). A more in-depth study on the bactericidal eﬀect of ozonated oil encapsulated in liposomes will have to be carried out to evaluate its full potential, including the evaluation of its oxidative eﬀect on the ocular surface, which is particularly delicate and unique. The data from this study cannot be compared with those from other studies reported in the lit- erature, since the only other study that evaluated the eﬃcacy of an ozonated product against MDR bacteria applied a dif- ferent methodology. Moreover, a previous study assessed an oil for skin application, as opposed to an ophthalmic pro- duct. The ionization characteristics of the two ozonated lipids were also not known for comparison. From the data collected in this study, it can be con- cluded that the ozonated oil has an inhibitory and there- fore bacteriostatic eﬀect on MDR bacteria. However, it is necessary to assess if it also possesses bactericidal prop- erties. Re-colonization may be a result of the bacterial growth outside the zone of inhibition, while those origin- ally present in the zone are killed (bactericidal eﬀect). Alternatively, only the growth of bacteria present in the zone of inhibition may be inhibited (bacteriostatic eﬀect). Bactericidal properties could make the action of the pro- duct even more eﬀective. With the exception of Streptococcus spp., all MDR strains that were selected were very diﬃcult to treat with antibiotics. It is clear from the literature that the incorrect and unwarranted use of antibiotic eye drops has resulted in the emergence of MDR strains, meaning that even infections that appear to be trivial can become diﬃcult to treat. According to the principles of evidence- based medicine, it is type A evidence that in the majority of cases of conjunctivitis (suspected or conﬁrmed), without other ocular and/or systemic comorbidities, antibiotic treat- ment is not necessary since the condition is self-limiting and generally resolves within 7–10 days. However, it is also type A evidence that the use of antibiotics can reduce the duration of the disease, allowing faster reintegration into work or school. Therefore, the administration of broad- spectrum antibiotics to treat bacterial conjunctivitis may be considered reasonable. A key limitation of our study is largely related to its qualitative nature, thus limiting the applicability in clin- ical practice. Another clear limitation of our study is that we performed an in vitro experiment, which may not reﬂect exactly the real situation in vivo. We believe that in-depth research into the character- istics of the oil and its ozonation may be the key to opti- mizing and improving antimicrobial therapy. The use of ozonated oil in ophthalmology, such as in preoperative prophylaxis, can be useful, as antibiotics are now likely to fail or lead to the development of new drug resistance. This approach to therapy can broaden the spectrum of action and prevent the emergence of drug-resistant strains. The possibility of providing clinicians with a new tool, as an alternative to antibiotics, can be fundamental in bacterial endophthalmitis prophylaxis for cataract surgery and intra- vitreal treatment and in improving therapeutic strategies, broadening the antibiotic spectrum, and avoiding viral and fungal superinfections. There is a signiﬁcant controversy around empirical topical therapy, and references are often made to the local pattern of resistance, cost, dosage, and other patient-related factors, such as allergies and compliance [45]. Therefore, it is possible to understand how the inappropriate use of anti- biotics in a self-limiting and widespread pathology can potentially cause the emergence of bacterial resistance. Therefore, antiseptics can be extremely valuable, providing at least partial protection against the development of resis- tance, while also shortening the course of the disease and allowing patients to return to their normal activities. 4 Discussion They demonstrated a 100% reduction in the bacterial load of both isolates after 15 min [43]. Also, Zanardi et al. used ozonated oil preparations to eliminate cutaneous infections. These authors reported a potent bactericidal eﬀect of ozo- nated oils with a reduction of cutaneous infection exu- dates [44]. The present study is the ﬁrst to test the eﬀect of ozo- nated oil encapsulated in liposomes against MDR bacteria (MRSA, MRSE, and PA MDR) and Streptococcus spp., which are implicated in both superﬁcial and deep eye diseases. Interestingly, liposomal ozonated oil has long-lasting eﬀects on Streptococcus spp., which are responsible for a large proportion of ocular infections. References [16] Ting DSJ, Settle C, Morgan SJ, Baylis O, Ghosh S. A 10 years analysis of microbiological proﬁles of microbial keratitis: the north east England study. Eye. 2018;32(8):1416–7. doi: 10.1038/s41433-018-0085-4. [1] Wilcox HD. Characterization of the normal microbiota of the ocular surface. Exp Eye Res. 2013;117:99–105. doi: 10.1016/ j.exer.2013.06.003. [17] Green M, Apel A, Stapleton F. Risk factors and causative organisms in microbial keratitis. Cornea. 2008;27(1):22–7. doi: 10.1097/ICO.0b013e318156caf2. [2] Dong Q, Brulc JM, Iovieno A, Bates B, Garoutte A, Miller D, et al. Diversity of bacteria healthy at humans conjunctiva. Invest Ophthalmol Vis Sci. 2011;52(8):5408–13. doi: 10.1167/ iovs.10-6939. [18] Pachigolla G, Blomquist C, Cavangh HD. Microbial keratitis pathogens and antibiotic susceptibilities: a 5 years review of cases at an urban country hospital in north texas. Eye Contact lens. 2007;33(1):45–9. doi: 10.1097/ 01.icl.0000234002.88643.d0. [3] Shin H, Price K, Albert L, Dodick J, Park L, Dominguez- Bello MG. Changes in the eye microbiota associated with contact lens wearing. mBIO. 2016;7(2):e00198–216. doi: 10.1128/mBio.00198-16. [19] Attisano C, Cibinel M, Strani G, Panepinto G, Pollino C, Furfaro G. Severe ocular bacterial infections: a retrospective study over 13 years. Ocul Immunol Inﬂamm. 2017;25(6):826–9. doi: 10.1080/09273948.2016.1181767. [4] Ozkan J, Nielsen S, Diez-Vives C, Coroneo M, Thomas T, Willcox M. Temporal stability and composition of the ocular surface microbiome. Sci Rep. 2017;7(1):9880. doi: 10.1038/ s41598-017-10494-9. [20] Moconey TP, Park J. Microbiological isolates and antibiotic sensitivities in culture- proven endophthalmitis: a 15 years review. Br J Ophthalmol. 2014;98(11):1492–7. doi: 10.1136/ bjophthalmol-2014-305030. [5] Foulsham W, Coco G, Amouzegar A, Chauhan SK, Dana R. When clarity is crucial: regulating surface immunity. Trends Immunol. 2018;39(4):288–301. doi: 10.1016/j.it.2017.11.007. [6] Allansmith M. Immunology of tears. Int Ophthalmol Clin. 1973;13(1):47–72. doi: 10.1097/00004397-197301310-00006. [21] Durand ML. Endophthalmitis. Clin Microbiol Infec. 2013;19(3):227–34. doi: 10.1111/1469-0691.12118. [7] Dor M, Eperon S, Lalive PH, Guex-Crosier Y, Hamedani M, Salvisberg C, et al. Investigation of the global protein content from human tears. Exp Eye Res. 2019;179:64–74. doi: 10.1016/ j.exer.2018.10.006. [22] Masterton RG. The new treatment paradigm and the role of carbapenems. Int J Antimicrob Agents. 2009;33(2):105–10. doi: 10.1016/j.ijantimicag.2008.07.023. [23] Asbell PA, Colby KA, Deng S, McDonnell P, Meisler DM, Raizman MB. Ocular TRUST: nationwide antimicrobial sus- ceptibility patterns in ocular isolates. Am J Ophthalmol. 2008;145(6):951–8. doi: 10.1016/j.ajo.2008.01.025. [8] Kijstra A, Kuizega A, van Der Velde M, Haeringen NJ. Gel electrophoresis of human tears reveals various forms of tear lactoferrin. Curr Eye Res. 1989;8(6):581–8. doi: 10.3109/ 02713688908995757. References [24] Asbel PA, Sham DF, Shaw M, Draghi DC, Brown NP. Increasing prevalence of methicillin resistance in serious ocular infec- tions caused by staphylococcus aureus in the united states: 2000 to 2005. J Cataract Refract Surg. 2008;34(5):814–8. doi: 10.1016/j.jcrs.2008.01.016. [9] Grandi G, Bianco G, Boattini M, Scalabrin S, Iannaccone M, Fea A, et al. Bacterial etiology and antimicrobial resistance trends in ocular infections: a 30 –year study, Turin area, Italy. Eur J Ophthalmol. 2021;31(2):405–14. doi: 10.1177/ 1120672119896419. [25] Bandello F, Coassin M, Di Zazzo A, Rizzo S, Biagini I, Pozdeyeva N, et al. One week of levoﬂoxacin plus dexa- methasone eye drops for cataract surgery: an innovative and rational therapeutic strategy. Eye. 2020;34(11):2112–22. doi: 10.1038/s41433-020-0869-1. [10] Teweldemedhin M, Gebreyesus H, Atsbaha AH, Asgedom SW, Saravanan M. Bacterial proﬁle of ocular infections: a sys- tematic review. BMC Ophthalmol. 2017;17(1):212. doi: 10.1186/ s12886-017-0612-2. [11] Giardini F, Grandi G, De Sanctis U, Eandi C, Machetta F, Pollino C, et al. In vitro susceptibility to diﬀerent topical anti- biotics of bacterial isolates from patients with conjunctivitis. Ocul Immunol Inﬂamm. 2011;19(6):419–21. doi: 10.3109/ 09273948.2011.624289. [26] Asbel PA, Pandit RT, Sanﬁlippo CM. Antibiotic resistence rates by geographic region among ocular pathogens collected during the ARMOR surveillance study. Ophthalmol Ther. 2018;7(2):417–29. doi: 10.1007/s40123-018-0141-y. [27] Bertino JS. Impact of antibiotic resistence in the management of ocular infections: the role of current and future antibiotics. Cilin Ophthalmol. 2009;3:507–21. doi: 10.2147/opth.s5778. [12] Azari AA, Barney NP. Conjunctivitis a systemic review of diagnosis and treatment. JAMA. 2013;310(16):1721–9. doi: 10.1001/jama.2013.280318. [28] Tognetto D, Pastore MR, Guerin GM, Decorti G, Franzin M, Lagatolla C, et al. Bactericidal activity of three diﬀerent anti- septic ophthalmic preparations as surgical prophylaxis. Graefes Arch Clin Exp Ophthalmol. 2022;260(1):289–93. doi: 10.1007/s00417-021-05361-3. [13] Bremon-Gignac D, Chiambaretta F, Milazzo S. European per- spective on topical ophthalmic antibiotics:current and evol- ving options. Ophthalmol Eye Dis. 2011;3:29–43. doi: 10.4137/ OED.S4866. [14] Peng MY, Cevallos V, Mc Leod SD, Lietman TM, Rose- Nussbaumer J. Bacterial keratitis: isolate organisms and antibiotic resistence patterns in San Francisco. Cornea. 2018;37(1):84–7. doi: 10.1097/ICO.0000000000001417. [29] Gupta G, Mansi B. Ozone therapy in periodontics. J Med Life. 2012;5(1):59–67. [30] Pietrocola G, Ceci M, Preda F, Poggio C, Colombo M. Evaluation of the antibacterial activity of a new ozonated olive oil against oral and peridontal pathogens. J Clin Exp Dent. 2018;10(11):e1103–8. doi: 10.4317/jced.54929. [15] Sand D, She R, Shulman IA, Schur M, Hsu HY. Giuseppe Grandi et al. Funding information: No funding was received. The increase in bacterial resistance, including bac- teria that cause ocular pathologies, has been demon- strated by the “Ocular Tracking Resistance in U.S. Today” Study (TRUST) [23] and the “Antibiotic Resistance Among Ocular Pathogens in the United States” study (ARMOR) [24,26]. This becomes an even more serious problem in the case of infections of the anterior and posterior segments of the eye and in preoperative prophylaxis. Therefore, the search for new antibacterial compounds capable of preventing the onset of resistance has become increas- ingly important. The possibility of using ozonated oil in Author contributions: G.G., R.C., and E.Z.: collected the samples and writing original draft; R.C., C.P., and R.P.: car- ried out the experiments and editing; G.S, S.D., and C.C.: designed the concept of study and revised the manuscript. Conﬂict of interest: Sergio Davinelli become Editor in Open Medicine from March, but this fact hasnt impacted the peer-review process. Data availability statement: The data are available from the corresponding author upon reasonable request. 1062 1062  Giuseppe Grandi et al. References Microbial ker- atitis in los angeles country hospital institute and the los angeles country hospital experience. Ophthalmology. 2015;122(5):918–24. doi: 10.1016/j.ophtha.2014.11.027. [31] Martinez SG, Al-Dalain SM, Menendez S, Re L, Giuliani A, Candelario-Jalil E. Therapeutic eﬃcacy of ozone in patients 1063 Antimicrobial activity of ozonated oil eye drops with diabetic food. Eur J Pharm. 2005;523(1–3):151–61. doi: 10.1016/j.ejphar.2005.08.020. and permeability of oral microorganisms. Oral Microbiol Immunol. 2004;19(4):240–6. doi: 10.1111/j.1399- 302X.2004.00146.x. [32] Fizpatrick E, Holland OJ, Vanderleilie JJ. Ozone therapy for the treatment of cronin wounds: A systematic review. Int Wound J. 2018;15(4):633–44. doi: 10.1111/iwj.12907. [39] Moureau S, Violleau F, Ali Haimuod-Lekhal D, Calmon A. Ozonation of sunﬂowers oils: Impact of experimental condi- tions on the composition and the antibacterial activity of ozonated oils. Chem Phys Lipids. 2015;186:79–85. doi: 10.1016/j.chemphyslip.2015.01.004. [33] Serio F, Alba M, Cozzolino G, Idolo A, Grassi T, Bagordo F, et al. Eﬃcacy of a dermatological gel based on ozonated sunﬂower seed oil (Oz.Or.Oil 30) on bedsores: A pilot study. Clin Derm Res J. 2017;2:2. [40] Carata E, Tenuzzo BA, Dini L. Powerful properties of ozonated extra virgin olive oil. In: Builders PF, editor. Herbal medicine. London: IntechOpen; 2018. #https://www.intechopen.com/ chapters/58960.doi: 10.5772/intechopen.73211. [34] Marchegiani A, Magagnini M, Cerquetella M, Troiano P, Franchini I, Franchini A. Preoperative topical liposomal ozone dispersion to reduce bacterial colonization in conjunctival sac and periocular skin: Preliminary study in dogs. Exp Eye Res. 2019;189:107848. doi: 10.1016/j.exer.2019.107848. [41] Zeng J, Lu J. Mechanisms of action involved in ozone-therapy in skin diseases. Int Immunopharmacol. 2018;56:235–41. doi: 10.1016/j.intimp.2018.01.040. [35] Celenza G, Iorio R, Cracchiolo S, Petricca S, Costagliola C, Cinque B. Antimycotic activity of ozonated oil in liposome eye drops against candida spp. Transl Vis Sci Technol. 2020;9(8):4. doi: 10.1167/tvst.9.8.4. [42] Harada T. Olive oil ozonide and its fungicidal quality. Bull Chem Soc Jpn. 1934;9:192–7. [43] Song M, Zeng Q, Xiang Y, Gao L, Huang J, Huang J. The anti- bacterial eﬀect of topical ozone on the treatment of MRSA skin infection. Mol Med Rep. 2017;17(2):2449–55. doi: 10.3892/ mmr.2017.8148. [36] Spadea L, Tonti E, Spaterna A, Marchegiani A. Use of ozone- based eye drops: A series of cases in veterinary and human spontaneous ocular pathologies. Case Rep Ophthalmol. 2018;9(2):287–98. doi: 10.1159/000488846. [44] Zanardi I, Burgassi S, Paccagnini E, Gentile M, Bocci V, Travagli V. What is the best strategy for enhancing the eﬀects of topically applied ozonated oils in cutaneous infections? Biomed Res Int. 2013;2013:702949. doi: 10.1155/2013/ 702949. References [37] Ugazio S, Vivian T, Binello A, Tagliapietra S, Dosio F. Ozonated oils as antimicrobial system in topical applications. Their characterisation, current applications, and advances in improved delivery tachniques. Molecules. 2020;25(2):334. doi: 10.3390/molecules25020334. [38] Nagayoshi M, Fukuizumi T, Kitamura M, Nagayoshi M, Fukuizumi T, Kitamura M. Eﬃcacy of ozone on survival [37] Ugazio S, Vivian T, Binello A, Tagliapietra S, Dosio F. Ozonated oils as antimicrobial system in topical applications. Their characterisation, current applications, and advances in improved delivery tachniques. Molecules. 2020;25(2):334. doi: 10.3390/molecules25020334. [45] Azari AA, Barney NP. Conjunctivitis a systematic review of diagnosis and treatment. JAMA. 2013;310(16):1721–9. doi: 10.1001/jama.2013.280318. [38] Nagayoshi M, Fukuizumi T, Kitamura M, Nagayoshi M, Fukuizumi T, Kitamura M. Eﬃcacy of ozone on survival
https://openalex.org/W2290260868	https://findresearcher.sdu.dk/ws/files/129761742/Novel_loci_and_pathways_significantly_associated_with_longevity.pdf	English	null	Novel loci and pathways significantly associated with longevity	Scientific reports	2,016	cc-by	14,508	University of Southern Denmark Novel loci and pathways significantly associated with longevity Novel loci and pathways significantly associated with longevity Zeng, Yi; Nie, Chao; Min, Junxia; Liu, Xiaomin; Li, Mengmeng; Chen, Huashuai; Xu, Hanshi; Wang, Mingbang; Ni, Ting; Li, Yang; Yan, Han; Zhang, Jin-Pei; Song, Chun; Chi, Li-Qing; Wang, Han-Ming; Dong, Jie; Zheng, Gu-Yan; Lin, Li; Qian, Feng; Qi, Yanwei; Liu, Xiao; Cao, Hongzhi; Wang, Yinghao; Zhang, Lijuan; Li, Zhaochun; Zhou, Yufeng; Wang, Yan; Lu, Jiehua; Li, Jianxin; Qi, Ming; Bolund, Lars; Yashin, Anatoliy; Land, Kenneth C; Gregory, Simon; Yang, Ze; Gottschalk, William; Tao, Wei; Wang, Jian; Wang, Jun; Xu, Xun; Bae, Harold; Nygaard, Marianne; Christiansen, Lene; Christensen, Kaare; Franceschi, Claudio; Lutz, Michael W; Gu, Jun; Tan, Qihua; Perls, Thomas; Sebastiani, Paola; Deelen, Joris; Slagboom, Eline; Hauser, Elizabeth; Xu, Huji; Tian, Xiao-Li; Yang, Huanming; Vaupel, James W Published in: Citation for pulished version (APA): Zeng, Y., Nie, C., Min, J., Liu, X., Li, M., Chen, H., Xu, H., Wang, M., Ni, T., Li, Y., Yan, H., Zhang, J.-P., Song, C., Chi, L.-Q., Wang, H.-M., Dong, J., Zheng, G.-Y., Lin, L., Qian, F., ... Vaupel, J. W. (2016). Novel loci and pathways significantly associated with longevity. Scientific Reports, 6, Article 21243. https://doi.org/10.1038/srep21243 Go to publication entry in University of Southern Denmark's Research Portal Terms of use This work is brought to you by the University of Southern Denmark. Unless otherwise specified it has been shared according to the terms for self-archiving. If no other license is stated, these terms apply: Terms of use This work is brought to you by the University of Southern Denmark. Unless otherwise specified it has been shared according to the terms for self-archiving. If no other license is stated, these terms apply: This work is brought to you by the University of Southern Denmark. Unless otherwise specified it has been shared according to the terms for self-archiving. If no other license is stated these terms apply: This work is brought to you by the University of Southern Denmark. Unless otherwise specified it has been shared according to the terms for self-archiving. • You may download this work for personal use only. Download date: 24. Oct. 2024 Novel loci and pathways significantly associated with longevity received: 25 August 2015 accepted: 20 January 2016 Published: 25 February 2016 received: 25 August 2015 accepted: 20 January 2016 Published: 25 February 2016 Yi Zeng1,2,,†, Chao Nie3,, Junxia Min4,, Xiaomin Liu3,, Mengmeng Li5, Huashuai Chen1,6, Hanshi Xu3, Mingbang Wang3, Ting Ni7, Yang Li8, Han Yan8, Jin-Pei Zhang8, Chun Song8, Li- Qing Chi8, Han-Ming Wang8, Jie Dong8, Gu-Yan Zheng8, Li Lin5, Feng Qian5, Yanwei Qi3,9, Xiao Liu3, Hongzhi Cao3, Yinghao Wang3, Lijuan Zhang3, Zhaochun Li3, Yufeng Zhou3, Yan Wang3, Jiehua Lu10, Jianxin Li10, Ming Qi4, Lars Bolund3,11, Anatoliy Yashin12, Kenneth C. Land12, Simon Gregory13, Ze Yang14, William Gottschalk15, Wei Tao16, Jian Wang3,17, Jun Wang3,18, Xun Xu3, Harold Bae19, Marianne Nygaard20, Lene Christiansen20, Kaare Christensen20, Claudio Franceschi21, Michael W. Lutz15, Jun Gu16, Qihua Tan20, Thomas Perls22, Paola Sebastiani23, Joris Deelen24, Eline Slagboom24, Elizabeth Hauser13, Huji Xu5, Xiao-Li Tian8,†, Huanming Yang3,17,† & James W. Vaupel25 Only two genome-wide significant loci associated with longevity have been identified so far, probably because of insufficient sample sizes of centenarians, whose genomes may harbor genetic variants associated with health and longevity. Here we report a genome-wide association study (GWAS) of Han Chinese with a sample size 2.7 times the largest previously published GWAS on centenarians. We identified 11 independent loci associated with longevity replicated in Southern-Northern regions of China, including two novel loci (rs2069837-IL6; rs2440012-ANKRD20A9P) with genome- wide significance and the rest with suggestive significance (P < 3.65 × 10−5). Eight independent SNPs overlapped across Han Chinese, European and U.S. populations, and APOE and 5q33.3 were replicated as longevity loci. Integrated analysis indicates four pathways (starch, sucrose and xenobiotic 1Center for the Study of Aging and Human Development and Geriatrics Division, Medical School of Duke University, Durham, North Carolina, USA. 2Center for Healthy Aging and Development Studies, National School of Development, Peking University, Beijing, China. 3BGI-Shenzhen, Shenzhen, China. 4The First Affiliated Hospital, Institute of Translational Medicine, School of Medicine, Zhejiang University, Hangzhou, China. 5Department of Rheumatology and Immunology, Shanghai Changzheng Hospital, The Second Military Medical University, Shanghai, China. 6Business School of Xiangtan University, Xiangtan, China. 7State Key Laboratory of Genetics Engineering & MOE Key Laboratory of Contemporary Anthropology, School of Life Sciences, Fudan University, Shanghai, China. 8Department of Human Population Genetics, Institute of Molecular Medicine, Peking University, Beijing, China. 9Department of Bio-Medical Engineering, School of Life Sciences, Anhui Medical University, Hefei, China. 10Department of Sociology, Peking University, Beijing, China. 11Department of Biomedicine, Aarhus University, Aarhus, Denmark. Novel loci and pathways significantly associated with longevity • You may download this work for personal use only y p y • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying this open access version y p y • You may not further distribute the material or use it for any profit-making activity or commercial gain Y f l di t ib t th URL id tif i thi i y p y • You may not further distribute the material or use it for any profit-makin y p y You may not further distribute the material or use it fo You may not further distribute the material or use it for any profit making activity or commercial gain • You may freely distribute the URL identifying this open access version y y p g • You may freely distribute the URL identifying this open access version If you believe that this document breaches copyright please contact us providing details and we will investigate your claim. Please direct all enquiries to puresupport@bib.sdu.dk If you believe that this document breaches copyright please contact us providing details and we will investigate your claim. Please direct all enquiries to puresupport@bib.sdu.dk Download date: 24. Oct. 2024 www.nature.com/scientificreports www.nature.com/scientificreports received: 25 August 2015 accepted: 20 January 2016 Published: 25 February 2016 Novel loci and pathways significantly associated with longevity Human longevity is a complex trait determined by both genetic and environmental factors, and the genetic influ- ence increases with the higher ages1–4. Prior studies showed that most centenarians compress disability period towards the ends of their lives, suggesting that they avoid diseases of aging and associated disability until old ages5–7. Furthermore, the offspring of centenarians have significantly better health compared to peers8,9. Hence, centenarian genomes may harbor genetic variants associated with health and longevity10. In recent years, a num- ber of genome-wide association studies (GWAS) in North America and Europe have been conducted on lon- gevity, but only two loci have been identified associated with longevity at a genome-wide significant level: the well-known TOMM40/APOE/APOC1 locus which is negatively associated with longevity4,10,11, and a second locus on chromosome 5q33.3, which was identified in a recent GWAS by Deelen et al.10.i qi y While interest in this topic is strong, the field is hindered by a lack of databases with both genotypic and phenotypic information and sufficiently large sample sizes of centenarians4,11. The largest sample size of cente- narians in GWAS studies published to date was 801, in which the study found only one genome-wide significant single nucleotide polymorphism (SNP)4. Much larger samples of centenarians and ethnically matched controls are needed for genome wide significant discoveries of genetic associations with longevity3,11. g gi g g y To expand the catalogue of longevity-associated loci and gain a better understanding of the influences of genes and biological pathways on longevity, we performed a GWAS using the samples derived from the Chinese Longitudinal Healthy Longevity Survey (CLHLS) (section M1 of Methods). The present study involved 2,178 Han Chinese centenarians, which is about 2.7 times the largest sample size of ever-published GWAS on centenarians4, and 2,299 middle-age controls after sample filtering, implying reasonably good power in our study (section M2 of Methods).i Our work also represents the first GWAS of longevity in Asian populations and in developing countries more generally. There were about five centenarians per million in China in the 1990s, compared with 50 per million in Western Europe in the same period12. Han Chinese centenarians may be more likely to have longevity-associated and/or disease-preventive genes than their centenarian counterparts in the West since they survived the brutal mortality regimes of the past when famine, wars, and starvation operated on birth cohorts of many millions. Novel loci and pathways significantly associated with longevity 12Population Research Institute, Duke University, Durham, North Carolina, USA. 13Duke Molecular Physiology Institute, Medical Center, Duke University, Durham, North Carolina, USA. 14National Institutes of Geriatrics, Beijing Hospital, Ministry of Health, Beijing, China. 15Department of Neurology, Medical Center, Duke University, Durham, North Carolina, USA. 16School of Life Sciences, Peking University, Beijing, China. 17James D. Watson Institute of Genome Sciences, Hangzhou, China. 18Department of Biology, University of Copenhagen, Copenhagen, Denmark. 19College of Public Health and Human Sciences, Oregon State University, USA. 20The Danish Aging Research Centre, Unit of Epidemiology, Biostatistics and Biodemography, Institute of Public Health, University of Southern Denmark, Odense, Denmark. 21Department of Experimental, Diagnostic and Specialty Medicine and Interdepartmental Centre ‘L. Galvani’, University of Bologna, Bologna, Italy. 22Geriatrics Section, Department of Medicine, Boston University School of Medicine, Boston, MA, USA. 23Department of Biostatistics, Boston University School of Public Health, Boston, MA, USA. 24Department of Molecular Epidemiology and Netherlands Consortium for Healthy Ageing, Leiden University Medical Center, Leiden, The Netherlands. 25Max Planck Institute for Demographic Research, Rostock, Germany. *These authors contributed equally to this work. †These authors jointly supervised this work. Correspondence and requests for materials should be addressed to Y.Z. (email: zengyi@duke. edu) or X.T. (email: tianxiaoli@pku.edu.cn) or H.Y. (email: yanghm@genomics.org.cn) Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 1 www.nature.com/scientificreports/ metabolism; immune response and inflammation; MAPK; calcium signaling) highly associated with longevity (P ≤ 0.006) in Han Chinese. The association with longevity of three of these four pathways (MAPK; immunity; calcium signaling) is supported by findings in other human cohorts. Our novel finding on the association of starch, sucrose and xenobiotic metabolism pathway with longevity is consistent with the previous results from Drosophilia. This study suggests protective mechanisms including immunity and nutrient metabolism and their interactions with environmental stress play key roles in human longevity. metabolism; immune response and inflammation; MAPK; calcium signaling) highly associated with longevity (P ≤ 0.006) in Han Chinese. The association with longevity of three of these four pathways (MAPK; immunity; calcium signaling) is supported by findings in other human cohorts. Our novel finding on the association of starch, sucrose and xenobiotic metabolism pathway with longevity is consistent with the previous results from Drosophilia. This study suggests protective mechanisms including immunity and nutrient metabolism and their interactions with environmental stress play key roles in human longevity. Novel loci and pathways significantly associated with longevity Unlike Western countries that received many international immigrants from other parts of the world resulting in relatively heterogeneous genetic compositions even within ethnic groups, China received very few international immigrants. Consequently, even though one might expect population substructures because of the long history of interaction with surrounding minority Chinese ethnic groups13, Han Chinese are relatively more homogenous in genetic composition compared to their Western counterparts. For example, it has been estimated that the average of genetic differences measured by F-statistics (FST)14 between Han Chinese population samples (FST = 0.002) is much lower than that among European populations (FST = 0.009)13. Thus, GWAS with large samples of Han Chinese centenarians and middle-age controls are expected to be instrumental for identifying genetic variants related to longevity. Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 Results P l ti We then conducted an evaluation/comparison analysis of our results with the data from longevity GWAS of European Union (EU) longevity genetics consortium and the U.S. New England centenarians study (NECS). To extend the foregoing SNP association analyses towards understanding of biological processes underlying longevity, we conducted pathway and network analyses. A flow chart of the consecutive analysis steps is depicted in Fig. 1. Association analyses of SNPs with longevity in Han Chinese. In the bi-directional discovery-evaluation analysis, we identified 11 independent loci (Table 1 and Fig. 2) that were replicated in the independent GWAS datasets of Southern and Northern regions of China. All 11 loci were associated with lon- gevity with p < 3.65 × 10−5 in the combined dataset, of which two loci, rs2069837 (chromosome 7p15.3, IL6, P = 1.80 × 10−9) and rs2440012 (chromosome 13q12.12, ANKRD20A9P, P = 3.73 × 10−8) reached genome-wide significance (P < 5 × 10−8) (Table 1 and Fig. 2). IL6 has been linked to longevity previously18,19, but rs2069837 identified in this study is a novel signal within the IL6 locus. The minor allele of rs2069837 is significantly less frequent among centenarians than middle-age individuals in Han Chinese (odds ratio = 0.61; P = 1.80 × 10−9), suggesting the effect of this locus on longevity is deleterious. This outcome is consistent with published findings that the IL6 gene functions as an inflammatory biomarker of functional decline and poor health outcomes includ- ing increased mortality risk20,21. The other novel SNP rs2440012 is located in ANKRD20A9P, a pseudogene that is affiliated with the long non-coding RNAs (lncRNA) class. The biological function of this particular non-coding transcript variant remains to be characterized.i p Among the other 9 replicated loci associated with longevity at a suggestive significant level (P < 3.65 × 10−5), the TOMM40/APOE/APOC1 locus is particularly interesting since its relationship with longevity is well known4,10,11, and we discuss it in more detail below. The remaining 8 novel replicated loci include MIR3156-3 (rs145672791, 21q11.2, 28 kb downstream), AKR1C2 (rs61856137, 10p15.1, 27 kb upstream), FAM13A (rs2704588, 4q22.1, intronic), BEND4 (rs1487614, 4p13, 114 kb upstream), EPHA6 (rs10934524, 3q11.2, 383 kb upstream), ASIC2 (rs11658235 and rs7212444, 17q12, intronic) and OLFM4 (rs9568833, 13q14.3, 200 kb downstream).h The meta-analysis results (last column in Table 1) are in full agreement with those of the combined analysis adjusted by a binary covariate of Southern and Northern regions, except the P values are slightly higher. Results P l ti Population characteristics and analytical strategy. Standard quality controls are detailed in section M2 of Methods. Based on existing literature on Chinese genetic studies15 and our principal component analysis (PCA) (section M4 of Methods, Supplementary Figs 1 and 2), we stratified the samples into two independent GWAS datasets of Southern and Northern regions, with 1,063/887 centenarians/controls in the Southern dataset, and 1,115/1,412 centenarians/controls in the Northern dataset (Supplementary Table 1). The genomic inflation factors (λ ) in the Southern, Northern and combined datasets were 1.022, 1.010 and 1.022, respectively, indicat- ing that the effects of population stratification on genetic analysis are well controlled (Supplementary Fig. 3)16. To minimize both false-positive and false-negative rates, we applied a novel bi-directional discovery-evaluation strategy17, which fully uses the two independent GWAS datasets of Southern and Northern regions by parallel analysis: one as the discovery dataset and the other as the evaluation dataset, and vice versa. As described in the section M5 of Methods, when analyzing two independent GWAS datasets, the classic uni-directional discov- ery-replication approach of assigning one GWAS dataset as discovery and another GWAS dataset as replication would result in a higher false-negative rate missing a substantial number of replicated SNPs that have a p-value higher than the threshold and lower than the nominal significance level in the discovery GWAS dataset but that reach the threshold significance level in the replication GWAS dataset. Our bi-directional discovery-evaluation strategy enables us to avoid the higher false-negative rate.i gy g g We defined a priori discovery threshold of P < 10−4 as the discovery threshold p-value based on recent rel- evant literature as cited and discussed in section M5 of Methods. We used a nominal significance of P < 0.05 in the evaluation stage. We performed GWAS using logistic regression implemented in PLINK (version 1.06) and adjusted for the top two eigenvectors and sex (section M6 of Methods). Using the SNPs identified in the bi-directional discovery-evaluation analyses, we performed a combined analysis comparing Han Chinese 2,178 centenarians with 2,299 mid-age controls, adjusted for the geographic stratification of Southern and Northern regions in addition to sex. In addition, we performed a meta-analysis, treating the two independent datasets from Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 2 www.nature.com/scientificreports/ Figure 1. A flow chart of the consecutive analysis steps. Figure 1. A flow chart of the consecutive analysis steps. Southern and Northern regions of China as two groups. Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 Results P l ti The 11 independent loci associated with longevity at p < 10−4 in the discovery and at least a nominal significance (p < 0.05) in the evaluation, using the independent GWAS datasets of Southern and Northern regions of China as discovery/evaluation. Note: The SNPs rs57681851 and rs7213812 were imputed, while the other 9 SNPs listed in this Table were genotyped. MAF; minor allele frequency. Table 1. The 11 independent loci associated with longevity at p < 10−4 in the discovery and at least a nominal significance (p < 0.05) in the evaluation, using the independent GWAS datasets of Southern and Northern regions of China as discovery/evaluation. Note: The SNPs rs57681851 and rs7213812 were imputed, while the other 9 SNPs listed in this Table were genotyped. MAF; minor allele frequency. Northern regions of China as discovery/evaluation. Note: The SNPs rs57681851 and rs7213812 were imputed, while the other 9 SNPs listed in this Table were genotyped. MAF; minor allele frequency. Figure 2. Manhattan plot showing the results of the association with longevity in the combined GWAS dataset. Figure 2. Manhattan plot showing the results of the association with longevity in the combined GWAS dataset. a p-value of p < 10−3 in women. Ten of the 11 SNPs achieved nominal significance (P < 0.05) in men and one SNP had a P = 0.0549 in men (Supplementary Table 1). The cross-gender similarities in effect direction/size on longevity of these 11 SNPs, which are replicated between discovery and evaluation datasets, is consistent with the results of the other GWAS on longevity10,4, and the gender differences of the not-replicated SNPs is out of scope of this article and will be investigated in our subsequent study. Comparison of current results with longevity GWAS studies of different ethnicities. To investi- gate the similarities and differences of genetic associations with longevity across the Han Chinese, European and U.S. ethnicities, we performed an evaluation/comparison analysis by using our Han Chinese GWAS as discovery and two European/American studies as evaluation. The two GWAS studies used for evaluation were the EU lon- gevity genetics consortium, with a population of 5,406 aged 90+ and 15,112 controls aged < 65 that included 14 studies from the Netherlands, Denmark, Iceland, Germany, Italy, the United Kingdom and Sweden10, and the U.S. Results P l ti Position Nearest gene Coded/ non- coded allele Southern region of China Northern region of China Southern-Northern Combined Meta analysis MAF (case/ control) P Odds ratio MAF (case/ control) P Odds ratio MAF (case/ control) P Odds ratio P rs2069837 7 22768027 IL6 (intronic) G/A 0.018/0.033 5.98E-03 0.582 0.086/0.134 1.00E-06 0.64 0.053/0.095 1.80E-09 0.61 4.05E-08 rs2440012 13 19440123 ANKRD20A9P (nc_exonic) G/C 0.050/0.092 1.38E-06 0.506 0.057/0.079 2.26E-03 0.69 0.054/0.084 3.73E-08 0.602 4.89E-08 rs145672791 21 14750023 MIR3156-3 (28 kb down- stream) A/G 0.003/0.011 5.08E-03 0.267 0.004/0.022 9.88E-06 0.203 0.004/0.018 8.95E-08 0.219 2.34E-07 rs61856137 10 5087978 AKR1C2 (27 kb upstream) T/G 0.019/0.032 9.85E-03 0.572 0.040/0.070 1.56E-05 0.549 0.029/0.056 1.60E-07 0.544 7.54E-07 rs2704588 4 89849772 FAM13A (intronic) C/T 0.004/0.013 4.32E-03 0.289 0.005/0.021 3.26E-05 0.237 0.004/0.018 2.38E-07 0.248 5.63E-07 rs1487614 4 42269480 BEND4 (114 kb upstream) T/C 0.112/0.146 1.85E-03 0.738 0.103/0.141 8.13E-05 0.707 0.107/0.143 2.87E-07 0.716 5.30E-07 rs10934524 3 96150160 EPHA6 (383 kb upstream) T/C 0.453/0.384 2.97E-05 1.354 0.470/0.431 4.76E-03 1.192 0.462/0.413 5.33E-07 1.266 1.16E-06 rs57681851 4 2290698 ZFYVE28 (intronic) G/T 0.187/0.136 7.05E-05 1.448 0.155/0.128 7.75E-03 1.256 0.170/0.131 1.83E-06 1.348 3.78E-06 rs7213812 17 31448649 ASIC2 (intronic) C/A 0.216/0.161 1.36E-05 1.45 0.176/0.152 2.84E-02 1.182 0.196/0.155 6.33E-06 1.29 6.25E-06 rs9568833 13 53827016 OLFM4 (200 kb downstream) T/C 0.145/0.193 7.85E-05 0.712 0.144/0.168 2.46E-02 0.836 0.144/0.177 1.77E-05 0.778 1.75E-05 rs405509 19 45408836 APOE (200 bp upstream) G/T 0.374/0.316 7.92E-05 1.32 0.308/0.279 2.56E-02 1.148 0.341/0.293 3.64E-05 1.21 1.85E-05 Table 1. The 11 independent loci associated with longevity at p < 10−4 in the discovery and at least a nominal significance (p < 0.05) in the evaluation, using the independent GWAS datasets of Southern and Northern regions of China as discovery/evaluation. Note: The SNPs rs57681851 and rs7213812 were imputed, while the other 9 SNPs listed in this Table were genotyped. MAF; minor allele frequency. Figure 2. Manhattan plot showing the results of the association with longevity in the combined GWAS dataset. SNP Chr. Results P l ti Position Nearest gene Coded/ non- coded allele Southern region of China Northern region of China Southern-Northern Combined Meta analysis MAF (case/ control) P Odds ratio MAF (case/ control) P Odds ratio MAF (case/ control) P Odds ratio P rs2069837 7 22768027 IL6 (intronic) G/A 0.018/0.033 5.98E-03 0.582 0.086/0.134 1.00E-06 0.64 0.053/0.095 1.80E-09 0.61 4.05E-08 rs2440012 13 19440123 ANKRD20A9P (nc_exonic) G/C 0.050/0.092 1.38E-06 0.506 0.057/0.079 2.26E-03 0.69 0.054/0.084 3.73E-08 0.602 4.89E-08 rs145672791 21 14750023 MIR3156-3 (28 kb down- stream) A/G 0.003/0.011 5.08E-03 0.267 0.004/0.022 9.88E-06 0.203 0.004/0.018 8.95E-08 0.219 2.34E-07 rs61856137 10 5087978 AKR1C2 (27 kb upstream) T/G 0.019/0.032 9.85E-03 0.572 0.040/0.070 1.56E-05 0.549 0.029/0.056 1.60E-07 0.544 7.54E-07 rs2704588 4 89849772 FAM13A (intronic) C/T 0.004/0.013 4.32E-03 0.289 0.005/0.021 3.26E-05 0.237 0.004/0.018 2.38E-07 0.248 5.63E-07 rs1487614 4 42269480 BEND4 (114 kb upstream) T/C 0.112/0.146 1.85E-03 0.738 0.103/0.141 8.13E-05 0.707 0.107/0.143 2.87E-07 0.716 5.30E-07 rs10934524 3 96150160 EPHA6 (383 kb upstream) T/C 0.453/0.384 2.97E-05 1.354 0.470/0.431 4.76E-03 1.192 0.462/0.413 5.33E-07 1.266 1.16E-06 rs57681851 4 2290698 ZFYVE28 (intronic) G/T 0.187/0.136 7.05E-05 1.448 0.155/0.128 7.75E-03 1.256 0.170/0.131 1.83E-06 1.348 3.78E-06 rs7213812 17 31448649 ASIC2 (intronic) C/A 0.216/0.161 1.36E-05 1.45 0.176/0.152 2.84E-02 1.182 0.196/0.155 6.33E-06 1.29 6.25E-06 rs9568833 13 53827016 OLFM4 (200 kb downstream) T/C 0.145/0.193 7.85E-05 0.712 0.144/0.168 2.46E-02 0.836 0.144/0.177 1.77E-05 0.778 1.75E-05 rs405509 19 45408836 APOE (200 bp upstream) G/T 0.374/0.316 7.92E-05 1.32 0.308/0.279 2.56E-02 1.148 0.341/0.293 3.64E-05 1.21 1.85E-05 Table 1. The 11 independent loci associated with longevity at p < 10−4 in the discovery and at least a nominal significance (p < 0.05) in the evaluation, using the independent GWAS datasets of Southern and Northern regions of China as discovery/evaluation. Note: The SNPs rs57681851 and rs7213812 were imputed, while the other 9 SNPs listed in this Table were genotyped. MAF; minor allele frequency. p ) Table 1. The 11 independent loci associated with longevity at p < 10−4 in the discovery and at least a nominal significance (p < 0.05) in the evaluation, using the independent GWAS datasets of Southern and Northern regions of China as discovery/evaluation. Note: The SNPs rs57681851 and rs7213812 were imputed, while the other 9 SNPs listed in this Table were genotyped. MAF; minor allele frequency. Figure 2. Manhattan plot showing the results of the association with longevity in the combined GWAS dataset. Table 1. Results P l ti Together, the 11 loci associated with longevity explained 3.38% of the variance in surviving to ages 100 + from middle-age, with each locus contributing from 0.39% (rs9568833-OLFM4) to 1.0% (rs2069837-IL6) of the variance based on effect estimates in the combined analysis, using the restricted maximum likelihood (REML) method22.i f y , g ( ) Notably, gender-specific association analysis for the 11 SNPs listed in Table 1 showed the same direction of effect in men and women with mostly very similar odds ratios (Supplementary Table 2). All of the 11 SNPs had Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 3 www.nature.com/scientificreports/ a p-value of p < 10−3 in women. Ten of the 11 SNPs achieved nominal significance (P < 0.05) in men and one SNP had a P = 0.0549 in men (Supplementary Table 1). The cross-gender similarities in effect direction/size on longevity of these 11 SNPs, which are replicated between discovery and evaluation datasets, is consistent with the results of the other GWAS on longevity10,4, and the gender differences of the not-replicated SNPs is out of scope of this article and will be investigated in our subsequent study. Comparison of current results with longevity GWAS studies of different ethnicities. To investi- gate the similarities and differences of genetic associations with longevity across the Han Chinese, European and U.S. ethnicities, we performed an evaluation/comparison analysis by using our Han Chinese GWAS as discovery and two European/American studies as evaluation. The two GWAS studies used for evaluation were the EU lon- gevity genetics consortium, with a population of 5,406 aged 90+ and 15,112 controls aged < 65 that included 14 SNP Chr. Results P l ti New England centenarian study (NECS) of 1,030 long-lived individuals (age 95 and older for males and 100 and older for females) and 368 controls with a mean age of 79.94.h ) g The results presented in Supplementary Table 3 show that among the 11 SNPs associated with longevity and replicated in Southern and Northern GWAS datasets of CLHLS, six SNPs had available data in at least one of the two other longevity GWAS studies, and one SNP is not available in the other two longevity GWAS but it has Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 4 www.nature.com/scientificreports/ a proxy SNP (r2 = 0.97) from the European population SNPs database (see Table 1 and its note (3)). The SNP rs405509 in APOE replicated in Southern and Northern GWAS datasets of CLHLS was also replicated in both EU (P = 2.75 × 10−06) and New England GWAS (P = 2.46 × 10−03), identifying a novel SNP-specific replication (Supplementary Table 3). ( pp y ) We conducted additional evaluation/comparison analysis of the other 723 SNPs associated with longevity at a suggestive significance level (P < 10−4) in the Han Chinese GWAS combined dataset. Among the 723 SNPs, 267 were available in the EU GWAS, and 37 were available in the New England GWAS. Of these, eight independent SNPs associated with longevity (P < 10−4) in the Han Chinese GWAS overlapped with at least nominal signifi- cance (P < 0.05) in at least one of the longevity GWAS in Europe (8 SNPs) and New England (2 SNP). Among these overlapped SNPs are rs405509 in APOE described above and rs4420638 in the TOMM40/APOE/APOC1 locus, which was replicated across continents with strong supporting evidence in all of the three GWAS datasets: P = 7.85 × 10−5 in the Han Chinese, P = 4.09 × 10−21 in the EU and P = 1.03 × 10−09 in the New England. , g In addition, we evaluated 54 SNPs associated with longevity (P < 10−4) discovered in the four previously published major studies of GWAS on longevity4,10,11,23 and found that 44 of them were available in Han Chinese GWAS. Results P l ti Among these 44 SNPs, thirty-nine were not associated with longevity in the Han Chinese (p > 0.05, data not shown but available upon request), and four SNPs in the TOMM40/APOE/APOC1 region reported in EU and New England GWAS replicated in the Han Chinese GWAS with P < x 10−4 (Supplementary Table 5). Furthermore, our Han Chinese GWAS identified an additional 10 SNPs with P < 10−4 in the TOMM40/APOE/ APOC1 region; and among them, 8 SNPs were in very high linkage disequilibrium (r2 = 1.0 or 0.99) with the significant SNPs reported in the EU and New England GWAS on longevity, and an additional 2 SNPs have not been reported before (Supplementary Table 5). Given the large number of associated SNPs in the TOMM40/ APOE/APOC1 region we performed a conditional analysis to identify independent association signals at this locus. The top independent association was rs405509 with P = 3.64 × 10−5. When conditioning on rs405509, we observed a secondary independent association at rs71352238 (Pconditional = 2.1 × 10−4) (Supplementary Fig. 4). After adjusting for rs405509 and rs71352238, we observed no other significant associations (P > 0.01) at this locus (Supplementary Fig. 4). These results demonstrate that the two independent associations in the TOMM40/APOE/ APOC1 account for all of the remaining associated signals in this region in the CLHLS GWAS.hi g g g The genome-wide significant longevity locus 5q33.3 (rs2149954, P = 1.74 × 10−8) reported in the EU GWAS10 showed the same direction of effect with P = 0.02 in the Han Chinese GWAS. We also found 4 SNPs of the reported longevity gene FOXO3 are associated with longevity with a P < 0.04 and odds ratio around 1.2 in Han Chinese (data not shown, but available upon request).hii p q The genome-wide significant locus identified in our Han Chinese GWAS, rs2069837 (P = 1.80 × 10−9), is located in the intronic region of the IL6 gene on chromosome 7p15.3 (Table 1). The SNP rs2069827 in IL6 that was reported previously as associated with longevity18,19 was not found in Han Chinese, but it is rather com- mon in European populations based on 1000genome annotation in HaploReg V2. Results P l ti Before the current study, starch, sucrose and xenobiotic metabolism and calcium signaling had not been iden- tified as being associated with human longevity in the European and American GWAS. The Chinese diet is high in carbohydrates, mainly starch and sucrose, and is usually low in fat. Sucrose is a heterodimer of glucose and fructose, while starch is a glucose polymer, and both molecules are degraded in the gut. The liberated glucose generates an insulin response. Fructose does not generate an insulin response but when it is consumed in excess it contributes to multiple chronic metabolic diseases. Fructose metabolism also can promote reactive oxygen species (ROS) formation, which leads to cellular dysfunction and aging33. Thus, our novel finding that the starch, sucrose and xenobiotic metabolism pathway is significantly associated with human longevity in Han Chinese is interpretable, because the favorable defensive genotypes carried out by the Chinese centenarians may interact l Among the four major pathways identified in our present study, p38 MAPK27 and immunity28 were linked to longevity previously. In particular, p38 MAPK acts downstream of the IL-6 receptor, and that may account for this association. p38 MAPK also acts upstream of FOXO, which, as discussed above, has been associated with longev- ity in previous GWAS. Moreover, we also note that p38 MAPK can be activated by AMPK, which is upstream of mTOR, and activation of mTOR has been shown to extend life span in C. elegans29, mice30 and yeast31. In com- mon with our results, mTOR signaling was identified as significant for the ageing phenotype by a recent study32. B f h d h d b b l d l l h d b d Before the current study, starch, sucrose and xenobiotic metabolism and calcium signaling had not been iden- tified as being associated with human longevity in the European and American GWAS. The Chinese diet is high in carbohydrates, mainly starch and sucrose, and is usually low in fat. Sucrose is a heterodimer of glucose and fructose, while starch is a glucose polymer, and both molecules are degraded in the gut. The liberated glucose generates an insulin response. Fructose does not generate an insulin response but when it is consumed in excess it contributes to multiple chronic metabolic diseases. Fructose metabolism also can promote reactive oxygen species (ROS) formation, which leads to cellular dysfunction and aging33. Results P l ti Similarly, our identified SNP rs2069837 has a MAF of 0.14 in Asian populations but 0.09 in the European population, indicating that the previously reported SNP rs2069827 is a European-specific longevity associated genotype and our identified SNP rs2069837 could be a Han Chinese-specific longevity associated genotype. The finding that these two genetic variants of IL6 contribute to longevity in opposite directions in Han Chinese versus Europeans emphasizes the impact of genetic heterogeneity on longevity across ethnicities. p g g y g y In brief, our comparative analysis indicates both considerable similarities and differences in genetic associa- tions with longevity between the Chinese, European and U.S. populations. Further cross-national meta-analysis is warranted to develop an in-depth understanding of the cross-ethnics genetic associations with human longevity. Pathway and network analyses. We conducted pathway analyses by applying i-GSEA4GWAS, an improved gene set enrichment analysis (GSEA) for GWAS24 (section M7 of Methods). Twenty-five canonical pathways were ranked as significantly enriched (FDR < 0.05 and corrected p-value ≤ 0.004) and associated with longevity (Supplementary Table 6). These can be functionally classified into 4 major pathways: starch, sucrose and xenobiotic metabolism (10 enriched pathways), immune response and inflammation (7 enriched pathways), MAPK (4 enriched pathways) and calcium signaling (2 enriched pathways), plus 2 other significant enriched pathways (see Supplement-Table 6). Our finding of the four major significant pathways for longevity is generally similar to the results from Alzheimer’s disease GWAS studies, in which over 20 genes with statistically-significant association signals25 were mapped to relatively few major significant pathways including lipid metabolism, inflammatory response, endocytosis and immune response26.i l y p y p Among the four major pathways identified in our present study, p38 MAPK27 and immunity28 were linked to longevity previously. In particular, p38 MAPK acts downstream of the IL-6 receptor, and that may account for this association. p38 MAPK also acts upstream of FOXO, which, as discussed above, has been associated with longev- ity in previous GWAS. Moreover, we also note that p38 MAPK can be activated by AMPK, which is upstream of mTOR, and activation of mTOR has been shown to extend life span in C. elegans29, mice30 and yeast31. In com- mon with our results, mTOR signaling was identified as significant for the ageing phenotype by a recent study32. Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 Results P l ti Thus, our novel finding that the starch, sucrose and xenobiotic metabolism pathway is significantly associated with human longevity in Han Chinese is interpretable, because the favorable defensive genotypes carried out by the Chinese centenarians may interact Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 5 www.nature.com/scientificreports/ Figure 3. Gene and Pathway networks in longevity traits. Stronger associations are represented by thicker lines. Protein-protein interactions are shown in blue, protein-chemical interactions and chemical-chemical interactions in green. Note: The green nodes mark 35 highly-represent genes in the four main pathways, the white nodes mark highly interacted genes with these 35 genes. The four main pathways are shown with colored rectangles. Figure 3. Gene and Pathway networks in longevity traits. Stronger associations are represented by thicker lines. Protein-protein interactions are shown in blue, protein-chemical interactions and chemical-chemical interactions in green. Note: The green nodes mark 35 highly-represent genes in the four main pathways, the white nodes mark highly interacted genes with these 35 genes. The four main pathways are shown with colored rectangles. with their high-carbohydrates diet to achieve extreme longevity. Furthermore, this novel finding is consistent with the hypothesis that dietary sources of carbohydrates, mainly sucrose and starch, may affect life span in Drosophilia34. p Altered calcium homeostasis contributes to neurodegenerative diseases of aging, including Alzheimer’s dis- ease (AD), and pharmacological inhibitors of calcium signaling have been shown to rescue structural plasticity defects in AD murine diseased neurons35. Our analysis suggests that calcium signaling has a potentially important impact on prolonging longevity.i Noteworthy, support for the relevance to longevity of three of the four major significant pathways found in our Han Chinese GWAS (MAPK, immunity and calcium signaling) has been most recently provided by Tan et al.36 as significant pathways in their longitudinal epigenome-wide association study (EWAS) of Danish twins with a mean age of 76 years at entry and, followed for 10 years. Interestingly, gene expression levels for the four significant pathways found in the Han Chinese GWAS were reported as differentially regulated during aging in C. elegans37. Discussion Here, we summarize and discuss our key findings. First, our GWAS of Han Chinese identified 11 independent loci associated with longevity and replicated in two independent GWAS datasets from Southern and Northern regions of China (Table 1). More specifically, we identified two novel loci (rs2069837-IL6 and rs2440012-ANKRD20A9P) that reached genome-wide significance, plus 9 suggestively significant loci (P < 3.64 × 10−5), including rs405509 in APOE, which is a well-known locus negatively associated with longevity42. Our strongest novel signal is for the rs2069837 SNP (P = 1.80 × 10−9) within the IL6 locus at 7p15.3, negatively associated with longevity in the Han Chinese, which is consistent with previous findings that the IL6 gene functions as an inflammatory biomarker of poor health outcomes20,21. An earlier Dutch prospective cohort study of longevity reported a different SNP in IL6, rs2069827, which was positively associated with longevity18. The analyses showed that the previously reported SNP rs2069827 is European-specific and our newly identified SNP rs2069837 could be a Han-Chinese-specific longevity associated genotype. Another strong novel locus that reached genome-wide significance in Han Chinese GWAS was ANKRD20A9P (ankyrin repeat domain 20 family, member A9), a pseudogene which is affiliated with the lncRNA class. The biological function of this particular variant remains to be characterized.i h g p Secondly, we identified 8 independent SNPs associated with longevity with P < 10−4 in Han Chinese GWAS that overlapped results from the European and/or New England longevity GWAS, with at least nominal signifi- cance (P < 0.05). We also confirmed 4 previously reported longevity-associated SNPs and identified 2 independ- ent associations with longevity in the well-known TOMM40/APOE/APOC1 LD region. Our findings provided additional new signals and replicated the negative association of the TOMM40/APOE/APOC1 LD region with exceptional longevity in Han Chinese, and confirmed 5q33.3 as a longevity locus that was previously identified in EU-GWAS. In short, we demonstrated both similarities and differentials in genetic associations with longevity across continents and ethnicities.hil Thirdly, at the pathway level, our study identified 4 major pathways that influence longevity, including starch, sucrose and xenobiotic metabolism, immunity, MAPK signaling and calcium homeostasis. Among these path- ways, p38 MAPK and immunity have been previously related to longevity in numerous studies27–31. Our net- work analysis indicates these 4 identified pathways are highly interconnected. Collectively, our findings strongly support the notion that longevity is a polygenic trait influenced by a complex interplay of multiple genes and pathways. Overlap between association signals and regulators of transcription. Overlap between association signals and regulators of transcription. By querying the RegulomeDB database40, which includes high-throughput, experimental datasets from ENCODE (Encyclopedia of DNA Elements data), we found rs405509 upstream of APOE exhibited a strong cis-effect on ZNF226 expres- sion in lymphoblastoid cell lines41. Three SNPs in two other genes also mapped to regulatory regions, including rs2069837 in IL6, and rs7212444 and rs8064775 in ASIC2. These variants showed evidence of transcription factor binding and DNase I hypersensitivity peaks. www.nature.com/scientificreports/ variants. The findings of our SNP analysis and this eQTL analysis suggest ASIC2 locus in brain cerebellum plays a role in mediating longevity phenotype.i variants. The findings of our SNP analysis and this eQTL analysis suggest ASIC2 locus in brain cerebellum plays a role in mediating longevity phenotype.i g g y p yp We also investigated whether the SNPs with a significance level of P < 10−3 in the 35 genes highly represented across pathways were associated with gene expression using eQTL data. By querying three publically availa- ble databases (GTEx, seeQTL, Chicago eQTL), we mapped multiple SNPs with highly significant cis-eQTLs in human tissue samples to three genes, including MAP3K1, HLA-B and HLA-DPB1 (supplementary Tables 10–12). Methods M1. Samples and data source. DNA samples and data for the present GWAS are from the Chinese Longitudinal Healthy Longevity Surveys (CLHLS), which were conducted in 1998, 2000, 2002, 2005, 2008, 2011 Results P l ti These consistent findings of the important pathways for both genetic association with longevity as well as differential functional regulation over the aging process in both humans and animal models suggest that the functional coordination of these 4 major pathways have a profound impact on longevity and aging.i j p y p p g y g g To further understand the interconnection of genes in the identified major pathways, we performed network analysis in STITCH38 (section M8 of Methods) and identified 35 genes (Supplementary Table 7) that are highly represented across the 25 enriched pathways. These 35 highly-represented genes include gene family members UGT1A (the gene cluster that defends against organic molecules, such as small molecule toxins) and HLA (the gene functions in immunity protecting against pathogens), which indicate potential defensive mechanisms that are important for longevity. As shown in the connectivity map (Fig. 3), we find that the four major pathways that mediate defensive mechanisms and have a role in longevity, are highly interconnected. For example, IL6, a top ranked gene in both our SNPs and pathway analyses, is a key gene in immune response and inflammation and activating p38 via IL6 receptor. The p38 MAPK pathway, in turn, mediates pathogen-specific responses by regu- lating expression of many protective genes related to immune response, contributing to longevity39. Involvement of the both of UGT1A gene cluster, which detoxifies small molecules, and the HLA family, lend further support to the notion that defensive mechanisms are critical for successful longevity. g y Taken together, our data strongly suggest that the longevity trait represents a complex interplay of multiple genes and pathways interacting with the environment, notably diet, that converge on specific biological processes. Analysis of regulation of gene expression by associated variants. We examined the potential effects of the 11 SNPs or their proxies (r2 ≧ 0.6) in our study on gene expression using several eQTL databases (section M9 of Methods). Two SNPs in moderate LD (r2 = 0.61) with rs11658235 (ASIC2), i.e. rs7224279 and rs11658301, showed significant trans-eQTL associations (P = 3.88 × 10−9) in human brain cerebellum tissue sam- ples (Supplementary Table 9), but we did not find statistically significant eQTLs for other longevity-associated Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 6 www.nature.com/scientificreports/ Discussion Interestingly, the three significant pathways (MAPK, immunity and calcium signaling) identified in our Han Chinese GWAS study were also most recently reconfirmed as significant pathways in the Danish longitu- dinal Epigenome-wide association study36. Our novel finding that the starch, sucrose and xenobiotic metabolism pathway is significantly associated with human longevity in Han Chinese is consistent with the previous animal model study in Drosophilia34. Together, these data strengthen our findings that combinations of SNPs or genes represented among GWAS associations could converge upon gene pathways and biological processes related to healthy aging. y g g Like other GWAS studies on longevity, the major limitation of the present study is a lack of physiologically relevant model systems for further biological and functional validation of our new findings. This study represents the largest GWAS of centenarians to date and provides a particularly useful reference for analysis of disease mit- igation or prevention genotypes, pathways and biological mechanisms. We believe that sharing the findings of this study will inspire biologists who are interested in aging research to conduct functional studies of the various identified genetic variants and pathways associated with longevity.hi i g p y g y The findings in the present study support the hypothesis that defensive mechanisms (such as immunity) and metabolism driven by diet in response to environmental stress may play key roles in longevity in the Han Chinese. Diet mediated mechanisms also suggest that genetic and epigenetic influences upon extreme survival can dif- fer according to culture and ethnicity. These findings may provide novel insights for expanding longevity theo- ries that have emphasized genes associated with maintenance of genome integrity (especially DNA repair) and fertility-related mechanisms43–45. Our new findings and previous studies about the reactive and adaptive nature of the immune system and about metabolism influenced by factors such as diet lead us to hypothesize that Han Chinese centenarians, who survived through the brutal past, may carry favorable defensive genotypes associated with longevity. Further studies are warranted for functional characterization of the new loci and interactive path- ways to elucidate the underlying mechanisms for the observed associations with longevity. www.nature.com/scientificreports/ The sample sizes of any minority group in the CLHLS data are too small for meaningful analysis, so we include Han Chinese samples only in the present study. g y , p y p y Descriptions of the samples and data sources of GWAS on longevity from the European Union Consortium and New England centenarians study (NECS) are given in the refs 4 and 10. M2. Genotyping and quality controls. After DNA extraction, all of the 2,578 centenarians and 2,387 middle-age controls were genotyped using the Illumina HumanOmniZhongHua-8 BeadChips, which was created by strategically selecting optimized tag SNP content from all three HapMap phases and the 1000 Genomes Project (1 kGP). It allows profiling of 900,015 SNPs per sample, including 600 k SNPs of common variants (MAF ≥ 5%), 290 k SNPs of rare variants (MAF < 5%) and 10 k SNPs existing only among Chinese and other Asian populations. In other words, 98.9% of the 900k SNPs of the Illumina HumanOmniZhongHua-8 BeadChip are internationally compatible, with 1.1% specific to Chinese and Asian populations; this provides coverage of about 81% of com- mon variants at r2 > 0.8 with MAF ≥ 5%, and about 60% of rare variants at r2 > 0.8 with MAF < 5%. Our selec- tion of this chip represents a state-of-the-art choice for GWAS in Asian populations to maximize international compatibility.f We randomly arranged cases and controls during sample batches, with negative controls (buffer water) and positive controls (YanHuang samples). For the sample filtering, 339 individuals for whom the genotypes were generated with a call rate less than 95% were excluded. We also computed identity-by-state probabilities for all subjects to search for any possible duplicates and kinship-related individuals among the samples, using PLINK (1.06)16. The 134 individuals who had identity-by-state probabilities with PI_HAT > 0.25 were excluded based on the IBD analysis implemented in PLINK.ii y p To complete the sample filtering data quality control procedures and investigate population stratification in our GWAS samples, we performed principal components analysis (PCA) to evaluate genomic correlations in our dataset on the basis of pairwise identity by state (IBS) for all of the successfully genotyped samples using PLINK (1.06)16. We used the filter option (MAF > 0.01 and genotyping rate > 0.9) as discussed above; about 80% of the SNPs were used to calculate IBS. www.nature.com/scientificreports/ and 2014 in a randomly selected half of the counties and cities in 22 out of 31 provinces in China. The CLHLS covers approximately 85% of the total population of China. We tried to interview all consented centenarians in the sampled counties and cities. For each centenarian interviewee, we recruited one nearby un-related middle-age control participant aged 40–59. “Nearby” is loosely defined – it could mean the same village or the same street if available, or in the same town or in the same sampled county or city46. In the present study, all of the DNA sam- ples of the middle-age controls were collected in the same county/city or different county/city but in the same province, as that of the centenarians. p Phenotype data were collected in the CLHLS using internationally standardized questionnaires adapted to the Chinese cultural and social context46. Extensive evaluations of the data quality of the CLHLS, including assess- ments of mortality rate, proxy use, non-response rate, sample attrition, reliability and validity of major health measures, and the rates of logically inconsistent answers, have shown that the data from the CLHLS surveys are of good quality47. The genetic samples and data from CLHLS were successfully used in prior published studies on candidate genes and gene-environment interactions relevant to longevity48–52.i g g g y A wide variety of international and Chinese studies53,54 have confirmed that age reporting of the Han Chinese oldest-old aged 80 + , including centenarians, is reasonably accurate; this is due to the Han Chinese cultural tradi- tion of memorizing one’s date of birth to determine dates of important life events such as engagement, marriage, starting to build a residential house, etc. The accuracy of age reporting in the CLHLS data was reconfirmed by an investigation, which compared various standard demographic indices of age reporting among the oldest-old and age distributions of centenarians between the CLHLS and comparable data from Sweden, Japan, England and Wales, Australia, Canada, the U.S., and Chile; this study concluded that the age reporting of the Han Chinese oldest-old, including centenarians, is as good as the average in developed countries55. g g g p Note that the Han Chinese comprise about 93% of the total population in China, with 53 Chinese minority groups comprising 7% of the total population. Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 Methods Methods M1. Samples and data source. DNA samples and data for the present GWAS are from the Chinese Longitudinal Healthy Longevity Surveys (CLHLS), which were conducted in 1998, 2000, 2002, 2005, 2008, 2011 Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 7 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Only 242 out of 743 (~30%) GWAS conducted from 2005 to 2011 considered chromosome X in their analyses according to the recent report of Wise et al.59. Wise et al. discussed several reasons for exclusion of chromosome X, including a lower proportion of genes on chromosome X and a lower coverage of chromosome X on current genotyping platforms compared with autosomal coverage. They also described a number of technical hurdles that might add to the reluctance to include chromosome X in a GWAS, including complications in genotype calling, imputation, selection of test statistics, and the lack of readily available implementations. Accordingly, we excluded the chromosome X. After quality filtering and cleaning, 818, 084 genotyped SNPs remained for association analysis in our final GWAS dataset. M3. Imputation. To further increase genome coverage, we performed imputation analysis to infer the gen- otypes of all SNPs (MAF ≥ 0.01) using IMPUTE software version 260 and the 1000 Genomes Project integrated phase 1 release as reference panel. SNPs with a quality score (Rsq) of < 0.9 were discarded before analysis. After standard GWAS quality-control filtering for subjects and SNPs as described above, we obtained data for 5, 595, 657 SNPs (818, 084 genotyped SNPs and 4, 777, 573 imputed SNPs) in 2, 178 centenarians cases and 2, 299 middle-age controls for the subsequent GWAS analyses. M4. Stratifying Southern and Northern regions for our discovery-evaluation analysis. China is considered to have geographically and traditionally distinct Southern and Northern regions that have existed for a few thousand years, although there is no legal or official administrative meaning to this division. As reviewed in Xu et al.15, previous analyses of anthropological, anatomical, linguistic, and genetic data (including classic markers, microsatellite DNA markers, mtDNA, and Y chromosome SNP markers) have consistently suggested the presence of a significant boundary between the Southern and Northern populations in China. Among the 22 provinces where the CLHLS was conducted, Liaoning, Jilin, Heilongjiang, Hebei, Beijing, Tianjin, Shanxi, Shandong, Shaanxi, Jiangsu, Anhui and Henan belong to the geographically and traditionally defined Northern region; Shanghai, Zhejiang, Fujian, Jiangxi, Hubei, Hunan, Guangdong, Sichuan, Guangxi, Chongqing, and Hainan belong to the geographically and traditionally defined Southern region.i i Our principle components analysis (described in section M2 above) also genetically confirmed the geographic and traditional cultural division between Southern and Northern regions in the Han Chinese GWAS dataset (Supplementary Fig. 2). www.nature.com/scientificreports/ Thus, we divided the entire sample into two independent GWAS datasets of Southern and Northern regions for our discovery-evaluation analysis. The Southern region dataset consists of 1,063 cen- tenarians and 887 middle-age controls and the Northern region dataset consists of 1,115 centenarians and 1,412 middle-age controls (Supplementary Table 1). Given the nature of our population analysis and due to the fact that all of the samples analyzed in this study belong to the same ethnic group of Han Chinese living in one country with the same culture, we also performed a combined analysis comparing all centenarians with all controls. The genomic inflation factors (λ ) in the Southern and Northern datasets and the combined dataset were 1.022, 1.010 and 1.022, respectively, suggesting that the association test statistics conformed to the underlying null distribution and would not require further adjustment for genomic control. M5. A bi-directional discovery-evaluation strategy. In the classic uni-directional discovery-replication approach, the entire sample is divided into two datasets with one dataset used as discovery and the other dataset as the replication or valuation. The top SNPs found in the discovery stage using a pre-determined P value thresh- old are analyzed in the evaluation dataset. The SNPs with a P value lower than the threshold (e.g. P < 1.0 × 10−4) in the discovery stage and nominal significance (e.g. P < 0.05) in the evaluation stage are identified as the signifi- cant/replicated SNPs, which are the final results of the analysis. This classic uni-directional discovery-evaluation approach is especially useful when the expensive GWAS serves as the first stage of discovery and the much less expensive second stage of evaluation genotypes only the top SNPs with a P value lower than the threshold found in the first stage. However, when analyzing two available independent GWAS datasets, the uni-directional discovery-evaluation approach of assigning one GWAS dataset as discovery and another GWAS dataset as evalu- ation would have a higher false-negative rate, missing a substantial number of significant/replicated SNPs which have a p-value higher than the threshold and lower than the nominal significance level in the discovery GWAS dataset but reach the threshold significance level in the evaluation GWAS dataset. gi To avoid the high false-negative rate and to fully utilize all information in the two independent Southern and Northern GWAS datasets available to us, we applied a novel bi-directional discovery-evaluation strategy17 to search for consistent associations of SNPs with longevity. www.nature.com/scientificreports/ Our principle components analysis (PCA) showed that 9 individuals (all from Fujian province) were genetically separated from the total population of the 22 provinces (including Fujian) where the CLHLS was conducted, so these 9 individuals were excluded from our GWAS. Based on the PCA results, we also excluded 6 individuals who had serious deviations from the genetically-derived geographic membership in Southern or Northern populations.t After excluding these 15 outliers, our PCA showed that the cases and controls in this study were of Han Chinese ancestry and were well matched, without evidence of gross population stratification. In general, cases and controls were not separated on the basis of the first two principal components, and the cases and controls are evenly distributed in clusters of PC1 vs. PC2, PC1 vs. PC3 and PC2 vs. PC3 (see Supplementary Fig. 1). We com- puted the inflation factor (λ ), a metric describing genome-wide inflation in association statistics56, both before or after correcting for stratification using PLINK (1.06)16. Values of λ after correcting along 0, 1, 2, 3, or 4 eigen- vectors are 1.109, 1.0241, 1.0235, 1.0226, or 1.0221, respectively, demonstrating that the correction is important and the top two eigenvectors correct nearly all of the stratification that can be corrected using the 4 eigenvectors. Consequently, in the subsequent GWAS regression models, we adjusted for the top two eigenvectors to minimize the effects of population stratification, following the approach adopted in the refs 54 and 57. f p pi g pp p Finally, after various stages of sample filtering, 2,178 centenarians (mean age 102.7 ± 3.49 (SD)) and 2,299 middle-age controls (mean age 48.4 ± 7.44 (SD)) were included in the subsequent GWAS dataset (Supplementary Table 1). We also conducted quality-control filtering of the GWAS data from the total 4,477 indi- viduals. Following the standard adopted in the published studies of GWAS using the same 900 k SNPs Illumina HumanOmniZhongHua-8 BeadChips58 as the one used in present study, SNPs with call rates of less than 90% were removed from our GWAS dataset; SNPs were also excluded if they had a MAF less than 1% or if there was significant deviation from Hardy-Weinberg equilibrium in the samples, defined as P < 10−5. SNPs on the X and Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 8 www.nature.com/scientificreports/ Y chromosomes and mitochondria were removed from the initial statistical analysis. Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 www.nature.com/scientificreports/ In fact, the “case-control association analysis on longev- ity” approach uses the same basic idea of FAD and produces exactly the same results of estimates as FAD does48.h y pp p y The method of case-control association analysis (or FAD) on longevity, which uses long-lived individuals as cases and ethnically/geographically matched middle-age persons as controls, is based on fundamental demo- graphic insight that the prevalence of a genetic variant in a population can change with age even though no individuals can change their fixed attribute of the genetic variant, and that therefore much can be learned about the impact of the genetic variant on longevity. Ideally, we would use complete cohort data to compare the distribu- tions of the genetic variants of members of the same cohort at two points of time in their life span: when they were young versus when some of them reach age 100 + , while all of the other cohort members died before age 100. However, such long-term and complete follow-up GWAS or candidate gene genotype data for the members of the same birth cohort at young ages and ages 100 + are not currently available for any population. Thus, various pre- viously published association studies on longevity used cross-sectional datasets to compare long-lived individuals (centenarians and/or nonagenarians) as cases and middle-aged as controls, observed in the same period. Such cross-sectional cases/controls analyses are based on two assumptions: (1) the initial distribution of the genetic variants does not differ substantially between the long-lived and middle-age cohorts, which is generally reason- able, because the basic genetic structure would not change substantially in about 50 years within the same ethnic population. (2) The basic genetic profiles of the non-migrants do not differ substantially from those migrants who are of the same ethnicity as non-migrants. This assumption is also generally reasonable, especially for stud- ies of Han Chinese because China received remarkably few immigrants of other ethnicities in the past decades. With these two assumptions, one may intuitively understand that the proportion of genetic variants which are positively (or negatively) associated with longevity are significantly higher (or lower) among the centenarians, compared to the middle-aged controls, because those who have the favorite genetic profile have better chance to survive to age 100 + , while those with less favorite genetic profile could not reach age 100. www.nature.com/scientificreports/ Such an intuitive mechanism based on the two assumptions outline above has been proven mathematically48. Note that in the middle aged controls, selective mortality is still very low and on the other hand it is impossible for natural selection to change the gene pool within a short period of several decades. Thus, genotype frequency in the middle-aged controls could well represent population genotypes. However, the findings based on the study design of using centenarians as cases and middle-aged as controls may be driven by the gene-environment inter- actions that lead to selective mortality for the centenarians. For example, the Chinese centenarians survived the brutal mortality regimes of the past when famine, wars, and starvation operated on their large birth cohorts. This may imply that these long-lived individuals who successfully passed through the hard life conditions and high mortality of earlier decades may carry genotypes that interact with environmental factors such as life stress, cer- tain diet or behavior, which lead to exceptional longevity. Thus, the genotypes associated with longevity identified through comparing centenarians and middle-age controls may serve as candidates for further studies on effects of gene-environment interactions on longevity and health to determine whether they are found in other populations and cohorts that have been subjected to similar or different environmental factors. f We performed GWAS analysis using logistic regression as implemented in PLINK (1.06)16. To minimize the effects of population stratification, we adjusted for the top two eigenvectors, which corrected nearly all of the stratification that can be corrected57. In the combined data analysis, we also adjusted for the sex and geographic stratification of Southern and Northern regions. In addition, we performed the meta-analysis, treating the two independent datasets from Southern and Northern regions of China as two groups and employing the METAL software65. t For the SNPs associated with longevity and replicated in the datasets of the Southern and Northern regions, we also conducted gender-specific analyses and tested whether the differences in the effect sizes measured by the odds ratio were statistically significant between the two genders. We used Woolf’s test66 to calculate P values to measure the gender differences in association with longevity (Supplementary Table 2). www.nature.com/scientificreports/ More specifically, the two GWAS datasets from Southern and Northern regions were analyzed with a parallel and bi-directional procedure: one is assigned as a discovery dataset and another as an evaluation dataset, and vice versa17. In our case, we first assigned and ana- lyzed the Southern region GWAS dataset as discovery and Northern region GWAS dataset as evaluation, and then analyzed the Northern region GWAS dataset as discovery and Southern region GWAS dataset as evaluation (see Fig. 1). In this way, we capture all SNPs associated with longevity with a p-value lower than the threshold in the discovery stage and nominal significance in the evaluation stage.hi i The recent literature on GWAS discovery-evaluation analysis indicate that researchers defined a priori dis- covery threshold of P < 10−5 or P < 10−4, or P < 10−3, depending on the circumstances of the research topics and the characteristics of the datasets used61–63. In the present study of GWAS on longevity, we use exceptionally long-lived centenarians as cases and middle-aged adults as controls. Furthermore, we seek to identify SNPs which may individually have small effects, but may have large and important joint effects on the extremely complex trait of exceptional longevity. Thus, we believe that selecting the modest a priori discovery threshold of P < 10−4 is a reasonable choice. More specifically, we defined the a priori discovery threshold of P < 10−4 in the discovery stage in our bi-directional discovery-evaluation analyses using the independent GWAS datasets of Southern and Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 9 www.nature.com/scientificreports/ Northern rejoins of China, and in further international comparative analysis among the longevity GWAS datasets of the CLHLS, European Union and New England. We used a nominal significance level of p < 0.05 in the evalu- ation stage, following the usually adopted practice. Northern rejoins of China, and in further international comparative analysis among the longevity GWAS datasets of the CLHLS, European Union and New England. We used a nominal significance level of p < 0.05 in the evalu- ation stage, following the usually adopted practice. M6. Case-control study design and the statistical analysis. The logistic regression approach (named as “case-control association analysis on longevity”) or the “Fixed-Attributes Dynamics (FAD) Method”64 and its extension48 is often used to estimate the associations between genetic variants and longevity, through comparing long-lived individuals and middle-age controls4,10,11,51,52. Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 www.nature.com/scientificreports/ Woolf’s test is a method for testing the heterogeneity of 2 × 2 contingency tables over multiple strata, which is sensitive when the odds ratios (OR) of two genders are in opposite directions (the allele is protective for one sex and risky for the other), or in the same direction but of significantly different magnitude of effect size.h gi yf gf The methods of statistical analyses using the GWAS datasets from EU longevity genetics consortium and NECS were described in the refs 4 and 10, respectively4,10. M7. Pathway analysis. High-throughput single-SNP association analysis per se does not directly produce biologically functional findings because SNPs and genes work in an intricate network of interactions through biological pathways. In addition, each of the significant SNPs normally has a small effect on a complex trait such as longevity. Thus, pathway analysis is an important follow-up for GWAS studies to provide mechanistic insight about the underlying biology of the phenotype of interest. By identifying well-annotated pathways that map to the lists of significant genes identified in a GWAS, biochemical hypotheses can be enumerated and tested67. Thus, to further investigate the large number of SNPs produced by our GWAS in the context of biological processes, we conducted pathway and network analysis. We applied an algorithm known as improved gene set enrichment analysis for GWAS (i-GSEA4GWAS) to place variants associated with longevity within curated pathways and functional categories68. SNPs with p < 0.05 in our combined GWAS dataset were mapped to genes if within a Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 10 www.nature.com/scientificreports/ 20 kb distance (upstream or downstream). For a given SNP, if multiple genes were located within this range, the closest gene was selected and assigned the association P value. Since multiple SNPs can map to the same gene, a SNP label permutation was used to reduce biases caused by larger loci having a disproportionately higher number of SNPs. We did not mask the non-MHC/xMHC region and set the gene size to minimum 10 and maximum 200, based on the filter criteria recommended by Ramanan et al.69 The program uses the list of SNP-mapped genes to filter the collection of pathways/gene sets to obtain candidate pathways/gene sets. In total, 127, 092 SNPs associ- ated with longevity with a p < 0.05 were used and they were mapped to 8,085 genes. www.nature.com/scientificreports/ These were limited to curated pathways derived from multiple resources such as KEGG70, REACTOME71 and BioCarta (http://www.biocarta. com/), and functional annotations extracted from the Gene Ontology database72. A modified version of the GSEA procedure was performed, adjusted for multiple testing using false discovery rate (FDR), and pathways/gene sets with corrected p-value < 0.05 and FDR < 0.05 regarded as associated with longevity traits68. M8. Gene interaction network analysis. To investigate whether and how the pathways may inter- act to affect longevity, we performed gene interaction network analysis. We first sorted the genes which were highly-represented across the 25 enriched pathways. Then we selected 35 genes which were constituents of 2 or more enriched pathways, and used the human protein-chemical interactions database STITCH 4.073 to investigate the interaction network among significant genes and among genes and chemicals. The STITCH database accu- mulates data from multiple sources and contains a confidence value for each interaction. The confidence scores range between 0 and 1, with a default cutoff value of 0.15. To ensure that the results of our gene interaction net- work analysis attain a high confidence level, we generated evaluation datasets for STITCH in which we required a confidence score of at least 0.4 (median value)73. M9. Expression Quantitative Trait Loci (eQTL) analysis. We examined the potential effects of the SNPs, identified as associated with longevity at a suggestive significance level in our GWAS, on gene expression, using the three publicly available expression Quantitative Trait Loci (eQTL) databases: (1) the Genotype-Tissue Expression (GTEx) data derived from multiple cell and tissue types (lymphoblastoid cell lines, brain tissue and human fibroblasts)74; (2) SeeQTL database from the University of North Carolina at Chapel Hill75; and (3) eQTL resources from the Gilad/Pritchard group at University of Chicago (http://eqtl.uchicago.edu/Home.html). In total, 127, 092 SNPs with P < 0.05 in the combined datasets of our GWAS were included to examine their effects on gene expression in using these three eQTL database resources. M10. Statements of Approval for Human Subjects Informed Consent and in accordance with relevant guidelines. The Research Ethics Committees of Duke University and Peking University granted approval for the Protection of Human Subjects for the Chinese Longitudinal Healthy Longevity Survey, including collection of DNA sample used for present study. The survey respondents who contributed their DNA samples gave informed consent before participation. www.nature.com/scientificreports/ All of the GWAS experiments and methods of analyses in present study were performed in accordance with relevant guidelines and regulations. 1. Hjelmborg, J. et al. Genetic influence on human lifespan and longevity. Hum Genet. 119, 312–321 (2006).if References j g Jl p g y , ( ) 2. Tan, Q. et al. Analyzing age-specific genetic effects on human extreme age survival in cohort-based longitudinal studies. European Journal of Human Genetics. 21, 451–454 (2013).i f 3. Sebastiani, P. et al. Increasing Sibling Relative Risk of Survival to Older and Older Ages and the Importance of Precise De “Aging,” “Life Span,” and “Longevity”. J. Gerontol. A. Biol. Sci. Med. Sci. doi: 10.1093/gerona/glv020 (2015). 4. Sebastiani, P. et al. Genetic signatures of exceptional longevity in humans. PLoS One 7, e29848 (2012). , g p g y , ( ) 5. Christensen, K. et al. Exceptional longevity does not result in excessive levels of disability. Proc. Natl. Acad. Sci. USA 105 13274–13279 (2008).h ( ) 6. Hitt, R., Young-Xu, Y. & Perls, T. Centenarians: The older you get, the healthier you’ve been. Lancet, 354, 652 (1999). ns: The older you get, the healthier you’ve been. Lancet, 354, 652 (19 , R , g , ,h y g , y , , ( ) 7. Andersen, S., Sebastiani, P., Deworkis, D., Feldman, L. & Perls, T. Health span approximates life span amongst supercentenarians: Compression of morbidity at the approximate limit of life span. J Gerontol Biol Sci. 67, 395–405 (2012). 8. Zeng, Y. et al. Health consequences of familial longevity influence among the Chinese elderly. J. Gerontol. A. Biol. Sci. Med. S p p y pp p 8. Zeng, Y. et al. Health consequences of familial longevity influence among the Chinese elderly. J. Gerontol. A. Biol. Sci. Med. Sci. 68, 473–82 (2013). 9. Terry, D. F. et al. Reduced all-cause, cardiovascular and cancer mortality in centenarian offspring. J. Amer. Geriatr. Soc. 52, 2074–2076 (2004).i 0. Deelen, J. et al. Genome-wide association meta-analysis of human longevity identifies a novel locus conferring survival beyond 90 years of age. Hum. Mol. Genet. 23, 4420–4432 (2014). y g 11. Newman, A. B. et al. A meta-analysis of four genome-wide association studies of survival to age 90 years or older: the Cohorts for Heart and Aging Research in Genomic Epidemiology Consortium. J. Gerontol. A Biol. Sci. Med. Sci. 65, 478–487 (2010).i y g 11. Newman, A. B. et al. A meta-analysis of four genome-wide association studies of survival to age 90 years or older: the Cohorts for Heart and Aging Research in Genomic Epidemiology Consortium. J. Gerontol. A Biol. Sci. Med. Sci. References Towards Deeper Research and Better Policy for Healthy Aging—Using the Unique Data of Chinese Longitudinal Healthy Longevity Survey. China Economic Journal 5, 131–149 (2012). g y y 7. Goodkind, D. Review on the book Healthy Longevity in China: Demographic, Socioeconomic, and Psychological Dimensions Population Studies 63, 1–7 (2009).f 8. Zeng, Y. et al. Effects of FOXO genotypes on longevity: a biodemographic analysis. J. Gerontol. A Biol. Sci. Med. Sci. 65, 1285–1299 (2010). ( ) 49. Zeng, Y. et al. Interactions between Social/behavioral factors and ADRB2 genotypes may be associated with health at advanced ages in China. BMC geriatrics 13, 91–91 (2013).i 50. Zeng, Y. et al. GxE Interactions between FOXO Genotypes and Tea Drinking Are Significantly Associated with Cognitive Disability at Advanced Ages in China. J. Gerontol. A Biol. Sci. Med. Sci. 70(4), 426–433. doi: 10.1093/gerona/glu060. (2015). 1. Li, Y. et al. Genetic association of FOXO1A and FOXO3A with longevity trait in Han Chinese populations. Human molecula genetics 18, 4897–4904 (2009).ffi g 52. Zhao, L. et al. Common genetic variants of the β 2-adrenergic receptor affect its translational efficiency and are associated with human longevity. Aging Cell, 11, 1094–1101 (2012).hf g y g g , , ( ) 53. Coale, A. J. & Li, S. The effect of age misreporting in China on the calculation of mortality rates at very high ages. Demography 28, 293–301 (1991). 53. Coale, A. J. & Li, S. The effect of age misreporting in China on the calculation of mortality ra 293–301 (1991). ( ) 54. Wang, Z. et al. Age validation of Han Chinese centenarians. GENUS 54, 123–141 (1998). 54. Wang, Z. et al. Age validation of Han Chinese centenarians. GE 55. Zeng, Y. & Gu, D. Reliability of age reporting among the Chinese oldest-old in the CLHLS datasets. In: Zeng, Y. et al. (eds.). Healthy longevity in China: Demographic, socioeconomic, and psychological dimensions. pp. 61–80. Dordrecht, The Netherlands: Springer Publisher, (2008). 56. Devlin, B. & Roeder, K. Genomic control for association studies. Biometrics 55, 997–1004 (1999). 57. Price, A. L. et al. Discerning the ancestry of European Americans in genetic association studies. PLoS genetics 4, e236 (2008). h l d d h l l et al. Discerning the ancestry of European Americans in genetic as 58. Chen, Y. et al. Common variants near ABCA1 and in PMM2 are associated with primary open-angle glaucoma. Nature genetic 1115–1119 (2014). 59. Wise, A. L. et al. References Coagu at o a d act vat o ol a ato y pat ways t e deve op e t o u ct o a decline and mortality in the elderly. Am. J. Med. 114, 180–187 (2003). 22 Y J l C SNP l i l i f h h i bili f h h i h N G 42 6 69 (2010) y y ( ) 22. Yang, J. et al. Common SNPs explain a large proportion of the heritability for human height. Nat. Genet. 42, 565–569 (2010).i 3. Broer, L. et al. GWAS of Longevity in CHARGE Consortium Confirms APOE and FOXO3 Candidacy. J. Gerontol. A. Biol. Sci. Med Sci. 70, 110–118 (2015).i 4. Zhang, K. et al. i-GSEA4GWAS: a web server for identification of pathways/gene sets associated with traits by applying an improved gene set enrichment analysis to genome-wide association study. Nucleic Acids Res. 38, W90–W95 (2010).i 5. Lambert, J. C. et al. Meta-analysis of 74, 046 individuals identifies 11 new susceptibility loci for Alzheimer’s disease. Nat. Genet. 45 1452–1458 (2013). ( ) 26. Karch, C. M. et al. Alzheimer’s disease genetics: from the bench to the clinic. Neuron. 83, 11–26 (2014). 27. Asai, T. et al. MAP kinase signaling cascade in Arabidopsis innate immunity. Nature 415, 977–983 (2002). 28. Naumova, E., et al. Immunogenetics of ageing. Int. J. Immunogenet. 38, 373–381 (2011).h 29. Apfeld, J. et al. The AMP-activated protein kinase AAK-2 links energy levels and insulin-like signals to lifespan in C. elegans. G Dev. 18, 3004–3009 (2004). 30. Anisimov, V. N. et al. Metformin slows down aging and extends life span of female SHR mice. Cell Cycle 7, 2769–2773 (2008).h 30. Anisimov, V. N. et al. Metformin slows down aging and extend 31. Yao, Y. et al. Proteasomes, Sir2, and Hxk2 Form an Interconnected Aging Network That Impinges on the AMPK/Snf1-Regu Transcriptional Repressor Mig1. PLoS Genet. 11, e1004968 (2015). 32. Li, X. et al. Integrated genomic approaches identify major pathways and upstream regulators in late onset Alzheimer’s disease. Sci. Rep. 5, 12393 (2015).ff p ( ) 33. Semchyshyn, H. M. et al. Fructose and glucose differentially affect aging and carbonyl/ oxidative stress parameters in Saccharomyces cerevisiae cells. Carbohydr. Res. 346, 933–938 (2011).ii y 4. Troen, A. M. et al. Lifespan modification by glucose and methionine in Drosophila melanogaster fed a chemically defined diet. Age 29, 29–39 (2007). 35. Zhang, H. et al. References Calcium signaling, excitability, and synaptic plasticity defects in mouse model of Alzheimer’s disease. J. Alzheimers. Dis. 45, 561–580 (2015).t 36. Tan, Q. H., Christensen, k. & Christiansen, L. Epigenetic drift in the aging genome: a ten-year follow-up in an elderly cohort of twins. Paper presented at the International Conference on “Interdisciplinary Research on Long-term Care and Healthy Aging”, May 22-23, 2015, Hangzhou, China. (2015). g 37. He, K. et al. Dynamic regulation of genetic pathways and targets during aging in Caenorhabditis elegans. AGING 6, 215–230 (2 h l f h l h d l d ( 37. He, K. et al. Dynamic regulation of genetic pathways and targets during aging in Caenorhabditis elegans. AGING 6, 215–230 (2014). 38. Kuhn, M. et al. STITCH 4: integration of protein-chemical interactions with user data. Nucleic. Acids. Res. 42, D401–D407 (2014). 37. He, K. et al. Dynamic regulation of genetic pathways and targets during aging in Caenorhabditis elegans. AGING 6, 215–230 (2014). 38. Kuhn, M. et al. STITCH 4: integration of protein-chemical interactions with user data. Nucleic. Acids. Res. 42, D401–D407 (2014). y g g p y g g g g g 38. Kuhn, M. et al. STITCH 4: integration of protein-chemical interactions with user data. Nucleic. Acids. Res. 42, D401–D407 (2014). . Kuhn, M. et al. STITCH 4: integration of protein-chemical intera l l l f g p 9. Troemel, E. R. et al. p38 MAPK regulates expression of immune response genes and contributes to longevity in C. elegans. PLoS Genet. 2, e183 (2006). 40. Boyle, A. P. et al. Annotation of Functional Variation in Personal Genomes Using RegulomeDB. Genome Research 22, 1790–1797 (2012). 41. Montgomery, S. B. et al. Transcriptome genetics using second generation sequencing in a Caucasian population. Nature 464, 773–777 (2010). 42. Rebeck, G. W. et al. Reduced apolipoprotein epsilon 4 allele frequency in the oldest old. Alzheimer’s patients and cognitively no individuals. Neurology 44, 1513–1516 (1994).h 43. Kirkwood, T. B. L. & Kowald, A. The free‐radical theory of ageing–older, wiser and still alive. Bioessays 34, 692–700 (2012). d h l h f ( ) h 44. Baudisch, A. & Vaupel, J. W. Getting to the root of aging. Science 338, 618 (2012). 5. Sun, F. et al. Extended maternal age at birth of last child and women’s longevity in the Long Life Family Study. Menopause 22, 26–31 (2015). 6. Zeng, Y. References 65, 478–487 (2010).i g g gy 12. Jeune, B. In search of the first centenarians. In: Jeune, B. & Vaupel, J. (eds.) Exceptional longevity: from prehistory to the present. Odense University Press. 11–24 (1995). 12. Jeune, B. In search of the first centenarians. In: Jeune, B. & Vaupel, J. (eds.) Exceptional longevity: from prehistory to the present. Odense University Press. 11–24 (1995). y 3. Xu, S. & Jin, L. A Genome-wide Analysis of Admixture in Uyghurs and a High-Density Admixture Map for Disease-Gene Discovery Am. J. Hum. Genet. 83, 322–336 (2008). ( ) 14. Weir, B. S. & Hill, W. G. Estimating F-statistics. Annu. Rev. Genet. 36, 721–750 (2002). 14. Weir, B. S. & Hill, W. G. Estimating F-statistics. Annu. Rev. Genet. 36, 721–750 (2002). 14. Weir, B. S. & Hill, W. G. Estimating F-statistics. Annu. Rev. Genet. 36, 721–750 (2002). 15 X S l G i di i f l i b f H Chi d i i li i i i i di A J H 15. Xu, S. et al. Genomic dissection of population substructure of Han Chinese and its implication in association studies. Am. J. H Genet. 85, 762–774 (2009). 6. Purcell, S. et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet. 81 559–575 (2007). 7. Jia, P. et al. Network-assisted investigation of combined causal signals from genome-wide association studies in schizophrenia. PLoS Comput Biol. 8, e1002587 (2012). Comput Biol. 8, e1002587 (2012). 18. Soerensen, M. et al. Evidence from case–control and longitudinal studies supports associations of genetic variation in APOE, CETP, and IL6 with human longevity. Age 35, 487–500 (2013). g y g , ( ) 19. Christiansen, L. et al. Paraoxonase 1 polymorphisms and survival. Eur J Hum Genet. 12, 843–847 (2004). 19. Christiansen, L. et al. Paraoxonase 1 polymorphisms and survival. Eur J Hum Genet. 12, 843 847 (2004). 20. Bonafè, M. et al. A gender—dependent genetic predisposition to produce high levels of IL-6 is detrimental for longevity. Eur. J. Immunol. 31, 2357–2361 (2001). Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 11 www.nature.com/scientificreports/ 21. Cohen, H. J., Harris, T. & Pieper, C. F. Coagulation and activation of inflammatory pathways in the development of funct decline and mortality in the elderly. Am. J. Med. 114, 180–187 (2003). . Co e , . J., a s, . & epe , C. . Acknowledgements g This study is supported by the National Key Basic Research Program of China (2013CB530700) (XLT, TN, YZ), National Natural Science Foundation of China (71233001, 71490732) (YZ), NIA/NIH (R01 AG023627, YZ; Pepper Center Grant P30AG028716, ERH; U19AG023122, TP, PS). The EU GWAS on longevity was performed by the EU Longevity Genetics Consortium undertaken within the framework of European Union’s Seventh Framework Programme (FP7/2007-2011) under grant agreement no 259679 (IDEAL). We are grateful to Xiaomin Shi, Zhaoxue Yin, Jieshi Luo and Yuebin Lv from China Center for Diseases Control and Prevention; Yuzhi Liu, Chunyuan Zhang, Yun Zhou and Zhenzhen Zheng from Peking University; Zhenyu Xiao, Liqun Tao, Qin Xu and Ye Yuan from Chinese Center for Aging Science Research; and Jie Zhan from China Social Science Academy for their contributions to the Chinese Longitudinal Healthy Longevity Survey (CLHLS) and the DNA sample collections, and we thank all interviewees and their families for their voluntary participation in thr CLHLS study. We thank Xiaopan Liu, Qiuyan Liao, Jingzhi Chen, Hongcheng Zhou and Mingrui Li from China National GeneBank-Shenzhen for their contributions of sample storage, DNA extractions and genotyping. We are grateful for the thoughtful comments provided by Jessica Sautter. We are grateful for the following scholars from EU Longevity Genetics Consortium (including GEHA project consortium) for their contributions to the EU longevity GWAS database which was used for the look-up/comparative analysis across Han Chinese, European and U.S. ethnicities in this article: Erik B. van den Akker, Maris Alver, Karen Andersen-Ranberg, Kristin L. Ayers, Marian Beekman, Anna M. Bennet, Vladyslav Bezrukov, Hélené Blanché, Stefan Böhringer, Dorret I. Boomsma, Marie E. Breen, Ruud van der Breggen, Linda Broer, Amke Caliebe, Miriam Capri, Elisa Cevenini, Joanna C. Collerton, Heather J. Cordell, Anton J.M. de Craen, Paolina Crocco, Serena Dato, Karen Davies, Luca Deiana, Cornelia M. van Duijn, Krista Fischer. Friederike Flachsbart, Ian Ford, Pilar Galan, Jutta Gampe, Paolo Garagnani, Eco J.C. de Geus, Efsthatios Gonos, Daníel F. Guðbjartsson, Jennifer Harrow, Diana van Heemst, Bastiaan T. Heijmans, Femke-Anouska Heinsen, Quinta Helmer, Antti Hervonen, Albert Hofman, Jouke-Jan Hottenga, Jeanine J. Houwing-Duistermaat, Bernard Jeune, Palmi V. Jonsson, J.Wouter Jukema, Yoichiro Kamatani, Thomas B.L. Kirkwood, Peter Kristensen, Mark Lathrop, Doris Lechner, Alberta Leon, Patrik Magnusson, Carmen Martin-Ruiz, Susan E. Mcnerlan, Andres Metspalu, Evelin Mihailov, Alberto Montesanto, Simon P. Mooijaart, Anne Murphy, Almut Nebel, Christopher P. Nelson, Ellen A. Acknowledgements Nohr, Giuseppe Passarino, Lavinia Paternoster, Pier Giuseppe Pelicci, Markus Perola, Iris Postmus, Michel Poulain, Irene Maeve Rea, Josè Remacle, Fernando Rivadeneira, Jean Marie Robine, Giuseppina Rose, Owen A. Ross, Stefano Salvioli, Nilesh J. Samani, Naveed Sattar, Gary I. Saunders, Stefan Schreiber, Ewa Sikora, Thorkild I.A. Sørensen, Liana Spazzafumo, Hreinn Stefánsson, Kári Stefánsson, David J. Stott, Riin Tamm, Henning Tiemeier, Olivier Toussaint, Stella Trompet, Hae-Won Uh, André G. Uitterlinden, Alexander Viktorin, Rudi G.J. Westendorp, Gonneke Willemsen. References 5 S lli P F & W i ht F A QTL A h bl d t b f h QTL Bi i f ti 28 451 452 (2012) i 73. Kuhn, M. et al. STITCH 4: integration of protein-chemical interactions with user data. Nucleic. Acids. Res. 42, D401–D407 (2014). 74 Th G t Ti E i C ti Th G t Ti E i (GTE ) j t N t G t 45 580 585 (2013) 3. Kuhn, M. et al. STITCH 4: integration of protein chemical interactions with user data. Nucleic. Acids. Res. 42, D401 D407 (2014). 4. The Genotype-Tissue Expression Consortium. The Genotype-Tissue Expression (GTEx) project. Nat Genet. 45, 580–585 (2013). 5. Sullivan, P. F. & Wright, F. A. seeQTL: A searchable database for human eQTLs. Bioinformatics 28, 451–452 (2012). Author Contributions Y.Z., H.Y., X.T., J.W.V., C.N. and E.H. designed the study. Y.Z., J.W.V., J.L. and J.X.L. designed and supervised CLHLS which recruited the participants. C.N., X.L., M.W.,Y.F.Z., Y.W., L.Z., Y.W. and Z.L. contributed to sample management and performed the genotyping. C.N., X.L., H.C., H.X., Y.Q., X.L., H.C., X.X., P.S. and Y.F.Z. performed bioinformatics quality controls and the statistical analysis. W.T., Y. L., H.Y. J.Z., C.S., L.C., H.W., J.D., G.Z. and W.T. performed the DNA extractions. H.Y., J.Z. and C.S. performed independent double-check of the main parts of the GWAS data analysis. Y.Z., C.N., J.M. and X.L. wrote the initial draft, with critical revisions by J.W.V., X.T., H.Y., E.H., M.L., J.M., Q.T., E.S., J.D., P.S., T.P., H.B., C.F., K.L., J.G., W.G., J.W., J.U.W., M.N., L.C. and Z.Y. All authors revised and gave final approval of this manuscript. References eXclusion: toward integrating the X chromosome in genome-wide association analyses. Am. J. Hum. Genet. 92, 643–647 (2013). 60. Marchini, J. et al. A new multipoint method for genome-wide association studies by imputation of genotypes. Nat. Genet. 39, 906–913 (2007). 61. Chen, Y. et al. Common variants near ABCA1 and in PMM2 are associated with primary open-angle glaucoma. Nature Genetic 1115–1119 (2014).i 62. International Multiple Sclerosis Genetics Consortium (IMSGC). Analysis of immune-related loci identifies 48 new susceptibility variants for multiple sclerosis. Nat Genet. 45, doi: 10.1038/ng.2770 (2013).i 63. Saxena, R. et al. Genome-wide association study identifies a novel locus contributing to type 2 diabetes susceptibility in Sikhs of Punjabi origin from India. Diabetes. 62, 1746–55 (2013). i Punjabi origin from India. Diabetes. 62, 1746–55 (2013). j g 4. Zeng, Y. & Vaupel, J. W. Association of late childbearing with healthy longevity among the oldest-old in China. Population Studies 58, 37–53 (2004).fi 65. Willer, C. J., Li, Y. & Abecasis, G. R. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics. 26, 2190–2191 (2010). ( ) 66. Woolf, B. On estimating the relation between blood group and disease. Ann. Hum. Genet. 19, 251–253 (1955). Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 12 www.nature.com/scientificreports/ 67. K at , . et al. e yea s o pat way a a ys s: cu e t app oac es a d outsta d g c a e ges. oS Comput iol. 8, e 00 375 ( 0 ). 68. Zhang, K. et al. i-GSEA4GWAS: a web server for identification of pathways/gene sets associated with traits by applying an improved gene set enrichment analysis to genome-wide association study. Nucleic Acids Res. 38, W90–W95 (2010). p g p y g g 70. Kanehisa, M. & Goto, S. KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res. 28, y y oshi-Tope, G. et al. Reactome: a knowledgebase of biological pathw p g g p y 72. Ashburner, M. et al. Gene ontology: tool for the unification of biology. Nat. Genet. 25, 25–29 (2000) gyi gy 3. Kuhn, M. et al. STITCH 4: integration of protein-chemical interactions with user data. Nucleic. Acids. Res. 42, D401–D407 (2014). 4. The Genotype-Tissue Expression Consortium. The Genotype-Tissue Expression (GTEx) project. Nat Genet. 45, 580–585 (2013). Additional Information Supplementary information accompanies this paper at http://www.nature.com/srepi Competing financial interests: The authors declare no competing financial interests. Competing financial interests: The authors declare no competing financial interests. How to cite this article: Zeng, Y. et al. Novel loci and pathways significantly associated with longevity. Sci. Rep. , 21243; doi: 10.1038/srep21243 (2016). How to cite this article: Zeng, Y. et al. Novel loci and pathways significantly associated with longevity. Sci. Rep. , 21243; doi: 10.1038/srep21243 (2016). This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ Scientific Reports \| 6:21243 \| DOI: 10.1038/srep21243 13
https://openalex.org/W3212563905	https://hal-cea.archives-ouvertes.fr/cea-03533664/document	English	null	Deep transfer learning for blended source identification in galaxy survey data	Astronomy & astrophysics	2,022	cc-by	9,881	Deep transfer learning for blended source identification in galaxy survey data S. Farrens, A. Lacan, A. Guinot, A. Z. Vitorelli To cite this version: S. Farrens, A. Lacan, A. Guinot, A. Z. Vitorelli. Deep transfer learning for blended source identification in galaxy survey data. Astronomy and Astrophysics - A&A, 2022, 657, pp.A98. 10.1051/0004- 6361/202141166. hal-03533667 To cite this version: S. Farrens, A. Lacan, A. Guinot, A. Z. Vitorelli. Deep transfer learning for blended source identification in galaxy survey data. Astronomy and Astrophysics - A&A, 2022, 657, pp.A98. 10.1051/0004- 6361/202141166. hal-03533667 1. Introduction biases can be introduced by unidentiﬁed blends coming from sources below the detection threshold (Hoekstra et al. 2017; Euclid Collaboration 2019). It is therefore necessary to develop an appropriate procedure for dealing with blends in survey data. The blending of sources, that is, the apparent overlapping of extended objects in 2D images, has a signiﬁcant impact on measurements of the morphological and structural properties of galaxies, in particular for ground-based surveys. Objects in close proximity to one another can easily be mistaken for a single source, which can lead to signiﬁcant detection and/or measure- ment biases depending on the depth of the survey. The process of handling blended sources in astrophysical images can be broadly divided into four major problems: (a) the detection of objects in the image, (b) the classiﬁcation of those objects as either single sources or blended sources, (c) the segmentation of pixels from blended sources (often referred to as ‘deblending’), and (d) the rejection of those objects that cannot be easily included in the scientiﬁc analysis. This paper focuses on the problem of classifying objects already detected with standard source extraction software as either blends or sin- gle sources. Unlike the problem of segmentation, which has been abundantly addressed in the literature (Joseph et al. 2016; Melchior et al. 2018; Reiman & Göhre 2019), little eﬀort has been made to ﬁnd reliable and automated methods for identi- fying blended sources in survey data. For weak gravitational lensing analysis, it is essential to understand how and to what degree blending impacts shear and photometric redshift measurements (Mandelbaum 2018). This problem becomes even more critical in the current epoch of high- precision cosmology, where systematic eﬀects need to be care- fully accounted for. This is particularly important for reliable comparisons between late-time probes, like weak lensing, and early-time probes, such as the CMB. Blended sources make up a signiﬁcant fraction of the observed sources in survey images: >30% for the Dark Energy Survey (DES, Samuroﬀet al. 2018), >50% (up to >60% and >70% for the Deep and UltraDeep layers, respectively) for Hyper Suprime-Cam (HSC, Bosch et al. 2018), and >60% for Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST, Sanchez et al. 2021). Simply removing all the blends that have been identiﬁed would signiﬁcantly reduce the sam- ple size and could potentially lead to entangled biases in the shear correlation function (Hartlap et al. 2011). Open Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ⋆In the spirit of reproducible research, all code and data needed to reproduce the results in this paper have been made publicly available on GitHub (https://github.com/CosmoStat/BlendHunter) without any restrictions. ABSTRACT We present BlendHunter, a proof-of-concept deep-transfer-learning-based approach for the automated and robust identiﬁcation of blended sources in galaxy survey data. We take the VGG-16 network with pre-trained convolutional layers and train the fully connected layers on parametric models of COSMOS images. We test the eﬃcacy of the transfer learning by taking the weights learned on the parametric models and using them to identify blends in more realistic Canada-France Imaging Survey (CFIS)-like images. We compare the performance of this method to SEP (a Python implementation of SExtractor) as a function of noise levels and the separation between sources. We ﬁnd that BlendHunter outperforms SEP by ∼15% in terms of classiﬁcation accuracy for close blends (<10 pixel separation between sources) regardless of the noise level used for training. Additionally, the method provides consistent results to SEP for distant blends (≥10 pixel separation between sources) provided the network is trained on data with noise that has a relatively close standard deviation to that of the target images. The code and data have been made publicly available to ensure the reproducibility of the results. Key words. techniques: image processing – methods: numerical – methods: data analysis – gravitational lensing: w Deep transfer learning for blended source identiﬁcation in galaxy survey data⋆ S. Farrens1 , A. Lacan2 , A. Guinot1, and A. Z. Vitorelli1 1 AIM, CEA, CNRS, Université Paris-Saclay, Université Paris Diderot, Sorbonne Paris Cité, 91191 Gif-sur-Yvette, France e-mail: samuel.farrens@cea.fr 1 AIM, CEA, CNRS, Université Paris-Saclay, Université Paris Diderot, Sorbonne Paris Cité, 91191 Gif-sur-Yvette, France e-mail: samuel.farrens@cea.fr niversité Paris-Saclay, Université Paris Diderot, Sorbonne Paris Cité, 91191 Gif-sur-Yvette, France 2 Université Paris-Saclay, CNRS, CEA, Astrophysique, Instrumentation et Modélisation de Paris-Saclay, 91191 Gif-sur-Yvette, France is-Saclay, CNRS, CEA, Astrophysique, Instrumentation et Modélisation de Paris-Saclay, 91191 Gif-sur-Yvette, 2 Université Paris-Saclay, CNRS, CEA, Astrophysique, Instrumentation et Modélisation de Paris-Saclay, 91191 Gif-sur-Yvette, France Received 23 April 2021 / Accepted 13 October 2021 Received 23 April 2021 / Accepted 13 October 2021 HAL Id: hal-03533667 https://hal.science/hal-03533667v1 Submitted on 18 Jan 2022 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Astronomy & Astrophysics A&A 657, A98 (2022) https://doi.org/10.1051/0004-6361/202141166 c⃝S. Farrens et al. 2022 Astronomy & Astrophysics A&A 657, A98 (2022) https://doi.org/10.1051/0004-6361/202141166 c⃝S. Farrens et al. 2022 2.1. Blend deﬁnition In order to label our training data, we ﬁrst needed to make a clear deﬁnition of a blend. Typically, this would be some measure of the amount of overlap between the light proﬁles of the individual sources that constitute the blend. For the purposes of this work, we assume a simple scenario in which isolated (i.e., unblended) sources are composed of a single galaxy centred within a postage stamp of size 51 × 51 pixels (9.5 arcsec × 9.5 arcsec). We then deﬁne a blend as a postage stamp (of the same size) contain- ing two galaxies, one at the centre and a secondary source at a random position. This deﬁnition is fairly agnostic in that we do not require any speciﬁc overlap between the light proﬁles. This choice was made in order to gauge classiﬁcation accuracy as a function of distance between the two sources. Transfer learning, in particular deep transfer learning, is a machine-learning approach whereby network weights obtained by training on a given data set are applied to another distinct but similar data set. This can help to prevent over-ﬁtting to the train- ing data and can signiﬁcantly reduce the time required to apply the method to new data sets. Deep transfer learning has been used in a variety of problems in astrophysics in recent years, including: the classiﬁcation of compact star clusters (Wei et al. 2020), the separation of low-surface-brightness galaxies from artefacts (Tanoglidis et al. 2021), the classiﬁcation of planetary nebulae (Awang Iskandar et al. 2020), and the deblending of galaxy images (Arcelin et al. 2021). Given the size of the postage stamps, we expect that standard source-extraction software, such as SExtractor, will be easily able to identify and separate sources near the borders but may struggle as the sources get closer together. To better highlight this speciﬁc problem, we additionally separate blends into two categories: ‘close blends’ and ‘distant blends’. We deﬁne close blends as postage stamps in which the two sources are separated by less than ten pixels and distant blends as those in which they are separated by ten or more pixels. This work introduces a proof-of-concept deep transfer learn- ing approach, called BlendHunter (BH), for the automated and robust identiﬁcation of blended sources in single-band galaxy survey data. This method incorporates a convolutional neural network (CNN) trained on a large database of natural images and a fully connected layer for classiﬁcation. 1. Introduction Additional Traditional methods for identifying blended sources, such as SExtractor (Bertin & Arnouts 1996), rely on ﬁxed thresh- olds to detect multiple peaks in the light intensity proﬁles of the objects. While this approach may work for a reasonable fraction of the sources, it lacks the ﬂexibility needed to handle blends with a greater discrepancy in the brightness and/or size between the individual objects. This means that a large number of blended sources, which may have a non-negligible impact on the scien- tiﬁc analysis, could be missed. Machine-learning techniques, in particular deep-learning architectures, have been shown to be incredibly success- ful when applied to complicated classiﬁcation problems (see e.g., Kotsiantis 2007; Lecun et al. 2015; Srinivas et al. 2016). ⋆In the spirit of reproducible research, all code and data needed to reproduce the results in this paper have been made publicly available on GitHub (https://github.com/CosmoStat/BlendHunter) without any restrictions. A98, page 1 of 9 A&A 657, A98 (2022) However, the eﬀectiveness of these tools can be diﬃcult to gauge without reliably labelled data. In real astrophysical images, it is not known a priori if the underlying signal for a given detection comes from a single object or indeed from the combination of several. This necessitates the use of simulated data in order to generate a reliable training set, where the number and diversity of overlapping sources are perfectly known. Nevertheless, this introduces the possibility of over-ﬁtting the network to proper- ties speciﬁc to the simulation, i.e., overly simpliﬁed galaxy mod- els. This in turn can cause the network to be very sensitive to artefacts and more complex structures seen in real observations. Developing simulated galaxy images that contain all of the prop- erties expected in real data is extremely challenging and time consuming. Therefore, the application of networks well suited to transfer learning represents an interesting approach to miti- gate this problem. ally prepared a sample of realistic CFIS-like galaxy images to test the eﬃcacy of transferring the learned weights to a similar but unseen data set. This section details how the training and testing data sets were generated. 2.2. The COSMOS catalogue We use the Cosmological Evolution Survey (COSMOS, Scoville et al. 2007) catalogue as the basis from which to derive our simplistic training data and more realistic testing data. COS- MOS was chosen as it provides a representative sample of galax- ies in terms of size, ellipticity, luminosity, and morphology. COSMOS is a catalogue of Hubble Space Telescope observa- tions of 1.64 deg2 on the sky with very accurate photometry and morphology. This catalogue contains high-resolution images (0.05 arcsec pixel−1) with a very small point spread function (PSF) of 0.01 arcsec. 2.1. Blend deﬁnition Simple parametric models are used to train the fully connected layers and the learned weights are in turn used to identify blended sources in more realistic images. To avoid any spurious correlations, blended images are pro- duced by simulating the galaxies individually and then artiﬁ- cially combining them. For simplicity, we only consider blends consisting of two sources. Real images may well contain cases involving more than two sources as well as various artefacts that would make classiﬁcation more complicated. We leave these issues to be addressed in future work and here focus on testing the applicability of transfer learning to our simple test case. This paper is organised as follows. The following section introduces the properties of the training and testing data. Section 3 presents the transfer learning approach used and how it was trained to identify blended galaxy images. Section 4 pro- vides results on how this approach compares to the state of the art. Finally, conclusions are presented in Sect. 5. 1 https://www.cfht.hawaii.edu/Science/CFIS/ 3. Deep transfer learning framework for blend identiﬁcation The objective of transfer learning is to learn a set of weights from a given training set and then to transfer these to a related but independent problem. This is interesting from the perspective of blend identiﬁcation given that we do not have a large sample of labelled data and training with a small sample of known blends may signiﬁcantly bias the learned weights. Additionally, the use of pre-trained weights considerably reduces the time required to train a network, thereby making the application of a deep learn- ing approach to galaxy survey data more feasible. For half of the sample, a second galaxy model is placed at a random position within the postage stamp to produce blends according to the deﬁnition provided in Sect. 2.1. We pad all of the images with zeros (7 pixels in every direction) to avoid issues when the secondary source is close to the border of the postage stamp. The complete sample is comprised of 80 000 noiseless, padded postage stamps, half of which are isolated galaxies and the other half are blends. For the purposes of this work we test the applicability of transfer learning to the problem of blend identiﬁcation in a two-step process. In the ﬁrst step, CNN weights learned from natural images are used to extract generic features from the galaxy images. The fully connected layers are then trained on galaxy image features derived from simple parametric models. The objective being to capture the general properties of blended images and not speciﬁc features of the individual galaxies. In the second step, the weights learned from simple parametric models are applied to more realistic CFIS-like images to test the classi- ﬁcation accuracy. The ﬁnal step in producing our simulated training sets is to add random Gaussian noise. In order to test how sensitive blend identiﬁcation is to noise, we generate seven diﬀerent noise stan- dard deviations (σnoise = 5, 10, 15, 20, 25, 30, 35). We addition- ally created ten realisations of the noise for each value of σnoise in order to test the stability of the training. Each realisation of each noise level is treated as an independent training set. In other words, we train the network 70 times and obtain 70 sets of weights. The architecture we choose to implement our deep transfer problem was that of VGG-16. 2.4. Parametric model training data In order to keep our training samples as simple and generic as possible, we use a series of parametric models derived from ﬁts to the COSMOS sample described in Sect. 2.2. This constitutes a range of simulated galaxies with diﬀerent sizes, shapes, and ellipticities. The parametric models are then convolved with the CFIS-like PSF described in Sect. 2.3. Each image corresponds to a 51 × 51 pixel postage stamp with a convolved galaxy model at the centre. 3. Deep transfer learning framework for blend identiﬁcation For simplicity we refer to our spe- ciﬁc VGG-16 setup for the problem of blend identiﬁcation as BlendHunter. 2.3. Point spread function The impulse response or PSF of an instrument encompasses all of the aberrations induced by the optical system along with other distortions arising from the atmosphere, and so on. The PSF induces blurring in observed images, which artiﬁcially increases the size of sources and can cause their light proﬁles to overlap creating blends. random position in the postage stamp and then all the postage stamps are zero padded in the same way. Finally, Gaussian noise with σnoise = 14.5 was added to the images to replicate the S/N of CFIS data. We model the optical PSF as a Moﬀat proﬁle with β = 4.765 (Trujillo et al. 2001) and the PSF ellipticity is drawn from real optical variations of the Canada-France-Hawaii Tele- scope (CFHT) derived from real CFIS data (Guinot et al. 2021). Atmospheric turbulence is modelled by a Kolmogorov proﬁle (Tatarski 2016) with random ellipticity drawn from a Gaussian distribution with average µ = 0 and standard deviation σ = 0.01. The ﬁnal PSF is obtained by convolving the two models and has an average size of 0.65 arcsec. Given the relatively small size of the postage stamps, spatial variations of the PSF are neglected. This CFIS-like PSF model is simple but suﬃciently realistic for the purposes of this work. The ﬁnal sample of 80 000 realistic postage stamps is con- siderably more complex than the parametric models described in Sect. 2.4 and more closely approximate the conditions expected in real images. For example, the light proﬁles do not necessarily have central symmetry because of star forming regions or com- plex morphology. Out of the 80 000 postage stamps, 4838 cor- respond to close blends (i.e., the sources are less than ten pixels apart) and the remaining 35 162 correspond to distant blends. Table 1 summarises all of the data sets used for training and testing. Nisolated is the number of postage stamps containing iso- lated sources, Nblended is the number of postage stamps contain- ing blended sources, σnoise is the amount of noise added to the postage stamps, and Nσ real is the number of realisations for each noise level. 2. Data Data set Parametric training set Realistic testing set Nisolated 40 000 40 000 Nblended 40 000 40 000 σnoise 5, 10, 15, 20, 25, 30, 35 14.5 Nσ real 10 1 Table 1. Summary of the data used for training and testing. extensively in several weak lensing challenges, such as GREAT3 (Mandelbaum et al. 2014). The package provides all the tools necessary to work with COSMOS data. 2. Data Supervised machine-learning approaches, such as that presented in this work, require accurately labelled and representative train- ing data. These labels often correspond to properties that cannot be directly and/or reliably measured from observed data. Simula- tions, on the other hand, provide a controlled environment where these properties are known a priori. ( ) In particular, we use a subset of this catalogue selected and processed for weak lensing purposes (Mandelbaum et al. 2012). This subset of images have a negligible amount of noise and the PSF has been deconvolved. This makes it possible to re-sample the images on a larger pixel scale and to convolve them with a diﬀerent PSF in order to mimic observations from another instru- ment. Finally, the galaxies have been ﬁtted by either a Sérsic pro- ﬁle or a bulge + disc proﬁle for which the bulge is represented by a De Vaucouleurs proﬁle (de Vaucouleurs 1948) and the disc by an exponential proﬁle (see Mandelbaum et al. 2012, for details). The subset of processed COSMOS data has been made available through the Galsim software package (Rowe et al. 2015). GalSim is widely used in the astrophysics commu- nity to simulate and manipulate galaxy images and was used To train our network, we opted to produce a set of simulated single-band (i.e., monochromatic) galaxy images from a para- metric model. This allowed us to control the fraction of blended images in the whole sample and the separation of sources in the blended images. The choice of single-band images was made to restrict the learning to pixel features (i.e., no colour informa- tion), which could potentially be of interest for the r-band of the Canada-France Imaging Survey (CFIS)1, part of the Ultraviolet Near-Infrared Optical Northern Survey (UNIONS), or eventu- ally the Euclid visible band (Cropper et al. 2012). We addition- The subset of processed COSMOS data has been made available through the Galsim software package (Rowe et al. 2015). GalSim is widely used in the astrophysics commu- nity to simulate and manipulate galaxy images and was used 1 https://www.cfht.hawaii.edu/Science/CFIS/ A98, page 2 of 9 S. Farrens et al.: Deep transfer learning for blended source identiﬁcation in galaxy survey data Table 1. Summary of the data used for training and testing. 2.5. Realistic testing data We generate a sample of realistic CFIS-like images to test our transfer learning approach. To do so, we take the real denoised and deconvolved COSMOS images described in Sect. 2.2, crop them to 51 × 51 pixel postage stamps, and convolve them with the CFIS-like PSF described in Sect. 2.3. The ﬂux is rescaled in order to reproduce a 300 s exposure at the 3.6m CFHT telescope. The images are re-sampled at the resolution of the CFIS survey (0.187 arcsec). Similarly to the training images, blends are cre- ated for half of the sample by adding a secondary source at a 3.2.2. Fully connected layers In the second phase, we train the fully connected layers (i.e., the last three weight layers) on top of the previously stored features. These layers consist of a layer with linear transformation, fol- lowed by a layer with dropout and ﬁnally a layer with rectiﬁed linear unit (ReLu) activation. We chose this non-linear activa- tion function, widely used with deep neural networks, because it allows for faster convergence. We used dropout alongside data augmentation to avoid over-ﬁtting (Srivastava et al. 2014). This is a way of cancelling some activations to prevent the network from learning random correlations in the images. We employed a dropout rate of 0.1 for our experiments. For the training, we pro- ceeded with batches of 250 images at a time over 500 epochs. An epoch corresponds to passing the entire training sample through the network and a batch size of 250 proved to be the most com- putationally eﬃcient for the gradient descent. The VGG-16 network was chosen for the work presented here for several reasons. Firstly, the network can be implemented with weights pre-trained on the ImageNet database (Deng et al. 2009) in order to save computation time and resources. The diversity of this data set has allowed the network to learn a variety of generic image features applicable to most image- classiﬁcation tasks. Finally, VGG-16 has already been applied in a variety of astrophysical problems, including galaxy mor- phology classiﬁcation (Wu et al. 2019; Zhu et al. 2019), glitch classiﬁcation in gravitational wave data (George et al. 2018), and coronagraph image classiﬁcation (Shan et al. 2020). The VGG- 16 network was implemented in Python via the TensorFlow- Keras neural network API (Chollet et al. 2015; Abadi et al. 2015), which includes ImageNet pre-trained weights. p y g We use the binary cross entropy loss function in Eq. (1) : BCE = −1 N N X i=1 yi · log(p(yi)) + (1 −yi) · log(1 −p(yi)), (1) (1) where yi is the true label (1 for blended, 0 otherwise) for the ith image and p(yi) is the probability returned by the network of the image being blended or not. We selected this loss function to optimise as it is typically used in this type of binary classiﬁcation problem. p The Adam algorithm (Kingma & Ba 2014) was used to minimise our loss function. 3.1. The VGG-16 network VGG-16 is a deep convolutional network with 16 weight lay- ers developed by the Visual Geometry Group (VGG) at the Uni- versity of Oxford (Simonyan & Zisserman 2014). The network was ranked ﬁrst in the ImageNet Large-Scale Visual Recognition A98, page 3 of 9 A98, page 3 of 9 A98, page 3 of 9 A&A 657, A98 (2022) 224 x 224 x 64 112 x 112 x 128 56 x 56 x 256 28 x 28 x 512 14 x 14 x 512 1 x 1 x 4096 7 x 7 x 512 1 x 1 x 1 Convolution + ReLU Max pooling Fully connected + ReLU Softmax Fig. 1. Visual representation of the VGG-16 network. Convolutional layers with ReLU activation are shown in solid blue, max pooling lay- ers are shown in brick-pattern red, fully connected layers are shown as green bars, and the output softmax layer is shown as the last box in pink. this part of the network is not trained with our data and can be seen as a simple feature extractor. As ImageNet weights are used, the VGG-16 network expects RGB images as inputs. Therefore, we rescaled the simulated images to the range i ∈Z : 0 ≤i ≤255 and the monochro- matic pixel values were repeated across the three RGB channels. The ﬁnal images are saved as portable network graphics (PNG) ﬁles which are fed into the network. Given the relatively small amount of training data, we addi- tionally implemented data augmentation in order to obtain more diversiﬁed features from the convolutional layers of the network. Augmenting our simulated images simply means creating addi- tional images with minor changes such as ﬂips, translations, or rotations. Speciﬁcally, we used the Keras image pre-processing modules to include a shear range and zoom range of 0.2, as well as horizontal ﬂipping. We note that data augmentation was not used on the test images. Fig. 1. Visual representation of the VGG-16 network. Convolutional layers with ReLU activation are shown in solid blue, max pooling lay- ers are shown in brick-pattern red, fully connected layers are shown as green bars, and the output softmax layer is shown as the last box in pink. g Figure 2 shows some examples of features extracted from one of the simulated images with the convolutional layers. The earlier convolutional blocks provide more general features while the later blocks provide more speciﬁc features. 3.2. Training and validation We trained BlendHunter in two phases using the 70 sets of 80 000 simulated images described in Sect. 2.4. We divided each data set in the following way: 36 000 images were used for train- ing, 36 000 for validation, and the remaining 8000 for testing. 3.2.2. Fully connected layers In contrast to a classic stochas- tic gradient descent (SGD) algorithm, it computes individual learning rates for each weight and is relatively eﬃcient in the optimisation of deep neural networks. In order to adapt the learning rate, Adam requires the estimations of the ﬁrst (the mean) and second (the uncentred variance) moments of the gra- dient using exponentially moving averages (Eqs. (2) and (3), respectively), with gt the objective function gradient at time- step t, and β1 and β2 the exponential decay rates for the moment estimates: 3.1. The VGG-16 network Challenge (ILSVRC) in 2014 (Russakovsky et al. 2015). The main feature of this architecture was the increased depth of the network compared to the state of the art at the time. In VGG-16, three-channel images (RGB) are passed through ﬁve blocks of convolutional layers, where each block is composed of increas- ing numbers of 3×3 ﬁlters. The stride (i.e., the amount by which the ﬁlter is shifted) is ﬁxed to 1, while the convolutional layer inputs are padded such that the spatial resolution is preserved after convolution (i.e., the padding is 1 pixel for 3 × 3 ﬁlters). The blocks are separated by max-pooling (i.e., down-sampling) layers. Max-pooling is performed over 2 × 2 windows with a stride of 2. The ﬁve blocks of convolutional layers are followed by three fully connected layers. The ﬁnal layer is a soft-max layer that outputs class probabilities. The full network architec- ture used is shown in Fig. 1. A98, page 4 of 9 3.2.1. Convolutional layers (4) blend identiﬁcation and many of these steps cannot easily be iso- lated; however we tried to make the comparison as fair and con- sistent as possible. At the end of each epoch, both training and validation losses were computed, and the weights were updated every time the validation loss decreased. Training was stopped when the vali- dation loss had not decreased after ten epochs. The network con- verged after around 70 epochs on average. SEP implements a multi-thresholding technique to decom- pose detected sources into subsources (when possible). This method takes two input parameters: the number of bins to decompose the light proﬁle and the minimum contrast value between the main peak and a given subpeak. The contrast is evaluated based on the ﬂux of each peak (see Fig. 2 in Bertin & Arnouts 1996). The set of parameters we used for SEP is shown in Table 2 (all the other parameters are kept to their default values). The value of 0.005 for DEBLEND_MINCONT (the minimum contrast parameter for deblending) may be considered a little high compared to that commonly used in the literature, but here we chose to favour reliable identiﬁcation over increas- ing the number of blends found at the cost of also increasing the number of spurious detections. We began training with a learning rate of 1×10−3. As choos- ing the right learning rate can be challenging, we decided to reduce the learning rate by a factor of 0.5 every time the val- idation loss did not decrease after ﬁve epochs. A small learn- ing rate would make it possible to avoid big jumps in gradient descent. Otherwise, in this case, it could fail to converge and set- tle around a local minimum. No weight decay was implemented in this phase. Tuning deep neural network hyperparameters with a consid- erable amount of parameters to learn can prove to be very time- consuming. This is why we focused on the hyperparameters that would have the most impact on the results. Several tests were made such as changes to the regularisation, weight initialisation, dropout rate, learning rate, and optimiser. However, no signif- icant improvement in accuracy was observed. Switching to the SGD optimiser or increasing the dropout rate led to worse per- formance overall. Given our loose deﬁnition of blended sources (see Sect. 2.1), we chose not to rely exclusively on the deblending ﬂags provided by SEP. 3.2.1. Convolutional layers In the ﬁrst phase, the we simply initialise the convolutional lay- ers (i.e., the ﬁrst 13 weight layers) with the pre-trained ImageNet weights. As mentioned in the previous section, the purpose of using pre-trained weights is to save processing time. Therefore, (2) mt = β1 · mt−1 + (1 −β1) · gt (2) vt = β2 · vt−1 + (1 −β2) · g2 t (3) mt = β1 · mt−1 + (1 −β1) · gt vt = β2 · vt−1 + (1 −β2) · g2 t (2) (3) (3) S. Farrens et al.: Deep transfer learning for blended source identiﬁcation in galaxy survey data Image (65x65) Block 1 (224x224) Block 2 (112x112) Block 3 (56x56) Block 4 (28x28) Block 5 (14x14) Fig. 2. Examples of features extracted from the VGG-16 convolutional layers pre-trained with ImageNet weights. Leftmost panel: input postage stamp containing blended sources. The following panels show example features extracted from various convolution blocks. Block 2 (112x112) Fig. 2. Examples of features extracted from the VGG-16 convolutional layers pre-trained with ImageNet weights. Leftmost panel: input postage stamp containing blended sources. The following panels show example features extracted from various convolution blocks. Fig. 2. Examples of features extracted from the VGG-16 convolutional layers pre-trained with ImageNet weights stamp containing blended sources. The following panels show example features extracted from various convolutio Table 2. SEP parameter settings. Table 2. SEP parameter settings. Parameter Value THRESH_TYPE RELATIVE DETECT_THRESH 1.5 DETECT_MINAREA 5 FILTER Y FILTER_NAME kernel_3x3.conv (default) DEBLEND_NTHRESH 32 DEBLEND_MINCONT 0.005 Adam takes advantage of both RMSprop (Tieleman & Hinton 2012) and AdaGrad (Duchi et al. 2011) methods. In addition to being robust and less time-consuming, it can therefore be applied to a wider selection of optimisation problems. It also requires almost no tuning of its parameters. We set β1 and β2 to their default values, respectively 0.9 and 0.999. Finally, the weights are updated according to Eq. (4), where wt are the fully con- nected network weights at time-step t, η is the step size, ˆmt and ˆvt are the bias-corrected estimators of the ﬁrst and second moments, and ϵ is a value set to 10−8 to prevent division by 0. wt = wt−1 −η · ˆmt √ˆvt + ϵ . (4) wt = wt−1 −η · ˆmt √ˆvt + ϵ . 4. Results If a single source is detected: If a single source is detected: – If the source is ﬂagged as a blend and is at the expected posi- tion, the image is labelled as a blend. 3.2.1. Convolutional layers Instead, we also check that the sources are found at the right positions (within a two pixel radius) to make sure we do not extract noise features. Additionally, as some of the sources in the postage stamp do not technically overlap, when SEP correctly identiﬁes the number of sources (i.e., 1 or 2) we take this as a correctly labelled postage stamp. The network takes approximately 630 s to train on a sample of 80 000 images using a standard Intel(R) Core(TM) i7-6900K CPU (3.20 GHz, 32GB). The process for labelling postage stamps as either isolated or blended sources using SEP can be summarised as follows: 4.1. SExtractor benchmark We compared the performance of BlendHunter with SExtractor, as the latter is the most widely used tool in the community for identifying and handling blended sources in astronomical images. The objective was to test the reliability of our approach compared to the state of the art. Speciﬁcally, we made use of the SExtractor Python wrapper SEP (Barbary 2016) for our tests. We note that SEP does many things beyond – Otherwise, the image is labelled as an isolated source. If two sources are detected: – If both sources are detected at the right positions, the image is labelled as a blend (this stands even if one of the two sources is itself detected as a blend). – If the sources are incorrectly identiﬁed, the image is labelled as an isolated source. A98, page 5 of 9 A98, page 5 of 9 A98, page 5 of 9 A&A 657, A98 (2022) A&A 657, A98 (2022) 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 5 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 10 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 15 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 20 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 25 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 30 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 35 BH SEP Gain on SEP Fig. 3. Classiﬁcation accuracy of BlendHunter (black solid line) vs. SEP (blue dashed line) as a function of separation between sources. Results are from the blended samples of the COSMOS parametric model testing sets. Each panel shows one realisation of a given noise standard deviation, σnoise. If more than two sources are detected: ing procedure described in Sect. 3.2 and apply them to the test sample of 8000 parametric model postage stamps (see Sect. 2.4) to obtain classiﬁcation labels. – If at least the right number of sources are detected at their expected positions, the image is labelled as a blend. – If at least the right number of sources are detected at their expected positions, the image is labelled as a blend. – If the number of detected sources is larger than the number of true sources, the image is ﬂagged as a SEP failure and is not included in the analysis. We note that this occurred for less than 1% of the images. Figure 3 shows the classiﬁcation accuracy of Blend Hunter (black solid line) versus SEP (blue dashed line) as a function of separation between sources. Results are from the blended samples of the COSMOS parametric model testing sets (i.e., sets of 4000 postage stamps). Each panel shows one realisa- tion of a given noise standard deviation, σnoise. The blue shaded area shows the gain in accuracy of BlendHunter with respect to SEP. For close blends (i.e., where the two sources are less than ten pixels apart), BlendHunter signiﬁcantly outperforms SEP, showing gains as high as ∼80% in classiﬁcation accuracy. A98, page 6 of 9 4.1. SExtractor benchmark The blue shaded area shows the gain in accuracy of BlendHunter with respect to SEP. 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 5 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 10 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 15 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 20 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 25 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 30 BH SEP Gain on SEP 0 5 10 15 20 25 30 35 Distance between objects (pixels) 0 20 40 60 80 100 Accuracy (%) noise = 35 BH SEP Gain on SEP Fig. 3. Classiﬁcation accuracy of BlendHunter (black solid line) vs. SEP (blue dashed line) as a function of separation between sources. Results are from the blended samples of the COSMOS parametric model testing sets. Each panel shows one realisation of a given noise standard deviation, σnoise. The blue shaded area shows the gain in accuracy of BlendHunter with respect to SEP. If more than two sources are detected: S. Farrens et al.: Deep transfer learning for blended source identiﬁcation in galaxy survey data 5 10 15 20 25 30 35 Noise Standard Deviation ( noise) 75 80 85 90 95 100 Accuracy % BlendHunter SEP 5 10 15 20 25 30 35 Training Noise Standard Deviation ( noise) 10 5 0 5 10 15 20 25 30 35 Relative Acc. (BlendHunter - sep) % All cases Close blends (distance < 10 px) Distant blends (distance 10 px) Fig. 5. Relative classiﬁcation accuracy for BlendHunter with respect to SEP on the realistic CFIS-like postage stamps, with σnoise = 14.5. The points are taken from the average relative classiﬁcation accuracy from the ten training noise realisations and the error bars from the stan- dard deviation. For close blends (red dot-dashed line), BlendHunter outperforms SEP for any level of noise in the training data. Fig. 4. Overall classiﬁcation accuracy of BlendHunter (black solid line) versus SEP (blue dashed line) with respect to σnoise on the COS- MOS parametric model testing set. The points are taken from the aver- age accuracy from ten realisations of each noise level and the error bars from the standard deviation. For distant blends, the two techniques appear to be consistent to within a few percent and correctly label the majority of the postage stamps. noise level that diﬀers. The overall accuracy of SEP in each case is: 91% over the whole data set, 69% for close blends, and 94% for distant blends. These results indicate that BlendHunter is not overly sen- sitive to the training without perfect knowledge of the noise in the target sample. This is a strong indication that the transfer learning has been successful and is promising for the prospect of applying this approach to real survey data. Additionally, the per- formance of BlendHunter is noticeably more robust for close blends, where galaxy proﬁles signiﬁcantly overlap. It should be stressed that these results are based on the relatively simple blend deﬁnition provided in Sect. 2.1, which does not take into account all of the complexities that could be encountered for real extended sources. In Fig. 4 we display the overall accuracy as a function of the standard deviation of the noise for BlendHunter and SEP on the full parametric model testing set (i.e., sets of 8000 postage stamps). The points are taken from the average classiﬁcation accuracy from the ten noise realisations and the error bars from the standard deviation. 4.3. Results from realistic CFIS-like images Here we test the weights obtained from the parametric mod- els with various noise levels and apply them to the test sample of 80 000 realistic CFIS-like postage stamps (see Sect. 2.5) to obtain classiﬁcation labels. Figure 5 shows the relative classiﬁcation accuracy as func- tion of the standard deviation of the training noise (i.e., the noise level of the sample used for training) for BlendHunter with respect to SEP on the realistic CFIS-like postage stamps. Points show the average relative classiﬁcation accuracy corre- sponding to the ten noise realisations for each training noise level and the error bars are the standard deviation. SEP is sim- ply run once on the full data set. For the whole set of images (i.e., all 80 000 postage stamps), we can see that BlendHunter is able to slightly outperform SEP (only by a few percent) around σnoise = 14.5, which is the true noise standard devia- tion of the CFIS-like sample. The classiﬁcation accuracy then degrades if the network is trained on noise levels that diﬀer sig- niﬁcantly. In the subset of close blends (4838 postage stamps), BlendHunter signiﬁcantly outperforms SEP regardless of the training conditions. This is consistent with the results from the parametric model data. Finally, for distant blends (35 162 postage stamps), BlendHunter matches the performance of SEP when trained on the same noise level as the target images and under-performs by a maximum of 5% when trained on a The results show that BlendHunter is signiﬁcantly better at identifying true blends than SEP regardless of the noise level used for training. This is particularly noticeable for the sample of close blends. However, BlendHunter also produces more incorrect labels for isolated sources and this worsens as the train- ing noise level increases beyond that of the target sample. S. Farrens et al.: Deep transfer learning for blended source identiﬁcation in galaxy survey data As expected, both approaches perform almost perfectly for low noise levels and their performance drops oﬀas the noise increases. Overall, the two approaches appear fairly consistent, however BlendHunter drops oﬀless rapidly for higher noise levels indicating it may be slightly more robust in higher noise regimes. We additionally examined the confusion matrices for the pre- dictions from the realistic CFIS-like test set. Figure 6 shows the fraction of blends correctly classiﬁed as blends (i.e., true positives) or isolated sources correctly classiﬁed as isolated sources (i.e., true negatives) as a function of the training noise standard deviation. The ﬁrst panel shows the performance of BlendHunter compared to SEP for the sample of close blends (green lines) along with an equivalent number (4838) of isolated sources (red lines). The second panel shows the equivalent plot for distant blends with the remaining sample of isolated sources. The third panel shows the results for all 80 000 postage stamps. S. Farrens et al.: Deep transfer learning for blended source identiﬁcation in galaxy survey data 5 10 15 20 25 30 35 Noise Standard Deviation ( noise) 75 80 85 90 95 100 Accuracy % BlendHunter SEP Fig. 4. Overall classiﬁcation accuracy of BlendHunter (black solid line) versus SEP (blue dashed line) with respect to σnoise on the COS- MOS parametric model testing set. The points are taken from the aver- age accuracy from ten realisations of each noise level and the error bars from the standard deviation. 5 10 15 20 25 30 35 Training Noise Standard Deviation ( noise) 10 5 0 5 10 15 20 25 30 35 Relative Acc. (BlendHunter - sep) % All cases Close blends (distance < 10 px) Distant blends (distance 10 px) Fig. 5. Relative classiﬁcation accuracy for BlendHunter with respect to SEP on the realistic CFIS-like postage stamps, with σnoise = 14.5. The points are taken from the average relative classiﬁcation accuracy from the ten training noise realisations and the error bars from the stan- dard deviation. For close blends (red dot-dashed line), BlendHunter outperforms SEP for any level of noise in the training data. 5 10 15 20 25 30 35 Training Noise Standard Deviation ( noise) 10 5 0 5 10 15 20 25 30 35 Relative Acc. (BlendHunter - sep) % All cases Close blends (distance < 10 px) Distant blends (distance 10 px) Fig. 5. Relative classiﬁcation accuracy for BlendHunter with respect to SEP on the realistic CFIS-like postage stamps, with σnoise = 14.5. The points are taken from the average relative classiﬁcation accuracy from the ten training noise realisations and the error bars from the stan- dard deviation. For close blends (red dot-dashed line), BlendHunter outperforms SEP for any level of noise in the training data. 5 10 15 20 25 30 35 Training Noise Standard Deviation ( noise) 10 5 0 5 10 15 20 25 30 35 Relative Acc. (BlendHunter - sep) % All cases Close blends (distance < 10 px) Distant blends (distance 10 px) Fig. 5. Relative classiﬁcation accuracy for BlendHunter with respect to SEP on the realistic CFIS-like postage stamps, with σnoise = 14.5. The points are taken from the average relative classiﬁcation accuracy from the ten training noise realisations and the error bars from the stan- dard deviation. For close blends (red dot-dashed line), BlendHunter outperforms SEP for any level of noise in the training data. 4.2. Results from parametric models In the following, we deﬁne classiﬁcation accuracy as the per- centage of postage stamps correctly labelled as either isolated or blended sources. Here we take the weights obtained via the train- S. Farrens et al.: Deep transfer learning for blended source identiﬁcation in galaxy survey data 5. Conclusions This work is based on data obtained as part of the Canada-France Imaging Survey, a CFHT large program of the National Research Council of Canada and the French Centre National de la Recherche Scientiﬁque. Based on observations obtained with MegaPrime/MegaCam, a joint project of CFHT and CEA Saclay, at the Canada-France-Hawaii Telescope (CFHT) which is operated by the National Research Council (NRC) of Canada, the Institut National des Science de l’Univers (INSU) of the Centre National de la Recherche Scientiﬁque (CNRS) of France, and the University of Hawaii. This research used the facili- ties of the Canadian Astronomy Data Centre operated by the National Research Council of Canada with the support of the Canadian Space Agency. AZV would like to thank LSC/PMR/EP at University of São Paulo for providing additional computing power. The authors also wish to thank Xinyu Wang and Alexandre Bruckert for initial eﬀorts that were expanded upon in this work. DIM-ACAV. This work is based on data obtained as part of the Canada-France Imaging Survey, a CFHT large program of the National Research Council of Canada and the French Centre National de la Recherche Scientiﬁque. Based on observations obtained with MegaPrime/MegaCam, a joint project of CFHT and CEA Saclay, at the Canada-France-Hawaii Telescope (CFHT) which is operated by the National Research Council (NRC) of Canada, the Institut National des Science de l’Univers (INSU) of the Centre National de la Recherche Scientiﬁque (CNRS) of France, and the University of Hawaii. This research used the facili- ties of the Canadian Astronomy Data Centre operated by the National Research Council of Canada with the support of the Canadian Space Agency. AZV would like to thank LSC/PMR/EP at University of São Paulo for providing additional computing power. The authors also wish to thank Xinyu Wang and Alexandre Bruckert for initial eﬀorts that were expanded upon in this work. parametric models derived from COSMOS data for various lev- els and realisations of Gaussian random noise. Comparison with the community standard SEP (a Python implementation of SExtractor) on 70 sets of 8000 parametric models demonstrates that BlendHunter is signiﬁcantly more sensitive to blends when the images are very noisy or when galaxy pairs are very close (<10 pixels). In the low-noise dis- tant blend regime, the performance of BlendHunter appears to be consistent with that of SEP. Acknowledgements. The authors wish to acknowledge the COSMIC project funded by the CEA DRF-Impulsion call in 2016, the CrossDisciplinary Program on Numerical Simulation (SILICOSMIC project in 2018) of CEA, the French Alternative Energies and Atomic Energy Commission. The Euclid Collabora- tion, the European Space Agency and the support of the Centre National d’Etudes Spatiales. This work was also supported by the ANR AstroDeep project – grant 19-CE23-0024-01. This work has made use of the CANDIDE Cluster at the Insti- tut d’Astrophysique de Paris and made possible by grants from the PNCG and the References Abadi, M., Agarwal, A., Barham, P., et al. 2015, TensorFlow: Large-Scale Machine Learning on Heterogeneous Systems, software available from tensorflow.org Arcelin, B., Doux, C., Aubourg, E., Roucelle, C., & LSST Dark Energy Science Collaboration 2021, MNRAS, 500, 531 Awang Iskandar, D. N. F., Zijlstra, A. A., McDonald, I., et al. 2020, Galaxies, 8, 88 BlendHunter notably outperforms SEP by 5−15% for blends in which the galaxies are separated by less than ten pixels. This is an interesting result as these are precisely the cases that generate biases in galaxy shape measurements (MacCrann et al. 2022). Barbary, K. 2016, J. Open Source Softw., 1, 58 Bertin, E., & Arnouts, S. 1996, A&AS, 117, 393 Barbary, K. 2016, J. Open Source Softw., 1, 58 Bertin, E., & Arnouts, S. 1996, A&AS, 117, 393 Bosch J Armstrong R Bickerton S et al 2018 PASJ 70 S5 Barbary, K. 2016, J. Open Source Softw., 1, 58 Bertin, E., & Arnouts, S. 1996, A&AS, 117, 393 Bosch, J., Armstrong, R., Bickerton, S., et al. 2018, PASJ, 70, S5 Chollet, F., et al. 2015, Keras https://keras.io. , , g, , , , , Chollet, F., et al. 2015, Keras https://keras.io. Cropper, M., Cole, R., James, A.,et al. 2012, in Space Telescopes and Instrumentation 2012: Optical, Infrared, and Millimeter Wave, eds. M. C. Overall the results are promising and indicate that it may be possible to adapt this approach to more accurately identify blended sources in real survey data. The next steps in this direc- tion would entail the generation of more realistic testing data that contain some artefacts and images with more than two sources. Further work is also required to reduce the number of false nega- tives, i.e., incorrect labels for isolated sources. Finally, additional tests should be performed to determine if and to what degree the use of pre-trained weights in the CNN layers can help prevent over-ﬁtting the network to the training sets. We leave the inves- tigation and implementation of these steps for future work. Clampin, G. G. Fazio, H. A. MacEwen, J. Oschmann, & M. Jacobus, SPIE Conf. Ser., 8442, 84420V de Vaucouleurs, G. 1948, Ann. Astrophys., 11, 247 Deng, J., Dong, W., Socher, R., et al. 2009, Proc. CVPR Duchi, J., Hazan, E., & Singer, Y. 2011, J. Mach. Learn. Res., 12, 2121 Euclid Collaboration (Martinet, N., et al. 2019, A&A, 627, A59 George, D., Shen, H., & Huerta, E. 5. Conclusions We additionally produced a sample of 80 000 more realis- tic CFIS-like images derived from real COSMOS data. Results demonstrate that the BlendHunter weights, learned on the parametric models, can successfully be transferred to this more complex data set and achieve >90% classiﬁcation accuracy, out- performing SEP by a few percent on the full sample when the appropriate weights are used. The results also indicate that BlendHunter is capable of achieving results that are roughly consistent with those of SEP even when weights are used that have been trained with a signiﬁcantly diﬀerent noise standard deviation. 5. Conclusions We present a proof-of-concept deep transfer learning approach for the automated and robust identiﬁcation of blended sources in galaxy survey data called BlendHunter. This technique uses convolution layers pre-trained on natural images from Ima- geNet, thus signiﬁcantly reducing the time required for train- ing. The fully connected layers are then trained on simple A98, page 7 of 9 A98, page 7 of 9 A&A 657, A98 (2022) A&A 657, A98 (2022) 0 10 20 30 Training Noise Standard Deviation noise 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 Fraction Close Blend Set SEP, True Positives SEP, True Negatives BH, True Positives BH, True Negatives 0 10 20 30 Training Noise Standard Deviation noise 0.30 0.35 0.40 0.45 0.50 Fraction Distant Blend Set 0 10 20 30 Training Noise Standard Deviation noise 0.30 0.35 0.40 0.45 0.50 Fraction All Postage Stamps Fig. 6. Classiﬁcation confusion matrices of BlendHunter (solid lines) compared to SEP (dot-dashed lines) for the realistic CFIS-like test data. Each panel shows the fraction of blends correctly classiﬁed as blends (i.e., true positives, green lines) or isolated sources correctly classiﬁed as isolated sources (i.e., true negatives, red lines) as a function of the training noise standard deviation. Left panel: sample of close blends along with an equivalent number of isolated sources. Middle panel: sample of distant blends along with the remaining isolated sources. Right panel: results for the full set of postage stamps. 0 10 20 30 Training Noise Standard Deviation noise 0.30 0.35 0.40 0.45 0.50 Fraction Distant Blend Set 0 10 20 30 Training Noise Standard Deviation noise 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 Fraction Close Blend Set SEP, True Positives SEP, True Negatives BH, True Positives BH, True Negatives 0 10 20 30 Training Noise Standard Deviation noise 0.30 0.35 0.40 0.45 0.50 Fraction All Postage Stamps Fig. 6. Classiﬁcation confusion matrices of BlendHunter (solid lines) compared to SEP (dot-dashed lines) for the realistic CFIS-like test data. Each panel shows the fraction of blends correctly classiﬁed as blends (i.e., true positives, green lines) or isolated sources correctly classiﬁed as isolated sources (i.e., true negatives, red lines) as a function of the training noise standard deviation. Left panel: sample of close blends along with an equivalent number of isolated sources. Middle panel: sample of distant blends along with the remaining isolated sources. Right panel: results for the full set of postage stamps. DIM-ACAV. References A. 2018, Phys. Rev. D, 97 Guinot, A., Kilbinger, M., Farrens, S., et al. 2021, A&A, submitted Hartlap, J., Hilbert, S., Schneider, P., & Hildebrandt, H. 2011, A&A, 528, A51 Hoekstra, H., Viola, M., & Herbonnet, R. 2017, MNRAS, 468, 3295 Joseph, R., Courbin, F., & Starck, J.-L. 2016, A&A, 589, A2 oseph, R., Courbin, F., & Starck, J.-L. 2016, A&A, 589, A2 Kingma, D. P., & Ba, J. 2014, Adam: A Method for Stochastic Optimization, cite arxiv:1412.6980Comment: Published as a conference paper at the 3rd International Conference for Learning Representations, San Diego, 2015 Kotsiantis, S. 2007, Super. Mach. Learn.: Rev. Class. Tech., 31, 249 Kotsiantis, S. 2007, Super. Mach. Learn.: Rev. Class. Tech., 31, 249 Lecun, Y., Bengio, Y., & Hinton, G. 2015, Nature, 521, 436 MacCrann, N., Becker, M. R., McCullough, J., et al. 2022, MNRAS, 509, 3371 Lecun, Y., Bengio, Y., & Hinton, G. 2015, Nature, 521, 436 MacCrann, N., Becker, M. R., McCullough, J., et al. 2022, MNRAS, 509, 3371 Mandelbaum, R. 2018, ARA&A, 56, 393 Mandelbaum, R., Hirata, C. M., Leauthaud, A., Massey, R. J., & Rhodes, J. 2012, MNRAS, 420, 1518 Mandelbaum, R., Rowe, B., Bosch, J., et al. 2014, ApJS, 212, 5 Melchior, P., Moolekamp, F., Jerdee, M., et al. 2018, Astron. Comput., 24, 129 Reiman, D. M., & Göhre, B. E. 2019, MNRAS, 485, 2617 A98, page 8 of 9 S. Farrens et al.: Deep transfer learning for blended source identiﬁcation in galaxy survey data Tanoglidis, D., ´Ciprijanovi´c, A., & Drlica-Wagner, A. 2021, Astron. Comput., 35 Rowe, B. T. P., Jarvis, M., Mandelbaum, R., et al. 2015, Astron. Comput., 10, 121 Tatarski, V. I. 2016, Wave propagation in a turbulent medium (Courier Dover Publications) Russakovsky, O., Deng, J., Su, H., et al. 2015, Int. J. Comput. Vis., 115, 211 y g p amuroﬀ, S., Bridle, S. L., Zuntz, J., et al. 2018, MNRAS, 475, 452 Sanchez, J., Mendoza, I., Kirkby, D.P., Burchat, P.R., & LSST Dark En Science Collaboration 2021, JCAP, 2021, 043 Tieleman, T., & Hinton, G. 2012, Lecture 6.5–RmsProp: Divide the gradient by a running average of its recent magnitude (COURSERA: Neural Networks for Machine Learning) Scoville, N., Aussel, H., Brusa, M., et al. 2007, ApJS, 172, 1 p Shan, J.-H., Feng, L., Yuan, H.-Q., et al. 2020, Chin. Astron. Rowe, B. T. P., Jarvis, M., Mandelbaum, R., et al. 2015, Astron. Comput., 10, 121 Russakovsky, O., Deng, J., Su, H., et al. 2015, Int. J. Comput. Vis., 115, 211 Samuroﬀ, S., Bridle, S. L., Zuntz, J., et al. 2018, MNRAS, 475, 4524 Sanchez, J., Mendoza, I., Kirkby, D.P., Burchat, P.R., & LSST Dark Energy Science Collaboration 2021, JCAP, 2021, 043 Scoville, N., Aussel, H., Brusa, M., et al. 2007, ApJS, 172, 1 Shan, J.-H., Feng, L., Yuan, H.-Q., et al. 2020, Chin. Astron. Astrophys., 44, 507 Simonyan, K., & Zisserman, A. 2014, CoRR [arXiv:1409.1556] Srinivas, S., Sarvadevabhatla, R. K., Mopuri, K. R., et al. 2016, Front. Robot. AI, 2, 36 Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., & Salakhutdinov, R. 2014, J. Mach. Learn. Res., 15, 1929 Tanoglidis, D., ´Ciprijanovi´c, A., & Drlica-Wagner, A. 2021, Astron. Comput., 35 Tatarski, V. I. 2016, Wave propagation in a turbulent medium (Courier Dover Publications) Tieleman, T., & Hinton, G. 2012, Lecture 6.5–RmsProp: Divide the gradient by a running average of its recent magnitude (COURSERA: Neural Networks for Machine Learning) Trujillo, I., Aguerri, J., Cepa, J., & Gutiérrez, C. 2001, MNRAS, 328, 977 Wei, W., Huerta, E. A., Whitmore, B. C., et al. 2020, MNRAS, 493, 3178 Wu, C., Wong, O. I., Rudnick, L., et al. 2019, MNRAS, 482, 1211 Zhu, X.-P., Dai, J.-M., Bian, C.-J., et al. 2019, Ap&SS, 364, 55 References Astrophys., 44, 507 Si K & Zi A 2014 C RR [ Xi 1409 1556] g Trujillo, I., Aguerri, J., Cepa, J., & Gutiérrez, C. 2001, MNRAS, 328, 977 p Shan, J.-H., Feng, L., Yuan, H.-Q., et al. 2020, Chin. Astron. Astrophys., 44, 507 Simonyan, K., & Zisserman, A. 2014, CoRR [arXiv:1409.1556] Shan, J.-H., Feng, L., Yuan, H.-Q., et al. 2020, Chin. Astron. A Simonyan, K., & Zisserman, A. 2014, CoRR [arXiv:1409.1556] Wei, W., Huerta, E. A., Whitmore, B. C., et al. 2020, MNRAS, 493, 3178 Srinivas, S., Sarvadevabhatla, R. K., Mopuri, K. R., et al. 2016, Front. Robot. AI, 2, 36 Wu, C., Wong, O. I., Rudnick, L., et al. 2019, MNRAS, 482, 1211 Zhu, X.-P., Dai, J.-M., Bian, C.-J., et al. 2019, Ap&SS, 364, 55 Wu, C., Wong, O. I., Rudnick, L., et al. 2019, MNRAS, 482, 1211 Zhu, X.-P., Dai, J.-M., Bian, C.-J., et al. 2019, Ap&SS, 364, 55 Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., & Salakhutdinov, R. 2014, J. Mach. Learn. Res., 15, 1929 Wu, C., Wong, O. I., Rudnick, L., et al. 2019, MNRAS, 482, 1211 Zhu, X.-P., Dai, J.-M., Bian, C.-J., et al. 2019, Ap&SS, 364, 55 A98, page 9 of 9
https://openalex.org/W4361911431	https://figshare.com/articles/journal_contribution/Supplementary_Table_2_from_Myeloid_Cells_Obtained_from_the_Blood_but_Not_from_the_Tumor_Can_Suppress_T-cell_Proliferation_in_Patients_with_Melanoma/22455888/1/files/39907038.pdf	English	null	Supplementary Table 1 from Myeloid Cells Obtained from the Blood but Not from the Tumor Can Suppress T-cell Proliferation in Patients with Melanoma	null	2,023	cc-by	340	Supplementary Table 2. Percent inhibition of T cell proliferation exerted by tumor myeloid-derived b Supplementary Table 2. Percent inhibition of T cell proliferation exerted by tumor myeloid-derived b Supplementary Table 2. Percent inhibition of T cell proliferation exerted by tumor myeloid-derived subsets upplementary Table 2. Percent inhibition of T cell proliferation exerted by tumor mye b t Supplementary Table 2. Percent inhibition of T cell proliferation exerted by tumor myeloid-derived subsets Tumor P#1 P#2 P#3 P#4 P#5 P#6 MEAN SEM Phenotype isolated Lin- CD11b+ -0.8 ± 0.1 1.8 ± 0.3 -0.7 ± 0.1 -1.4 ± 0.9 0.1 ± 0.3 1.3 ± 0.1 0.1 0.5 CD14+ -1.1 ± 0.1 1.4 ± 0.1 ND ND -1.6 ± 0.1 2.1 ± 0.4 0.2 0.9 CD14- CD15hi -0.5 ± 0.1 0.2 ± 0.3 ND ND ND ND -0.2 0.4 CD14- CD15int ND 1.4 ND ND ND ND 1.4 Tumor-derived myeloid subsets from (n=6) were stained and sorted and their suppressive activity was tested in a co-culture with autologous T cells activated with anti-CD3/anti-CD28 beads in duplicates T cell proliferation was assessed by quantifying the percent of T cells that had T cells activated with anti-CD3/anti-CD28 beads in duplicates. T cell proliferation was assessed by quantifying the percent of T cells that had not diluted the CFSE at day 4 and calculating the percent inhibition of T cell proliferation. Average percent inhibition ± SEM are shown. Average of the cumulative data from the 6 tumorsis also shown. Tumor-derived myeloid subsets from (n=6) were stained and sorted and their suppressive activity was tested in a co-culture with autologous T cells activated with anti-CD3/anti-CD28 beads in duplicates T cell proliferation was assessed by quantifying the percent of T cells that had T cells activated with anti-CD3/anti-CD28 beads in duplicates. T cell proliferation was assessed by quantifying the percent of T cells that had not diluted the CFSE at day 4 and calculating the percent inhibition of T cell proliferation. Average percent inhibition ± SEM are shown. Average of the cumulative data from the 6 tumorsis also shown.
https://openalex.org/W2599995607	https://vestnik.kemsu.ru/jour/article/download/2118/2020	Russian	null	THE CATEGORY OF "MUNDANE CONSCIOUSNESS" IN PSYCHOLOGY	Vestnik Kemerovskogo gosudarstvennogo universiteta	2,017	cc-by	4,274	КАТЕГОРИЯ «ОБЫДЕННОЕ СОЗНАНИЕ» В ПСИХОЛОГИИ Обы- денное сознание при этом представляют как совокуп- ность теоретических и эмпирических знаний, представ- лений и чувств групп людей. пример, в работах В. В. Розова и М. Я. Грушина обыден- ное сознание является составляющей частью обществен- ной психологии и включает в себя те взгляды и представ- ления, которые были сформированы в результате много- векового опыта людей [2]. Такие философы, как В. А. Ядов, В. В. Лобас, Ю. М. Серов, считают обществен- ную психологию элементом обыденного сознания. Обы- денное сознание при этом представляют как совокуп- ность теоретических и эмпирических знаний, представ- лений и чувств групп людей. В целом обыденное сознание оценивалось в большей степени отрицательно. В своих работах Б. А. Грушин, Т. А. Кузьмин, В. А. Штофф, В. Ж. Келле и др. отмечали, что обыденное сознание носит дотеоретический характер и отражает в значительной степени негативные, напол- ненные предрассудками и заблуждениями, ложные воз- зрения. В то время как теоретическое и научно- философское сознание систематизировано и рациональ- но. Однако обыденное сознание не может быть бессис- темной совокупностью различных фрагментов сознания. В целом обыденное сознание оценивалось в большей степени отрицательно. В своих работах Б. А. Грушин, Т. А. Кузьмин, В. А. Штофф, В. Ж. Келле и др. отмечали, что обыденное сознание носит дотеоретический характер и отражает в значительной степени негативные, напол- ненные предрассудками и заблуждениями, ложные воз- зрения. В то время как теоретическое и научно- философское сознание систематизировано и рациональ- но. Однако обыденное сознание не может быть бессис- темной совокупностью различных фрагментов сознания. В целом обыденное сознание оценивалось в большей степени отрицательно. В своих работах Б. А. Грушин, Т. А. Кузьмин, В. А. Штофф, В. Ж. Келле и др. отмечали, что обыденное сознание носит дотеоретический характер и отражает в значительной степени негативные, напол- ненные предрассудками и заблуждениями, ложные воз- зрения. В то время как теоретическое и научно- философское сознание систематизировано и рациональ- но. Однако обыденное сознание не может быть бессис- темной совокупностью различных фрагментов сознания. Оно содержит в себе когнитивные, эмоциональные и во- левые компоненты, которые неразрывно связаны и пред- ставлены навыками, привычками, стереотипами, обы- чаями, предрассудками, заблуждениями и суевериями. Обыденное сознание является отражением многовеко- вого опыта, формирующегося в повседневной жизни и практике [1]. Феномен обыденного сознания анализировался в русле марксистской философии и отождествлялся ею с массовым сознанием и общественной психологией. Во- просам соотношения обыденного и массового сознания посвящены работы Е. К. Макаровой, Б. Я. Грушина, В. С. Ремжа. КАТЕГОРИЯ «ОБЫДЕННОЕ СОЗНАНИЕ» В ПСИХОЛОГИИ Надежда П. Андрюшкова1, @ 1 Донецкий национальный университет, Украина, 83001, г. Донецк, ул. Университетская, 24 @ аndruchkovа@rаmbler.ru Аннотация: Статья посвящена анализу феномена обыденного сознания как кате- гории, которая имеет во многом противоречивую и парадоксальную природу, чем вызывает интерес. Описаны основные подходы изучения обыденного созна- ния и взгляды на его природу в философских и в психологических концепциях. Проанализировано соотношение обыденного сознания со смежными понятиями «массовое сознание», «обыденная психология» «групповое сознание», «общест- венное сознание». Описана взаимосвязь обыденного сознания с культурой и его место в жизни личности как культурного компонента, неотъемлемой характери- стики любого исторического типа сознания. Делается вывод, что обыденное соз- нание выступает промежуточным звеном между субъектом и средой, в которой он находится, а также между личностью и обществом. Помимо этого, обыденное сознание выполняет не только отражательную функцию, но также выступает регулятором социального и, в частности, группового поведения. Поступила в редакцию 28.06.2016. Принята к печати 17.11.2016. Поступила в редакцию 28.06.2016. Принята к печати 17.11.2016. Ключевые слова: сознание, бы- тие, обычаи, предрассудки, куль- тура. Ключевые слова: сознание, бы- тие, обычаи, предрассудки, куль- тура. Для цитирования: Андрюшкова Н. П. Категория «обыденное сознание» в психологии // Вестник Кемеровского госу- дарственного университета. 2017. № 1. С. 90 – 93. DOI: 10.21603/2078-8975-2017-1-90-93. Проблема обыденного сознания в силу его роли в ре- гуляции социального поведения человека является акту- альной и вызывает интерес своей парадоксальной приро- дой. Особую актуальность эта тема приобретает в период социальных изменений, которые отражаются на повсе- дневной жизни людей. Обыденное сознание, как отмеча- ют исследователи, не только отражает реальность и за счет этого обеспечивает человеку эффективное приспо- собление к действительности, успешное решение жиз- ненных проблем, но вместе с тем продуцирует искаже- ния, неточности, противоречия в отражении. Отражая общественное бытие сквозь призму повседневной жизни и практического опыта людей, обыденное сознание спо- собно отражать социальный контекст развития общества. Таким образом, при учете представлений о явлениях и особенностях обыденного сознания может быть эффек- тивней использован человеческий фактор, также это даст более четкие представления состояний массового созна- ния и может использоваться для разработки механизмов регулирования общественных отношений. пример, в работах В. В. Розова и М. Я. Грушина обыден- ное сознание является составляющей частью обществен- ной психологии и включает в себя те взгляды и представ- ления, которые были сформированы в результате много- векового опыта людей [2]. Такие философы, как В. А. Ядов, В. В. Лобас, Ю. М. Серов, считают обществен- ную психологию элементом обыденного сознания. Вестник Кемеровского государственного университета 2017  № 1 Вестник Кемеровского государственного университета 2017  № 1 КАТЕГОРИЯ «ОБЫДЕННОЕ СОЗНАНИЕ» В ПСИХОЛОГИИ Отмечалось, что обыденное сознание – эле- мент массового, в котором отражается конкретное обще- ственное бытие, преломленное через призму обществен- ной жизни. С другой стороны, массовое сознание форми- руется стихийно, и на его формирование влияют и рав- ноценно воздействуют два фактора: обыденное и теоре- тическое сознание. Однако, как отмечает Л. Г. Судас, мы не можем отождествлять массовое и обыденное созна- ние, несмотря на то, что и то и другое тесно связаны и являются продуктами духовной жизни людей [1]. Оно содержит в себе когнитивные, эмоциональные и во- левые компоненты, которые неразрывно связаны и пред- ставлены навыками, привычками, стереотипами, обы- чаями, предрассудками, заблуждениями и суевериями. Обыденное сознание является отражением многовеко- вого опыта, формирующегося в повседневной жизни и практике [1]. Оно содержит в себе когнитивные, эмоциональные и во- левые компоненты, которые неразрывно связаны и пред- ставлены навыками, привычками, стереотипами, обы- чаями, предрассудками, заблуждениями и суевериями. Обыденное сознание является отражением многовеко- вого опыта, формирующегося в повседневной жизни и практике [1]. В. М. Межуев отмечает, что даже в условиях научно- технического прогресса большинство людей используют сознание, прямо противоположное научному. И вопреки доводам науки можно верить в предсказания, пророчест- ва, нечистую силу и др. Следовательно, причину этому следует искать не только в слабости научных доказа- тельств, но и в самом бытии людей, которые питают и поддерживают эту веру [3]. В более поздних исследованиях, если выйти за рамки марксистской традиции, обыденное сознание перестает противопоставляться научно-теоретическому (Э. Гуссерль, Также обыденное сознание связывают с общественной психологией, однако характеризуют связь по-разному. Встречаются прямо противоположные мнения. Так, на- 90 Вестник Кемеровского государственного университета 2017  № 1 Б. Ф. Ломов, А. Н. Леонтьев, Б. Ф. Поршнев, В. С. Му- хина, С. Л. Рубинштейн и др. Б. Ф. Ломов, А. Н. Леонтьев, Б. Ф. Поршнев, В. С. Му- хина, С. Л. Рубинштейн и др. Б. Ф. Ломов, А. Н. Леонтьев, Б. Ф. Поршнев, В. С. Му- хина, С. Л. Рубинштейн и др. Б. Ф. Ломов, А. Н. Леонтьев, Б. Ф. Поршнев, В. С. Му- хина, С. Л. Рубинштейн и др. М. Хайдегер, М. Вартовский, М. Бунге, К. Поппер, К. Ясперс и др.). Признается, что оно носит более широ- кий характер и содержит в себе помимо познавательного отношения к миру еще и соционормативное, эстетическое и практическое [4]. М. Хайдегер, М. Вартовский, М. Бунге, К. Поппер, К. Ясперс и др.). Признается, что оно носит более широ- кий характер и содержит в себе помимо познавательного отношения к миру еще и соционормативное, эстетическое и практическое [4]. КАТЕГОРИЯ «ОБЫДЕННОЕ СОЗНАНИЕ» В ПСИХОЛОГИИ То есть существование индивидуальной психики в форме сознательного отражения невозможно вне общественного сознания [12]. Б. Ф. Ломов отмечал, что общественное сознание существует не над индивидами, не между ними, а непо- средственно в них. Общественные идеи, взгляды и на- строения – это ни что иное как формирующиеся в про- цессе развития общества идеи и взгляды конкретных лю- дей, живущих и действующих в конкретных историче- ских условиях [13; 14]. Некоторые авторы, в частности С. Лурье, рассмат- ривают общественное сознание как присущее каждому человеку. Индивидуальным же сознанием обладает не- значительное количество людей. Личностное сознание предполагает изменение всей психики и внесение в нее новых качеств, а это, как отмечает С. Лурье, очень трудо- емкий и тяжелый процесс [15]. Общепринятого четкого теоретического объяснения обыденного сознания в психологии нет, оно рассматри- вается через сходные и пересекающиеся категории. Так, например, рассматривая обыденное сознание, Е. В. Улы- бина отмечает, что оно расположено на стыке различных психологических образований индивидуальной и обще- ственной природы. Опираясь на существующее в психо- логической литературе понимание данного феномена, Е. В. Улыбина отметила, что обыденное сознание зани- мает положение медиатора по отношению к обществен- ному и индивидуальному сознанию, к мифологическому и научному отражению мира, а также к бессознательному и рефлексивному сознанию [7]. К. А. Абульханова-Славская отмечает, что общест- венное сознание первично и передает индивидуальному сознанию определенные нормы, правила, обычаи, тради- ции, которые отличают и выделяют конкретное общество от любого другого [16; 6]. В своих работах Б. Ф. Поршнев также утверждает, что формирование общественного сознания первобытно- го общества первично по отношению к индивидуальному сознанию. Анализируя грамматику языков народов мира, Б. Ф. Поршнев определил, что формирование образа «мы» позднее формирование образа «они». Он отмечал, что эта тенденция универсальна и присутствует в любых группах на любом уровне взаимодействия [17]. Чаще встречаются работы, в которых обыденное соз- нание рассматривается через феномены группового и об- щественного сознания. Говоря о групповом сознании, Б. Шефер и Б. Шлендер отмечают, что это – субъективные представления людей об основных особенностях группы, к которой они принадлежат. Групповое сознание рассмат- ривается как результат протекающего в историческое вре- мя социального взаимодействия [8]. Групповое сознание проявляется в когнитивных аспектах – общих знаниях, а также в аффективных – общих чувствах и эмоциональ- ных реакциях на определенные объекты [8; 9]. Таким образом, согласно распространенным в психо- логии точкам зрения относительно общественного, кол- лективного сознания, оно является базовым, первичным по отношению к индивидуальному, и оказывает на него значительное влияние. КАТЕГОРИЯ «ОБЫДЕННОЕ СОЗНАНИЕ» В ПСИХОЛОГИИ Одним из первых, обративших внимание на пробле- мы общественного сознания, был Э. Дюркгейм. Он выде- лял у человека два сознания: индивидуальное, характер- ное для определенного субъекта, и коллективное, кото- рое является общим для целой социальной группы и вы- ражается через коллективные представления. Обыденное сознание – самостоятельная сфера жизни личности. Теоретическое сознание не заменяет собой в полной мере обыденного сознания; структура послед- него многокомпонентна и многослойна, отражает в себе различные области: природа, общество, человеческая жизнь. Оно содержит в себе разнообразные эмпириче- ские знания, суждения и мнения об окружающей дейст- вительности, восприятия из народного эпоса, народной агрономии, народной медицины. Это тот уровень знаний, который используется человечеством повседневно, на основании которого обеспечивается приспособление к действительности. Таким образом, обыденное созна- ние – это некий культурный компонент, неотъемлемая характеристика любого исторического типа сознания. Феномены общественного и группового сознания рас- сматривались также В. Вундтом, которым вводится поня- тие психологии народов. Он отмечает, что в пределах од- ной народности существует некоторое собирательное соз- нание, которое включает в себя различные представления, чувства, а также зависит от развития речи, мифа, обычаев и различных форм духовного общения [3; 10]. Л. С. Выготский обосновывал идею о том, что кол- лективное сознание – более раннее, чем индивидуальное, и индивидуальное формируется на его основе. Таким образом, структура общественных связей и отношений является источником структуры и связей индивидуально- го сознания [11]. Л. С. Выготский обосновывал идею о том, что кол- лективное сознание – более раннее, чем индивидуальное, и индивидуальное формируется на его основе. Таким образом, структура общественных связей и отношений является источником структуры и связей индивидуально- го сознания [11]. В целом обыденное сознание, как его характеризует большинство исследователей, – это неоднородная, слож- ная, многослойная, противоречивая, стихийно сложив- шаяся совокупность теоретически необобщенных знаний, представлений, житейских суждений, чувств и настрое- ний, порождаемых массовым опытом, влиянием соци- альной среды, ее обычаями и традициями. Оно способст- вует групповому единству, позволяет подтверждать ис- тинность и значимость принятых мнений. Как указывает Д. Кэмпбелл, универсальными ценностями человеческо- го общества является принятие характерных для опреде- ленной группы норм, идеалов, знаний [5]. Так формиру- ется взаимоподкрепляемая связь, если мнение принято в обществе и рассматривается как правильное, значит оно верное [6]. Следовательно, на любом уровне культу- ры для макро и микрогрупп существуют определенные мнения, которые не нуждаются в дополнительных дока- зательствах, а принимаются на веру. А. Н. Леонтьев также подчеркивал, что индивидуаль- ное сознание необходимо рассматривать как продукт, возникающий в ходе становления и развития общества. КАТЕГОРИЯ «ОБЫДЕННОЕ СОЗНАНИЕ» В ПСИХОЛОГИИ Это дает основания полагать, что обыденное сознание, присущее индивиду, является ско- рее продуктом общественного сознания. Некоторыми авторами обыденное сознание рассмат- ривается как содержащее в себе нерациональные модели действительности. Так, например, рассматривая феноме- Исследованиями общественного сознания занима- лись Э. Дюркгейм, В. Вундт, С. Лурье, Л. С. Выготский, 91 Вестник Кемеровского государственного университета 2017  № 1 ны обыденного сознания, С. Московичи отмечает, что каждому человеку присущи представления, которые за- висят от принадлежности к определенной социальной группе и определяют его некритичное и в некотором смысле нерациональное восприятие окружающей дейст- вительности. С. Московичи обращает внимание на сход- ство между обыденной психологией и психологией бес- сознательного в психоаналитической концепции [9]. ны обыденного сознания, С. Московичи отмечает, что каждому человеку присущи представления, которые за- висят от принадлежности к определенной социальной группе и определяют его некритичное и в некотором смысле нерациональное восприятие окружающей дейст- вительности. С. Московичи обращает внимание на сход- ство между обыденной психологией и психологией бес- сознательного в психоаналитической концепции [9]. принимается именно так, как она названа, категорирована и описана в языке. Обыденное сознание является регулятором поведе- ния человека, оно представляет не всегда рационально освоенные накопленные знания, во-первых, это те зна- ния, которые не зависимы от научно-теоретической мысли и являются личным жизненным опытом каждого человека. Помимо этого, это также знания не только индивидуальные, но и коллективные, накопленные мно- гими поколениями и усваиваемые в процессе социали- зации. Именно этот опыт служит основой для обыден- ного знания, делает согласованными практические дей- ствия людей. Н. Л. Мусхелишвили описывает противоречия, суще- ствующие в обыденном сознании. Он отмечает, что про- тиворечия могут даже не восприниматься человеком, они относятся к разным фрагментам и аспектам картины мира, потребность в сопоставлении этих фрагментов отсутствует [18]. Таким образом появляются определенные искажения в сведениях о мире. Искажения могут в большей или меньшей степени осознаваться, но принадлежность к оп- ределенной группе и культуре позволяет обеспечивать удобные механизмы объяснения различных событий и мира в целом [4; 18]. Но следует сказать, что в связи с развитием научно- технического прогресса доля осмысленных теоретиче- ских знаний в обыденном сознании достаточно велика. Это знания из области философии, права, политики, всех форм духовной культуры. Таким образом, можно говорить о том, что обыден- ное сознание, являясь в значительной мере продуктом общественного сознания, является неким посредником, системой отношений между человеком и социальной средой. Оно, выполняя двойственную функцию – и от- ражательную, и регулирующую, обусловливает подчас продуктивное, а подчас противоречивое взаимодействие человека с социумом. КАТЕГОРИЯ «ОБЫДЕННОЕ СОЗНАНИЕ» В ПСИХОЛОГИИ По этой причине обыденное соз- нание носит проблемный характер, что определяет не- обходимость его научного исследования в психологии. Изучение обыденного сознания также может помочь выяснить причины возникновения и пути распростране- ния различных социальных явлений, в частности пред- рассудков, суеверий, предубеждений. Интересное мнение относительно обыденного созна- ния выдвинуто В. А. Шкуратовым, который главенст- вующее место в обыденном сознании отводит языку. Он характеризует обыденное сознание как поле чистой вербальности, связанное с рациональным и эмоциональ- но-аффективным уровнями [19]. В каждом естественном языке отражается определенный способ восприятия ми- ра, который является общепринятым и преимущественно присущ всем носителям данного языка [3]. Таким обра- зом, существует определенная языковая картина мира, которая носит в большей степени неосознаваемый харак- тер. Реальность для носителя определенного языка вос- Литература р ур 1. Судас Л. Г. Массовое сознание: в поисках новой парадигмы. М., 1996. 111 с. 1. Судас Л. Г. Массовое сознание: в поисках новой парадигмы. М., 1996. 111 с. 2. Петренко В. Ф. К проблеме психологии сознания // Вопросы философии. 2010. № 11. С. 57 – 74. 2. Петренко В. Ф. К проблеме психологии сознания // Вопросы философии. 2010. № 11. 3. Межуев В. М. Идея культуры: очерки по философии культуры. М., 2006. 346 с. 4. Мохова С. Ю. Кризис современной российской культуры: личность в пространстве социальных взаимо и преобразований // Психология в экономике и управлении. 2010. № 1(3). С. 81 – 90. р р у р ( ) 5. Кэмпбелл Д. Тысячеликий герой = The Hero with thousand faces / пер. с англ. А. П. Хомик. К., М.: Ваклер; Рефл- бук; АСТ, 1997. 384 с. р р у р ( ) 5. Кэмпбелл Д. Тысячеликий герой = The Hero with thousand faces / пер. с англ. А. П. Хомик. К., М.: Ваклер; Рефл- бук; АСТ, 1997. 384 с. 6 Ф ф б / б В В М М Н 2014 928 р р у р ( ) 5. Кэмпбелл Д. Тысячеликий герой = The Hero with thousand faces / пер. с англ. А. П. Хомик. К., М.: Ваклер; Рефл- бук; АСТ, 1997. 384 с. 6. Философия: учебник для вузов / под общ. ред. В. В. Миронова. М.: Норма, 2014. 928 с. 7. Улыбина Е. В. Обыденное сознание: структура и функции. М.: Смысл, 2001. 266 с. 8. Шеффер Б., Шледер Б. Социальная идентичность и групповое сознание как медиаторы межгруппового поведе- бук; АСТ, 1997. 384 с. 6. Философия: учебник для вузов / под общ. ред. В. В. Миронова. М.: Норма, 2014. 928 с. 7. Улыбина Е. В. Обыденное сознание: структура и функции. М.: Смысл, 2001. 266 с. 8 Ш фф Б Ш Б С ф у д у д щ р д р р , 7. Улыбина Е. В. Обыденное сознание: структура и функции. М.: Смысл, 2001. 266 с. Шледер Б. Социальная идентичность и групповое сознание как медиаторы межгруппового поведе- кий журнал. 1993. Т. 1. № 1. С. 34 – 56. ур 9. Московичи С. От коллективных представлений к социальным // Вопросы социологии. 1992. Т. 1. № 2. С. 83 – 96. 10. Вундт В. Проблемы психологии народов. М.: Академический проект, 2010. 136 с. 9. Московичи С. От коллективных представлений к социальным // Вопросы социологии. 1992. Т. 1. № 2. 19. Шкуратов В. А. Историческая психология. Ростов н/Д., 1994. 224 с. Литература С. 83 – 96. 10. Вундт В. Проблемы психологии народов. М.: Академический проект, 2010. 136 с. 9. Московичи С. От коллективных представлений к социальным // Вопросы социологии. 1992. Т. 1. № 2. С. 83 – 96. 10. Вундт В. Проблемы психологии народов. М.: Академический проект, 2010. 136 с. 11. Выготский Л. С. Психология развития как феномен культуры. М.: Институт практической психологии; Воро- неж: МОДЭК , 1996. 512 с. у р р р 11. Выготский Л. С. Психология развития как феномен культуры. М.: Институт практической психологии; Воро- неж: МОДЭК , 1996. 512 с. у р р р 11. Выготский Л. С. Психология развития как феномен культуры. М.: Институт практической психологии; Воро- неж: МОДЭК , 1996. 512 с. 12. Леонтьев А. Н. Деятельность. Сознание. Личность. М.: Политиздат, 1975. 304 с. б // Д д , 13. Ломов Б. Ф. Общение и социальная регуляция поведения индивида // Психологические проблемы социальной регуляции поведения. М., 1976. С. 64 – 93. 13. Ломов Б. Ф. Общение и социальная регуляция поведения индивида // Психологические проблемы социальной регуляции поведения. М., 1976. С. 64 – 93. р у , 14. Полежаев Д. В. Ментальность личности и внутренние механизмы динамики социальной направленности // Credo new. 2000. № 2(20). С. 60 – 67. р у 14. Полежаев Д. В. Ментальность личности и внутренние механизмы динамики социальной направленности // Credo new. 2000. № 2(20). С. 60 – 67. 15. Лурье С. В. Психологическая антропология: история, современное состояние, перспективы. М.: Академиче- ский Проект, 2003. 624 с. 15. Лурье С. В. Психологическая антропология: история, современное состояние, перспективы. М.: Академиче- ский Проект, 2003. 624 с. 16. Абульханова-Славская К. А. Деятельность и психология личности. М.: Наука, 1980. 335 с. 17. Поршнев Б. Ф. Социальная психология и история. М., 1966. 166 с. 16. Абульханова Славская К. А. Деятельность и психология личности. М.: Наука, 1980. 335 с. 17. Поршнев Б. Ф. Социальная психология и история. М., 1966. 166 с. швили Н. Л., Шрейдер Ю. А. Сознание как предмет метапсихологии // Системные исследования. М проблемы. Ежегодник. 1989 – 1990. М.: Наука, 1991. С. 190 – 204. 19. Шкуратов В. А. Историческая психология. Ростов н/Д., 1994. 224 с. 19. Шкуратов В. А. Историческая психология. Ростов н/Д., 1994. 224 с. 19. Шкуратов В. А. Историческая психология. Ростов н/Д., 1994. 224 с. 92 Вестник Кемеровского государственного университета 2017  № 1 References 1. Sudаs L. G. Mаssovoe soznаnie v poiskаkh novoi pаrаdigmy [Mass consciousness in search of new paradigm]. Moscow, 1996, 111. 2. Petrenko V. F. K probleme psikhologii soznаniiа [To the problem of psychology of consciousness]. Voprosy filosofii = Russian Studies in Philosophy, no. 11 (2010): 57 – 74. p y, ( ) 3. Mezhuev V. M. Ideiа kultury: ocherki po filosofii kultury [Ideas of culture. Essays on philosophy of culture]. Moscow, 2006, 346. p y ( ) Mezhuev V. M. Ideiа kultury: ocherki po filosofii kultury [Ideas of culture. Essays on philosophy of cult 46 , 4. Mokhovа S. Yu. Krizis sovremennoi rossiiskoi kultury: lichnost' v prostrаnstve sotsiаlnykh vzаimodeistvii i preobrаzovаnii [Crisis of the modern Russian culture: personality in space of social co-operations and transformations]. Psikhologiiа v ekonomike i uprаvlenii = Psychology in Economics and management, no. 1(3) (2010): 81 – 90. g p y gy g ( ) ( ) 5. Campbell J. Tysiachelikii geroi [The Hero with thousand faces]. Kiev; Moscow: Vakler; Refl-buk; AST, 1997, 384. 6. Filosofiiа [Philosophy]. Ed. Mironov V. V. Moscow: Norma, 2014, 928. 7 Ul bi Y V Ob d k f k [O di i i d f i ] M g p y gy g ( ) ( ) 5. Campbell J. Tysiachelikii geroi [The Hero with thousand faces]. Kiev; Moscow: Vakler; Refl-b p y gy g ( ) ( ) ell J. Tysiachelikii geroi [The Hero with thousand faces]. Kiev; Moscow: Vakler; Refl-buk; AST, 1997, 3 iа [Philosophy] Ed Mironov V V Moscow: Norma 2014 928 5. Campbell J. Tysiachelikii geroi [The Hero with thousand faces]. Kiev; Moscow: Vakler; Refl-buk; AST, 1997, 384. 6. Filosofiiа [Philosophy]. Ed. Mironov V. V. Moscow: Norma, 2014, 928. 7. Ulybinа Ye. V. Obydennoe soznаnie: strukturа i funktsii [Ordinary consciousness is а structure and functions]. Moscow: 6. Filosofiiа [Philosophy]. Ed. Mironov V. V. Moscow: Norma, 2014, 928. 7. Ulybinа Ye. V. Obydennoe soznаnie: strukturа i funktsii [Ordinary consciousness is а structure and functions]. Moscow: Smysl, 2001, 266. f p y 7. Ulybinа Ye. V. Obydennoe soznаnie: strukturа i funktsii [Ordinary consciousness is а structure and functions]. Moscow: Smysl, 2001, 266. 8. Shefer B., Shleder B. Sotsial'naia identichnost' i gruppovoe soznanie kak mediatory mezhgruppovogo povedeniia [So- cial identity and group consciousness as mediators intergroup behavior]. Psikhologicheskii zhurnal = Psychological journal, 1, no. 1 (1993): 34 – 56. ( ) 9. Moskovichi S. Nаdejdа P. Andryushkova1, @ Nаdejdа P. Andryushkova1, @ 1 Donetsk Nаtionаl University, 24, Universitetskaya St., Donetsk, Ukraine, 83001 @ аndruchkovа@rаmbler.ru Received 28.06.2016. Accepted 17.11.2016. Received 28.06.2016. Accepted 17.11.2016. Abstract: The article analyzes the phenomenon of ordinary consciousness as a cate- gory, whose largely contradictory and paradoxical nature evokes interest. It describes the main approaches to the study of ordinary consciousness and views on its nature in philosophical and psychological concepts. The article analyzes the ratio of ordi- nary consciousness with the related concepts of "mass consciousness", "common psychology", "group consciousness", "social consciousness". It describes the rela- tionship of ordinary consciousness with culture and its place in the life of an individ- ual as a cultural component, an inherent trait of any historical type of consciousness. The conclusion is that ordinary consciousness acts as an intermediary between the subject and the environment in which it is located, as well as between the indi- vidual and society. In addition to this, ordinary consciousness has not only a reflec- tive function, but also acts as a regulator of social and particularly group behavior. Received 28.06.2016. Accepted 17.11.2016. Keywords: consciousness, being, customs, prejudices, culture. For citation: Andryushkova N. P. Kategoriia «obydennoe soznanie» v psikhologii [The Category of "Mundane Conscious- ness" in Psychology]. Bulletin of Kemerovo State University, 2017; (1): 90 – 93. (In Russ.) DOI: 10.21603/2078-8975- 2017-1-90-93. For citation: Andryushkova N. P. Kategoriia «obydennoe soznanie» v psikhologii [The Category of "Mundane Conscious- ness" in Psychology]. Bulletin of Kemerovo State University, 2017; (1): 90 – 93. (In Russ.) DOI: 10.21603/2078-8975- 2017-1-90-93. ( ) 19. Shkuratov V. A. Istoricheskaia psikhologiia [Historical psychology]. Rostov-on-Don, 1994, 224. 18. Muskhelishvili N. L., Shreider Iu. A. Soznanie kak predmet metapsikhologii [Consciousness as a subject of metapsy- chology]. Sistemnye issledovaniia. Metodologicheskie problemy = System studies. Methodological problems, Yearbook. 1989 – 1990 (1991): 190 – 204. 17. Porshnev B. F. Sotsial'naia psikhologiia i istoriia [Social Psychology and History]. Moscow, 1966, 166. References Ot kollektivnykh predstаvleniy k sotsiаlnym [From collective presentations to sociаl one]. Voprosy sotsi- ologii = Questions of sociology, 1, no. 2 (1992): 83 – 96. Q f gy, , ( ) Wundt W. Problemy psikhologii narodov [Problems of the psychology of nations]. Moscow: Akademi 36. 11. Vygotsky L. S. Psikhologiia razvitiia kak fenomen kultury [Developmental psychology as a phenomenon of culture]. Moscow: Institut prakticheskoi psikhologii; Voronezh: NPO 'MODEK', 1996, 512. 12. Leontev A. N. Deiatelnost'. Soznanie. Lichnost' [Activities. Consciousness. Personality]. Moscow: Politizdat, 1975, 304. 13. Lomov B. F. Obshchenie i sotsial'naia reguliatsiia povedeniia individa [Communication and social regulation of indi- vidual behavior]. Psikhologicheskie problemy sotsial'noi reguliatsii povedeniia [Psychological problems of social regulation of behavior]. Moscow, 1976, 64 – 93. 14. Polezhаev D. V. Mentаlnost lichnosti i vnutrennie mekhаnizmy dinаmiki sotsiаlnoi nаprаvlennosti [Mentаlity of per- sonality аnd internal mechanisms of dynamics of social orientation]. Credo new, no. 2(20) (2000): 60 – 67. 15 Lur'e S V Psikhologicheskaia antropologiia: istoriia sovremennoe sostoianie perspektivy [Psychological anthropol 15. Lur'e S. V. Psikhologicheskaia antropologiia: istoriia, sovremennoe sostoianie, perspektivy [Psycho : history, modern condition, prospects]. Moscow: Akademicheskii Proekt, 2003, 624. gy y, , p p ] , , 16. Abulkhanova-Slavskaya K. A. Deiatel'nost' i psikhologiia lichnosti [Activities and Personality Psychology]. Moscow: Nauka, 1980, 335. 17. Porshnev B. F. Sotsial'naia psikhologiia i istoriia [Social Psychology and History]. Moscow, 1966, 166. 18. Muskhelishvili N. L., Shreider Iu. A. Soznanie kak predmet metapsikhologii [Consciousness as a subject of metapsy- chology]. Sistemnye issledovaniia. Metodologicheskie problemy = System studies. Methodological problems, Yearbook. 1989 – 1990 (1991): 190 – 204. 93
W2320044145.txt	https://zenodo.org/record/1525924/files/article.pdf	de		Beiträge zur Kenntnis der Gallenfarbstoffe.	Hoppe-Seyler's Zeitschrift für Physiologische Chemie	1,914	public-domain	6,411	Beiträge zur Kenntnis der Gallenfarbstoffe. Von William Küster. VII. Mitteilung. Über die Einwirkung von Bisenchlorid auf Bilirubin und über die Aufarbeitung von Gallensteinen. Von W. Küster, K. Reibung und R. Schmiedel. (Aus dem dritten chemischen Laboratorium der Technischen Hochschule, Stuttgart.) (Der Redaktion zugegangen am 3. April 1914.) In der fünften Mitteilung1) über die Gallenfarbstoffe wurde die Aufspaltung des Bilirubins durch Alkalien beschrieben und, der Ansicht Staedelers folgend, der die Bildung von Biliverdin aus Bilirubin mit dem Entstehen von Indigo aus Indigweiß verglichen hat, die Frage aufgeworfen, ob man das Bilirubin zu den indigo'iden Farbstoffen, resp. deren LeukoVerbindungen rechnen darf.2) Dabei war zu berücksichtigen, daß die Pa) Diese Zeitschrift, Bd. 59, S. 63 (1909). ) Für diesen Vergleich sprach gerade die Möglichkeit, beide Körper durch Alkalien aufspalten zu können; auch habe ich bereits die aus dem Indigo entstehende Anthranilsäure mit der aus dem Biliverdin hervorgehenden Hämatinsäure in Parallele gesetzt. Zieht man dann aus den Konstitutionsformeln der genannten Säuren einen Schluß auf die Komplexe, aus denen sie hervorgehen, so ergibt sich, daß der die Hämatinsäure liefernde ein Kohlenstoffatom mehr enthalten muß, als der, aus dem die Anthranilsäure entsteht. Da letzterer aber einen Pyrrolring vorstellt, muß jener ein Pyridinring sein, der sich aus einem der vier Pyrrolkomplexe des Hämatins bilden könnte, vielleicht dadurch, daß ein Methin, das im Hämatin die Verbindung zweier Pyrrolringe bewirkt, in den Ring aufgenommen wird. Vgl. die Formulierung des Hämatins, Diese Zeitschrift, Bd. 82, S. 469 (1912). Die von H. Fischer aufgestellte Formel für das Bilirubin (Diese Zeitschrift, Bd. 89, S. 262 [1914]) trägt m. E. diesen Verhältnissen keine Rechnung. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM Beiträge zur Kenntnis der Gallenfarbstoffe. VII. 59 rallele zwischen beiden Farbstoffen nur haltbar blieb, wenn der Übergang von Bilirubin C32H3606N4 in Biliverdin C32H3608N4 > nicht einfach in einer Aufnahme von zwei Atomen Sauerstoff bestand, sondern in einer Aufnahme von Wasser und von zwei Atomen Sauerstoff unter Wiederabspaltung von Wasser im Sinne der Gleichungen: 1. C32H3606N4 + H20 = C32H3807N4, 2. C32H3S07N4 + 2 0 = C32H3008N4 + H20, d. h. das Bilirubin mußte in zwei Komplexen das Wasser und ein Atom Sauerstoff einlagern, während aus den beiden anderen Komplexen die Wegnahme zweier Atome" Wasserstoff erfolgte.1) Es durfte also keineswegs das ganze Bilirubinmolekül mit dem Indigweiß verglichen werden, sondern nur die Hälfte desselben. Eine wenn auch noch nicht ganz sichere Bestätigung dieser Vorstellungen wird nun durch die Einwirkung von Eisenchlorid auf eine Suspension von Bilirubin in konzentrierter Essigsäure erbracht,, welche Reaktion den Gegenstand der nachfolgenden Untersuchung bildet. Bei einer qualitativen Probe wurde die Beobachtung gemacht, daß das genannte Oxydationsmittel Bilirubin in einen grünen Farbstoff überführt, und es war also zu untersuchen, ob diese mit Farbvertiefung verknüpfte Oxydation in einer Wegnahme von Wasserstoff oder in einer Aufnahme von Sauerstoff besteht Die Analyse von Präparaten verschiedener Herstellungen ergab nun Werte, die mit der Formel C32H3407N4 für das bisher allerdings nicht deutlich krystallisiert erhaltene Oxydationsprodukt in Übereinstimmung stehen, wonach wir es also mit einem Dehydrooxybilirubin zu tun haben, d. h. es dürfte zu gleicher Zeit eine Wegnahme zweier Wasserstoffatome und eine Aufnahme von Sauerstoff erfolgt sein. Der neue Körper unterscheidet sich vom Biliverdin durch seine Schwerlöslichkeit in Alkohol, er gibt die Gmelin-Tiedemannsche Reaktion, d. h. die essigsaure Lösung wird durch Salpetersäure rot, er ist also viell ) Aus den Untersuchungen H. Fischers folgt, daß im Bilirubin mindestens drei von den vier im Hämatin vorhandenen Pyrrolkomplexen noch enthalten sind. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM 60 W. K ü s t e r , K. Reibung und R. Schmiedel, leicht identisch mit dem das grüne Stadium dieser Reaktion bedingenden Farbstoff, er gibt auch die Diazoreaktion, aber augenscheinlich kein Kondensationsprodukt mit Formaldehyd, eine Reaktion, die beim Bilirubin unter dem Einfluß konzentrierter Schwefelsäure zu einem intensiv blau gefärbten Körper führt. Beim Stehen einer alkalischen Lösung des Dehydrooxybilirubins an der Luft findet eine sichtbare Farbveränderung nicht statt, beim Kochen mit Alkalien wird etwa ein Viertel des vorhandenen Stickstoffs als Ammoniak abgespalten, also nicht mehr als aus dem Bilirubin, dabei findet Abspaltung von Hämatinsäure statt. Normal zusammengesetzte Salze wurden vom Dehydrooxybilirubin nicht erhalten, bemerkenswert erscheint, daß höchstens drei Atome Silber aufgenommen werden, während das Bilirubin eine Silberverbindung mit fast vier Atomen Metall gibt. Die Reduktion mit Natriumamalgam scheint nicht zum Hemibilirubin zu führen; zwar wird die braungrüne alkalische Lösung unter dem Einfluß des Reduktionsmittels fast farblos, doch konnte das Entstehen eines komplexen Kupfersalzes hier nicht beobachtet werden. Das in Essigsäure gelöste Dehydrooxybilirubin wird durch Natriumamalgam anscheinend nicht reduziert,1) wohl aber durch Zinkstaub und durch schweflige Säure und zwar nur bis zu,, einem die Lösung rot färbenden Körper, der in Chloroform nicht löslich und daher mit Bilirubin nicht identisch ist. Die Oxydation des Bilirubins durch Eisenchlorid führt nun nicht nur zum Dehydrooxybilirubin, sondern auch noch zu einem zweiten Körper, der durch die Unlöslichkeit seiner Alkalisalze in Wasser charakterisiert ist. Ihm kommt wahrscheinlich die Formel (G16H1605N2) zu; wir haben es also ebenfalls mit einem Körper zu tun, der durch Wegnahme von Wasserstoff und ) Es möge erwähnt sein, daß es mir nicht gelungen ist, mit den von mir geprüften Natriumamalgampräparaten eine Reduktion des p-Benzochinons in essigsaurer Lösung zu erreichen, dieselbe Lösung wurde von Zinkstaub sofort entfärbt. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM Beiträge zur Kenntnis der Gallenfarbstoffe. VII. 61 Aufnahme von Sauerstoff aus dem Bilirubin entstanden .ist. Wir haben diesen Körper seiner Farbe wegen «Bilinigrin» genannt. Seine Bildung geht Hand in Hand mit einer bemerkenswerten Erscheinung. Es bildet sich nämlich zunächst eine lose Verbindung zwischen Bilirubin, Dehydrooxybilirubin und Bilinigrin, was besonders ins Auge fällt, wenn man zu der Suspension des Bilirubins in Eisessig auf einmal viel Eisenchlorid fügt. Es sind dann die Ausbeuten an Dehydrooxybilirubin herabgesetzt, der größte Teil des Farbstoffs hat sich in eine schwarze Masse verwandelt, die auch an kochende Essigsäure schließlich nur noch Spuren von Farbstoff abgibt. .Derselben läßt sich aber durch Chloroform unverändertes Bilirubin entziehen, worauf an Essigsäure auch wieder Dehydrooxybilirubin abgegeben wird, während Bilinigrin zurückbleibt, nachdem die Extraktion mit Chloroform und Essigsäure-wiederholt worden ist. Dieses Verhalten erweckt den Anschein, daß in dem zunächst in Essigsäure nicht löslichen Körper eine Verbindung von Bilirubin mit seinen Oxydationsprodukten vorliegt, die an die Verbindungen chinhydronartigen Charakters erinnert. Auch^ durch eine verdünnte Sodalösüng wird die Verbindung nur langsam zerlegt, so daß es mehrstündiger Ein- Wirkung bedarf, ehe das Bilirubin in Lösung zu gehen beginnt, wobei die Unlöslichkeit des Bilinigrinnatriums jedenfalls eine Rolle spielt. Wahrscheinlich handelt es sich aber doch nur um eine Adsorption, z. B. war es nicht möglich, in .Chloroform gelöstes Bilirubin durch Chinon zu oxydieren, auch ist €s bisher nicht gelungen, das Bilirubin zu acetylieren oder zu benzoylieren. Ähnliche Beobachtungen macht man ja auch schon bei der Aufarbeitung von Gallensteinen. Wie ich bereits beschrieben habe, läßt sich nach der Wegnahme des Bilirubins mit Hilfe von Chloroform durch konzentrierte Essigsäure, mit der vorher bereits bis zur Erschöpfung extrahiert worden war, von neuem grüner Farbstoff herauslösen, worauf Chloroform wieder Bilirubin entzieht. Bei einer neuen Aufarbeitung von Gallensteinen wurde nun so verfahren, daß nach der Einwirkung von verdünnter Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM 62 W. Küster, K. R e i b u n g und R. Schmiede], Essigsäure mit Chloroform extrahiert wurde, wobei ein hellorangerot gefärbtes Bilirubin erhalten wurde. Nachdem dann durch konzentrierte Essigsäure das sogenannte Choleprasin entfernt worden war, wurde erneut mit Chloroform extrahiert und nunmehr ein prachtvoll rotbraun gefärbtes Bilirubin erhalten. Demnach scheint letzteres an das Choleprasin gebunden gewesen zu sein und ist erst durch die siedende konzentrierte Essigsäure aus dieser Bindung herausgelöst worden. Vielleicht handelt es sich hier also doch um eine chemische Bindung, die durch die Essigsäure unter Einlagerung von Wasser gesprengt wird. In diesem Falle könnte der Befund dahin gedeutet werden, daß die Umwandlung der eisenhaltigen Komponente des Blutfarbstoffs in Gallenfarbstoff noch im Verbände mit dem Globin erfolgt, da ja das Choleprasin als ein Eiweißabkömmling aufgefaßt werden muß. Übrigens enthält auch das rohe Choleprasin immer noch etwas Bilirubin, das jenem durch Chloroform in Gestalt eines grünen Farbstoffs entzogen wird, der bei erneuter Einwirkung von Essigsäure das Bilirubin hinterläßt. Wenn dann die Aufarbeitung der Gallensteine soweit getrieben worden ist, daß weder Chloroform noch Eisessig organische Substanz mehr herausnehmen, so hinterbleibt ein Körper, der mit dem «Bilinigrin» insofern Ähnlichkeit hat, als auch er durch die ünlöslichkeit seiner Alkaliverbindungen ausgezeichnet ist. Es dürfte also auch in diesem «Bilihumin» ein Oxydationsprodukt des Bilirubins auf natürlichem Wege enthalten sein, und die Bildung der Gallensteine könnte demgemäß damit beginnen, daß sich das unlösliche Natriumsalz dieses Bilihumins zunächst abscheidet und daß auf diesen einmal gebildeten Kern weitere Ablagerungen erfolgen. Experimenteller Teil. Die Oxydation von Bilirubin durch Eisenchlorid wurde meistens in der Weise bewirkt, daß eine Aufschwemmung von z. B. 5 g Bilirubin in 500 ccm Eisessig auf dem Wasserbade erhitzt wurde, worauf 20 ccm der offizinellen Eisenchloridlösung, entsprechend zwei Atomen Sauerstoff auf die Molekel Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM Beiträge zur Kenntnis der Gallenfarbstoffe. VII. 63 Bilirubin, zugegeben wurden, nachdem sich gezeigt hatte, daß ein Überschuß von Eisenchlorid in Beziehung auf die Ausbeuten an grünem Oxydationsprodukt keinen Vorteil bringt und daß die Verwendung von wasserfreiem Eisenchlorid einen anderen Verlauf der Reaktion nicht zur Folge hat. Bis zur Vollendung der Oxydation, d. h. bis zum völligen Verschwinden unveränderter Bilirubinpartikel muß nunmehr einige Stunden auf dem Wasserbade erhitzt werden, worauf durch ein Kreppfilter filtriert und der Rückstand 8—10mal mit Eisessig ausgekocht, wurde, bis letzterer fast farblos ablief. Von den vereinigten Farbstofflösungen wurde der Eisessig unter vermindertem Druck bis etwa auf x /io des ursprünglichen Volumens abdestilliert und die konzentrierte Lösung in Wasser gegossen, wobei sich der Farbstoff abschied. Der Zusatz eines Elektrolyten wie Ammoniak oder auch Salzsäure beschleunigte und vervollständigte das Ausfallen des Farbstoffs. Er wurde dann filtriert, mit verdünnter Salzsäure eisenfrei und mit Wasser säurefrei gewaschen, zur Reinigung erneut in heißem Eisessig gelöst und das Lösungsmittel bis auf einen kleinen Rest abdestilliert, worauf der Farbstoff wieder durch Eintragen in Wasser gefällt wurde. Der schwarze in Eisessig unlösliche Rückstand wurde eisen- und säurefrei gewaschen und nach dem Trocknen mit Chloroform extrahiert, wobei namentlich bei einigen Versuchen, bei denen ein Überschuß von Eisenchlorid verwendet worden war, ganz erhebliche Mengen unveränderten Bilirubins ententfernt wurden, was dafür spricht, daß, wenn sofort eine größere Menge des Oxydationsprodukts gebildet wird, eine entsprechend große Menge von Bilirubin niedergerissen wird, d. h. sich in einer durch Essigsäure nicht trennbaren Verbindung mit dem Oxydationsprodukt ausscheidet. Denn nach der Extraktion durch Chloroform nahm dann auch Essigsäure wieder erhebliche Mengen des grünen Farbstoffs auf. Schließlich hinterblieb eine schwarze Masse, die weder an Chloroform noch an Essigsäure mehr als Spuren organischer Substanz abgab, und die als Bilinigrin bezeichnet werden soll. Zur Übersicht diene die folgende Tabelle: Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM 64 Versuchsn u mmor W. Küster, K. Reihling und R. Schmiedel, Bilirubin Eis- essig Ausbeute 0-Atomc auf eine graner in % doe echwar- nicht verEisenchlorid Mol. Biliverwen- zer ' ändertes Farb- deten BiliBili- Rückrubin stoff rubins stand rubin 12,5g 1 rohes, 2,8g 500 500 8,4 > 2 reines, 2,0 > 24,5 > 3 rohes, 8,5 > 500 » 8 » 500 4 27 » » 6 » 500 24 ccm ^ 5 6 » 5 » 500 20 » > 3 » 300 15 » 7 8 3 » 300 18 » 9 5 > 500 30 » 10 5 » 500 40 » 5 > 500 20 » 11 12 5 » 500 20 » 7 7 5 6 2 2 2,5 3 3 4 2 2 1,0 g 36 '/o zusammen 1,4 g 0,5 > 0,5 > 25 °/o 2 » 2,8 » 33°/o 2,25g 2,65» 33°/o 0,55 > 3,3» vorp 9 1,5 > handen 2,4 g 48°/o 1,7 > 1,4 » 47 > 1,2 > immer vor1,38» 46°/o 1,2 > handen, I.D. 2,5 > 50°/o 1,65 > 6—7% 2,4 » 48 °/o 1,7 wieder ge2,6 » 52 °/o 1,7 > wonnen 2,7 » 54°/o 1,8 > Wie aus der Tabelle hervorgeht, betragen die Ausbeuten an grünem Farbstoff im besten Falle 54?°/o des angewandten Bilirubins und zwar bei Verwendung einer nur zwei Atomen Sauerstoff auf die Molekel des Bilirubins gleichkommenden Menge von Eisenchlorid. Aus der Analyse der erhaltenen Produkte geht aber hervor, daß bedeutend mehr als zwei Atome Sauerstoff aufgenommen werden. Somit dürfte die sich bildende Ferroverbindung auch hier nur die Rolle eines Sauerstoff übertragenden Mittels spielen. 1. 0,1468 g Subst. (120 °): 0,3504 g C02, 0,0786 g H20 und Spuren Fe20s. 0,1520 > : 12,7 ccm N, 724,5 mm B., 13°. 2. 0,1583 > (120°): 0,3794 g C02, 0,0875 g H20 u. 0,0012 g Asche, worin Fe203. 0,1072 » » : 9,2 ccm N, 723 mm B., 16°. 3. u. 4. 0,1332 » » (100°): 0,3179 g C02, 0,0742 g H20 u. 0,0013 g Asche, worin Fee03. 0,1414 » » : 11 ccm N, 749 mm B., 12°. 3. u. 4. Versuch, die durch Essigsäure zuletzt extrahierten Anteile. 0,1488 g Subst. (100°): 0,3543 g C02, 0,0795 g H20, 0,0014 g Fe203. 5.0,1708» » (110°): 0,4072» » 0,0885» » 0,0008» » 0,1378 » » (110°): 10,8 ccm N bei 11° und 739 mm B. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM Beiträge zur Kenntnis der Gallenfarbstoffe. VII. 6.—12. 0,3135 g 0,2493» 0,2065» 0,1566 » CSiHM07N4. Subst. > » » 65 : 0,7480 g CO, und 0,1693 g H20. : 0,5965 > » » 0,1380 > > : 0,4915 » » » 0,1143» » : 12,9 ccm N bei 18° und 750 mm. B. Ber.: 65,5e/o C Gef. : 1. 65,1 2. 65,86 3. u. 4. 65,73 3. u. 4. 65,55 5. 65,33 6.— 12. 65,07 65,26 64,92 5,8°/o H 5,95 6,20 6,26 6,00 5,8 6,0 6,15 6,15 9,55 > N 9,4 9,6 9,19 — 9,1 9,4 — — Aus diesen auf aschefreie Substanz bezogenen Analysen läßt sich die Formel C32HS407N4 berechnen, -wenn auch C32H3607N4 nicht ganz ausgeschlossen ist, die 6,12 °/o H verlangen würde. Sie erscheint mir aber weniger wahrscheinlich, und ich möchte daher das neue Oxydationsprodukt des Bilirubins als ein Dehydrooxybilirubin ansprechen. Es löst sich reichlich nur in konzentrierter Essigsäure mit grüner Farbe, verliert aber beim Lagern auch für dieses Mittel die Löslichkeit, in kaltem Alkohol ist es kaum löslich, etwas besser — mit blaugrüner Farbe — in siedendem und scheidet sich aus der heißen Lösung in deutlich abgegrenzten Partikelchen beim Erkalten ab. In anderen organischen Lösungsmitteln (Äther, Chloroform, Dimethylanilin), sowie in Wasser, ist es unlöslich. In Alkalien, auch in Natriumbicarbonat, ist es löslich und zwar in Natronlauge und Ammoniak mit grünbrauner Farbe. Bei einigen quantitativen Versuchen zeigte es sich, daß zwei Moleküle Alkali nicht ausreichen, um ein Molekül Dehydrooxybilirubin in- l°/oige wässerige Lösung zu bringen. Möglichst neutrale Lösungen des Farbstoffs in verdünnter Kalilauge werden durch Baryum-, Magnesium- und Silbersalze nicht vollständig gefällt, dies gelang nur mit Ferroammonsulfat. Doch enthielt keine der Fällungen den berechneten Metallgehalt, wahrscheinlich wurde ein kleiner Teil des Kaliums nicht ersetzt. Das Ferro- und das Silbersalz waren in Alkalien kolloid löslich. Bemerkenswert ist, daß höchstens drei Atome Silber Hoppe-Seyler's Zeitschrift f. physiol. Chemie. XCI. 1 5 Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM 66 W. Küster, K. Reihling und R, Schmiedel, aufgenommen werden, während Bilirubin ein Salz mit fast 4? Atomen gibt. l g DehydrooxybiJirubin vom fünften Versuch wurde mit 100 g Wasser und einem kleinen Überschuß von Ammoniak gelöst und die filtrierte Lösung im Vakuum unter Schwefelsäure belassen, bis das überschüssige Ammoniak abgedampft war, worauf durch 200 ccm 1 /2°/oiger Silbersulfatlösung gefällt und der entstandene Niederschlag abfiltriert wurde. Das Filtrat war hier nur schwach gelb gefärbt. Beim Auswaschen ging der voluminöse Niederschlag in dem Moment in kolloide Lösung, als die Elektrolyten entfernt waren. Er wurde im Vakuum und für die Analyse noch bei 100° getrocknet, wobei ein Gewichtsverlust von 6,28 °/o stattfand. 0,2420 g Substanz (100°): 0,3743 g C02, 0,0832 g H20 und 0,0773 g Ag1) Ber. für C32H3lAg307N4: 42,46°/o C, 3,4°/o H, 35,64°/o Ag ' Gef.: 42,18 °/o > 3,8 °/o » 32,21 °/o > Dieses Silbersalz ließ sich durch Cyankalium in Lösung bringen, durch Kalilauge entstand eine kolloide Lösung, aus welcher Jodkalium Jodsilber fällte. Durch weitere Einwirkung von Eisenchlorid auf die essigsaure Lösung des Dehydrooxybilirubins findet eine allmählich sich vollziehende Umwandlung in einen 'in Essigsäure unlöslichen Körper statt, der mit «Bilinigrin» identisch sein dürfte, da er wie dieses unlösliche Alkalisalze bildet.2) Einige Reaktionen des Dehydrooxybilirubins sind bereits in der Einleitung beschrieben worden. II. Über das Bilinigrin. Mit «Bilinigrin» haben wir den schwarzen, in Essigsäure unlöslichen Körper bezeichnet, der bei der Oxydation des Bilirubihs durch Eisenchlorid entsteht und nach der Entfernung unveränderten Bilirubins durch Chloroform und des Dehydrooxybilirubins durch Essigsäure zurückbleibt. Dieser Rückstand erwies sich auch in allen anderen gebräuchlichen organischen 1 ) Die Asche enthielt etwas Eisen, sie wurde in Salpetersäure gelöst und das Silber als AgCl gefällt, gefunden 0,1026 g. 2 ) Vgl. S. 12. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM Beiträge zur Kenntnis der Gallenfarbstoffe. VII. 67 Lösungsmitteln als nicht löslich, eine krystallinische Struktur zeigte er nicht. Die Analyse des wie atigegeben isolierten Körpers wies auf eine durch die Formel (C16H1605N2) wiederzugebende Zusammensetzung.1) I. 0,1645 g Substanz 0,3656 g CO, und 0,0740 g HS0. 0,1400 > 12 ccm N bei 22° und 744 mm B. II. 0,2402 » 0,5341 g C02 und 0,1095 g H80. 0,1280 » 10,3 ccm N bei 21° und 748 mm B. III. 0,2493 » 0,5553 g C0ä und 0,1169 g H20. 0,1163 » 9,5 ccm N bei 18° und 746 mm B. Ber. ( 6 1605 2) : 60,7°/o C 5,06°/o H 8,86°/o-N Gef.: 1.60,61 4,98 9,4 II. 60,64 5,06 8,96 III. 60,75 5,21 9,2 Diese Präparate gaben indessen an Alkalien noch etwas Substanz ab, und es kann daher die Formel nicht als endgültig bewiesen angesehen werden.2) Es ist nämlich bemerkenswert und wichtig wegen der Analogie mit dem Bilihumin, d. h. mit dem bei der Aufarbeitung der Gallensteine als in jedem Lösungsmittel unlöslichen Rückstande, daß das Bilinigrin unlösliche Alkalisalze bildet. Wir machten diese Beobachtung bei dem Versuch, die letzteren zu gewinnen, und stellten zunächst fest, daß der bei der Einwirkung von Sodalösung erhaltene Farbstoff aschehaltig geworden war und dann, daß nach dem Schütteln von 0,5 g Bilinigrin mit 10 ccm n-Na2C03-Lösung und 50 ccm Wasser in den Filtraten nur noch 8,4 ccm n-Alkali vorhanden waren. 0,5 g Bilinigrin hatten also zur Neutralisation 1,6 ccm n-Alkali oder 0,0368g Na verbraucht, woraus sich berechnet, daß ein Grammatom Natrium in 312 g des Farbstoffs enthalten war, was mit dem durch obige Formel repräsentierten Äquivalentgewicht (316) merkwürdig gut übereinstimmte. Es wurden nun 5,4 g des rohen Bilinigrins mit überschüssiger Natrium) Die geringe Aschenmenge ist abgezogen worden. . ) Auch nach anscheinend erschöpfender Behandlung mit Chloroform gab ein. Präparat von Bilinigrin an Soda noch Bilirubin und zwar erst nach längerer Einwirkung ab, welche Beobachtung dafür spricht, daß beide Farbstoffe sich enger miteinander verbinden können.T 5* Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM 68 W. K ü s t e r , K. Reibung und R. Schmiedel, carbonatlösung geschüttelt und das unlösliche Salz mit Wasser ausgekocht, wonach 5,42 g zurückblieben. 1,02 g dieses Salzes gaben dann an HC1 0,1763 g NaCl ab = 0,0693 g Na. Bcr. CieHlGNa06N2: 6,8°/o Na. Gef.: 6,7 °/o Na. Auf gleichem Wege wurde das Kaliumsalz dargestellt und aus 0,5 g Büinigrin 0,53 g des Salzes erhalten. 0,530 g Substanz, mit Salzsäure zersetzt, gaben 0,1112 g KC1 = 0,0582 g K. Ber. C16H16K06N2: 11,02°/o K. Gef.: 10,98% K. Beim Schütteln dieser unlöslichen Salze mit einer Silbernitratlösung trat Silber an Stelle der Alkalien, ein vollständiger Ersatz wurde bei zweimaliger Einwirkung erreicht. Ein solches Silbersalz gab bei der Analyse folgende Werte: 0,2173 g Substanz (100°): 0,3627 g C0.2, 0,0754 g H20 und 0,0567 g Ag. Ber. C16H16Ag06N2: 45,39 °/o C, 3,55 °/o H, 25,53 °/o Ag. Gef.: 45,52°/o » 3,86% » 26,09> > Zu bemerken ist noch, daß beim Schütteln von Büinigrin mit sehr verdünnten Alkalicarbonatlösungen nur die Hälfte der Kaliummenge aufgenommen wurde, die aus einer konzentrierteren Lösung eintrat. Ein ganz analoges Verhalten wurde beim Bilihumin festgestellt. Die Analyse eines Kaliumsalzes, das aus Büinigrin durch Einwirkung einer ^lo-n-Kaliumcarbonatlösung hervorging, ergab: 0,4357 g Substanz lieferten 0,0410 g KC1 = 0,0215 g K = 4,94 °/o K. Der aus den Alkalisalzen durch Einwirkung von Salzsäure wiedergewonnene Farbstoff gab dann bei der Analyse von den aus der Formel G16H1605N2 berechneten abweichende Werte. Es ist nicht ausgeschlossen, daß die Salzsäure das Büinigrin selbst angegriffen hatte, wie es denn auch bei der Einwirkung von stärkeren Alkalien zu einer Zersetzung unter Abspaltung von Ammoniak kommt. 0,778 g Büinigrin wurden mit 250 ccm l°/oiger Natronlauge längere Zeit gekocht, wobei vollständige Lösung nicht erzielt wurde, hierbei destillierten zunächst 0,6 ccm n-NHs, am zweiten Tage noch 0,29 ccm über. Diese entsprechen 0,01246 g N oder, da 0,07 g N vorhanden waren, etwa 18°/o des vorhandenen Stickstoffs. Das Büinigrin spaltet also wohl ebenfalls wie Bilirubin und Dehydrooxybüirubin 1U des in ihm enthaltenen Stickstoffs ab, aber sehr viel langsamer. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM Beiträge zur Kenntnis der Gallenfarbstoffe. VII. 69 Das Auftreten von Hämatinsäüre konnte bei dieser Zersetzung nicht festgestellt werden. Bei weiterer Einwirkung von Eisenchlorid auf Dehydrooxybilirubin wurde in sehr geringer Menge ein Körper erhalten, der ein unlösliches Natriumsalz gab, aus dessen Analyse vielleicht die Identität mit Bilinigrin gefolgert werden kann. 2,7 g Dehydrooxybilirubin vom Versuch 3 und 4 wurden in essigsaurer Lösung 4 Stunden nach Zusatz von 10 ccm offizineller Eisenchloridlösung erwärmt, wobei sich ein Niederschlag in der Menge von 0,45 g bildete. Er wurde säureund eisenfrei gewaschen und dann mit 5°/oiger Sodalösung behandelt, wobei 0,16 g eines unlöslichen Natriumsalzes zurückblieben. 0,1417 g Substanz (120°) gaben durch Behandlung mit verd. Salzsäure 0,0220 g NaCl; die Substanz enthielt mithin 6,14 °/o Na. III. Die Aufarbeitung von Gallensteinen ist des öfteren von mir behandelt worden. Den früher gemachten Beobachtungen möchte ich noch hinzufügen, daß sich die Extraktion des Pulvers in der von H. Fischer 1 ) beschriebenen Röhre an Stelle des Soxhletschen Extraktionsapparates sehr bewährt hat, und daß dem mit Eisessig extrahierten «Choleprasin* stets Bilirubin beigemengt ist. Wird nämlich das getrocknete rohe Choleprasin mit Chloroform ausgezogen, so geht eine grün gefärbte Substanz in Lösung, die beim Aufnehmen in Essigsäure Bilirubin hinterläßt. Bei einer vor 11/2 Jahren vorgenommenen Aufarbeitung von 600 g Gallensteinen wurde nun so verfahren, daß nach der Behandlung mit Äther, Wasser, verdünnter Essigsäure, Äther und Alkohol der jetzt frei gemachte Teil des Bilirubins durch Chloroform extrahiert wurde, wobei 97g eines hellorangerot gefärbten rohen Präparates gewonnen wurden, das sich im Extraktionskolben abgesetzt hatte, leichter lösliche Teile waren nur in sehr geringer Menge (3,5 g) vorhanden. Die ') Diese Zeitschrift, Bd. 73, S. 216 (1911. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM 70 W. Küster, K. Reibung und R. Schmiedel, Reinigung des rohen Bilirubins erfolgte hier nur durch eine Behandlung mit konzentrierter Essigsäure in der Wärme und durch Aus waschen mit Alkohol, wonach 91,5 g hinterblieben. Die Analyse dieses Präparates ergab keine befriedigende Übereinstimmung mit den aus der Formel C32H3606N4 berechneten Werten: 0,1496 g Substanz (110°): 0,3619 g C02, 0,0788 g H20. Spuren Asche = 65,98 °/o C u. 5,86 °/o H. 0,1153 g Substanz (110°): 9,7 ccm N bei 16,5° u. 745 mm B = 9,6 °/o N. 0,2003 » > (110°): 0,0016 g AgCi (Carius) = ca. 0,2 °/o Cl. Schwefel ließ sich nicht nachweisen. Ein Teil des Präparates wurde darauf in verdünnter Natronlauge gelöst und die übrigens restlos filtrierende Lösung durch verdünnte Schwefelsäure gefällt; dieses S04-frei gewaschene und im Vakuum getrocknete Präparat stellte muschlig brechende, fast schwarzrote Stücke vor und sah im gepulverten Zustande dunkelrotbraun aus. Eine Reinigung war auf diese Weise aber nicht geglückt. 0,1597 g Substanz (120°): 0,3868 g C02, 0,0793 g H20, 0,0006 g Asche = 66,3°/o C und 5,54°/<> H. Erst eine Umkrystallisation nach folgendem Verfahren führte zum Ziel. Es wurde je l g Bilirubin und Chinin mit 100 ccm Chloroform bis zur Lösung geschüttelt, was etwa im Verlauf einer halben Stunde erreicht war, worauf filtriert und etwa die Hälfte des Chloroforms vorsichtig abdestilliert wurde. Die konzentrierte Lösung wurde in 250 ccm siedenden Eisessigs eingetragen. Läßt man nun auf dem Wasserbade stehen, bis das. Chloroform fa,st verdunstet ist, so krystallisiert das Bilirubin in rhombischen Säulen und in langgestreckten Nadeln aus. Es wird von der dunkelgrünen Lösung abfiltriert, mit Essigsäure und mit Alkohol bis zum Verschwinden der Chininreaktionen ausgewaschen und im Vakuum getrocknet. Hierbei wurden aus l g Rohbilirubin durchschnittlich 0,7 g reines Bilirubin von rotbrauner Farbe erhalten. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM Beiträge zur Kenntnis der Gallenfarbstoffe. VII; 71 L 0,1247 g Substanz (120°): 0,3067 g C02, 0,0713 g H20. 0,1345 > >(120°): 11,2 ccm N bei 15° und 745 mm B. II. 0,2972 » > : 0,7311 g C02 und 0,1672 g , .1) * 0,0689 > » : 6,2 ccm N bei 11° und 720 mm B. CS2H8eOeN4. Ber.: 67,16 °/o C, 6,29 °/o H, 9,79 °/o N Gef.: 1. 67,14 °/o > 6,35 °/o » 9,56 °/o > . II. 67,09 °/o > 6,26 °/o > 10,3 °/o > Aus .dem anfangs hellorange gefärbten Rohprodukt war also, was die Farbe betrifft, einmal ein amorphes Präparat von dunkelroter Farbe, das andere Mal ein krystallisierles, braunrotes Bilirubin hervorgegangen. Es ist daher wohl nicht. angängig, wegen der verschiedenen Farbe verschiedene Namen zu geben, wie es M. Piettre 2 ) getan hat, der ein Biliflavin vom Bilirubin unterscheidet, zumal wenigstens unser Rohprodukt Beimengungen enthielt, und die Angabe von Piettre, daß aus Bilirubin bei der alkalischen Hydrolyse Fettsäuren entstehen, vermuten läßt, daß ein nicht gereinigter Farbstoff verwendet wurde. Schließlich wurde auch die Leichtlöslichkeit des Ammoniakanlagerungproduktes in Methylalkohol3) mit Erfolg zur Reinigung des Rohbilirubins benutzt, eine Methode, die den Vorteil bietet, mit verhältnismäßig hoch konzentrierten Bilirubinlösungen arbeiten zu können. l g Rohbilirubin wurde mit 10 ccm Methylalkohol überschichtet und auf die Suspension trockenes Ammoniakgas geleitet, wodurch sich der Farbstoff augenblicklich löst. Die ohne Rückstand filtrierte 'Lösung wurde dann mit Äther gefällt und der hellrot gefärbte Niederschlag abgesaugt und mit Äther gewaschen.4) Beim Trocknen im Vakuum färbte er sich dunkelrot und erschien an den Rändern wie geschmolzen. Es tritt ) Das zweite Präparat stammte von einem änderen Rohbilirubin her. «) C. r. de l'Acad. des sciences, Bd. 148, S. 1213; Chem. Oentralbl., 1909, II, S. 135. 8 ) Diese Zeitschrift, Bd. 82, S. 476 (1912). 4 ) Aus den Filtraten schied sich beim Stehen in der Kälte zunächst noch etwas Substanz ab, die sich rasch grün färbte. Dann krystallisierte aber nach dem Abdestülieren des Äthers ein hellorange gefärbter Körper in sehr geringer Menge aus, dessen Analyse noch aussteht. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM 72 W. Küster, K. Reibung und R. Schmiedel, also augenscheinlich ein Verlust von Ammoniak ein, doch enthält das so getrocknete Präparat noch ein Molekül Ammoniak, das erst bei höherer Temperatur abgegeben wird.1) 0,2087 g Substanz (i. V): 21,4 ccm N bei 8° und 745 mm B. C82H8eOeN4NHa. Ber.: ll,88°/o N. Gef.: 12,13% N. 0,5283 g Substanz (i. V.) verloren bei 110° 0,0133 g = 2,51 °/o. Ber. f. l Mol. NH3: 2,88°/o. 0,1352 g Substanz (110°): 0,3332 g C02 und 0,0811 g H20: keine Asche. 0,1715 > > (110°): 14,8 ccm N, 6° und 744 mm B. CS2H3e06N4. Ber.: 67,16°/o C, 6,29°/o H, 9,79°/o N. Gef.: 67,21 °/o > 6,66°/o > 10,28°/o > Die weitere Aufarbeitung der Gallensteine lieferte dann 22,2 g eines prachtvoll rotbraun gefärbten rohen Bilirubins, nachdem mit Hilfe von siedendem Eisessig das Choleprasin herausgelöst worden war.2) Dieses rohe Bilirubin, das also vielleicht mit dem Choleprasin eine Verbindung bildet, die erst durch die heiße, konzentrierte Essigsäure gespalten wird, gab nun nach der Behandlung mit Essigsäure und Alkohol sofort, also ohne Umkrystallisation, auf die Formel C32H3606N4 stimmende Werte. 0,1320 g Substanz (110°): 0,3230 g C02, 0,0721 g H20 und 0,0004 g Asche. 0,1423 » > (110°): 0,3498 > > 0,0823 > >. 0,1345 > > (110°): 11 ccm N bei 15° und 749 mm B. 0,1237 > · · > (110°): 10,5 > » > 15° > 745 > > C32H3606N4. Ber.: 67,16> C, 6,29°/o H, 9,79°/o N. Gef.: 66,95°/o > 6,1 > > 9,4°/o > 67,04> > 6,43°/o > 9,74% > Ein kleinerer Teil des Gallensteinpulvers, dem die Hauptmenge des Bilirubins und Choleprasins, wie soeben beschrieben, entzogen worden war, wurde nun in zwei Portionen noch so) 1. c., S. 476 wurde ein Gewichtsverlust von 10,22 °/o ermittelt, damals wurde die methylalkoholische Lösung im Vakuum zur Trockene gebracht. Durch das Fällen mit Äther dürfte es also bereits zur Ammoniakabspaltung gekommen sein. Die Fällung mit Äther dürfte aber vorzuziehen sein, wenigstens krystallisierte der beim Erhitzen des BilirubinAmmoniaks erhaltene Farbstoff aus Chloroform sehr schön. 2 ) Die gleiche Beobachtung, d. h. die Herausnahme eines hellorange gefärbten Bilirubins nach der Behandlung mit verdünnter Essigsäure, die Extraktion eines rotbraunen Bilirubins nach Behandlung mit heißer, konzentrierter Essigsäure haben wir neuerdings wieder machen können. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM Beiträge zur Kenntnis der Gallenfarbstoffe. 'VII. 73 lange abwechselnd mit Eisessig und Chloroform behandelt, bis diese beiden Lösungsmittel so gut wie keine Substanz mehr herausnahmen. Je 5 g dieses Materials wurden alsdann mit Sodalösung geschüttelt, vom Ungelösten filtriert und der Filterrückstand mit heißem Wasser gewaschen, bis nur noch Spuren von Farbstoff in Lösung gingen und das Ablaufende gegen Methylrot alkalische Reaktion nicht mehr zeigte. Es hinterblieben 4 resp. 453 g, die reichliche Mengen von Natrium enthielten, wie aus der Aschenanalyse hervorging. Es enthielt nämlich der nicht mit Soda behandelte Teil: L 0,5 resp. II 1,12 °/o unlösliche Asche (hauptsächlich Eisenphosphat). L 0,6347 g Substanz gaben 0,0032 g Asche. . 0,4085 > > » 0,0046 » > Der mit Soda behandelte Teil enthielt: L 0,6 °/o unlösliche und 7,1 > lösliche Asche. II. 1,37 resp. 9,52 °/o. I. 0,4923 g Substanz: 0,0379g Gesamtasche, wovon 0,0030g unlöslich. II. 0,5253 > > : 0,0572 > > > 0,0072 > Beim Behandeln der beiden Natriumsalze mit verdünnter Salzsäure wurde außer geringen Mengen organischer Substanz das Natrium entfernt. Natriumsalz I enthielt 4,46 °/o N, II. 4,0 °/o Na. L 0,2317 g Substanz gaben 0,0270 g NaCl, daraus 0,0319 g Na2S04 (statt 0,0328 g). U. 2,45 > > > 0,25 > > Ein anderer Teil des erschöpfend extrahierten Gallensteinpulvers wurde mit 5°/oiger Kalilauge behandelt, wobei von je 5 g: L 4,35, II. 4,3 g nach dem Auswaschen hinterblieben. Die beiden Präparate gaben an 5°/oige Kalilauge so gut wie keine Substanz mehr ab und enthielten: L 6,98°/o K. II. 6,2'/o K. I. 1,2638 g Substanz: 0,1685 g KC1. II. 1,4065 Substanz: 0,1661 g KC1. Durch Schütteln des Gallensteinpulvers mit ^lo-n-Kaliumcarbonatlösung wurde dagegen ein Kaliumsalz mit nur 3,05 °/o K erhalten. . 0,4715 g Substanz: 0,0275 g KCL Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM 74 W. Küster, K. Reibung und R. Schmiede!, Wurden die Kaliumsalze mit einer Silbernitratlösung geschüttelt, so fand ein Ersatz des Kaliums durch Silber statt, doch gelang eine vollständige Verdrängung des Alkalis nicht. Durch Behandlung mit Cyankaliumlösung konnte alsdann wieder das gesamte Silber herausgenommen und durch Kalium ersetzt werden, aber nur wenn frisch hergestellte Silbersalze verwendet wurden. Die aus den Natriumsalzen durch Einwirkung stark verdünnter Salzsäure wieder zurückgewonnene organische Substanz erwies sich als nicht mehr völlig identisch mit dem Ausgangsmaterial, insofern jetzt konzentriertere Laugen wieder einen kleinen Teil in Lösung überführten. I. 0,1635 g Substanz (120°): 0,3609 g'C02, 0,0882 g H20 und eine sehr geringe Menge Asche. (120°): 19,4 ccm N bei 13° und 742,6 mm B. 0,2090 (120°): 19,0 » » » 10° > 746 > > 0,2038 II. 0,1462 (120°): 0,3161 g C0„ 0,0805 g H20 u. 0,0030 g Asche (120°): 18,3 ccm N bei 12° und 748,5 mm B. 0,1825 (120°): 11,4 » > > 6° > 747 > > 0,1157 Aus diesen Analysen berechnet sich: L 60,26 °/o G, 6,0 °/o H, 10,7 und 10,96 °/o N. II. 60,2 °/o > 6,25 °/o > 11,7 » 11,7 > » Es konnte somit der höhere Stickstoffgehalt gegenüber dem «Bilimgrin» in Beziehung stehen zu der Tatsache, daß das Bilinigrin mehr Alkali aufzunehmen imstande ist, als das «Bilihumin», d. h. das letztere konnte ein Ammoniumsalz sein, das weder durch konzentrierte Essigsäure noch durch Sodalösung zerlegt worden war. So wurde denn ein weiterer Teil des erschöpfend mit Chloroform und Eisessig behandelten Gallensteinpulvers nach Verlauf von 3 /4 Jahren in Arbeit genommen, wobei es sich zunächst zeigte, daß jetzt an Essigsäure wieder Substanz abgegeben wurde. Sie erwies sich als «Choleprasin», denn bei der Behandlung mit Salzsäure wurde sie zerlegt und die salzsaure Lösung gab beim Eindampfen denselben an Fleischexträkt erinnernden Geruch wie das Choleprasin bei- der gleichen Zersetzung. Als dann mit Essigsäure wieder erschöpfend extrahiert worden war, wobei etwa 10°/o des verwendeten Gallenstein- Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM Beiträge zur Kenntnis der Gallenfarbstoffe. VII. 75 pulvers herausgenommen wurden, wurde der Rest mit verdünnter Salzsäure bei Z. T. behandelt.1) In der Tat ließ sich leicht nachweisen, daik hierbei Ammoniak abgespalten wurde, die Menge elben warvaber sehr gering: aus 2,465 g wurden nur 0,0284 g Platinsalmiak erhalten. Immerhin war insofern eine bemerkenswerte EinWirkung erfolgt, als nunmehr bei der Behandlung m i Alkalien Salze entstanden, deren Metallgehalt mit dem übereinstimmte^ den die Salze des Bilinigrins aufweisen. 2 g Substanz wurden mit 5°/oiger Sodalösung geschüttelt, wobei etwa 0,1 g in Lösung ging, die Ausbeute an erschöpfend ausgewaschenem Natriümsalz betrug 1,85 g. 1,7055 g Sbst. enthielten 0,11704 g Na (gef. 0,3613 g Na2S04) = 6,86 °/o NaWeitere 2 g Substanz wurden mit 0,7°/oiger Pottaschelösung behandelt, wodurch ebenfalls etwa 5°/o gelöst wurden. Ausbeute 1,8 g unlösliches Kaliumsalz. 1,7912 g Subst enthielten 0,1004g K (gef. 0,0,2240 g K2S04 = 5,6°/o K.) Es ist bemerkenswert, daß hier die Mengen Kalium .und Natrium im Verhältnis von l Atom Kalium zu 2 Atomen Natrium /» QA . QQ stehen, da ' ' = 5,76 ist, wie denn auch beim Bilinigrin ) Bei höherer Temperatur wird die Substanz durch Salzsäure stark angegriffen. 8 ) Die Zerlegung der Salze wurde mit verdünnter Salzsäure bewirkt und hierbei eine farblose Lösung erhalten, die beim Eindampfen neben KC1 resp. NaCl organische Substanz zurückließ, wie sich bei gelindem Glühen des Rückstandes herausstellte. Wurde jetzt das Alkalichlorid mit Wasser entfernt und die verbleibende Asche geglüht, so hinterblieb Ferriphophat Beim Eindampfen der Lösung des Alkalichlorids und Glühen des Rückstandes zeigte sich, daß immer noch etwas organische Substanz beigemengt war, welche abermals Ferriphosphat hinterließ. Aus 1,7055 g des Natriumsalzes wurden z. B. 0,0220 -f 0,0086 = 0,0306 oder 1,8 °/o erhalten, während die Menge der organischen Substanz schätzungsweise 0,1 + 0,04 g betrug. Es geht also zunächst ein Körper in Lösung, der aus organischer Materie verbunden mit Ferriphosphat besteht. Endlich ist zu bemerken, daß auch das nach zweimaligem Glühen erhaltene Alkalichlorid noch eine Beimengung, vielleicht von Ammoniumchlorid, enthalten muß. 0,198 g KC1 gaben nämlich nur 0,2240 g I^SO, statt 0,231 g. 0,309 > NaCl > > > 0,3613 » Na2S04 > 0,376 >. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM 76 W. Küster, K. Reibung und R. Schmiedel, gefunden wurde, daß beim Bebandeln mit sehr verdünnter Carbonatlösung nur die Hälfte der Menge des Metalls aufgenommen wird, die aus konzentrierteren Lösungen übertritt. Das aus dem Natrium resp. Kaliumsalz zurückgewonnene «Bilihumin» zeigte folgende Zusammensetzung: 1. 60,22 °/o C, 6,17 °/o H, 11,52 > N (auf aschefreie Substanz berechnet), 2. 60,23% > 5,6% > 10,96% > ( desgl. ), 0,1687 g Substanz (100°): 0,3672 g C02, 0,0923 g H20 und 0,0024 g Asche (Ferriphosphat), 0,1450 > > (100°): 0,3161 > > 0,0777 > > und 0,0018 g Ascbe (Ferripbosphat), 0,1923 > » (100°): 18,4 ccra N bei 6° und 741,5 mm B. 0,1580 > > (100°): 14 > > * 2° > 750 > > Durch die Behandlung mit der verdünnten Salzsäure war also die Zusammensetzung nicht wesentlich geändert worden, obgleich viel mehr Natrium aufgenommen worden war, als vor der Behandlung. Durch langes Erhitzen mit 25°/oiger Salzsäure wurde dagegen ein beträchtlicher Teil dieses «Bilihumins» in Lösung gebracht. Nachdem dann die salzsaure Lösung eingedampft, der Rückstand in Wasser gelöst, die filtrierte Lösung wiederum verdampft und das Verfahren oftmals wiederholt worden war, ergab sich ein auch in Alkohol löslicher Rückstand. Diese alkoholische Lösung wurde in Äther eingetragen, wobei eine Fällung entstand. Doch enthielt auch die alkoholisch-ätherische Lösung Substanz und zwar ein phosphorhaltiges Fett, das nicht in Äther, wohl aber in Aceton löslich war. Die Fällung wurde mit Aceton aufgenommen, wobei Chlorammonium zurückblieb, während nach dem Verdampfen des Acetons ein Sirup erhalten wurde, dessen Geruch an Fleischexträkt erinnerte. Es handelte sich also um dieselben Spaltprodukte, wie sie auch aus dem Gholeprasin erhalten wurden. Der durch Behandeln des «Bilihumins» mit konzentrierter Salzsäure nicht in Lösung zu bringende Anteil gab bei der Analyse folgende Werte: 62,73% C, 5,48 °/o H, 9,8 °/o N (auf aschefreie Substanz berechnet). 0,1438 g Substanz (110 °): 0,3292 g C02, 0,0706 g H20, 0,0007 g Asche. 0,1437 > > (110°): 12,5 ccm N bei 14° und 726 mm B. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM Beiträge zur Kenntnis der Gallenfarbstoffe. VII. 77 \ Die Asche enthielt nur Eisen, keine Phosphorsäure. Wurde deA gegen Salzsäure beständige Teil des Bilihumins auf konzenWierte Schwefelsäure gestreut, so trat langsam eine Auflösung einband jetzt bewirkte ein Tropfen Salpetersäure eine Rotfarbting. Danach liegt in diesem Teil ein Derivat des Bilirubins vor, dem wohl das «Bilihumin» die Fähigkeit verdankt, unlösliche Salze mit den Alkalien zu bilden. Die Ähnlichkeit in dieser Beziehung mit dem Bilinigrin, das wir durch Oxydation des Bilirubins erhielten, könnte dann vielleicht dafür sprechen, daß die Umwandlung des Hämatins im Organismus über das Stadium des Bilirubins hinaus gehen kann. Und zwar immer noch im Verbände mit Umwandlungsprodukten des Globiüs, denn das «Bilihumin» enthält, wie wir sahen, auch Anteile, die zu Aminosäuren hydrolysiert werden können. Nachschrift bei der Korrektur. Die Seite 14 angegebene Methode der Reinigung des Rohbilirubins ist in letzter Zeit 'dadurch wesentlich vereinfacht und verbessert worden, daß die filtrierte methylalkoholammoniakalische Lösung des Bilirubins in eine Kältemischung gestellt wird, wobei sich prachtvoll krystallisiertes BilirubinAmmonium abscheidet. Stuttgart, am 2. April 1914. Brought to you by \| Purdue University Libraries Authenticated Download Date \| 6/10/15 5:02 AM
https://openalex.org/W2980087448	https://www.e-publicacoes.uerj.br/index.php/rdc/article/download/36974/32487	Portuguese	null	Contribuição sindical facultativa: Da análise da (in)constitucionalidade da cobrança compulsória	Revista de Direito da Cidade	2,019	cc-by	11,508	Resumo O presente estudo tem como objetivo analisar em que medida o posicionamento recente do STF quanto à constitucionalidade da extinção da obrigação de cobrança compulsória da contribuição sindical encontra respaldo em leis estrangeiras. Temos como objetivos secundários sopesar a legislação criminal que tipifica algumas condutas no sentido da filiação sindical ou a não observância de normas trabalhistas, bem como, analisar as três modalidades de contribuições destinadas aos sindicatos. E, por fim, debateremos sobre a decisão do STF em confronto (ou consonância) com outras decisões internacionais que discutem o mesmo tema. O método utilizado foi o indutivo- comparativo, numa perspectiva de revisão bibliográfica. Por ora, o que se conclui é que a dispensa da obrigatoriedade de recolhimento da contribuição sindical é uma transformação praticamente natural dentro do direito trabalhista, se comparada com outras legislações internacionais, que apesar da dispensa continuam atentos aos direitos sociais dos trabalhadores. Palavras-chave: Direito do trabalho, sindicatos, contribuição sindical, (in)constitucionalidade, direito estrangeiro. Optional union contribution: From the analysis of (un)constitutionality of compulsory collection 1 Professora Doutora - Departamento de Direito - Faculdade CESUSC - Complexo de Ensino Superior de Santa Catarina - SC-401, n. 9301 - Santo Antonio de Lisboa, Florianópolis - SC, 88050-001, Brasil. Pós-doutora e Doutora em Ciências Humanas, PPGICH, UFSC, Florianópolis-SC. Mestre em Patrimônio Cultural e Sociedade, Univille, Joinville-SC. Especialista em Direito Civil e Processo Civil, ACE – Joinville. Advogada atuante em SC. A pesquisa que resultou neste artigo contou com apoio do Grupo de Pesquisa em Direito Internacional da faculdade CESUSC. Instituição: Complexo de Ensino Superior de Santa Catarina - CESUSC, Santa Catarina. BrasilE-mail: christianekalb@hotmail.com Revista de Direito da Cidade Revista de Direito da Cidade Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Contribuição sindical facultativa: Da análise da (in)constitucionalidade da cobrança compulsória _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 808 Revista de Direito da Cidade Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 is supported by foreign laws. As secondary objectives it aims to weigh the criminal legislation, that typifies some conduct in the sense of union membership or non-observance of labor standards, as well as to analyze the three types of contributions to the unions. Finally, we will discuss the decision of the Supreme Court in confrontation (or consonance) with other international decisions that discuss the same theme. The method used was the inductive-comparative, from a perspective of bibliographic review. The conclusion is that the waiver of the obligation to pay the union contribution is a virtually natural transformation within labor law, compared to other international legislation, which despite the waiver continue to pay attention to the social rights of workers. Keywords: Labor law, union, union contribution, (un) constitutionality, comparative law. 2 Artigo 1º da Lei 13.467/2017, o qual deu nova redação aos artigos 545, 578, 579, 582, 583, 587 e 602 da g q ç g Consolidação das Leis do Trabalho. g / , q Consolidação das Leis do Trabalho. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 809 2 Artigo 1º da Lei 13.467/2017, o qual deu nova redação aos artigos 545, 578, 579, 582, 583, 587 e 602 da Consolidação das Leis do Trabalho. Abstract The present study aims to analyze to what extent the recent position of the Supreme Court regarding the constitutionality of the extinction of the obligation of compulsory collection of union contribution 1 Professora Doutora - Departamento de Direito - Faculdade CESUSC - Complexo de Ensino Superior de Santa Catarina - SC-401, n. 9301 - Santo Antonio de Lisboa, Florianópolis - SC, 88050-001, Brasil. Pós-doutora e Doutora em Ciências Humanas, PPGICH, UFSC, Florianópolis-SC. Mestre em Patrimônio Cultural e Sociedade, Univille, Joinville-SC. Especialista em Direito Civil e Processo Civil, ACE – Joinville. Advogada atuante em SC. A pesquisa que resultou neste artigo contou com apoio do Grupo de Pesquisa em Direito Internacional da faculdade CESUSC. Instituição: Complexo de Ensino Superior de Santa Catarina - CESUSC, Santa Catarina. BrasilE-mail: christianekalb@hotmail.com _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 808 _____ 808 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Revista de Direito da Cidade Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 expressões: “desde que por eles devidamente autorizados”, “desde que prévia e expressamente autorizadas”, “autorização prévia e expressa”, “está condicionado à autorização prévia e expressa dos que participarem de uma determinada categoria econômica ou profissional”, etc; em seus dispositivos que se referiam à contribuição sindical, o debate ressurgiu com ainda mais força. Nesse sentido, o presente artigo tem a intenção de analisar em que medida o posicionamento recente do STF quanto à constitucionalidade da extinção da obrigação de cobrança compulsória da contribuição sindical encontra respaldo em legislações estrangeiras. Para tanto, apesar de todo o mérito envolvido nas questões sociais e políticas engendradas nesse levante, além do viés histórico que poderia aqui ser abordado, porém por ora não será o foco dessa pesquisa, afinal tal decisão afetará diretamente as entidades laborais presentes em praticamente todas as cidades do país, temos como objetivos secundários sopesar a legislação criminal que tipifica algumas condutas no sentido da filiação sindical (na acepção de um constrangimento ilegal) ou a não observância de normas trabalhistas, bem como, analisar as três modalidades de contribuições arrecadadas pelos sindicatos, com fulcro especial na contribuição sindical, foco desse estudo. E por fim, debateremos sobre a decisão do STF em confronto (ou consonância) com outras decisões/normas internacionais advindas do direito estrangeiro. O método utilizado para tanto é de revisão bibliográfica, através de análise de livros, artigos científicos e pesquisas acadêmicas que discutam sobre o assunto, bem como, houve pesquisa em arquivos jornalísticos tanto da imprensa nacional como internacional, blogs e outros websites que tratam do assunto. A metodologia de abordagem é do tipo indutiva, vez que se partiu de algo particular; como o foi por meio da decisão recente do STF, em relação a não obrigatoriedade de pagamento da contribuição sindical; para uma questão mais ampla, abrangendo outras legislações estrangeiras bem como, entrelaçar tal análise à possível tipificação penal de algumas condutas. Portanto, nesse sentido, o método de procedimento ou secundário utilizado é o comparativo, uma vez que analisa tanto a decisão do STF como legislações de outros países em formato de comparação. Esta pesquisa teve início junto ao Grupo de Pesquisa em Direito Internacional da Faculdade Cesusc, de Florianópolis-SC e se trata de uma pesquisa interdisciplinar, tanto dentro da Ciência Jurídica, como com um viés mais politizado. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 810 1 - Introdução O debate sobre a cobrança obrigatória da contribuição sindical, tecnicamente intitulada também de contribuição legal, é tema que remonta a criação dos sindicatos, das confederações e das centrais sindicais no início do século XX. Essas instituições fazem parte do cotidiano laboral das cidades e suas respectivas políticas públicas urbanas e ações de participação social, por isso, merecem um olhar mais atento das possíveis reverberações que virão ocasionar, após a modificação quanto à cobrança da contribuição sindical, para as entidades que tem o condão de proteger os direitos dos trabalhadores. Apesar de haver diversas fontes de renda destinadas aos sindicatos (e toda linha de entidades sindicais) a contribuição sindical sempre foi a que trouxe um maior contingente numérico para os mais de 15 mil sindicatos hoje ativos no país. Afinal, a contribuição sindical era cobrada de todos os trabalhadores com carteira assinada, todos os profissionais liberais e autônomos (com algumas pequenas exceções), independentemente se filiados a algum sindicato ou não. Assim, a população adulta ativa trabalhadora (como são chamados os contribuintes sindicais) era obrigada a pagar um dia de trabalho por ano de seu salário à entidade sindical. No entanto, com o recente posicionamento do Supremo Tribunal Federal, no Brasil, quanto à (in)constitucionalidade das mudanças legislativas trazidas pela Reforma Trabalhista em julho de 2017, em alguns artigos2 da Consolidação das Leis do Trabalho – CLT que tiveram o acréscimo das 2 Artigo 1º da Lei 13.467/2017, o qual deu nova redação aos artigos 545, 578, 579, 582, 583, 587 e 602 da Consolidação das Leis do Trabalho. Revista de Direito da Cidade No item 2, a seguir, se abordarão os crimes tipificados no Código Penal que atingem a organização do trabalho, com o objetivo de colocar em tela possíveis atos praticados no âmbito dos direitos sociais que são considerados crimes perante a lei criminal. 2 – Dos crimes contra a organização do trabalho 2 – Dos crimes contra a organização do trabalho _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 810 Revista de Direito da Cidade Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Os crimes tipificados no Código Penal que vão contra a organização do trabalho estão descritos nos artigos 197 ao 207. Estes delitos não são descritos e previstos junto à Consolidação de Leis Trabalhistas- CLT, pois se tratam de atos tipificados como crime e, nesse sentido possuem lei própria os descrevendo. Além disso, o Superior Tribunal Federal deferiu a medida cautelar da ADIN nº 3.684-0, com eficácia ex tunc, para atribuir interpretação conforme à Constituição da República ao seu artigo 114, incisos I, IV e IX, declarando que, no âmbito da jurisdição da Justiça do Trabalho, não compete processar e julgar ações penais (STF, Pleno, ADIn 3.684-0/DF, Rel. Min. Cezar Peluso, DJ 03.08.2007). Apesar da Emenda Constitucional n. 45 de dezembro de 2004, ter alterado a competência da Justiça Federal comum ou da Justiça Estadual para processar e julgar as ações penais para apuração de crimes contra a organização do trabalho, ampliando a competência da Justiça do Trabalho para julgar todas as controvérsias decorrentes das relações de trabalho, prevalece o entendimento de que a Justiça do Trabalho, mesmo depois da Emenda Constitucional 45/2004, não tem competência criminal. Antes de adentrar na especificação de algumas ações delituosas, é importante apontar que caso qualquer dos atos aqui descritos for praticado de forma individual, atingindo um único trabalhador ou um grupo de trabalhadores, a competência para julgá-los será da Justiça Estadual. Porém, se a conduta for praticada contra uma categoria profissional determinada ou atingir um interesse coletivo, a competência será da Justiça Federal. Tal premissa está prevista no artigo 109 da Constituição da República que compete ao Juízes Federais processar e julgar os crimes contra a organização do trabalho. “Art. 109 - Aos juízes federais compete processar e julgar: VI - (...) os crimes contra a organização do trabalho” (BRASIL, 1988). _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 811 Revista de Direito da Cidade Tal entendimento comunga com a redação da Súmula 115 do extinto Tribunal Federal de Recursos que diz: "Compete à Justiça Federal processar e julgar os crimes contra a organização do trabalho, quando tenham por objeto a organização geral do trabalho ou direitos dos trabalhadores considerados coletivamente” (NORAT, 2012). É importante mencionar também que tanto a CLT como o CP são normas positivadas sob o governo de Getúlio Vargas, destarte, sob pressão de um regime ditatorial que tentava aproximar empregados e empregadores numa exaltação bizarra do nacionalismo trabalhista. Dito isso, os crimes tipificados nos artigos 199 e 203 do CP são os que nos interessam aqui nesse estudo, uma vez que são os atos delituosos relacionados (o primeiro deles) à quebra de liberdade de escolha em se associar ou não a uma entidade sindical e (o segundo) à frustração de observação de lei trabalhista. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 811 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 O artigo 199, portanto, prevê a quebra de liberdade de associação profissional ou sindical. O direito à livre associação está previsto também no art. 5º XVII Constituição Federal (BRASIL, 1988) e nesse sentido se trata de um direito social. Assim dizem os artigos: Art. 199 – Constranger alguém, mediante violência ou grave ameaça, a participar ou deixar de participar de determinado sindicato ou associação profissional: Pena - detenção, de um mês a um ano, e multa, além da pena correspondente à violência (BRASIL, 1940). Art. 5º [...] Inciso XVII - é plena a liberdade de associação para fins lícitos, vedada a de caráter paramilitar (BRASIL, 1988). O tipo objetivo praticado nessa conduta é o constrangimento ilegal praticado com violência (física contra a pessoa) e grave ameaça, exercida para participar ou deixar de participar de sindicato ou associação profissional. O artigo 8º da Constituição Federal também dispõe que “é livre a associação profissional ou sindical...” (BRASIL, 1988). O sujeito ativo pode ser qualquer pessoa pertencente ou não a sindicato ou associação, e se caso for funcionário público pode incorrer ainda no artigo 3º, f., da Lei nº 4.898/1965. O sujeito passivo pode ou não ser sócio ou associado, mas deve ser obrigado a participar ou deixar de participar de determinado sindicato ou associação. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 812 Revista de Direito da Cidade Ainda há o constrangimento ilegal neste artigo, mas com o fim de compelir a liberdade de associação ou não à entidade em questão. Esse tipo penal permite a ocorrência da tentativa (MIRABETE, 2005, p. 387-388). É importante frisar que o ato de constrangimento ilegal obrigue a pessoa a tomar ou deixar de tomar parte de sindicato determinado (GONÇALVES, 2018, p.554), pois se a intenção for apenas a obrigação de associação de qualquer sindicato, o crime será de constrangimento ilegal, em sua modalidade simples. Em razão do crime, acima mencionado, penalizar suas condutas em menos de 2 anos, i.e., um ano de prisão (detenção), o art. 76 da Lei do JECRIM – 9.099/1995 (BRASIL, 1995) possibilita a Transação Penal. A Transação Penal seria um acordo realizado entre o cidadão/delinquente que cometeu algum delito de menor potencial ofensivo e o promotor de justiça. Por esse acordo entabulado, a pessoa se compromete, como forma de se redimir pelo fato praticado, a ajudar uma entidade carente, com dinheiro ou prestar serviços a certa comunidade. Por outro lado, o representante do Ministério Público se compromete a arquivar o processo, sem que haja julgamento do mérito, assim que a prestação for cumprida. Portanto, aqui não se discute se o acusado cometeu efetivamente o delito ou não. Existe também a possibilidade da Suspensão Condicional do Processo. Trata-se também de um acordo previsto no art. 89 da Lei JECRIM (BRASIL, 1995), em que o cidadão/delinquente, antes Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Revista de Direito da Cidade mesmo do oferecimento da denúncia pelo promotor de justiça, tendo cometido algum crime penalizado com prisão igual ou menor de um ano pode suspender o processo por dois a quatro anos. Para obtenção desse benefício, o investigado terá que cumprir algumas condicionantes, como por exemplo: reparar do dano, quando possível, não frequentar determinados lugares, como casas de tolerância, ficar proibido de ausentar-se da comarca onde reside, sem autorização do juiz, realizar o comparecimento pessoal e obrigatório ao juízo, mensalmente, para informar e justificar suas atividades, podendo o juiz ainda especificar outras condições a que fica subordinada a suspensão, desde que adequadas ao fato e à situação pessoal do acusado (BRASIL, 1995). O delito de frustração de direito assegurado por Lei Trabalhista, por sua vez, está descrito no art. Revista de Direito da Cidade 203, do Código Penal, senão vejamos: O delito de frustração de direito assegurado por Lei Trabalhista, por sua vez, está descrito no art. 203, do Código Penal, senão vejamos: Frustrar, mediante fraude ou violência, direito assegurado pela legislação do trabalho: Pena – detenção de um ano a dois anos, e multa, além da pena correspondente à violência. [...] § 2º A pena é aumentada de um sexto a um terço se a vítima é menor de dezoito anos, idosa, gestante, indígena ou portadora de deficiência física ou mental (BRASIL, 1940). Este tipo penal assegura os direitos trabalhistas previstos na Constituição Federal, na CLT e em leis esparsas e de acordo com Mirabete, “estão incluídos na proteção os direitos obtidos por meio das convenções e dissídios, uma vez que eles são previstos em lei” (2005). Portanto, se trata de uma norma penal em branco. O sujeito ativo, portanto, é aquele que impede a realização do direito do trabalho, normalmente o empresário / patrão. O sujeito passivo é a pessoa lesada no direito trabalhista, normalmente o trabalhador. A violência a que descreve o artigo é a física, sendo descartada a ameaça, que configuraria o crime próprio de ameaça, previsto no art. 147 do CP. Pode ocorrer também por meio de fraude, ou seja, quando há indução ao erro ou simplesmente não observar tal direito trabalhista. O parágrafo segundo, acrescentado pela Lei 9.777/1998, trata da causa de aumento de pena, caso a vítima for menor de 18 anos, idosa, gestante, indígena ou portadora de deficiência física ou mental. Capez menciona o posicionamento de Damásio a respeito da idade mínima em que se considera uma pessoa idosa, que diz “nem sempre a idade mínima da vítima representa, por si só, circunstância capaz de exasperar a pena. É possível que tenha mais de sessenta anos de idade e seja portadora de condições físicas normais (...)” (2003, p.563). _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 813 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Feitas as considerações iniciais sobre a possibilidade de se cometer crimes caso alguém obrigue outro a se associar ou deixar de se associar a algum sindicato – art. 199 CP -, bem como caso venha a infringir algum direito trabalhista – art. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 814 Revista de Direito da Cidade 203 CP -, portanto, esclarecida a possibilidade de se tipificar tais condutas, agora é o momento de se aclarar as diversas modalidades contributivas em favor das associações sindicais. 3 - Das contribuições destinadas às associações sindicais 3 - Das contribuições destinadas às associações sindicais 3 - Das contribuições destinadas às associações sindicais Após meses de debates acalorados, desde a entrada em vigor da Reforma Trabalhista (Lei n. 13.467/17) em meados de 2017, o Plenário do Supremo Tribunal de Justiça declarou, no final do mês de junho de 2018, por 6 votos a 3 que é constitucional a extinção do desconto obrigatório da contribuição sindical no salário dos trabalhadores. Portanto, o desconto destinado a financiar os sindicatos passou a ser opcional e somente através de autorização prévia e expressa do trabalhador, independentemente se este for filiado ou não ao sindicato que lhe poderá ser cobrado. Mas antes de adentrar ao debate sobre o processo decisório acima mencionado é importante esclarecer as diferenças existentes entre os diversos tipos de contribuição que se destinam aos sindicatos, a saber: a contribuição assistencial (e sua mensalidade), a contribuição confederativa e a contribuição sindical. Havendo outras fontes de renda dos sindicatos como a contribuição negocial, as "contribuições extraordinárias para a construção de nova sede" ou para o "fundo de greve" etc; ou ainda as contribuições ordinárias com o fim de ser uma "contribuição para auxílio mútuo", dentre outras várias vistas na pesquisa que não serão nesse momento analisadas a fundo, por não serem o cerne do debate. A contribuição assistencial trata-se de uma mensalidade ou também chamada de taxa assistencial cobrada pelos sindicatos aos participantes associados. E tem por finalidade custear as negociações coletivas que venham a beneficiar toda a categoria profissional contemplada por aquele sindicato. Assim, de acordo com artigo 513, alínea ‘e’ da CLT é prerrogativa dos sindicatos impor contribuições dos que participam das categorias econômicas ou profissionais ou das profissões liberais representadas (BRASIL, 1943). Nesse mesmo sentido é o art. 545 da CLT que prevê a obrigação dos empregadores descontar a contribuição assistencial, quando autorizados pelo empregado. Tal contribuição beneficia somente a entidade que a estipulou, assim não há qualquer repasse desses valores ao sistema confederativo. Apesar da previsão legal estipular a prerrogativa de cobrança pelos sindicatos da contribuição assistencial, o Precedente Normativo n. 119 do TST, publicado pela primeira vez em 1998 prevê que: _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 814 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 A Constituição da República, em seus arts. 3 Art. 5º inciso XX CF – “Ninguém poderá ser compelido a associar-se ou a permanecer associado”. 4 Art. 8º inciso V CF – “Ninguém será obrigado a filiar-se ou a manter-se filiado a sindicato”. 5 Sum.666 STF: “A contribuição confederativa de que trata o art. 8º, IV, da Constituição, só é exigível dos filiados ao sindicato respectivo”. 6 Art. 8º inciso IV CF – “a assembleia geral fixará a contribuição que, em se tratando de categoria profissional, será descontada em folha, para custeio do sistema confederativo da representação sindical respectiva, independentemente da contribuição prevista em lei”. 3 Art. 5º inciso XX CF – “Ninguém poderá ser compelido a associar-se ou a permanecer associado”. ________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 81 ao sindicato respectivo . 6 Art. 8º inciso IV CF – “a assembleia geral fixará a contribuição que, em se tratando de categoria profissional, será descontada em folha, para custeio do sistema confederativo da representação sindical respectiva, independentemente da contribuição prevista em lei”. 3 Art. 5º inciso XX CF – “Ninguém poderá ser compelido a associar-se ou a permanecer associado” 4 Art. 8º inciso V CF – “Ninguém será obrigado a filiar-se ou a manter-se filiado a sindicato”. 3 - Das contribuições destinadas às associações sindicais 5º, XX3 e 8º, V4, assegura o direito de livre associação e sindicalização. É ofensiva a essa modalidade de liberdade cláusula constante de acordo, convenção coletiva ou sentença normativa estabelecendo contribuição em favor de entidade sindical a título de taxa para custeio do sistema confederativo, assistencial, revigoramento ou fortalecimento sindical e outras da mesma espécie, obrigando trabalhadores não sindicalizados. Sendo nulas as estipulações que inobservem tal restrição, tornam-se passíveis de devolução os valores irregularmente descontados (BRASIL, 1998, 2014). Portanto, de acordo com esse PN, embora não tenha força legal, sendo apenas uma orientação jurisprudencial para fundamentação de julgamentos de casos assemelhados; e observados os ditames constitucionais dos arts. 5º, inciso XX e 8º, inciso V, e ainda a Súmula 666 do STF5; o desconto da contribuição assistencial precisa ser autorizado pelo empregado associado ou não ao sindicato. Essa súmula foi transformada na Súmula vinculante de n. 40 a seguir exposta. Dessa forma, a contribuição assistencial somente poderá ser descontada do empregado quando este for afiliado ao sindicato correspondente e quando autorizado expressamente tal desconto, após ter sido notificado pela associação sindical. Caso contrário, de acordo com todo esse aparato normativo acima exposto, não poderá haver qualquer desconto. Por outro norte, ainda há a contribuição confederativa, que é considerada um direito fundamental não necessitando de regulamentação por estar prevista no art. 8º, inciso IV da CF6. Esta contribuição é devida apenas aos empregados filiados ao sindicato, não podendo ser cobrada dos empregados que não sejam associados. Esse é o entendimento desde 1996 do STF que pacificou a questão através do RE 198.092-3, firmando o sentido de que “a contribuição confederativa, instituída em assembleia geral, CF art. 8º IV, distingue-se da contribuição sindical, instituída por lei, com caráter tributário, CF art.149, assim compulsória. A primeira é compulsória apenas para os filiados do sindicato” (BRASIL, 1996 apud VIANNA, 2009, p.510). O debate a respeito da obrigatoriedade do desconto das contribuições já encontrava celeumas desde 1996, conforme se pode averiguar a seguir: A questão a saber é se a denominada contribuição confederativa, inscrita no art. 8º, IV, da Constituição Federal, fixada pela assembleia geral, é devida pelos empregados não filiados ao sindicato. Noutras palavras, se apresenta ela caráter de compulsoriedade, vale dizer, se é obrigatório o seu pagamento por empregados não filiados ao sindicato. 7 Tal expressão – “imposto sindical” – estaria tecnicamente incorreto, vez que tal tributo se trata de uma contribuição e não um imposto, conforme dispõe o Código Tributário Brasileiro. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 816 3 - Das contribuições destinadas às associações sindicais (...) Primeiro que tudo, é preciso distinguir a contribuição X CF – “Ninguém poderá ser compelido a associar-se ou a permanecer associado”. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 815 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 sindical, contribuição instituída por lei, de interesse das categorias profissionais - art. 149 da Constituição - com caráter tributário, assim compulsória, da denominada contribuição confederativa, instituída pela assembleia geral da entidade sindical - C.F., art. 8º, IV. A primeira, conforme foi dito, contribuição parafiscal ou especial, espécie tributária, é compulsória. A segunda, entretanto, é compulsória apenas para os filiados do sindicato. (RE 198.092, Relator Ministro Carlos Velloso, Segunda Turma, julgamento em 27.8.1996, DJ de 11.10.1996) Portanto com base nesse entendimento e também na Súmula vinculante n. 40 (tendo como precedente representativo a Súmula 666) “A contribuição confederativa de que trata o art. 8º, IV, da Constituição Federal, só é exigível dos filiados ao sindicato respectivo”. Tal súmula vinculante também foi objeto de pedido de cancelamento, conforme publicação do STF de 2015, senão vejamos: Pedido de cancelamento da Súmula Vinculante 40: Pedido de cancelamento da Súmula Vinculante 40: (...) deve-se destacar que os argumentos trazidos pela proponente são idênticos àqueles debatidos no julgamento do RE 194.603/SP, Relator o Ministro Marco Aurélio, e Redator para o acórdão Ministro Nelson Jobim, cujo resultado foi paradigma para a elaboração da Súmula 666 do Supremo Tribunal Federal, convertida, posteriormente, na Súmula Vinculante 40. Portanto, o mero descontentamento ou divergência quanto ao conteúdo de verbete vinculante não propicia a reabertura de debate sobre matéria devidamente sedimentada por esta Corte. Ademais, a proponente não se desincumbiu da exigência constitucional de apresentar decisões reiteradas do Supremo Tribunal Federal para suportar o seu pedido de cancelamento da Súmula Vinculante 40, com o que também não se verifica a necessária adequação formal da presente proposta. (PSV 117, Relator Ministro Ricardo Lewandowski, decisão monocrática, julgamento em 2.9.2015, DJe de 9.9.2015) Portanto, mesmo após o pedido de cancelamento do PN 119, o STF manteve o seu entendimento no sentido de apenas poder haver desconto da contribuição confederativa dos empregados afiliados ao sindicato e que expressamente o autorizem, i.e., na mesma linha de cobrança da contribuição assistencial. Por último, a contribuição sindical, objeto principal de análise desse artigo, chamada também de contribuição legal ou “imposto sindical”7, está prevista no art. 578 da CLT e no art. 149 (3ª parte) da CF – que denomina o tributo de forma genérica como: contribuição especial. Essa contribuição de interesse das categorias profissionais e econômicas (conforme menciona a Constituição Federal) até a reforma trabalhista implementada em 2017 era descontada obrigatoriamente de todos os trabalhadores, de suas folhas de pagamento, sempre no mês de 7 Tal expressão – “imposto sindical” – estaria tecnicamente incorreto, vez que tal tributo se trata de uma contribuição e não um imposto, conforme dispõe o Código Tributário Brasileiro. Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 março, sendo correspondente à remuneração de um dia de trabalho, quaisquer que fosse a forma de remuneração e quer fossem ou não os empregados associados ao sindicato. Além disso, devia o empregador verificar junto ao sindicado correspondente se não havia determinação de prazo diverso para o seu recolhimento (VIANNA, 2009, p. 498). Tal contribuição possuía respaldo na OIT através da Convenção 98, ratificada pelo Brasil em 1953 que dispõe sobre o direito de sindicalização e pelo Comitê de liberdade sindical por meio do Informe 138, caso 631 (VIANNA, 2009, p.509). 8 Foram distribuídas por prevenção e encontram-se apensadas ao presente feito 18 (dezoito) ações diretas. São elas: ADI 5912; ADI 5923; ADI 5859; ADI 5865; ADI 5813; ADI 5887; ADI 5913; ADI 5810; ADI 5811; ADI 5888; ADI 5815; ADI 5850; ADI 5900; ADI 5945; ADI 5885; ADI 5892 e ADI 5806. Foi também apensada, pelas mesmas razões, a ADC 55 (BRASIL, 2018). Pedido de cancelamento da Súmula Vinculante 40: No entanto, em nenhum momento previa-se a obrigatoriedade da contribuição sindical, uma vez que no seu preâmbulo [Convenção 87 e 98 da OIT] diz que busca garantir “a afirmação do princípio da liberdade de associação sindical”. Assim vejamos: Verbete 112. Em um caso em que a lei dispunha a cobrança de uma cotização de solidariedade, pelo sistema de desconto da remuneração de trabalhadores não filiados a organização sindical parte em um contrato coletivo, porém que desejavam acolher-se de suas disposições (cotização fixada em não mais de 2/3 das cotizações pagas pelos trabalhadores sindicalizados da mesma categoria), o Comitê estimou que o sistema, ainda que não esteja coberto pelas normas internacionais do trabalho, não parece por si mesmo incompatível com os princípios de liberdade sindical (COMITÊ..., 1976, p. 44-45 apud AROUCA, 2012, p. 69). Verbete 112. Em um caso em que a lei dispunha a cobrança de uma cotização de solidariedade, pelo sistema de desconto da remuneração de trabalhadores não filiados a organização sindical parte em um contrato coletivo, porém que desejavam acolher-se de suas disposições (cotização fixada em não mais de 2/3 das cotizações pagas pelos trabalhadores sindicalizados da mesma categoria), o Comitê estimou que o sistema, ainda que não esteja coberto pelas normas internacionais do trabalho, não parece por si mesmo incompatível com os princípios de liberdade sindical (COMITÊ..., 1976, p. 44-45 apud AROUCA, 2012, p. 69). O verbete acima citado aceitava a cobrança de ‘cotização de solidariedade’ de empregados não filiados ao sindicato, no entanto, após a reforma de 2017 essa contribuição também deixou de ser obrigatória. Assim, a não ser que autorizada expressamente pelo trabalhador, não poderá ser descontada de sua folha de pagamento, independentemente se filiados ou não a qualquer sindicato. 4 – Da ADI 5.794 STF que pôs fim à contribuição sindical _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 817 4 – Da ADI 5.794 STF que pôs fim à contribuição sindical A decisão ADI 5.794 do STF pôs fim às reivindicações de várias entidades sindicais8 que pediam pela inconstitucionalidade da referida extinção da contribuição sindical, uma vez que tal aniquilamento representa 80% de suas arrecadações. O que poderia, afirmam as entidades, atingir suas atividades negociativas em prol dos direitos dos trabalhadores. Durante o julgamento pelo Supremo, o ministro Luiz Fux, que abriu a divergência na sessão, disse que “não é possível tomar capital para financiar sindicato sem o consentimento do empregado”. Ele reconheceu que mesmo leis ordinárias podem tratar sobre a contribuição sindical, pois nenhum comando na Constituição fixa a compulsoriedade da cobrança. Os ministros Alexandre 8 Foram distribuídas por prevenção e encontram-se apensadas ao presente feito 18 (dezoito) ações diretas. São elas: ADI 5912; ADI 5923; ADI 5859; ADI 5865; ADI 5813; ADI 5887; ADI 5913; ADI 5810; ADI 5811; ADI 5888; ADI 5815; ADI 5850; ADI 5900; ADI 5945; ADI 5885; ADI 5892 e ADI 5806. Foi também apensada, pelas mesmas razões, a ADC 55 (BRASIL, 2018). _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 817 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 de Moraes, Luís Roberto Barroso, Gilmar Mendes, Marco Aurélio e a presidente Cármen Lúcia seguiram o entendimento de Fux (CONJUR, 2018, web). de Moraes, Luís Roberto Barroso, Gilmar Mendes, Marco Aurélio e a presidente Cármen Lúcia seguiram o entendimento de Fux (CONJUR, 2018, web). Nesse sentido, Alexandre de Moraes afirmou no julgamento que a extinção da cobrança obrigatória não afasta a liberdade sindical e que não lhe parece razoável haver mais de 16 mil sindicatos no país, sendo que apenas 20% dos empregados são realmente sindicalizados. Isso mostraria um “vácuo de representatividade” (BRASIL, 2018). Por outro lado, o relator da ação, ministro Luiz Edson Fachin, ficou vencido ao votar pela obrigatoriedade da contribuição. Ele sustentou que a Constituição de 1988 foi precursora no reconhecimento de direitos nas relações entre capital e trabalho, entre eles, a obrigatoriedade do imposto para custear o movimento sindical. Conforme o relator, o modelo de sindicalismo criado pela Constituição sustenta-se em um tripé formado por unicidade sindical, representatividade obrigatória e custeio das entidades sindicais por meio de um tributo. Este último é a contribuição sindical, expressamente autorizada pelo artigo 149, e imposta pela parte final do inciso IV, do artigo 8º, ambos da Constituição da República. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 818 4 – Da ADI 5.794 STF que pôs fim à contribuição sindical No governo Fernando Collor de Mello, por sua vez, chegou-se a encaminhar ao Congresso Nacional projeto de lei nesse sentido, que se perdeu, por falta de empenho. Da mesma forma, quando a Presidência da República foi ocupada por Fernando Henrique Cardoso, anunciou- se que repousava em sua mesa uma minuta de medida provisória dispondo sobre a tardia extinção do tributo, que, apesar da mudança de nome, não perdera a sua natureza (GUNTHER, 2017). Após a Constituição Federal de 1988, os sindicatos foram reconhecidos como instituições de direito privado, colocando fim com qualquer celeuma a respeito de sua natureza jurídica. Consequentemente, puderam ser criados e constituídos de forma livre. Portanto, não há sentido em perceberem contribuições tributárias. A manutenção da cobrança do “imposto sindical” mesmo dos trabalhadores não vinculados a determinado sindicato se mostrava uma violação da liberdade sindical (prevista no art. 199 do CP, já citado no item 2), já que estavam obrigados ao recolhimento, sendo deles descontado o dia de trabalho. Conforme afirmou na relatoria da ADI, o ministro Edson Fachin (apesar de vencido em seu voto e, portanto, contrário a extinção da contribuição) observa que a doutrina especializada, atenta a este fenômeno, afirma que: Ora, as entidades sindicais foram acostumadas, durante várias décadas, a conviver com esse modelo do dinheiro fácil, e é certo que o hábito do cachimbo costuma deixar a boca torta. Presenciamos um caso real, no qual o sindicato tinha cerca de 4.500 associados e, por pura falta de interesse, esse número acabou sendo reduzido para menos de 500 associados. Agora, é necessário fazer o caminho inverso, e para isso será necessário algum tempo, para que os sindicatos se reestruturem e possam sair à luta, mostrando serviço para os integrantes da categoria e mostrando que efetivamente existe vantagem em ser associado à entidade sindical. Além do mais, a acomodação que se viu foi da direção do sindicato, e não se mostra coerente fazer com que toda a categoria pague por isso (DANTAS JÚNIOR, 2017, p. 283) Por todo o exposto, o que nos parece é que transformar a contribuição sindical que parecia ter natureza jurídica de tributo especial para um espírito de contribuição negocial facultativa, a partir da interpretação do art. 4 – Da ADI 5.794 STF que pôs fim à contribuição sindical Os ministros Dias Toffoli e Rosa Weber acompanharam o relator. (CONJUR, 2018, web). A revogação da cobrança obrigatória da contribuição sindical, chamado também de “imposto sindical”, não nos parece inconstitucional, como também não parece para a Associação Brasileira de Emissoras de Rádio e TV (Abert) que ajuizou ação declaratória de constitucionalidade ADC 55. A instituição desta contribuição teve respaldo dentro de um sindicalismo de Estado, perpetrado na Era Vargas, especialmente através da publicação da CLT em 1943. Nesse período, os sindicatos detinham pouca autonomia, se fazendo necessário uma intervenção estatal para poderem existir, e de certa forma se tornaram instituições supletivas do Estado. Para alguns poucos doutrinadores os sindicatos eram vistos como instituições de direito público. “A Constituição de 1946 atribuiu-lhe funções delegadas de Poder Público, mas apesar dessas atribuições, tornaram-se pessoas jurídicas de direito privado porque não são criados pelo Estado, sua criação não decorre de lei e o seu ato constitutivo é a assembleia que aprova os estatutos, a diretoria provisória e a sua fundação” (DELGADO, 2008, p.77). Por essa teoria, os interesses do sindicato confundem-se com os próprios interesses peculiares do Estado. Conforme ensina Amauri Mascaro Nascimento (2000), após a inscrição do princípio da liberdade sindical na Constituição Federal de 1988, restaram poucos adeptos a esta teoria na doutrina pátria. Em geral, o sindicato tem a natureza de pessoa jurídica de direito público apenas nos regimes totalitários (SIQUEIRA, 2005). Mas o debate em si quanto à possibilidade de extinção da cobrança compulsória remonta ao período de vigência do regime militar, quando Marechal Castelo Branco anunciou iniciativa de Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 medida legal tendente a acabar com o “imposto” (contribuição sindical). Essa expectativa acabou não se concretizando. No governo Fernando Collor de Mello, por sua vez, chegou-se a encaminhar ao Congresso Nacional projeto de lei nesse sentido, que se perdeu, por falta de empenho. Da mesma forma, quando a Presidência da República foi ocupada por Fernando Henrique Cardoso, anunciou- se que repousava em sua mesa uma minuta de medida provisória dispondo sobre a tardia extinção do tributo, que, apesar da mudança de nome, não perdera a sua natureza (GUNTHER, 2017). medida legal tendente a acabar com o “imposto” (contribuição sindical). Essa expectativa acabou não se concretizando. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 819 4 – Da ADI 5.794 STF que pôs fim à contribuição sindical 8º da CF e da confirmação da alterações dos artigos da CLT, extinguindo por vez a sua cobrança, mesmo dos afiliados, caso esses não autorizem a sua cobrança, sem ter criado uma regra de transição para que os sindicatos existentes buscassem novas fontes de financiamento e de renda antes da cessação completa do recebimento das contribuições, fez com que a reforma trabalhista confirmada na decisão da ADI 5.794 do STF gerasse uma renúncia fiscal que serviu de enfraquecimento dos quase 16 mil sindicatos já constituídos, podendo ter os prejudicado e muito. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 819 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Por outro lado, tal decisão pode colocar em debate o grande número de entidades sindicais e sua duvidosa necessidade em confronto com outras realidades políticas e sociais do globo. 5 – Do direito estrangeiro sobre a contribuição sindical Analisando os dados numéricos sobre os sindicatos existentes e vigentes no Brasil, trata-se de quase 12 mil sindicatos de trabalhadores e mais de 5 mil de empregadores (IBGE, 20189). Dos sindicatos ativos de trabalhadores, 75% eram filiados a uma das centrais sindicais existentes, e dos que eram filiados, 30% estavam na Central Única dos Trabalhadores (CUT), 22% na Força Sindical, 14% na Nova Central Sindical dos Trabalhadores (NTSC) e outros 14% na União Geral dos Trabalhadores (UGT). Os demais se distribuíam por outras oito centrais sindicais (CARDOSO, 2014). Colocando em xeque o número expressivo de entidades sindicais espalhadas pelo país, ao “passo que no Reino Unido há 168, nos EUA, 130 e na Argentina, 91” (CONJUR, 2018, web), buscamos analisar algumas outras realidades políticas e sociais, a exemplo do que trouxe KREIN, GIMENEZ, SANTOS (2018) em especial na Europa e na América no que concerne a (des)obrigação de pagamento da contribuição sindical. Serrano (2006, p. 34) aponta como primeira grande utilidade das análises de direito comparado / estrangeiro, a possibilidade de indicar as normas jurídicas afins e ou casos concretos nas legislações nacionais e estrangeiras, com o objetivo de confrontá-las para determinar as analogias e as diferenças existentes entre sistemas e institutos, bem como avaliar o desenvolvimento e aproximação das legislações ou instituições jurídicas de diversos países, formando assim "o novo Direito Positivo Contemporâneo”. Ao mesmo tempo, ressaltam-se também as críticas ao estudo comparativo. 9 Os dados constantes no website do IBGE são datados de 2001. Não há dados mais atuais. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 820 4 – Da ADI 5.794 STF que pôs fim à contribuição sindical Segundo Marc Ancel (1980, p 16) três críticas principais foram levantadas contra toda pesquisa jurídica comparativa: em primeiro lugar, o direito nacional seria suficientemente difícil de ser conhecido com clareza para que se possa complicar com os sistemas estrangeiros. Em segundo lugar, o direito comparado seria uma fonte constante de confusão. Por fim, se afirma que o direito de um país faz parte do seu patrimônio nacional e, portanto, deve ser preservado a despeito de qualquer comparação. O Boletim Informativo Consultor Jurídico (CONJUR, 2018, web) publicou que a Suprema Corte dos Estados Unidos também extinguiu a contribuição sindical obrigatória, numa decisão que envolve os sindicatos dos servidores públicos, mas se aplica a todo o país. O pronunciamento se refere a uma lei do estado de Illinois, tendo sido questionada a partir de um “processo iniciado por Mark Janus, funcionário do Departamento de Saúde de Springfield, no estado de Illinois, que alegou Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 que se opõe à contribuição sindical compulsória. Janus justificou em sua queixa que era forçado a contribuir, embora não integrasse o sindicato” (EM, 2018, web). Apesar de iniciado os autos naquele Estado, a decisão também declarou inconstitucionais leis de outros 22 estados que favoreciam os sindicatos. A decisão teve como fundamento o direito de expressão dos cidadãos. Melo (2018) comenta a decisão: Por 5 votos a 4, a corte decidiu que os membros de um sindicato de servidores públicos de Illinois, parte do processo em julgamento, não são obrigados a pagar contribuições para cobrir despesas de negociações coletivas, de defesa dos interesses dos trabalhadores e outras atividades do sindicato, como a de financiamento de campanhas políticas. Nos EUA, mais de um terço dos servidores públicos e apenas 7% dos empregados de empresas privadas são sindicalizados – o que significa que o sindicalismo no país já é naturalmente mais fraco do que outros países (MELO, 2018). No Brasil, conforme já mencionado anteriormente, o Supremo Tribunal Federal reconheceu regra semelhante ao manter dispositivos da reforma trabalhista que extinguem a contribuição sindical. Sendo relevante frisar que nos quase 16 mil sindicatos, apenas 20% dos empregados são realmente filiados, o que também demonstra um esvaziamento representativo, conforme afirmou Alexandre de Moraes, ministro do STF, durante o julgamento da ADIN. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 821 4 – Da ADI 5.794 STF que pôs fim à contribuição sindical A decisão da suprema corte norte-americana esclareceu que todos os empregados que optarem por não se afiliar aos sindicatos e não pagar contribuições sindicais irão se beneficiar, de qualquer forma, dos benefícios de negociações coletivas. O mesmo ocorre no Brasil, vez que há o efeito erga omnes das normas e acordos que resultam de negociações coletivas, conforme previsto no artigo 611 da CLT. “A autoaplicabilidade das normas coletivas para toda a categoria profissional, bem como o reconhecimento constitucional dos acordos e convenções coletivas (artigo 7º, XXIX, da CF/1988) também reforçam a importância da função das entidades sindicais na negociação coletiva”. (PEREIRA NETO, 2017, p. 60-1). Assim, os não contribuintes/filiados foram chamados no processo norte-americano de “caroneiros” (free-riders), por obterem gratuitamente os serviços dos sindicatos, financiados pelos filiados. O voto da maioria, escrito pelo ministro conservador Samuel Alito, justifica a decisão de desobrigar os empregados do pagamento da contribuição sindical com base no direito de expressão, protegido pela Primeira Emenda da Constituição dos EUA. É notório que nos EUA muitos sindicatos angariam recursos para financiar campanhas eleitorais e não seria justo alguém, no final das contas, financiar campanha para determinado candidato que nem sempre será seu escolhido. O Departamento de Justiça publicou nota destacando que "nenhum funcionário do setor público _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 821 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 deveria ser forçado a financiar um sindicato contra sua vontade e apoiar uma mensagem política com a qual poderia discordar" (EM, 2018, web). Janus é especialista em apoio à criança do Departamento de Saúde e Serviços Familiares do Estado de Illinois. E por mais de 40 anos - desde que a Suprema Corte decidiu no Conselho de Educação de 1977 o caso de Abood vs. Detroit - as leis em Illinois e outros 21 estados permitem que um funcionário do governo não se junte a um sindicato. No entanto, eles devem pagar uma taxa de “parte justa (fair share)” para cobrir os custos de representá-los. Essa taxa não cobre as despesas políticas ou de lobby dos sindicatos. Os advogados da Janus argumentaram em seu comunicado que a taxa paga pela Janus “subsidia” uma “organização de advocacia” que a Janus não apoia. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 822 4 – Da ADI 5.794 STF que pôs fim à contribuição sindical Janus pagava cerca de US $ 47 por mês para o sindicato, em comparação com as contribuições de US $ 60, de acordo com o porta-voz do Conselho da AFSCME, Anders Lindall. Há especulações de que Janus vencendo, grupos financiados por conservadores anti-sindicatos de doadores iniciarão campanhas para persuadir funcionários do governo a se desligarem dos sindicatos. A equipe jurídica da Janus inclui a Fundação Nacional de Defesa Legal pelo Direito ao Trabalho e o Liberty Justice Center, em Chicago, controlado pelo Illinois Policy Institute, cujos financiadores incluem uma fundação controlada pelo mega-doador conservador do Lake Forest, Richard Uihlein (SWEET, 2018, web). Poucos são os países que ainda hoje mantém o sistema de contribuição obrigatória, tendo como único exemplo na América Latina, o Equador e no continente africano, o Egito. Os sindicatos dos países europeus e o norte-americano, por sua vez, não possuem tal obrigatoriedade e são ferrenhos defensores dos direitos sociais, demonstrando que a obrigatoriedade não tem relação com o enfraquecimento da proteção dos mesmos. Se tomarmos por base o sindicalismo europeu (berço do movimento sindical mundial), não encontraremos a obrigatoriedade de contribuição sindical, já que lá os sindicatos mantêm suas estruturas com mensalidade voluntárias. Portanto, o Brasil não é o único a implementar a facultatividade da cobrança sindical, uma vez que países como Argentina, Alemanha, Colômbia, Grécia, Reino Unido possuem contribuição sindical facultativa, ou seja, o empregado possui a liberdade de dizer se quer ou não ser assistido por um sindicato e consequentemente se deseja contribuir (SILVA, 2018). Em parte, como solução à renda dos sindicatos, muitos países que tornaram facultativo o pagamento da contribuição sindical, optaram pela contribuição "negocial" ou de “taxa / cota de solidariedade”. Analisando algumas realidades, vemos que na Argentina, o financiamento do sistema sindical é formado por cotizações ordinárias e extraordinárias dos seus filiados, além de _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 822 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 contribuições de solidariedade. Tal é o que estabelece o art. 37 da Lei 23.55110 (ARGENTINA, Decreto nº 467/88, atualizada em maio de 2018). Os empregadores, a teor do art. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 823 10 Disponível em: < http://servicios.infoleg.gob.ar/infolegInternet/anexos/20000-24999/20993/texact.htm> Acesso em: 04 jul 2018. 11 Disponível em: < https://www.boe.es/buscar/act.php?id=BOE-A-1985-16660 > Acesso em: 4 jul 2018 12 Disponível em: < http://www.comune.jesi.an.it/MV/leggi/l300-70.htm > Acesso em: 04 jul 2018. [Regras relativas à proteção da liberdade e dignidade dos trabalhadores, à liberdade sindical e à atividade sindical no local de trabalho e às regras de colocação, publicadas no diário oficial]. 13 Disponível em: < https://dre.pt/pesquisa/-/search/632906/details/maximized > Acesso em: 04 jul 2018. Revista de Direito da Cidade Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Já na Itália, a Lei 300/7012 (Legge 20 maggio 1970, n. 300 - Norme sulla tutela della libertà e dignità dei lavoratori, della libertà sindacale e nell'attività sindacale nei luoghi di lavoro e norme sul collocamento, pubblicata nella gazzetta ufficiale n. 131 del 27 maggio 1970, atualizado em outubro de 2016), em seu art. 26, estabelece que os trabalhadores têm direito de recolher contribuições e desenvolver atividades visando angariar novos adeptos de suas organizações sindicais, no interior dos locais de trabalho, sem prejudicar o andamento das atividades na empresa. Ao mesmo tempo, fixa que as associações sindicais de trabalhadores têm o direito de receber, por meio do desconto em folha de pagamento, as contribuições sindicais dos trabalhadores, mediante as modalidades estabelecidas pelos contratos coletivos de trabalho que garantam o segredo em relação à quantia destinada pelo trabalhador a cada associação sindical (PESSOA, 2007). O financiamento do sistema sindical em Portugal é levado a efeito pelo Código do Trabalho13 - Lei n. 99/2003, que prevê a necessidade de declaração de autorização ou o pedido expresso do _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 824 13 Disponível em: < https://dre.pt/pesquisa/-/search/632906/details/maximized > Acesso em: 04 jul 201 4 – Da ADI 5.794 STF que pôs fim à contribuição sindical 38 da mesma lei, são obrigados a atuar como agentes de retenção dos valores destinados ao sindicato, sendo que tal retenção é mediada por uma resolução do Ministério do Trabalho específica para cada sindicato. No caso de ausência de retenção pelo empregador, ou ausência de repasse dos valores retidos, o próprio empregador se torna o devedor direto das quantias ao sindicato obreiro, operando-se a mora de pleno direito (PESSOA, 2007). O caso da Espanha é eloquente. Além da contribuição de solidariedade (canon de negociación), os sindicatos recebem subvenções pela participação em órgãos públicos responsáveis pelo seguro-desemprego, aposentadoria e pensões, saúde e segurança e outros; pela prestação de serviços junto a entidades governamentais que cuidam da formação profissional, supervisão da previdência complementar, convênios de saúde, etc. Além disso, recebem recursos das empresas: créditos horários para atividade sindical, ajuda às comissões intra-empresariais, atividades educacionais, esportivas, culturais e outras (MARTINS, 1998). Além disso, o art. 11 da Lei Orgânica da liberdade sindical espanhola (ESPANHA, Ley Orgánica 11/1985, de 2 de agosto, de Libertad Sindical, atualizada em julho de 2011)11, estabelece que os convênios coletivos poderão estabelecer cláusulas pelas quais os trabalhadores contribuam para os sindicatos representados na comissão negociadora, fixando um cânon econômico e regulando as modalidades de seu pagamento. Ao teor do mesmo artigo, porém, em todo caso se respeitará a vontade individual do trabalhador, que deverá expressar-se por escrito na forma e prazos que determinar a norma coletiva. Este mesmo dispositivo prevê ainda o desconto da cota sindical sobre os salários pelo empregador e a subsequente transferência dos valores para o sindicato (PESSOA, 2007). Na França, por exemplo, os sindicatos recebem verbas do governo para participar de centenas de colegiados tripartites de níveis local, regional e nacional; contam com verbas das prefeituras e subvenções do Ministério do Trabalho (formação profissional, saúde e segurança); ganham recursos de ministérios de outras áreas para funções específicas na educação, saúde, agricultura, etc.; e recebem subvenções contínuas do Primeiro Ministro para o funcionamento do IRES (Institut de Recherches Économiques et Sociales). No agregado, tais recursos representam cerca de 40% da receita total dos sindicatos franceses (LANDIER, LABBÉ, 2000). Referências ALVES, Aggeu. Dos crimes contra a organização do trabalho. Revista da Faculdade de Direito, Universidade de São Paulo, São Paulo, v. 53, p. 370-389, jan. 1958. ISSN 2318-8235. Disponível em: <http://www.revistas.usp.br/rfdusp/article/view/66301/68911>. Acesso em: 28 maio 2018. doi:http://dx.doi.org/10.11606/issn.2318-8235.v53i0p370-389. 1958 ANCEL, Marc. Utilidade e Métodos do Direito Comparado. Trad. Sérgio José Porto. Porto Alegre: Fabris, 1980. ARBEITSVERTRAG, .org. Der Arbeitsvertrag – das Arbeitsrecht definiert die Schranken. Disponível em: < https://www.arbeitsvertrag.org/> Acesso em: 04 jul 2018. 2018. AROUCA, José Carlos. A FLEXIBILIZAÇÃO DA CONVENÇÃO 87 DA OIT. Revista do TRT da 2ª Região. In: Direito Coletivo – Doutrina. São Paulo, n. 12/2012, p. 31-79. 2012. CAPEZ, Fernando. Curso de Direito Penal: parte especial. v.2, 2 ed. São Paulo: Saraiva, 2003. CARDOSO Adalberto Nota Técnica: Os sindicatos no Brasil In: Mercado de Trabalho. IPEA. n 56 ARBEITSVERTRAG, .org. Der Arbeitsvertrag – das Arbeitsrecht definiert die Schranken. Disponível em: < https://www.arbeitsvertrag.org/> Acesso em: 04 jul 2018. 2018. ARBEITSVERTRAG, .org. Der Arbeitsvertrag – das Arbeitsrecht definiert die Schranken. Disponível em: < https://www.arbeitsvertrag.org/> Acesso em: 04 jul 2018. 2018. AROUCA, José Carlos. A FLEXIBILIZAÇÃO DA CONVENÇÃO 87 DA OIT. Revista do TRT da 2ª Região. In: Direito Coletivo – Doutrina. São Paulo, n. 12/2012, p. 31-79. 2012. CAPEZ, Fernando. Curso de Direito Penal: parte especial. v.2, 2 ed. São Paulo: Saraiva, 2003. ARBEITSVERTRAG, .org. Der Arbeitsvertrag – das Arbeitsrecht definiert die Schranken. Disponível em: < https://www.arbeitsvertrag.org/> Acesso em: 04 jul 2018. 2018. AROUCA, José Carlos. A FLEXIBILIZAÇÃO DA CONVENÇÃO 87 DA OIT. Revista do TRT da 2ª Região. In: Direito Coletivo – Doutrina. São Paulo, n. 12/2012, p. 31-79. 2012. CAPEZ, Fernando. Curso de Direito Penal: parte especial. v.2, 2 ed. São Paulo: Saraiva, 2003. CARDOSO, Adalberto. Nota Técnica: Os sindicatos no Brasil. In: Mercado de Trabalho. IPEA. n.56, fev 2014, pp.21-27. 2014 _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 824 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 DANTAS JÚNIOR, Aldemiro Rezende. Fim da Contribuição Sindical Obrigatória – Consequências para a entidades sindicais e categorias representadas, in Revista do Tribunal Regional do Trabalho da 3ª Região, Belo Horizonte, nov. 2017, p. 271-287. 2017. DELGADO, Maurício Godinho. Direito Coletivo do Trabalho. 3ª edição. São Paulo: LTr, 2008. EM. com.br. Internacional: Suprema Corte dos EUA impõe derrota a sindicatos do setor público. 2 Jul 2018. Belo Horizonte. Disponível em: <https://www.em.com.br/app/noticia/internacional/2018/06/27/interna_internacional,969863/s uprema-corte-dos-eua-impoe-derrota-a-sindicatos-do-setor-publico.shtml> Acesso em: 04 jul 2018. Referências 2018. DANTAS JÚNIOR, Aldemiro Rezende. Fim da Contribuição Sindical Obrigatória – Consequências para a entidades sindicais e categorias representadas, in Revista do Tribunal Regional do Trabalho da 3ª Região, Belo Horizonte, nov. 2017, p. 271-287. 2017. DELGADO, Maurício Godinho. Direito Coletivo do Trabalho. 3ª edição. São Paulo: LTr, 2008. EM. com.br. Internacional: Suprema Corte dos EUA impõe derrota a sindicatos do setor público. 2 Jul 2018. Belo Horizonte. Disponível em: <https://www.em.com.br/app/noticia/internacional/2018/06/27/interna_internacional,969863/s uprema-corte-dos-eua-impoe-derrota-a-sindicatos-do-setor-publico.shtml> Acesso em: 04 jul 2018. 2018. GONÇALVES, Victor Eduardo Rios. Direito Penal Esquematizado: Parte Especial. 8.ed. coord. Pedro Lenza. São Paulo: Saraiva Educação, 2018. GUNTHER, Luiz Eduardo. O fim da contribuição sindical obrigatória: a crônica de uma morte anunciada, In: DALLEGRAVE NETO, José Affonso; KAJOTA, Ernani (Coord). Reforma Trabalhista ponto a ponto. São Paulo: Ltr, 2017. IBGE. Número de sindicatos no país. Disponível em: <https://www.ibge.gov.br/estatisticas- novoportal/sociais/administracao-publica-e-participacao-politica/9092-sindicatos-indicadores- sociais.html?=&t=o-que-e> Acesso em: 03 jul 2018, 2001-2018. JESUS, Damásio de. Direito Penal: parte especial: v. 3. 16. ed. São Paulo: Saraiva, 2007. KREIN, J., GIMENEZ, D., SANTOS, A. Dimensões críticas da reforma trabalhista no Brasil. Organizadores: José Dari Krein, Denis Maracci Gimenez, Anselmo Luis dos Santos. – Campinas, SP : Curt Nimuendajú, 2018. 304 p. LANDIER, Hubert; LABBÉ, Daniel. Les Organisations Syndicales en France, Paris: Les Editions Liasons, 2000. MARTINS, Sérgio Pinto. Contribuições Sindicais, São Paulo: Editora Atlas, 1998. MELO, João Ozorio de. Liberdade de Expressão: Suprema corte dos EUA também acaba com GONÇALVES, Victor Eduardo Rios. Direito Penal Esquematizado: Parte Especial. 8.ed. coord. Pedro Lenza. São Paulo: Saraiva Educação, 2018. GUNTHER, Luiz Eduardo. O fim da contribuição sindical obrigatória: a crônica de uma morte anunciada, In: DALLEGRAVE NETO, José Affonso; KAJOTA, Ernani (Coord). Reforma Trabalhista ponto a ponto. São Paulo: Ltr, 2017. IBGE. Número de sindicatos no país. Disponível em: <https://www.ibge.gov.br/estatisticas- novoportal/sociais/administracao-publica-e-participacao-politica/9092-sindicatos-indicadores- sociais.html?=&t=o-que-e> Acesso em: 03 jul 2018, 2001-2018. IBGE. Número de sindicatos no país. Disponível em: <https://www.ibge.gov.br/estatisticas- novoportal/sociais/administracao-publica-e-participacao-politica/9092-sindicatos-indicadores- sociais.html?=&t=o-que-e> Acesso em: 03 jul 2018, 2001-2018. JESUS, Damásio de. Direito Penal: parte especial: v. 3. 16. ed. São Paulo: Saraiva, 2007. KREIN, J., GIMENEZ, D., SANTOS, A. Dimensões críticas da reforma trabalhista no Brasil. Organizadores: José Dari Krein, Denis Maracci Gimenez, Anselmo Luis dos Santos. – Campinas, SP : Curt Nimuendajú, 2018. 304 p. LANDIER, Hubert; LABBÉ, Daniel. Les Organisations Syndicales en France, Paris: Les Editions Liasons, 2000. LANDIER, Hubert; LABBÉ, Daniel. Les Organisations Syndicales en France, Paris: Les Editions Liasons, 2000. MARTINS, Sérgio Pinto. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 825 Referências Contribuições Sindicais, São Paulo: Editora Atlas, 1998. d b d d d d bé b MARTINS, Sérgio Pinto. Contribuições Sindicais, São Paulo: Editora Atlas, 1998. MELO, João Ozorio de. Liberdade de Expressão: Suprema corte dos EUA também acaba com contribuição sindical obrigatória. In: Consultor Jurídico. 29 jun 2018. Disponível em: < https://www.conjur.com.br/2018-jun-29/suprema-corte-eua-acaba-contribuicao-sindical- obrigatoria> Acesso em: 03 jul 2018. 2018. MELO, João Ozorio de. Liberdade de Expressão: Suprema corte dos EUA também acaba com contribuição sindical obrigatória. In: Consultor Jurídico. 29 jun 2018. Disponível em: < https://www.conjur.com.br/2018-jun-29/suprema-corte-eua-acaba-contribuicao-sindical- obrigatoria> Acesso em: 03 jul 2018. 2018. MIRABETE, Julio Fabrini. Manual de Direito Penal. 23 ed. São Paulo: Atlas, 2005. MIRABETE, Julio Fabrini. Manual de Direito Penal. 23 ed. São Paulo: Atlas, 2005. NASCIMENTO, Amauri Mascaro, Compêndio de direito sindical. 2. ed., São Paulo: LTR, 2000. NASCIMENTO, Amauri Mascaro, Compêndio de direito sindical. 2. ed., São Paulo: LTR, 2000. SERRANO, Pablo Jiménez. Como utilizar o direito comparado para a Elaboração de Tese Científica. Rio de Janeiro: Forense, 2006. SERRANO, Pablo Jiménez. Como utilizar o direito comparado para a Elaboração de Tese Científica. Rio de Janeiro: Forense, 2006. SILVA, Gabriela Rangel. Contribuição Sindical: retorno ou retrocesso. 28 jun 2018. Disponível em: < http://silvaesilva.com.br/contribuicao-sindical-retorno-ou-retrocesso/ > Acesso em: 04 jul 2018. 2018. SIQUEIRA, Graciano Pinheiro de. Natureza jurídica e órgão registrador das entidades sindicais. Revista Jus Navigandi, ISSN 1518-4862, Teresina, ano 10, n. 818, 29 set. 2005. Disponível em: <https://jus.com.br/artigos/7355>. Acesso em: 3 jul. 2018. 2005. SIQUEIRA, Graciano Pinheiro de. Natureza jurídica e órgão registrador das entidades sindicais. Revista Jus Navigandi, ISSN 1518-4862, Teresina, ano 10, n. 818, 29 set. 2005. Disponível em: <https://jus.com.br/artigos/7355>. Acesso em: 3 jul. 2018. 2005. SWEET, Lynn. Unions, anticipating Janus case Supreme Court loss, jump-start recruiting drives. In: Chicago Sun Times. 28 jun 2018. Disponível em: <https://chicago.suntimes.com/news/public- sector-unions-supreme-court-decision-illinois-mark-janus-case-rauner/> Acesso em: 04 jul 2018. 2018. NASCIMENTO, Amauri Mascaro, Compêndio de direito sindical. 2. ed., São Paulo: LTR, 2000. NASCIMENTO, Amauri Mascaro, Compêndio de direito sindical. 2. ed., São Paulo: LTR, 2000. NORAT, Markus Samuel Leite. Competência para o processo e julgamento dos crimes contra a organização do trabalho. Cognitio Juris, João Pessoa, Ano II, n. 5, ago 2012. Disponível em <http://www.cognitiojuris.com/artigos/05/12.html>. Acesso em: 28 de Maio de 2018. 2018. NORAT, Markus Samuel Leite. Competência para o processo e julgamento dos crimes contra a organização do trabalho. Cognitio Juris, João Pessoa, Ano II, n. 5, ago 2012. Disponível em <http://www.cognitiojuris.com/artigos/05/12.html>. Acesso em: 28 de Maio de 2018. 2018. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 825 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 PEREIRA NETO, João Batista. O sistema brasileiro de unicidade sindical e compulsoriedade de representação. São Paulo: LTR, 2017. PESSOA, Flavia Moreira Guimarães. O financiamento das entidades sindicais na Argentina, Espanha, Itália e Portugal. Florianópolis: Busca Legis CCJ UFSC, 2007. RAUPP, Eduardo Caringi. O registro de entidades sindicais. In: Âmbito Jurídico, Rio Grande, VII, n. 17, maio 2004. Disponível em: <http://www.ambito- juridico.com.br/site/index.php?n_link=revista_artigos_leitura&artigo_id=3523>. Acesso em jul 2018. 2004. SERRANO, Pablo Jiménez. Como utilizar o direito comparado para a Elaboração de Tese Científica. Rio de Janeiro: Forense, 2006. SILVA, Gabriela Rangel. Contribuição Sindical: retorno ou retrocesso. 28 jun 2018. Disponível em: < http://silvaesilva.com.br/contribuicao-sindical-retorno-ou-retrocesso/ > Acesso em: 04 jul 2018. 2018. SIQUEIRA, Graciano Pinheiro de. Natureza jurídica e órgão registrador das entidades sindicais. Revista Jus Navigandi, ISSN 1518-4862, Teresina, ano 10, n. 818, 29 set. 2005. Disponível em: <https://jus.com.br/artigos/7355>. Acesso em: 3 jul. 2018. 2005. SWEET, Lynn. Unions, anticipating Janus case Supreme Court loss, jump-start recruiting drives. In: Chicago Sun Times. 28 jun 2018. Disponível em: <https://chicago.suntimes.com/news/public- sector-unions-supreme-court-decision-illinois-mark-janus-case-rauner/> Acesso em: 04 jul 2018. 2018. VIANNA, Claudia Salles Vilela. Manual prático das relações trabalhistas. São Paulo: LTr, 2009. Legislação e julgados pesquisados: PEREIRA NETO, João Batista. O sistema brasileiro de unicidade sindical e compulsoriedade de representação. São Paulo: LTR, 2017. PESSOA, Flavia Moreira Guimarães. O financiamento das entidades sindicais na Argentina, Espanha, Itália e Portugal. Florianópolis: Busca Legis CCJ UFSC, 2007. RAUPP, Eduardo Caringi. O registro de entidades sindicais. In: Âmbito Jurídico, Rio Grande, VII, n. 17, maio 2004. Disponível em: <http://www.ambito- juridico.com.br/site/index.php?n_link=revista_artigos_leitura&artigo_id=3523>. Acesso em jul 2018. 2004. Legislação e julgados pesquisados: ARGENTINA. Ley 23.551 - Decreto nº 467/1988, atualizada em maio de 2018. Disponível em: <http://servicios.infoleg.gob.ar/infolegInternet/anexos/20000-24999/20993/texact.htm> Acesso em: 04 jul 2018. BRASIL. Código Penal. Disponível em < http://www.planalto.gov.br/ccivil_03/decreto- lei/del2848.htm>. BRASIL. Código Penal. Disponível em < http://www.planalto.gov.br/ccivil_03/decreto- lei/del2848.htm>. _____. Consolidação das Leis Trabalhistas – CLT. Disponível em < http://www.planalto.gov.br/ccivil_03/decreto-lei/Del5452.htm> Acesso: 02 jun 2018, 1943. _____. Constituição da República Federativa do Brasil de 1988. Disponível em <http://www.planalto.gov.br/ccivil_03/constituicao/constituicao.htm>. Acesso: 02 jun 2018. Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 _____. Lei 7.783, de 28 de junho de 1989. Disponível em <http://www.planalto.gov.br/ccivil_03/leis/l7783.htm>. Acesso: 02 jun 2018. _____. Lei 9.099 de 26 de setembro de 1995. Disponível em: <www.planalto.gov.br/ccivil_03/Leis/L9099.htm> Acesso: 02 jun 2018. _____. Lei 4.898 de 9 de dezembro de 1965. Disponível em: <http://www.planalto.gov.br/ccivil_03/Leis/l4898.htm > Acesso: 02 jun 2018. _____. Lei 9.777 de 29 de dezembro de 1998. Disponível em: <https://presrepublica.jusbrasil.com.br/legislacao/103285/lei-9777-98 > Acesso: 02 jun 2018. _____. Lei 13467 de julho de 2017. Disponível em: <http://www.normaslegais.com.br/legislacao/Lei-13467-2017.htm > Acesso: 02 jun 2018. 2017. _____. Medida cautelar em ação direta de inconstitucionalidade 3.684-0 Distrito Federal. _____. Medida cautelar em ação direta de inconstitucionalidade 3.684-0 Distrito Federal. Disponível em <http://redir.stf.jus.br/paginadorpub/paginador.jsp?doc TP=TP&docID=415904>. STF, Pleno, ADIn 3.684-0/DF, Rel. Min. Cezar Peluso, DJ 03.08.2007. _____. Medida cautelar em ação direta de inconstitucionalidade 3.684-0 Distrito Federal. Disponível em <http://redir.stf.jus.br/paginadorpub/paginador.jsp?doc TP=TP&docID=415904>. STF, Pleno, ADIn 3.684-0/DF, Rel. Min. Cezar Peluso, DJ 03.08.2007. _____. RE 198.092, Relator Ministro Carlos Velloso, Segunda Turma, julgamento em 27.8.1996, DJ de 11.10.1996. _____. RE 198.092, Relator Ministro Carlos Velloso, Segunda Turma, julgamento em 27.8.1996, DJ de 11.10.1996. _____. RE 198.092, Relator Ministro Carlos Velloso, Segunda Turma, julgamento em 27.8.1996, DJ de 11.10.1996. _____. PSV 117, Relator Ministro Ricardo Lewandowski, decisão monocrática, julgamento em 2.9.2015, DJe de 9.9.2015. _____. PSV 117, Relator Ministro Ricardo Lewandowski, decisão monocrática, julgamento em 2.9.2015, DJe de 9.9.2015. _____. Ação Direta De Inconstitucionalidade 5.794 Distrito Federal. Voto Senhor Ministro Edson Fachin (Relator). ADIn 5794/DF, 2018. _____. Ação Direta De Inconstitucionalidade 5.794 Distrito Federal. Voto Senhor Ministro Edson Fachin (Relator). ADIn 5794/DF, 2018. Código Penal. Código de Processo Penal. Constituição Federal e Legislação Complementar/Organizadores Valdemar P. da Luz; Paulo Roberto Froes Toniazzo. 4 ed. Florianópolis: Conceito Editorial, 2009. ESPANHA. Ley Orgánica 11/1985, de 2 de agosto, de Libertad Sindical, atualizada em julho de 2011. Disponível em: < https://www.boe.es/buscar/act.php?id=BOE-A-1985-16660.> Acesso em: 4 jul 2018. ITÁLIA. Lei 300/70 - Legge 20 maggio 1970, n. Legislação e julgados pesquisados: 300 - Norme sulla tutela della libertà e dignità dei lavoratori, della libertà sindacale e nell'attività sindacale nei luoghi di lavoro e norme sul collocamento, pubblicata nella gazzetta ufficiale n. 131 del 27 maggio 1970. Disponível em: <http://www.comune.jesi.an.it/MV/leggi/l300-70.htm.> Acesso em: 4 jul 2018. PORTUGAL. Lei n.º 99/2003 – Código do trabalho. Disponível em: < https://dre.pt/pesquisa/- /search/632906/details/maximized> Acesso em: 04 jul 2018. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 827 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 trabalhador de dedução na retribuição das quotas sindicais. Fixa a lei ainda que nenhum trabalhador será obrigado a pagar quotas para sindicato em que não esteja inscrito, bem como que a aplicação do sistema de cobrança e entrega de quotas sindicais não pode implicar para o trabalhador qualquer discriminação, nem o pagamento de outras quotas ou indenizações, ou provocar-lhe sanções que, de qualquer modo, atinjam o seu direito ao trabalho (PESSOA, 2007). Por último, analisando a situação alemã, especialmente após o Plano Hartz (iniciado em 2002), junto a um website alemão intitulado Arbeitsvertrag – Contrato de emprego - embora se tenha que pagar para se unir a um sindicato, há muitos benefícios. Isso inclui o seguro de despesas legais, caso necessário ir à justiça, bem como se estiver em desacordo com o empregador em questões como a demissão, rescisão e direito às férias; tudo é realizado por um advogado de forma gratuita. E um seguro de acidentes recreativos também é concedido. Além disso, uma das tarefas dos sindicatos é a negociação de acordos coletivos. Como organizações de trabalhadores, eles têm muito mais poder do que um único empregado, e é por isso que os sindicatos podem alcançar efeitos muito mais benéficos para muitos. O fato de precisar pagar ou não uma taxa de associação em um sindicato geralmente depende da situação do trabalho do filiado. Quem tem que pagar e quanto deve ser verificado junto aos estatutos (ARBEITSVERTRAG, 2018). Ainda no website alemão há algumas informações atualizadas sobre o sistema sindical em outros países europeus: O maior sindicato de empregados hoje na Alemanha é o IG Metall. Tem mais de 2,3 milhões de membros em todo o mundo. Os sindicatos na França se comportam de maneira muito diferente em público do que seus equivalentes alemães. Legislação e julgados pesquisados: Os alemães são muito mais propensos a abandonar seus empregos e lutar por melhores condições. Na Bélgica, o nível de sindicalização é o mais alto da Europa, portanto, muitos trabalhadores são sindicalizados (ARBEITSVERTRAG, 2018). _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 828 5 - Considerações A decisão do STF quanto à constitucionalidade da extinção da obrigação de cobrança compulsória da contribuição sindical realmente encontra respaldo nas normas estrangeiras e já vem afetando diretamente as entidades sindicais dispostas em quase todas as cidades do Brasil, por isso a importância em se pesquisar e analisar essas políticas públicas urbanas que atingem em cheio a participação social, especialmente da classe trabalhadora. Descrever a legislação criminal que tipifica condutas com intento de forçar ou não a filiação sindical (no sentido de um constrangimento ilegal) ou a não observância de normas trabalhistas, auxiliou o entendimento de possíveis ações perpetradas no sentido de ferir direitos de uma população, via de regra, hipossuficiente. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 828 Revista de Direito da Cidade Revista de Direito da Cidade vol. 11, nº 2. ISSN 2317-7721 DOI: 10.12957/rdc.2019.36974 Além disso, analisar as três modalidades de contribuições destinadas aos sindicatos, com objetivo de focar a pesquisa na contribuição sindical se mostrou bastante interessante para compreensão da dinâmica da extinção de sua cobrança compulsória para um sistema de pagamento voluntário e necessariamente expresso pelo trabalhador. Tendo por base a análise feita nos primeiros itens desse estudo das condutas consideradas criminosas de acordo com a lei penal e também a descrição das contribuições recebidas pelas entidades sindicais, se observa que a declaração de facultatividade de recolhimento do imposto sindical, a partir do posicionamento recente do STF, coloca esse setor em fragilidade. Ao debater a decisão do STF em confronto (ou consonância) com outras decisões internacionais advindas do direito estrangeiro se percebe que, a representatividade em determinadas áreas econômicas se enfraquecerá em decorrência do fechamento de alguns sindicatos, mas também se perquire a real necessidade de tantos sindicatos dentro do território nacional, se comparados os números com outros países mencionados na pesquisa. Pela análise dos sistemas jurídicos estrangeiros apontados anteriormente, se verifica uma tendência, no âmbito internacional (em especial nos continentes europeu e americano), da abolição da cobrança compulsória de contribuições sindicais. Por outro lado, outra tendência é a fixação das chamadas quotas/taxas de solidariedade, que poderiam ser assimiladas à contribuição assistencial, mas cuja forma de cobrança dependerá do sistema de liberdade ou unicidade sindical adotado pela reforma sindical, uma vez que a cobrança poderá ser feita de todos os membros da categoria ou apenas por aqueles representados pelo sindicato que firmou a negociação coletiva. _________________________________________________________________ Revista de Direito da Cidade, vol. 11, nº 2. ISSN 2317-7721 pp. 808-829 829 Trabalho enviado em 30 de agosto de 2018 Aceito em 01 de fevereiro de 2019 5 - Considerações Tudo vai depender de como a reforma trabalhista, ocorrida em 2017 e confirmada em junho de 2018, será recebida tanto pelas entidades sindicais como pela população adulta ativa trabalhadora. Por fim, em todos os sistemas estudados, se percebeu que a contribuição associativa quando fortalecida e facultativa, além de se tornar uma grande fonte de rendimento dos sindicatos, provêm muitos benefícios, não só na manutenção dos direitos sociais dos trabalhadores de determinada categoria, mais vai além. E por isso, tem um papel tão importante frente à política e à ética nas negociações coletivas. Trabalho enviado em 30 de agosto de 2018 Aceito em 01 de fevereiro de 2019
https://openalex.org/W3134260138	http://elartu.tntu.edu.ua/bitstream/lib/35638/1/9123-%d0%a2%d0%b5%d0%ba%d1%81%d1%82%20%d1%81%d1%82%d0%b0%d1%82%d1%82%d1%96-16779-1-10-20210224.pdf	Ukrainian	null	ПІДВИЩЕННЯ ЕРГОНОМІЧНОСТІ ЕКОСУМКИ ДЛЯ ГОСПОДАРСЬКИХ І БУДІВЕЛЬНИХ МАТЕРІАЛІВ	Grail of science	2,021	cc-by	1,998	Міжнародний науковий журнал «Грааль науки» \| № 1 (Лютий, 2021) Міжнародний науковий журнал «Грааль науки» \| № 1 (Лютий, 2021) 209 DOI 10.36074/grail-of-science.19.02.2021.042 Авторські права захищені \| Creative Commons Attribution 4.0 International License Ігнатьєва Вікторія Борисівна канд. техн. наук, доцент, доцент кафедри будівельної механіки Тернопільський національний технічний університет імені Івана Пулюя, Україна Ігнатьєва Вікторія Борисівна канд. техн. наук, доцент, доцент кафедри будівельної механіки Тернопільський національний технічний університет імені Івана Пулюя, Україна ПІДВИЩЕННЯ ЕРГОНОМІЧНОСТІ ЕКОСУМКИ ДЛЯ ГОСПОДАРСЬКИХ І БУДІВЕЛЬНИХ МАТЕРІАЛІВ СЕКЦІЯ XVI. ТЕХНОЛОГІЇ ЛЕГКОЇ ТА ДЕРЕВООБРОБНОЇ ПРОМИС All rights reserved \| Creative Commons Attribution 4.0 International License All rights reserved \| Creative Commons Attribution 4.0 International License 2021 Бухта Вікторія Володимирівна магістр за спеціальністю будівництво та цивільна інженерія, інженер I категорії Тернопільський національний технічний університет імені Івана Пулюя, Україна Бухта Вікторія Володимирівна магістр за спеціальністю будівництво та цивільна інженерія, інженер I категорії Тернопільський національний технічний університет імені Івана Пулю Україна Бухта Вікторія Володимирівна магістр за спеціальністю будівництво та цивільна інженерія, інженер I категорії Тернопільський національний технічний університет імені Івана Пулю Україна ЕГКОЇ ТА ДЕРЕВООБРОБНОЇ ПРОМИСЛОВОСТІ Анотація. У статті ставиться завдання розробити конструкцію екосумки для перенесення господарських і будівельних матеріалів, яка задовольняє вимогам ергономічності. Авторами проведено аналіз основних недоліків існуючих конструкцій екосумок для перенесення господарських і будівельних матеріалів. Встановлено, що більшість конструкцій екосумок не дозволяють одночасне перенесення у них пляшок, банок, продуктів, господарських товарів або деяких будівельних матеріалів без втрати первісного зовнішнього вигляду деяких м’яких товарів. Також одночасне перенесення у екосумках деяких товарів суперечить санітарним нормам і сформованим культурним традиціям людства.Показана технологічна можливість підвищення показників ергономічності екосумки для перенесення господарських і будівельних матеріалів. Запропонована конструкція екосумки для перенесення господарських і будівельних матеріалів, яка відповідає вимогам ергономічності. Ключові слова: екосумка, господарська екосумка, ергономічність, універсальність, будівельні матеріали, господарські матеріали. Постановка проблеми. З розвитком суспільства зростають й вимоги споживачів до предметів особистого користування, а саме до конструкцій сумок призначених для перенесення господарських і будівельних матеріалів. Покупці віддають перевагу сумкам, які не тільки безпечні для здоров’я але і відповідають показникам ергономічності. Тому проблема підвищення ергономічності моделі сумки для перенесення господарських і будівельних матеріалів з урахуванням передового світового досвіду стосовно новітніх екологічно чистих розробок є актуальною. Аналіз останніх досліджень і публікацій. За період 1960-2020 р.р. у світі було видано близько 200 патентів на винаходи й корисні моделі, які відносяться до господарських сумок. На рис. 1 представлено розподіл кількості виданих 2021 International scientific journal «Grail of Science» \| № 1 (February, 2021) 210 0 LIGHT AND WOODWORKING INDUSTRY патентів на господарську сумку національним і іноземним заявникам п державам. Рис. 1. Розподіл кількості виданих патентів національним й іноземним SECTION XVI. TECHNOLOGIES OF LIGHT AND WOODWORKING IN Рис. 1. Розподіл кількості виданих патентів національним й іноземним заявникам по державам Рис. 1. Розподіл кількості виданих патентів національним й іноземним заявникам по державам Аналіз даних показує, що по кількості виданих охоронних документів держави розподілились таким чином: - національним заявникам: США – 9, Італія – 8, Німеччина – 5, Франція – 5, Україна – 3, Росія – 2, Велика Британія – 2, Канада – 2, Австрія – 2, Китай – 0, Бельгія – 0 охоронних документів; - іноземним заявникам: Велика Британія – 15, США – 14, Канада – 12, Німеччина – 10, Китай – 10, Австрія – 9, Франція – 7, Бельгія – 7, Італія – 6, Росія – 1, Україна – 0 охоронних документів.Кількісні показники у вигляді співвідношення охоронних документів, виданих іноземним заявникам, та охоронних документів національних заявників коригують цю послідовність і розподіляють держави таким чином: Велика Британія – 15/2=7,5; Канада – 12/2 = 6; Австрія – 9/2 = 4,5; Німеччина – 10/5 = 2; США – 14/9 = 1,55; Франція – 7/5 = 1,4; Італія – 6/8 = 0,75; Росія – 1/2 = 0,5. р Тобто систематизація масиву відібраних у процесі пошуку охоронних документів по кожній державі за національними та іноземними заявниками показала, що найбільшу кількість патентів, які відносяться до господарських сумок, в досліджуваний період отримали заявники США – 9 і Італії – 8 патентів. Це є свідченням того, що в США і Італії досить активно здійснюється фінансування НДР даного напряму. Українські заявники отримали значно меншу кількість охоронних документів. Всього у світі понад 50 виробників, які випускають господарські сумки. Ініціативу їх виготовлення підтримують все більше дизайнерських студій і 2021 Міжнародний науковий журнал «Грааль науки» \| № 1 (Лютий, 2021) 211 виробників, серед них такий відомий бренд, як Dolce & Gabbana. Відомими є екосумки, розроблені британською дизайнеркою Ані Хіндмарч. В Україні також є розробки сумок багаторазового використання. Так, в 2004 році було вирішено проблему перенесення винних пляшок. В пропонованій сумці кожна пляшка знаходиться в окремій захисній ємності. Також виробник передбачив варіанти конструкції сумок для перенесення 2-х, 4-х або 6-ти винних пляшок [1]. У 2005 році в офіційному бюлетені «Промислова власність» № 4 була запропонована господарська сумка, що містить горловину забезпечену зав'язкою і, принаймні, одну ручку [2]. СЕКЦІЯ XVI. ТЕХНОЛОГІЇ ЛЕГКОЇ ТА ДЕРЕВООБРОБНОЇ ПРОМИС Невирішені частини загальної проблеми. Проаналізовані варіанти господарських сумок не відповідають одночасно таким показникам як екологічність, універсальність та ергономічність. Недоліком всіх екосумок є те, що одночасне перенесення у них пляшок, банок, продуктів, господарських товарів або деяких будівельних матеріалів призводить до втрати первісного зовнішнього вигляду деяких м’яких товарів: пляшка нахиляється у бік і давить на прилеглі предмети. Авторські права захищені \| Creative Commons Attribution 4.0 International License Рис. 1. Розподіл кількості виданих патентів національним й іноземним заявникам по державам Також одночасне перенесення у екосумках таких предметів як пляшка з миючим засобом для унітазів або банка з рідкими будівельним матеріалом та продуктів харчування суперечить санітарним нормам і сформованим культурним традиціям людства. Недоліком існуючих екосумок, також є те, що вони мають або тільки одне відділення для перенесення однієї пляшки або банки, або взагалі, пристосовані для перенесення тільки пляшок або банок. Для перенесення продуктів, товарів першої необхідності, господарських товарів та деяких будівельних матеріалів потрібна ще одна сумка. Тобто всі варіанти господарських сумок призначені або тільки для перенесення пляшок і банок або тільки для перенесення продуктів, господарських товарів чи деяких будівельних матеріалів. Це означає, що споживачі повинні або мати декілька сумок, або порушувати санітарні норми і сформовані культурні традиції людства. Обидва варіанти незручні і ускладнюють життя споживачам. Мета статті. Метою даної роботи є пошук шляхів підвищення ергономічних показників екосумки для перенесення господарських і будівельних матеріалів. Реалізація поставленої мети передбачає вирішення наступного завдання: розробка конструкції екосумки для перенесення господарських і будівельних матеріалів. Виклад основного матеріалу. У якості ергономічної екосумки для перенесення господарських і будівельних матеріалів автори пропонують сумку (рис. 2) в корпусі якої є декілька внутрішніх перегородок. Сумка містить корпус, дно, ручку та мотузки. Поздовжня перегородка розділяє корпус сумки на два відділи. Один із відділів, призначений для перенесення будь-яких пляшок або банок, в які упаковані товари першої необхідності, господарські товари або будівельні матеріали. Другий відділ, призначений для перенесення не упакованих в ємності продуктів, товарів першої необхідності або будівельних матеріалів. У відділі, призначеному для перенесення пляшок або банок виконані одна або декілька поперечних перегородок, які створюють окремі 2021 International scientific journal «Grail of Science» \| № 1 (February, 2021) 212 212 SECTION XVI. TECHNOLOGIES OF LIGHT AND WOODWORKING INDUSTRY ємкості для кожної пляшки. На одній зі сторін корпусу сумки, у верхній його частині, виконані отвори. Крім того, у верхній частині внутрішньої перегородки, навпроти отворів в корпусі, виконані пази. Отвори в корпусі та пази у внутрішній перегородці призначені для кріплення мотузки. Мотузка почергово проходить крізь отвори корпусу та пази на внутрішній перегородці. Нею можна закріпити у сумці як одну так і декілька пляшок або банок. Закріплення пляшок мотузкою через пази та наявність поперечних перегородок захищає їх від пошкоджень, які можуть утворитися внаслідок зіткнення однієї пляшки з іншою. Таке рішення не дозволить пляшкам нахилятися у бік, тобто запобігає можливому виливанню вмісту пляшки або банки у випадку їх поганого закриття. Дно корпусу обладнане амортизуючим матеріалом, в якості якого можна обрати поролон. Амортизуючий матеріал зверху закритий вологонепроникаючою поверхнею. All rights reserved \| Creative Commons Attribution 4.0 International License rights reserved \| Creative Commons Attribution 4.0 International License 2021 Рис. 1. Розподіл кількості виданих патентів національним й іноземним заявникам по державам Пропонована конструкція екосумки для господарських і будівельних матеріалів підвищує її ергономічність, тобто робить екосумку привабливою для споживачів, і як наслідок, для виробників, які зможуть збільшити свої прибутки. Список використаних джерел: [1] Акчурін О. Р. (2004). Сумка господарська (Україна Патент № 3048). Укрпатент. Вилучено з: https://base.uipv.org/searchINV/search.php?action=vie wdetails&IdClaim=102537. [2] Михайлюк В. І. (2005). Еластична сумка (Україна Патент № 6080). Укрпатент. Вилучено з: https://base.uipv.org/searchINV/search.php?action=vie wdetails&IdClaim=100653 [3] Ігнатьєва В. Б. (2020). Розвиток трудового потенціалу шляхом інтенсифікації творчості. Фундаментальні та прикладні проблеми сучасних технологій: Матеріали Міжнародної науково-технічної конференції до 60-річчя з дня заснування Тернопільського національного технічного університету імені Івана Пулюя та 175- річчя з дня народження Івана Пулюя (с. 269). 14-15 травня 2020, Тернопіль, Україна. перенесенні пляшок, банок, продуктів, господарських товарів або будівельних матеріалів. При розробці сумки використовувався метод інтенсифікації творчої активності, описаний в роботі [3]. перенесенні пляшок, банок, продуктів, господарських товарів або будівельних матеріалів. При розробці сумки використовувався метод інтенсифікації творчої активності, описаний в роботі [3]. Висновки. Пропонована конструкція екосумки для господарських і будівельних матеріалів підвищує її ергономічність, тобто робить екосумку привабливою для споживачів, і як наслідок, для виробників, які зможуть збільшити свої прибутки. Рис. 1. Розподіл кількості виданих патентів національним й іноземним заявникам по державам Таким чином, дно корпусу сумки захищає дно пляшки від можливих пошкоджень при необережному користуванні екосумкою, наприклад при легкому ударі об тверду поверхню. SECTION XVI. TECHNOLOGIES OF LIGHT AND WOODWORKING IN Пропонована конструкція екосумки для перенесення господарських і будівельних матеріалів є універсальною і ергономічною. По-перше, є можливість одночасно переносити пляшки, банки, продукти, господарські товари та деякі будівельні матеріали. По-друге, виключається можливість пошкодження м’яких товарів при їх одночасному перенесенні з пляшками або банками, тобто зберігається первісний зовнішній вигляд продукції. По- третє, конструкція сумки дозволяє переносити не одну, а декілька пляшок одночасно. а) б) Рис. 1. Господарська сумка: а) загальний вигляд (з розрізом) господарської сумки; б) вигляд (з розрізом) господарської сумки у робочому стані 1 – корпус, 2 – дно, 3 – ручка, 4 – мотузка, 5 – перегородка, 6 – відділ для пляшок, 7 – відділ для продуктів і товарів першої необхідності, 8 – пази, 9 – отвори, 10 – перегородки 11 – ємкість для пляшки, 1 2 – амортизуючий матеріал, 13 – вологонепроникаюча поверхня б) а) а) б) Рис. 1. Господарська сумка: а) загальний вигляд (з розрізом) господарської сумки; б) вигляд (з розрізом) господарської сумки у робочому стані 1 – корпус, 2 – дно, 3 – ручка, 4 – мотузка, 5 – перегородка, 6 – відділ для пляшок, 7 – відділ для продуктів і товарів першої необхідності, 8 – пази, 9 – отвори, 10 – перегородки 11 – ємкість для пляшки, 1 2 – амортизуючий матеріал, 13 – вологонепроникаюча поверхня Рис. 1. Господарська сумка: а) загальний вигляд (з розрізом) господарської сумки; Рис. 1. Господарська сумка: а) загальний вигляд (з розрізом) господарської сумки; б) вигляд (з розрізом) господарської сумки у робочому стані 1 – корпус, 2 – дно, 3 – ручка, 4 – мотузка, 5 – перегородка, 6 – відділ для пляшок, 7 – відділ для продуктів і товарів першої необхідності, 8 – пази, 9 – отвори, 10 – перегородки 11 – ємкість для пляшки, 1 2 – амортизуючий матеріал, 13 – вологонепроникаюча поверхня Таким чином конструкція сумки розширяє її технологічні і споживчі можливості, що підвищує зручність її використання споживачем при All rights reserved \| Creative Commons Attribution 4.0 International License Міжнародний науковий журнал «Грааль науки» \| № 1 (Лютий, 2021) 213 213 СЕКЦІЯ XVI. ТЕХНОЛОГІЇ ЛЕГКОЇ ТА ДЕРЕВООБРОБНОЇ ПРОМИСЛОВОСТІ перенесенні пляшок, банок, продуктів, господарських товарів або будівельних матеріалів. При розробці сумки використовувався метод інтенсифікації творчої активності, описаний в роботі [3]. Висновки. Авторські права захищені \| Creative Commons Attribution 4.0 International License Список використаних джерел: СЕКЦІЯ XVI. ТЕХНОЛОГІЇ ЛЕГКОЇ ТА ДЕРЕВООБРОБНОЇ ПРОМИС [1] Акчурін О. Р. (2004). Сумка господарська (Україна Патент № 3048). Укрпатент. Вилучено з: https://base.uipv.org/searchINV/search.php?action=vie wdetails&IdClaim=102537. [2] Михайлюк В. І. (2005). Еластична сумка (Україна Патент № 6080). Укрпатент. Вилучено з: https://base.uipv.org/searchINV/search.php?action=vie wdetails&IdClaim=100653 XVI. ТЕХНОЛОГІЇ ЛЕГКОЇ ТА ДЕРЕВООБРОБНОЇ ПРОМИСЛОВОСТІ [3] Ігнатьєва В. Б. (2020). Розвиток трудового потенціалу шляхом інтенсифікації творчості. Фундаментальні та прикладні проблеми сучасних технологій: Матеріали Міжнародної науково-технічної конференції до 60-річчя з дня заснування Тернопільського національного технічного університету імені Івана Пулюя та 175- річчя з дня народження Івана Пулюя (с. 269). 14-15 травня 2020, Тернопіль, Україна. 2021 Авторські права захищені \| Creative Commons Attribution 4.0 International License
https://openalex.org/W2105276379	https://repository.up.ac.za/bitstream/2263/58353/1/Hussain_Common_2013.pdf	Latin	null	Common fixed point results for weak contractive mappings in ordered b-dislocated metric spaces with applications	Journal of inequalities and applications	2,013	cc-by	10,378	R ES EARCH Open Access Abstract We ﬁrst introduce a new concept of b-dislocated metric space as a generalization of dislocated metric space and analyze diﬀerent properties of such spaces. A fundamental result for the convergence of sequences in b-dislocated metric spaces is established and is employed to prove some common ﬁxed point results for four mappings satisfying the generalized weak contractive condition in partially ordered b-dislocated metric spaces. Moreover, some examples and applications to integral equations are given here to illustrate the usability of the obtained results. MSC: Primary 47H10; secondary 54H25 Keywords: coincidence point; common ﬁxed point; dislocated metric space; b-dislocated metric space; dominating and dominated maps; altering distance function Common ﬁxed point results for weak contractive mappings in ordered b-dislocated metric spaces with applications Nawab Hussain1, Jamal Rezaei Roshan2, Vahid Parvaneh3 and Mujahid Abbas4 Correspondence: jmlroshan@gmail.com 2Department of Mathematics, Qaemshahr Branch, Islamic Azad University, Qaemshahr, Iran Full list of author information is available at the end of the article *Correspondence: jmlroshan@gmail.com 2Department of Mathematics, Qaemshahr Branch, Islamic Azad University, Qaemshahr, Iran Full list of author information is available at the end of the article Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 1 Introduction and preliminaries The Banach contraction principle is one of the simplest and most applicable results of metric ﬁxed point theory. It is a popular tool for proving the existence of solution of prob- lems in diﬀerent ﬁelds of mathematics. There are several generalizations of the Banach contraction principle in literature on metric ﬁxed point theory [–]. Hitzler and Seda [] introduced the concept of dislocated topologies and named their corresponding gen- eralized metric a dislocated metric. They have also established a ﬁxed point theorem in complete dislocated metric spaces to generalize the celebrated Banach contraction prin- ciple. The notion of dislocated topologies has useful applications in the context of logic programming semantics (see []). Further useful results can be seen in [–]. Deﬁnition .[] Let X be a nonempty set. A mapping dl : X × X →[,∞) is called a dislocated metric (or simply dl-metric) if the following conditions hold for any x,y,z ∈X: (i) If dl(x,y) = , then x = y; (ii) dl(x,y) = dl(y,x); (iii) dl(x,y) ≤dl(x,z) + dl(z,y). (i) If dl(x,y) = , then x = y; (i) If dl(x,y) = , then x = y; (ii) dl(x,y) = dl(y,x); (ii) dl(x,y) = dl(y,x); (iii) dl(x,y) ≤dl(x,z) + dl(z,y). (iii) dl(x,y) ≤dl(x,z) + dl(z,y). (iii) dl(x,y) ≤dl(x,z) + dl(z,y). The pair (X,dl) is called a dislocated metric space or a dl-metric space. Note that when x = y, dl(x,y) may not be . Example .If X = R+ ∪{}, then dl(x,y) = x + y deﬁnes a dislocated metric on X. xample .If X = R+ ∪{}, then dl(x,y) = x + y deﬁnes a dislocated metric on X. ©2013 Hussain et al.; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Page 2 of 21 Deﬁnition .[] A sequence {xn} in a dl-metric space is called: () a Cauchy sequence if, given ε > , there exists n∈N such that for all n,m ≥n, we have dl(xm,xn) < ε or limn,m→∞dl(xn,xm) = , () convergent with respect to dl if there exists x ∈X such that dl(xn,x) →as n →∞. In this case, x is called the limit of {xn} and we write xn →x. A dl-metric space X is called complete if every Cauchy sequence in X converges to a point in X. A dl-metric space X is called complete if every Cauchy sequence in X converges to a point in X. Deﬁnition .A nonempty set X is called an ordered dislocated metric space if it is equipped with a partial ordering ⪯and there exists a dislocated metric dl on X. Deﬁnition .A nonempty set X is called an ordered dislocated metric space if it is equipped with a partial ordering ⪯and there exists a dislocated metric dl on X. Deﬁnition .Let (X,⪯) be a partially ordered set. Then x,y ∈X are called comparable if x ⪯y or y ⪯x holds. Deﬁnition .[] Let (X,⪯) be a partially ordered set. A self-mapping f on X is called dominating if x ⪯fx for each x in X. Example .[] Let X = [,] be endowed with the usual ordering, and let f : X →X be deﬁned by fx = n√x. Since x ≤x  n = fx for all x ∈X, therefore f is a dominating map. Deﬁnition .[] Let (X,⪯) be a partially ordered set. A self-mapping f on X is called dominated if fx ⪯x for each x in X. Example .[] Let X = [,] be endowed with the usual ordering, and let f : X →X be deﬁned by fx = xn for some n ∈N. Since fx = xn ≤x for all x ∈X, therefore f is a dominated map. xample .If X = R+ ∪{}, then dl(x,y) = x + y deﬁnes a dislocated metric on X. Example .[] Let X = [,] be endowed with the usual ordering, and let f : X →X be deﬁned by fx = xn for some n ∈N. Since fx = xn ≤x for all x ∈X, therefore f is a dominated map. In the following, we give the deﬁnition of a b-dislocated metric space. In the following, we give the deﬁnition of a b-dislocated metric space. Deﬁnition .Let X be a nonempty set. A mapping bd : X × X →[,∞) is called a b-dislocated metric (or simply bd-metric) if the following conditions hold for any x,y,z ∈X and s ≥: (bd) If bd(x,y) = , then x = y; (bd) bd(x,y) = bd(y,x); (bd) bd(x,y) ≤s(bd(x,z) + bd(z,y)). (bd) If bd(x,y) = , then x = y; (bd) bd(x,y) = bd(y,x); (bd) bd(x,y) ≤s(bd(x,z) + bd(z,y)). (bd) If bd(x,y) = , then x = y; (bd) If bd(x,y) = , then x = y; (bd) bd(x,y) = bd(y,x); (bd) bd(x,y) ≤s(bd(x,z) + bd(z,y)). (bd) bd(x,y) ≤s(bd(x,z) + bd(z,y)). The pair (X,bd) is called a b-dislocated metric space or a bd-metric space. It should be noted that the class of bd-metric spaces is eﬀectively larger than that of dl-metric spaces, since a bd-metric is a bl-metric when s = . Here, we present an example to show that in general a b-dislocated metric need not be a bl-metric. Example .Let (X,dl) be a dislocated metric space, and bd(x,y) = (bl(x,y))p, where p >  is a real number. We show that bd is a b-dislocated metric with s = p–. Obviously, conditions (bd) and (bd) of Deﬁnition .are satisﬁed. Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 3 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 If < p < ∞, then the convexity of the function f (x) = xp (x > ) implies that ( a+b )p ≤  (ap + bp). Hence, (a + b)p ≤p–(ap + bp) holds. Thus, for each x,y,z ∈X, we obtain that bd(x,y) = dl(x,y) p ≤ dl(x,z) + dl(z,y) p ≤p– dl(x,z) p + dl(z,y) p = p– bd(x,z) + bd(z,y) . So, condition (bd) of Deﬁnition .is also satisﬁed and bd is a bd-metric. So, condition (bd) of Deﬁnition .is also satisﬁed and bd is a bd-metric. However, if (X,dl) is a dislocated metric space, then (X,bd) is not necessarily a dislocated metric space. xample .If X = R+ ∪{}, then dl(x,y) = x + y deﬁnes a dislocated metric on X. For example, if X = R is the set of real numbers, then dl(x,y) = \|x\| + \|y\| is a dislocated metric, and bd(x,y) = (\|x\| + \|y\|)is a b-dislocated metric on R with s = , but not a dislocated metric on R. Recently, Sarma and Kumari [] established the existence of a topology induced by a dislocated metric which is metrizable with a family of sets {B(x,ε) ∪{x} : x ∈X,ε > } as a base, where B(x,ε) = {y ∈X : dl(x,y) < ε} for all x ∈X and ε > . Also, B(x,ε) = {y ∈X : dl(x,y) ≤ε} is a closed ball. On the similar lines, we show that each b-dislocated metric space on X generates a topol- ogy τbd whose base is the family of open bd-balls Bbd(x,ε) = y ∈X : bd(x,y) < ε . Deﬁnition .We say that a net (xα : α ∈) in X converges to x in (X,bd) and write limα∈ xα = x if limα∈ bd(xα,x) = . Note that the limit of a net in (X,bd) is unique. For A ⊆X, we write D(A) = {x ∈X : x is a limit of a net in (A,bd)}. Proposition .If A,B ⊆X, then (i) D(A) = ∅if A = ∅, (ii) D(A) ⊆D(B) if A ⊆B, (iii) D(A ∪B) = D(A) ∪D(B), (iv) D(D(A)) ⊆D(A). (iii) D(A ∪B) = D(A) ∪D(B), (iii) D(A ∪B) = D(A) ∪D(B), (iv) D(D(A)) ⊆D(A). Proof To prove (i), (ii) and (iii), we refer to []. To prove (iv), let x ∈D(D(A)). Suppose that for each α in , (xαβ : β ∈(α)) is a net in A such that xα = limβ∈(α) xαβ . Thus, for each positive integer i, there is αi ∈ such that bd(xαi,x) <  is, and βi ∈(αi) such that bd(xαiβi ,xαi) <  is. Take αiβi = γi for each i, then {γ,γ,γ,...} is a directed set γi < γj if i < j, and bd(xγi,x) ≤s(bd(xγi,xαi) + bd(xαi,x)) <  i . This implies that x ∈D(A). □ As a corollary, we have the following. Corollary .Let, for all A ⊂X, A = A ∪D(A). Then the operation A →A on P(X) sat- isﬁes Kuratowski’s closure axioms []: Corollary .Let, for all A ⊂X, A = A ∪D(A). xample .If X = R+ ∪{}, then dl(x,y) = x + y deﬁnes a dislocated metric on X. Then the operation A →A on P(X) sat- isﬁes Kuratowski’s closure axioms []: (i) ∅= ∅, (ii) A ⊂A, (iii) A = A, (iv) A ∪B = A ∪B. (i) ∅= ∅, (ii) A ⊂A, Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Page 4 of 21 Consequently, we have the following. Theorem .Let ϒ be the family of all subsets A of X for which A = A and τbd are the complements of members of ϒ. Then the τbd is a topology for X and the τbd-closure of a subset A of X is A. Deﬁnition .The topology τbd obtained in Theorem .is called the topology induced by bd and simply referred to as the bd-topology of X; and it is denoted by (X,bd,τbd). Deﬁnition .The topology τbd obtained in Theorem .is called the topology induced by bd and simply referred to as the bd-topology of X; and it is denoted by (X,bd,τbd). Now we state some propositions and corollaries in (X,bd,τbd) which can be proved fol- lowing similar arguments to those given in []. Proposition .Let A ⊆X. Then x ∈D(A) iﬀfor every δ > , Bδ(x) ∩A ̸= ∅. Proposition .Let A ⊆X. Then x ∈D(A) iﬀfor every δ > , Bδ(x) ∩A ̸= ∅. Corollary .x ∈A ⇐⇒x ∈A or Bδ(x) ∩A ̸= ∅, ∀δ > . Corollary .A set A ⊆X is open in (X,bd,τbd) if and only if for every x ∈A, there is δ >  such that {x} ∪Bδ(x) ⊆A. Proposition .If x ∈X and δ > , then {x} ∪Bδ(x) is an open set in (X,bd,τbd). Proposition .If x ∈X and δ > , then {x} ∪Bδ(x) is an open set in (X,bd,τbd). Corollary .If x ∈X and Vr(x) = Br(x) ∪{x} for r > , then the collection {Vr(x)\|x ∈X} is an open base at x in (X,bd,τbd). If bd is a b-metric and V = B(x), then τbd coincides with the metric topology. Proposition .(X,bd,τbd) is a Hausdorﬀspace. Proposition .(X,bd,τbd) is a Hausdorﬀspace. Proof If x,y ∈X and bd(x,y) s = r > , then Vr(x) ∩Vr(y) = ∅. □ Proof If x,y ∈X and bd(x,y) s = r > , then Vr(x) ∩Vr(y) = ∅. □ Proof If x,y ∈X and bd(x,y) s = r > , then Vr(x) ∩Vr(y) = ∅. □ □ Corollary .If x ∈X, then the collection {V(x)\|x ∈X} is an open base at x for (X,bd,τbd). Hence, (X,bd,τbd) is ﬁrst countable. Corollary .If x ∈X, then the collection {V(x)\|x ∈X} is an open base at x for (X,bd,τbd). Hence, (X,bd,τbd) is ﬁrst countable. Remark .The above corollary enables us to deal with sequences instead of nets. Remark .The above corollary enables us to deal with sequences instead of nets. Motivated by Proposition .in [], we have the following proposition for the b-dislo- cated metric space. Motivated by Proposition .in [], we have the following proposition for the b-dislo- cated metric space. Motivated by Proposition .in [], we have the following proposition for the b-dislo- cated metric space. Proposition .Let (X,bd) be a b-dislocated metric space. The following three conditions are equivalent: Proposition .Let (X,bd) be a b-dislocated metric space. The following three conditions are equivalent: are equivalent: (i) For all x ∈X, we have bd(x,x) = . (i) For all x ∈X, we have bd(x,x) = . (i) For all x ∈X, we have bd(x,x) = . (ii) bd is a b-metric. (iii) For all x ∈X and all r > , we have Br(x) ̸= ∅. (iii) For all x ∈X and all r > , we have Br(x) ̸= ∅. Proof We show that (iii) implies (i). Since B r s(x) ̸= ∅for all r > , there exists some y ∈X with bd(x,y) < r s. But for all y ∈X, we have bd(x,x) ≤sbd(x,y). Therefore, bd(x,x) < r for all r > . Hence, bd(x,x) = . □ Proof We show that (iii) implies (i). Since B r s(x) ̸= ∅for all r > , there exists some y ∈X with bd(x,y) < r s. But for all y ∈X, we have bd(x,x) ≤sbd(x,y). Therefore, bd(x,x) < r for all r > . Hence, bd(x,x) = . □ If (X,bd) is a b-dislocated metric space, then (X′,bd), where X′ = {x ∈X\|bd(x,x) = } is a b-metric space. Indeed, (X′,bd) is a b-dislocated metric space, so assertion now follows immediately from the above proposition. Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Page 5 of 21 Deﬁnition .A sequence {xn} in a b-dislocated metric space (X,bd) converges with respect to bd (bd-convergent) if there exists x ∈X such that bd(xn,x) converges to as n →∞. In this case, x is called the limit of {xn}, and we write xn →x. Proposition .Limit of a convergent sequence in a b-dislocated metric space is unique. Proof Let x and y be limits of the sequence {xn}. By properties (bd) and (bd) of Deﬁni- tion ., it follows that bd(x,y) ≤s(bd(xn,x) + bd(xn,y)) →. Hence, bd(x,y) = , and by property (bd) of Deﬁnition .it follows that x = y. Remark .The above corollary enables us to deal with sequences instead of nets. □ Proof Let x and y be limits of the sequence {xn}. By properties (bd) and (bd) of Deﬁni- tion ., it follows that bd(x,y) ≤s(bd(xn,x) + bd(xn,y)) →. Hence, bd(x,y) = , and by property (bd) of Deﬁnition .it follows that x = y. □ Deﬁnition .A sequence {xn} in a b-dislocated metric space (X,bd) is called a bd- Cauchy sequence if, given ε > , there exits n∈N such that for all n,m ≥n, we have bd(xm,xn) < ε or limn,m→∞bd(xn,xm) = . Deﬁnition .A sequence {xn} in a b-dislocated metric space (X,bd) is called a bd- Cauchy sequence if, given ε > , there exits n∈N such that for all n,m ≥n, we have bd(xm,xn) < ε or limn,m→∞bd(xn,xm) = . Proposition .Every convergent sequence in a b-dislocated space is bd-Cauchy. Proposition .Every convergent sequence in a b-dislocated space is bd-Cauchy. Proposition .Every convergent sequence in a b-dislocated space is bd-Cauchy. Proof Let {xn} be a sequence which converges to some x, and ε > . Then there exists n∈N with bd(xn,x) < ε s for all n ≥n. For m,n ≥n, we obtain bd(xn,xm) ≤s(bd(xn,x) + bd(xm,x)) < s ε s = ε. Hence, {xn} is bd-Cauchy. □ Deﬁnition .A b-dislocated metric space (X,bd) is called complete if every bd-Cauchy sequence in X is bd-convergent. Deﬁnition .A b-dislocated metric space (X,bd) is called complete if every bd-Cauchy sequence in X is bd-convergent. Deﬁnition .A b-dislocated metric space (X,bd) is called complete if every bd-Cauchy sequence in X is bd-convergent. The following example shows that in general a b-dislocated metric is not continuous. Example .Let X = N ∪{∞} and bd : X × X →R be deﬁned by Example .Let X = N ∪{∞} and bd : X × X →R be deﬁned by ⎧ ⎪⎪⎨  m +  n if m,n are even or mn = ∞, bd(m,n) = ⎧ ⎪⎪⎨ ⎪⎪⎩  m +  n if m,n are even or mn = ∞,  if m and n are odd and m ̸= n,  otherwise. Then it is easy to see that for all m,n,p ∈X, we have Then it is easy to see that for all m,n,p ∈X, we have Then it is easy to see that for all m,n,p ∈X, we have bd(m,p) ≤ bd(m,n) + bd(n,p) . Thus, (X,bd) is a b-dislocated metric space. Let xn = n for each n ∈N. Then Thus, (X,bd) is a b-dislocated metric space. Let xn = n for each n ∈N. Then Thus, (X,bd) is a b-dislocated metric space. Let xn = n for each n ∈N. Then bd(n,∞) =  n → as n →∞, that is, xn →∞, but bd(xn,) = ↛bd(∞,) as n →∞. bd(n,∞) =  n → as n →∞, We need the following simple lemma about the bd-convergent sequences in the proof of our main results. Lemma .Let (X,bd) be a b-dislocated metric with parameter s ≥. Suppose that {xn} and {yn} are bd-convergent to x, y, respectively. Then we have  sbd(x,y) ≤liminf n→∞bd(xn,yn) ≤limsup n→∞ bd(xn,yn) ≤sbd(x,y). Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 6 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Proposition .Every convergent sequence in a b-dislocated space is bd-Cauchy. Journal of Inequalities and Applications 2013, 2013:486 Page 6 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 In particular, if bd(x,y) = , then we have limn→∞bd(xn,yn) = = bd(x,y). Moreover, for each z ∈X, we have In particular, if bd(x,y) = , then we have limn→∞bd(xn,yn) = = bd(x,y). Moreover, for each z ∈X, we have  s bd(x,z) ≤liminf n→∞bd(xn,z) ≤limsup n→∞ bd(xn,z) ≤sbd(x,z).  s bd(x,z) ≤liminf n→∞bd(xn,z) ≤limsup n→∞ bd(xn,z) ≤sbd(x,z). In particular, if bd(x,z) = , then we have limn→∞bd(xn,z) = = bd(x,z). In particular, if bd(x,z) = , then we have limn→∞bd(xn,z) = = bd(x,z). Proof Using the triangle inequality in a b-dislocated metric space, it is easy to see that bd(x,y) ≤sbd(x,xn) + sbd(xn,yn) + sbd(yn,y) bd(x,y) ≤sbd(x,xn) + sbd(xn,yn) + sbd(yn,y) and bd(xn,yn) ≤sbd(xn,x) + sbd(x,y) + sbd(y,yn). Taking the lower limit as n →∞in the ﬁrst inequality and the upper limit as n →∞in the second inequality, the result follows. Similarly, using again the triangle inequality, the last assertion follows. □ Taking the lower limit as n →∞in the ﬁrst inequality and the upper limit as n →∞in the second inequality, the result follows. Similarly, using again the triangle inequality, the last assertion follows. □ Deﬁnition .[] Let f and g be two self-maps on a nonempty set X. If w = fx = gx, for some x in X, then x is called a coincidence point of f and g, where w is called a point of coincidence of f and g. Deﬁnition .[] Let f and g be two self-maps deﬁned on a set X. Then f and g are said to be weakly compatible if they commute at every coincidence point. Deﬁnition .[] Let f and g be two self-maps deﬁned on a set X. Then f and g are said to be weakly compatible if they commute at every coincidence point. Deﬁnition .Let (X,bd) be a b-dislocated metric space. Then the pair (f ,g) is said to be compatible if and only if limn→∞bd(fgxn,gfxn) = , whenever {xn} is a sequence in X so that limn→∞fxn = limn→∞gxn = t for some t ∈X. 2 Common ﬁxed point results Suppose that Suppose that = ψ : [,∞) →[,∞)\|ψ is a continuous non-decreasing function with ψ(t) = ⇔t =  = ψ : [,∞) →[,∞)\|ψ is a continuous non-decreasing function with ψ(t) = ⇔t =  and = ϕ : [,∞) →[,∞)\|ϕ is a lower semi-continuous function with ϕ(t) = ⇔t =  . = ϕ : [,∞) →[,∞)\|ϕ is a lower semi-continuous function with ϕ(t) = ⇔t =  . Theorem .Let (X,bd,⪯) be an ordered complete b-dislocated metric space, and let f , g, S and T be four self-maps on X such that (f ,g) and (S,T) are dominated and dominat- ing maps, respectively, with fX ⊆TX and gX ⊆SX. Suppose that for all two comparable elements x,y ∈X, ψ sbd(fx,gy) ≤ψ Ms(x,y) – ϕ Ms(x,y) (.) (.) ψ sbd(fx,gy) ≤ψ Ms(x,y) – ϕ Ms(x,y) Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 7 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 P http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Page 7 of 21 is satisﬁed, where is satisﬁed, where Ms(x,y) = max bd(Sx,Ty),bd(fx,Sx),bd(gy,Ty), bd(Sx,gy) + bd(fx,Ty) s , (.) (.) ψ ∈ and ϕ ∈. If for every non-increasing sequence {xn} and a sequence {yn} with yn ⪯ xn, for all n such that yn →u, we have u ⪯xn and either (a) (f ,S) are compatible, f or S is continuous and (g,T) is weakly compatible, or (a) (g,T) are compatible, g or T is continuous and (f ,S) is weakly compatible, then f , g, S and T have a common ﬁxed point. Moreover, the set of common ﬁxed points of f , g, S and T is well ordered if and only if f , g, S and T have one and only one common ﬁxed point. Proof Let xbe an arbitrary point in X. We deﬁne inductively the sequences {xn} and {yn} in X by Proof Let xbe an arbitrary point in X. We deﬁne inductively the sequences {xn} and {yn} in X by yn+= fxn = Txn+, yn+= gxn+= Sxn+, n = ,,,.... n+, yn+= gxn+= Sxn+, n = ,,,.... This can be done as fX ⊆TX and gX ⊆SX. By given assumptions, xn+⪯Txn+= fxn ⪯ xn and xn ⪯Sxn = gxn–⪯xn–. Thus, we have xn+⪯xn for all n ≥. 2 Common ﬁxed point results Suppose that Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Page 9 of 21 which yields that ϕ(bd(yn+,yn+)) ≤, or, equivalently, bd(yn+,yn+) = , a contradic- tion. which yields that ϕ(bd(yn+,yn+)) ≤, or, equivalently, bd(yn+,yn+) = , a contradic- tion. Hence, Ms(xn,xn+) ≤bd(yn,yn+). Since Ms(xn,xn+) ≥bd(yn,yn+), therefore, bd(yn+,yn+) ≤Ms(xn,xn+) = bd(yn,yn+). Following similar arguments to those given above, we have bd(yn+,yn+) ≤Ms(xn+,xn+) = bd(yn+,yn+). (.) (.) Therefore, {bd(yn,yn+)} is a non-increasing sequence and so there exists r ≥such that Therefore, {bd(yn,yn+)} is a non-increasing sequence and so there exists r ≥such that lim n→∞bd(yn–,yn) = lim n→∞Ms(xn,xn+) = r. Suppose that r > . As ψ bd(yn+,yn+) ≤ψ sbd(yn+,yn+) ≤ψ Ms(xn,xn+) – ϕ Ms(xn,xn+) , by taking the upper limit as n →∞, we obtain by taking the upper limit as n →∞, we obtain by taking the upper limit as n →∞, we obtain ψ(r) ≤ψ(r) – liminf n→∞ϕ Ms(xn,xn+) = ψ(r) – ϕ liminf n→∞Ms(xn,xn+) = ψ(r) – ϕ(r), a contradiction. Hence lim n→∞bd(yn–,yn) = . (.) (.) lim n→∞bd(yn–,yn) = . Now, we prove that {yn} is a bd-Cauchy sequence. To do this, it is suﬃcient to show that the subsequence {yn} is bd-Cauchy in X. Assume on the contrary that {yn} is not a bd-Cauchy sequence. Then there exists ε > for which we can ﬁnd subsequences {ymk} and {ynk} of {yn} so that nk is the smallest index for which nk > mk > k, Now, we prove that {yn} is a bd-Cauchy sequence. To do this, it is suﬃcient to show that the subsequence {yn} is bd-Cauchy in X. Assume on the contrary that {yn} is not a bd-Cauchy sequence. Then there exists ε > for which we can ﬁnd subsequences {ymk} and {ynk} of {yn} so that nk is the smallest index for which nk > mk > k, bd(ymk,ynk) ≥ε (.) bd(ymk,ynk) ≥ε (.) and and bd(ymk,ynk–) < ε. (.) (.) bd(ymk,ynk–) < ε. Using the triangle inequality and (.), we obtain that Using the triangle inequality and (.), we obtain that Using the triangle inequality and (.), we obtain that Using the triangle inequality and (.), we obtain that ε ≤bd(ymk,ynk) ≤sbd(ymk,ymk+) + sbd(ymk+,ynk). 2 Common ﬁxed point results Suppose that We will show that {yn} is bd-Cauchy. Suppose that bd(yn,yn+) > for every n. If not, then for some k, bd(yk,yk+) = , and from (.), we obtain ψ bd(yk+,yk+) ≤ψ sbd(yk+,yk+) = ψ sbd(fxk,gxk+) ≤ψ Ms(xk,xk+) – ϕ Ms(xk,xk+) , (.) ≤ψ Ms(xk,xk+) – ϕ Ms(xk,xk+) , (.) (.) where Ms(xk,xk+) = max bd(Sxk,Txk+),bd(fxk,Sxk),bd(gxk+,Txk+), bd(Sxk,gxk+) + bd(fxk,Txk+) s = max bd(yk,yk+),bd(yk+,yk),bd(yk+,yk+), bd(yk,yk+) + bd(yk+,yk+) s = max ,,bd(yk+,yk+), bd(yk,yk+) + bd(yk+,yk+) s = bd(yk+,yk+), (.) (.) since bd(yk,yk+) + bd(yk+,yk+) s ≤sbd(yk,yk+) + sbd(yk+,yk+) + sbd(yk,yk+) s = sbd(yk+,yk+) s < bd(yk+,yk+). Page 8 of 21 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 So, from (.) and (.), we obtain that So, from (.) and (.), we obtain that ψ bd(yk+,yk+) ≤ψ bd(yk+,yk+) – ϕ bd(yk+,yk+) , which gives ϕ(bd(yk+,yk+)) ≤and so yk+= yk+, which further implies that yk+= yk+. Thus, {yn} becomes a constant sequence, hence, yn is a Cauchy sequence. which gives ϕ(bd(yk+,yk+)) ≤and so yk+= yk+, which further implies that yk+= yk+. Thus, {yn} becomes a constant sequence, hence, yn is a Cauchy sequence. k b ( ) f h d bl f ( ) Now, take bd(yn,yn+) > for each n. As xn and xn+are comparable, so from (.) we have ψ bd(yn+,yn+) ≤ψ sbd(yn+,yn+) = ψ sbd(fxn,gxn+) ≤ψ Ms(xn,xn+) – ϕ Ms(xn,xn+) ≤ψ Ms(xn,xn+) . (.) (.) Hence bd(yn+,yn+) ≤Ms(xn,xn+), (.) (.) bd(yn+,yn+) ≤Ms(xn,xn+), where where xn+) = max bd(Sxn,Txn+),bd(fxn,Sxn),bd(gxn+,Txn+ bd(Sxn,gxn+) + bd(fxn,Txn+) s = max bd(yn,yn+),bd(yn+,yn),bd(yn+,yn+), bd(yn,yn+) + bd(yn+,yn+) s ≤max bd(yn,yn+),bd(yn+,yn+), sbd(yn,yn+) + sbd(yn+,yn+) + sbd(yn,yn+) s = max bd(yn,yn+),bd(yn+,yn+), sbd(yn,yn+) + sbd(yn+,yn+) s = max bd(yn,yn+),bd(yn+,yn+) . Ms(xn,xn+) = max bd(Sxn,Txn+),bd(fxn,Sxn),bd(gxn+,Txn+), bd(Sxn,gxn+) + bd(fxn,Txn+) s bd(Sxn,gxn+) + bd(fxn,Txn+) s = max bd(yn,yn+),bd(yn+,yn+), sbd(yn,yn+) + sbd(yn+,yn+) s If for some n, bd(yn+,yn+) ≥bd(yn,yn+) > , then (.) gives that Ms(xn,xn+) = bd(yn+,yn+) and from (.) we have ψ bd(yn+,yn+) ≤ψ sbd(yn+,yn+) ≤ψ Ms(xn,xn+) – ϕ Ms(xn,xn+) = ψ bd(yn+,yn+) – ϕ bd(yn+,yn+) , Hussain et al. 2 Common ﬁxed point results Suppose that Taking the upper limit as k →∞and using (.), we obtain (.) ε s ≤limsup k→∞ bd(ymk+,ynk). (.) ε s ≤limsup k→∞ bd(ymk+,ynk). Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Page 10 of 21 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Using the triangle inequality and (.), we have Using the triangle inequality and (.), we have ε ≤bd(ymk,ynk) ≤sbd(ymk,ynk–) + sbd(ynk–,ynk–) + sbd(ynk–,ynk) < εs + sbd(ynk–,ynk–) + sbd(ynk–,ynk). Taking the upper limit as k →∞and using (.), we obtain ε ≤limsup k→∞ bd(ymk,ynk) ≤εs. (.) (.) ε ≤limsup k→∞ bd(ymk,ynk) ≤εs. Also, ε ≤bd(ymk,ynk) ≤sbd(ymk,ynk–) + sbd(ynk–,ynk). Hence ε s ≤limsup k→∞ bd(ymk,ynk–). On the other hand, we have bd(ymk,ynk–) ≤sbd(ymk,ynk) + sbd(ynk,ynk–). So, from (.) and (.), we have bd(ymk,ynk–) ≤sbd(ymk,ynk) + sbd(ynk,ynk–). bd(ymk,ynk–) ≤sbd(ymk,ynk) + sbd(ynk,ynk–). So, from (.) and (.), we have limsup k→∞ bd(ymk,ynk–) ≤slimsup k→∞ bd(ymk,ynk) ≤εs. limsup k→∞ bd(ymk,ynk–) ≤slimsup k→∞ bd(ymk,ynk) ≤εs. Consequently, ε s ≤limsup k→∞ bd(ymk,ynk–) ≤εs. (.) Similarly, ε s≤limsup k→∞ bd(ymk+,ynk–) ≤εs. ε s≤limsup k→∞ bd(ymk+,ynk–) ≤εs. (.) (.) As xmk and xnk–are comparable, from (.) we have As xmk and xnk–are comparable, from (.) we have ψ sbd(ymk+,ynk) = ψ sbd(fxmk,gxnk–) ≤ψ Ms(xmk,xnk–) – ϕ Ms(xmk,xnk–) , where Ms(xmk,xnk–) = max bd(Sxmk,Txnk–),bd(fxmk,Sxmk),bd(gxnk–,Txnk–), bd(Sxmk,gxnk–) + bd(fxmk,Txnk–) s Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 11 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 11 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Page 11 of 21 = max bd(ymk,ynk–),bd(ymk+,ymk),bd(ynk,ynk–), bd(ymk,ynk) + bd(ymk+,ynk–) s . Taking the upper limit and using (.) and (.)-(.), we get ε + ε s s = min ε s , ε + ε s s ≤limsup k→∞ Ms(xmk,xnk–) k→∞ = max limsup k→∞ bd(ymk,ynk–),,, limsupk→∞bd(ymk,ynk) + limsupk→∞bd(ymk+,ynk–) s = max limsup k→∞ bd(ymk,ynk–),,, = max limsup k→∞ bd(ymk,ynk–),,, limsupk→∞bd(ymk,ynk) + limsupk→∞bd(ymk+,ynk–) s ≤max εs, εs + εs s = εs. ≤max εs, εs + εs s = εs. Hence, we have ε + ε s s ≤limsup k→∞ Ms(xmk,xnk–) ≤εs. (.) (.) Similarly, we can obtain Similarly, we can obtain ε + ε s s ≤liminf k→∞Ms(xmk,xnk–) ≤εs. 2 Common ﬁxed point results Suppose that (.) ε + ε s s ≤liminf k→∞Ms(xmk,xnk–) ≤εs. (.) As As ψ sbd(ymk+,ynk) = ψ sbd(fxmk,gxnk–) ≤ψ Ms(xmk,xnk–) – ϕ Ms(xmk,xnk–) , so, by taking the upper limit as k →∞, and from (.) and (.), we obtain ψ εs = ψ sε s ≤ψ slimsup k→∞ bd(ymk+,ynk) ≤ψ limsup k→∞ Ms(xmk,xnk–) – liminf k→∞ϕ Ms(xmk,xnk–) ≤ψ εs – ϕ liminf k→∞Ms(xmk,xnk–) ≤ψ εs – ϕ liminf k→∞Ms(xmk,xnk–) , ψ εs = ψ sε s ≤ψ slimsup k→∞ bd(ymk+,ynk) ≤ψ limsup k→∞ Ms(xmk,xnk–) – liminf k→∞ϕ Ms(xmk,xnk–) ≤ψ εs – ϕ liminf k→∞Ms(xmk,xnk–) ≤ψ εs – ϕ liminf k→∞Ms(xmk,xnk–) , which implies that ϕ liminf k→∞Ms(xmk,xnk–) = , Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 12 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 so liminfMs(xmk,xnk–) = , a contradiction to (.). Hence {yn} is a bd-Cauchy se- quence in X. Since X is complete, there exists y ∈X such that so liminfMs(xmk,xnk–) = , a contradiction to (.). Hence {yn} is a bd-Cauchy se- quence in X. Since X is complete, there exists y ∈X such that lim n→∞fxn = lim n→∞Txn+= lim n→∞gxn+= lim n→∞Sxn = y. Now, we show that y is a common ﬁxed point of f , g, S and T. Assume that (a) holds and S is continuous. Then Now, we show that y is a common ﬁxed point of f , g, S and T. Assume that (a) holds and S is continuous. Then Assume that (a) holds and S is continuous. Then lim n→∞Sxn+= Sy and lim n→∞Sfxn = Sy. Using the triangle inequality, we have Using the triangle inequality, we have bd(fSxn,Sy) ≤s bd(fSxn,Sfxn) + bd(Sfxn,Sy) . bd(fSxn,Sy) ≤s bd(fSxn,Sfxn) + bd(Sfxn,Sy) . Since the pair (f ,S) is compatible, limn→∞bd(fSxn,Sfxn) = . So, by taking the limit when n →∞in the above inequality, we have Since the pair (f ,S) is compatible, limn→∞bd(fSxn,Sfxn) = . So, by taking the limit when n →∞in the above inequality, we have n →∞in the above inequality, we have lim n→∞bd(fSxn,Sy) ≤s lim n→∞bd(fSxn,Sfxn) + lim n→∞bd(Sfxn,Sy) = . Hence, limn→∞fSxn = Sy. 2 Common ﬁxed point results Suppose that As Sxn+= gxn+⪯xn+, from (.) we obtain (.) ψ sbd(fSxn+,gxn+) ≤ψ Ms(Sxn+,xn+) – ϕ Ms(Sxn+,xn+) , (.) where Ms(Sxn+,xn+) = max bd Sxn+,Txn+ ,bd fSxn+,Sxn+ ,bd(gxn+,Txn+), bd(Sxn+,gxn+) + bd(fSxn+,Txn+) s . Now, by using Lemma ., we get Now, by using Lemma ., we get limsup n→∞ Ms(Sxn+,xn+) ≤max sbd(Sy,y),,, sbd(Sy,y) + sbd(Sy,y) s = sbd(Sy,y). limsup n→∞ Ms(Sxn+,xn+) ≤max sbd(Sy,y),,, sbd(Sy,y) + sbd(Sy,y) s = sbd(Sy,y). Hence, by taking the upper limit in (.) and using Lemma ., we obtain ψ sbd(Sy,y) = ψ s sbd(Sy,y) ≤ψ sbd(Sy,y) – ϕ sbd(Sy,y) ≤ψ sbd(Sy,y) – ϕ sbd(Sy,y) which gives ϕ(sbd(Sy,y)) ≤, or, equivalently, Sy = y. Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 13 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 13 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Now, since gxn+⪯xn+and gxn+→y as n →∞, then y ⪯xn+and from (.) we have ψ sbd(fy,gxn+) ≤ψ Ms(y,xn+) – ϕ Ms(y,xn+) , (.) ψ sbd(fy,gxn+) ≤ψ Ms(y,xn+) – ϕ Ms(y,xn+) , (.) where Ms(y,xn+) = max bd(Sy,Txn+),bd(fy,Sy),bd(gxn+,Txn+), bd(Sy,gxn+) + bd(fy,Txn+) s . Taking the upper limit as n →∞in (.) and using Lemma ., we have Taking the upper limit as n →∞in (.) and using Lemma ., we have ψ sbd(fy,y) = ψ s s bd(fy,y) ≤ψ bd(fy,y) – ϕ bd(fy,y) ≤ψ sbd(fy,y) – ϕ bd(fy,y) , which implies that ϕ(bd(fy,y)) ≤, so fy = y. which implies that ϕ(bd(fy,y)) ≤, so fy = y. Since f (X) ⊆T(X), there exists a point v ∈X such that fy = Tv. Suppose that gv ̸= Tv. Since v ⪯Tv = fy ⪯y, from (.) we have Since f (X) ⊆T(X), there exists a point v ∈X such that fy = Tv. Suppose that gv ̸= Tv. Since v ⪯Tv = fy ⪯y, from (.) we have ψ bd(Tv,gv) = ψ bd(fy,gv) ≤ψ Ms(y,v) – ϕ Ms(y,v) , (.) (.) where where Ms(y,v) = max bd(Sy,Tv),bd(fy,Sy),bd(gv,Tv), bd(Sy,gv) + bd(fy,Tv) s = bd(gv,Tv). 2 Common ﬁxed point results Suppose that Ms(y,v) = max bd(Sy,Tv),bd(fy,Sy),bd(gv,Tv), bd(Sy,gv) + bd(fy,Tv) s So, from (.) we have So, from (.) we have So, from (.) we have ψ bd(Tv,gv) ≤ψ bd(gv,Tv) – ϕ bd(gv,Tv) , a contradiction. Therefore gv = Tv. Since the pair (g,T) is weakly compatible, gy = gfy = gTv = Tgv = Tfy = Ty and y is the coincidence point of g and T. Since Sxn ⪯xn and Sxn → y as n →∞, it implies that y ⪯xn and from (.) we obtain a contradiction. Therefore gv = Tv. Since the pair (g,T) is weakly compatible, gy = gfy = gTv = Tgv = Tfy = Ty and y is the coincidence point of g and T. Since Sxn ⪯xn and Sxn → y as n →∞, it implies that y ⪯xn and from (.) we obtain ψ sbd(fxn,gy) ≤ψ Ms(xn,y) – ϕ Ms(xn,y) , (.) (.) where Ms(xn,y) = max bd(Sxn,Ty),bd(fxn,Sxn),bd(gy,Ty), bd(Sxn,gy) + bd(fxn,Ty) s . (.) (.) Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 14 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 14 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Page 14 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Taking the upper limit as n →∞in (.) and using Lemma ., we have Taking the upper limit as n →∞in (.) and using Lemma ., we have max  s bd(y,gy),bd(gy,Ty), s sbd(y,gy) ≤liminf n→∞Ms(xn,y) ≤limsup n→∞ Ms(xn,y) max  s bd(y,gy),bd(gy,Ty), s sbd(y,gy) ≤liminf n→∞Ms(xn,y) ≤limsup n→∞ Ms(xn,y) ≤max sbd(y,gy),bd(gy,Ty), s sbd(y,gy) = max sbd(y,gy),bd(gy,gy) ≤max sbd(y,gy),sbd(y,gy) = sbd(y,gy). (.) = sbd(y,gy). (.) (.) = sbd(y,gy). (.) Taking the upper limit as n →∞in (.) and using Lemma .and (.), we have ψ sbd(y,gy) = ψ s s bd(y,gy) ≤ψ limsup n→∞ Ms(xn,y) – liminf n→∞ϕ Ms(xn,y) ≤ψ sbd(y,gy) – ϕ liminf n→∞Ms(xn,y) ≤ψ sbd(y,gy) – ϕ liminf n→∞Ms(xn,y) , which implies that liminfn→∞Ms(xn,y) = , so we have y = gy. Therefore, fy = gy = Sy = Ty = y. The proof is similar when f is continuous. The proof is similar when f is continuous. Similarly, if (a) holds, then the result follows. 2 Common ﬁxed point results Suppose that Similarly, if (a) holds, then the result follows. Now, suppose that the set of common ﬁxed points of f , g, S and T is well ordered. We show that they have a unique common ﬁxed point. Assume on the contrary that fu = gu = Su = Tu = u and fv = gv = Sv = Tv = v, but u ̸= v. By assumption, we can apply (.) to obtain ψ sbd(u,v) = ψ sbd(fu,gv) ≤ψ sbd(fu,gv) ≤ψ Ms(u,v) – ϕ Ms(u,v) , ψ sbd(u,v) = ψ sbd(fu,gv) ≤ψ sbd(fu,gv) ≤ψ Ms(u,v) – ϕ Ms(u,v) , where Ms(u,v) = max bd(Su,Tv),bd(fu,Su),bd(gv,Tv), bd(Su,gv) + bd(fu,Tv) s = max bd(u,v),bd(u,u),bd(v,v), bd(u,v) + bd(u,v) s = max bd(u,v),bd(u,u),bd(v,v) ≤max bd(u,v),sbd(u,v),sbd(u,v) ≤sbd(u,v). Ms(u,v) = max bd(Su,Tv),bd(fu,Su),bd(gv,Tv), bd(Su,gv) + bd(fu,Tv) s = max bd(u,v),bd(u,u),bd(v,v), bd(u,v) + bd(u,v) s Page 15 of 21 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hence Hence ψ sbd(u,v) ≤ψ sbd(u,v) – ϕ Ms(u,v) . ψ sbd(u,v) ≤ψ sbd(u,v) – ϕ Ms(u,v) . So, we have Ms(u,v) = , a contradiction. Therefore u = v. The converse is obvious. So, we have Ms(u,v) = , a contradiction. Therefore u = v. The converse is obvious. □ In the following theorem, we omit the continuity assumption of f , g, T and S and replace the compatibility of the pairs (f ,S) and (g,T) by weak compatibility of the pairs, and we show that f , g, S and T have a common ﬁxed point on X. Theorem .Let (X,bd,⪯) be an ordered complete b-dislocated metric space, and f , g, S and T be four self-maps on X such that (f ,g) and (S,T) are dominated and dominating maps, respectively, with fX ⊆TX and gX ⊆SX, and TX and SX are bd-closed subsets of X. Suppose that for all two comparable elements x,y ∈X, ψ sd(fx,gy) ≤ψ Ms(x,y) – ϕ Ms(x,y) (.) (.) is satisﬁed, where Ms(x,y) = max bd(Sx,Ty),bd(fx,Sx),bd(gy,Ty), bd(Sx,gy) + bd(fx,Ty) s , ψ ∈ and ϕ ∈. 2 Common ﬁxed point results Suppose that If for every non-increasing sequence {xn} and a sequence {yn} with yn ⪯ xn, for all n such that yn →u, we have u ⪯xn, and the pairs (f ,S) and (g,T) are weakly compatible, then f , g, S and T have a common ﬁxed point. Moreover, the set of common ﬁxed points of f , g, S and T is well ordered if and only if f , g, S and T have one and only one common ﬁxed point. Proof Following the proof of Theorem ., there exists y ∈X such that Proof Following the proof of Theorem ., there exists y ∈X such that Proof Following the proof of Theorem ., there exists y ∈X such that lim k→∞bd(yk,y) = . (.) (.) lim k→∞bd(yk,y) = . lim k→∞bd(yk,y) = . Since T(X) is bd-closed and {yn+} ⊆T(X), therefore y ∈T(X). Hence, there exists u ∈X such that y = Tu and lim n→∞bd(yn+,Tu) = lim n→∞bd(Txn+,Tu) = . (.) (.) Similarly, there exists v ∈X such that y = Tu = Sv and Similarly, there exists v ∈X such that y = Tu = Sv and lim n→∞bd(yn,Sv) = lim n→∞bd(Sxn,Sv) = . (.) (.) lim n→∞bd(yn,Sv) = lim n→∞bd(Sxn,Sv) = . Now we prove that v is a coincidence point of f and S. Now we prove that v is a coincidence point of f and S. Since Txn+→y = Sv as n →∞, so, by assumption, Txn+⪯Sv. Therefore, from (.) we have ψ sbd(fv,gxn+) ≤ψ Ms(v,xn+) – ϕ Ms(v,xn+) , (.) (.) Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 16 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Page 16 of 21 where Ms(v,xn+) = max bd(Sv,Txn+),bd(fv,Sv),bd(gxn+,Txn+), bd(Sv,gxn+) + bd(fv,Txn+) s = max bd(Tu,Txn+),bd(fv,y),bd(gxn+,Txn+), bd(Sv,yn+) + bd(fv,Txn+) s . Ms(v,xn+) = max bd(Sv,Txn+),bd(fv,Sv),bd(gxn+,Txn+), bd(Sv,gxn+) + bd(fv,Txn+) s = max bd(Tu,Txn+),bd(fv,y),bd(gxn+,Txn+), bd(Sv,yn+) + bd(fv,Txn+) s . Taking the upper limit as n →∞and using (.)-(.) and Lemma ., we obtain that max bd(fv,y),  sbd(y,y),  sbd(y,y) ≤liminf n→∞Ms(v,xn+) ≤limsup n→∞ Ms(v,xn+) ≤max ,bd(fv,y),sbd(y,y), + sbd(fv,y) s ≤max bd(fv,y),sbd(fv,y) = sbd(fv,y). 2 Common ﬁxed point results Suppose that Therefore, fy = gy = Sy = Ty. which implies that Ms(y,y) = , so we have fy = gy. Therefore, fy = gy = Sy = Ty. Now, similar to the proof of Theorem ., indeed from (.)-(.), we have gy = y. Therefore, fy = gy = Sy = Ty = y, as required. The last conclusion follows similarly as in the proof of Theorem .. □ Now, we give an example to support our result. Now, we give an example to support our result. Now, we give an example to support our result. Example .Let X = [,∞) be equipped with the b-dislocated metric bd(x,y) = (x + y) where s = and suppose that ‘⪯’ is the usual ordering ≤on X. Obviously, (X,bd,≤) is an ordered complete b-dislocated metric space. Let f ,g,S,T : X →X be deﬁned as f (x) = ln + x  , g(x) = ln + x  , S(x) = ex – , T(x) = ex – . f (x) = ln + x  , g(x) = ln + x  , S(x) = ex – , T(x) = ex – . S(x) = ex – , T(x) = ex – . For each x ∈X, we have + x ≤ex and + x ≤ex, so f (x) = ln(+ x ) ≤x, g(x) = ln(+ x ) ≤x, x ≤ex – = S(x) and x ≤ex – = T(x). Thus, f and g are dominated and T and S are dominating with f (X) = g(X) = S(X) = T(X) = [,∞). Also, the pair (g,T) is compatible, g is continuous and (f ,S) is weakly compatible. Let the control functions ψ,ϕ : [,∞) → [,∞) be deﬁned as ψ(t) = bt and ϕ(t) = (b – )t, for all t ∈[,∞), where < b ≤ . Note that For each x ∈X, we have + x ≤ex and + x ≤ex, so f (x) = ln(+ x ) ≤x, g(x) = ln(+ x ) ≤x, x ≤ex – = S(x) and x ≤ex – = T(x). Thus, f and g are dominated and T and S are dominating with f (X) = g(X) = S(X) = T(X) = [,∞). Also, the pair (g,T) is compatible, g is continuous and (f ,S) is weakly compatible. Let the control functions ψ,ϕ : [,∞) → [,∞) be deﬁned as ψ(t) = bt and ϕ(t) = (b – )t, for all t ∈[,∞), where < b ≤ . 2 Common ﬁxed point results Suppose that (.) max bd(fv,y),  sbd(y,y),  sbd(y,y) ≤liminf n→∞Ms(v,xn+) ≤limsup n→∞ Ms(v,xn+) ≤liminf n→∞Ms(v,xn+) ≤limsup n→∞ Ms(v,xn+) n ∞ ≤max ,bd(fv,y),sbd(y,y), + sbd(fv,y) s ≤max bd(fv,y),sbd(fv,y) = sbd(fv,y). (.) ≤max ,bd(fv,y),sbd(y,y), + sbd(fv,y) s ≤max bd(fv,y),sbd(fv,y) = sbd(fv,y). (.) (.) Taking the upper limit as n →∞in (.) and using (.) and Lemma ., we obtain that ψ sd(fv,y) = ψ s s d(fv,y) ≤ψ sbd(fv,y) – ϕ liminf n→∞Ms(v,xn+) ≤ψ sbd(fv,y) – ϕ liminf n→∞Ms(v,xn+) , which implies that liminfn→∞Ms(v,xn+) = , so from (.) we obtain fv = y = Sv. which implies that liminfn→∞Ms(v,xn+) = , so from (.) we obtain fv = y = Sv. which implies that liminfn→∞Ms(v,xn+) = , so from (.) we obtain fv = y = Sv. As f and S are weakly compatible, we have fy = fSv = Sfv = Sy. Thus, y is a coincidence point of f and S. As f and S are weakly compatible, we have fy = fSv = Sfv = Sy. Thus, y is a coincidence point of f and S. As f and S are weakly compatible, we have fy = fSv = Sfv = Sy. Thus, y is a coincidence point of f and S. Similarly, it can be shown that y is a coincidence point of the pair (g,T). Now, we show that fy = gy. From (.) we have ψ sd(fy,gy) ≤ψ Ms(y,y) – ϕ Ms(y,y) , where where Ms(y,y) = max bd(Sy,Ty),bd(fy,Sy),bd(gy,Ty), bd(Sy,gy) + bd(fy,Ty) s = max bd(fy,gy),bd(fy,fy),bd(gy,gy), bd(fy,gy) + bd(fy,gy) s Page 17 of 21 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 = max bd(fy,gy),bd(fy,fy),bd(gy,gy) ≤max bd(fy,gy),sbd(fy,gy),sbd(fy,gy) = sbd(fy,gy). = max bd(fy,gy),bd(fy,fy),bd(gy,gy) ≤max bd(fy,gy),sbd(fy,gy),sbd(fy,gy) = sbd(fy,gy). So, we have So, we have ψ sd(fy,gy) ≤ψ Ms(y,y) – ϕ Ms(y,y) ≤ψ sbd(fy,gy) – ϕ Ms(y,y) ≤ψ sbd(fy,gy) – ϕ Ms(y,y) , ψ sd(fy,gy) ≤ψ Ms(y,y) – ϕ Ms(y,y) ≤ψ sbd(fy,gy) – ϕ Ms(y,y) ≤ψ sbd(fy,gy) – ϕ Ms(y,y) , which implies that Ms(y,y) = , so we have fy = gy. Now, we give an example to support our result. Corollary .Let (X,bd,⪯) be an ordered complete b-dislocated metric space. Let f and g be dominated self-maps on X. Suppose that for every two comparable elements x,y ∈X, Corollary .Let (X,bd,⪯) be an ordered complete b-dislocated metric space. Let f and g be dominated self-maps on X. Suppose that for every two comparable elements x,y ∈X, sbd(fx,gy) ≤Ms(x,y) – ϕ Ms(x,y) is satisﬁed, where Ms(x,y) = max bd(x,y),bd(fx,x),bd(gy,y), bd(x,gy) + bd(fx,y) s , and ϕ ∈. If for every non-increasing sequence {xn} and a sequence {yn} with yn ⪯xn, for all n such that yn →u, it implies that u ⪯xn, then f and g have a common ﬁxed point. Moreover, the set of common ﬁxed points of f and g is well ordered if and only if f and g have one and only one common ﬁxed point. Proof If we take S and T as the identity maps on X and ψ(t) = t for all t ∈[,∞), then from Theorem .it follows that f and g have a common ﬁxed point. □ Remark .As corollaries we can state partial metric space as well as b-metric space ver- sions of our proved results in a similar way, which extends recent results in these settings. Now, we give an example to support our result. Note that ψ sbd f (x),g(y) = b f (x) + g(y)  = b ln + x  + ln + y   ≤b x + y   =  b(x + y) ≤ ex – + ey –   = bd S(x),T(y) ≤M(x,y) = ψ M(x,y) – ϕ M(x,y) , x,y ∈X. ψ sbd f (x),g(y) = b f (x) + g(y)  = b ln + x  + ln + y   ≤b x + y   =  b(x + y) ≤ ex – + ey –   = bd S(x),T(y) ≤M(x,y) = ψ M(x,y) – ϕ M(x,y) , x,y ∈X. Thus, f , g, S and T satisfy all the conditions of Theorem .. Moreover, is a unique common ﬁxed point of f , g, S and T. Thus, f , g, S and T satisfy all the conditions of Theorem .. Moreover, is a unique common ﬁxed point of f , g, S and T. Thus, f , g, S and T satisfy all the conditions of Theorem .. Moreover, is a unique common ﬁxed point of f , g, S and T. Page 18 of 21 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Corollary .Let (X,bd,⪯) be an ordered complete b-dislocated metric space, and let f and g be two dominated self-maps on X. Suppose that for every two comparable elements x,y ∈X, ψ sbd(fx,gy) ≤ψ Ms(x,y) – ϕ Ms(x,y) is satisﬁed, where Ms(x,y) = max bd(x,y),bd(fx,x),bd(gy,y), bd(x,gy) + bd(fx,y) s , ψ ∈ and ϕ ∈. If for every non-increasing sequence {xn} and a sequence {yn} with yn ⪯ xn, for all n such that yn →u, we have u ⪯xn, then f and g have a common ﬁxed point. Moreover, the set of common ﬁxed points of f and g is well ordered if and only if f and g have one and only one common ﬁxed point. Proof Taking S and T as identity maps on X, the result follows from Theorem .. Proof Taking S and T as identity maps on X, the result follows from Theorem .. 3 Existence of a common solution for a system of integral equations d h f ll f l Theorem .Suppose that the following hypotheses hold: (i) K,K: [a,b] × [a,b] × R →R are continuous; (ii) for all t,r ∈[a,b] and x ∈X, we have Theorem .Suppose that the following hypotheses hold: (i) K,K: [a,b] × [a,b] × R →R are continuous; (ii) for all t,r ∈[a,b] and x ∈X, we have Theorem .Suppose that the following hypotheses hold: (i) K,K: [a,b] × [a,b] × R →R are continuous; (ii) for all t,r ∈[a,b] and x ∈X, we have x(t) ≤min b a K t,r,x(r) dr, b a K t,r,x(r) dr ; (iii) for all r,t ∈[a,b] and x,y ∈X with x ⪯y, we have K t,r,x(r) + K t,r,y(r) ≤ξ(t,r)ln + x(r) + y(r) p , where ξ is a continuous function satisfying where ξ is a continuous function satisfying sup t∈[a,b] b a ξ(t,r)p dr <  p–p(b – a)p–. Then the integral equations (.) have a common solution x ∈X. Then the integral equations (.) have a common solution x ∈X. 3 Existence of a common solution for a system of integral equations d h f ll f l 3 Existence of a common solution for a system of integral equations Consider the following system of integral equations: Consider the following system of integral equations: x(t) = b a K t,r,x(r) dr, x(t) = b a K t,r,x(r) dr, (.) (.) where b > a ≥. The purpose of this section is to present an existence theorem for a solution to (.) that belongs to X = C[a,b] (the set of continuous real functions deﬁned on [a,b]) by using the obtained result in Corollary .. Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 19 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Here, K,K: [a,b] × [a,b] × R →R. The considered problem can be reformulated in the following manner. Let f ,g : X →X be the mappings deﬁned by Let f ,g : X →X be the mappings deﬁned by fx(t) = b a K t,r,x(r) dr, gx(t) = b a K t,r,x(r) dr for all x ∈X and for all t ∈[a,b]. for all x ∈X and for all t ∈[a,b]. for all x ∈X and for all t ∈[a,b]. Then the existence of a solution to (.) is equivalent to the existence of a common ﬁxed point of f and g. According to Example ., X equipped with bd(u,v) = max t∈[a,b] u(t) + v(t) p bd(u,v) = max t∈[a,b] u(t) + v(t) p for all u,v ∈X, is a complete b-dislocated metric space with s = p–. We endow X with the partial ordering ⪯given by p p We endow X with the partial ordering ⪯given by We endow X with the partial ordering ⪯given by x ⪯y ⇐⇒ x(t) ≤y(t) for all t ∈[a,b]. Moreover, in [], it is proved that (X,⪯) is regular. Now, we will prove the following result. for all t ∈[a,b]. Moreover, in [], it is proved that (X,⪯) is regular. Now, we will prove the following result. for all t ∈[a,b]. Moreover, in [], it is proved that (X,⪯) is regular. Now, we will prove the following result. Now, we will prove the following result. Now, we will prove the following result. Author details 1Department of Mathematics, King Abdulaziz University, P.O. Box 80203, Jeddah, 21589, Saudi Arabia. 2Department of Mathematics, Qaemshahr Branch, Islamic Azad University, Qaemshahr, Iran. 3Young Researchers and Elite Club, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran. 4Department of Mathematics and Applied Mathematics, University of Pretoria, Lynwood road, Pretoria 0002, South Africa. Authors’ contributions ll h b d Authors’ contributions Authors’ contributions All authors contributed equally and signiﬁcantly in writing this article. All authors read and approved the ﬁnal manuscript. All authors contributed equally and signiﬁcantly in writing this article. All authors read and approved the ﬁnal manuscript Then the integral equations (.) have a common solution x ∈X. Proof From condition (ii), f and g are dominated self-maps on X. Let ≤p,q < ∞with  p +  q = . Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 Page 20 of 21 q pp http://www.journalofinequalitiesandapplications.com/content/2013/1/486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Now, let x,y ∈X be such that x ⪰y. From condition (iii), for all t ∈[a,b], we have Now, let x,y ∈X be such that x ⪰y. From condition (iii), for all t ∈[a,b], we have p–fx(t) + gy(t) p ≤p–p b a K t,r,x(r) + K t,r,x(r) dr p ≤p–p b a q dr  q b a K t,r,x(r) + K t,r,x(r) p dr  p p ≤p–p(b – a) p q b a ξ(t,r)p ln + x(r) + y(r) pp dr ≤p–p(b – a) p q b a ξ(t,r)p ln + bd(x,y) p dr ≤p–p(b – a) p q b a ξ(t,r)p ln + Ms(x,y) p dr = p–p(b – a)p– b a ξ(t,r)p dr ln + Ms(x,y) p < ln + Ms(x,y) p = Ms(x,y)p – Ms(x,y)p – ln + Ms(x,y) p . Hence, sbd(fx,gy) p = ssup t∈[a,b] fx(t) + gy(t) p ≤Ms(x,y)p – Ms(x,y)p – ln + Ms(x,y) p . sbd(fx,gy) p = ssup t∈[a,b] fx(t) + gy(t) p Taking ψ(t) = tp and ϕ(t) = tp –(ln(+t))p in Corollary ., there exists x ∈X, a common ﬁxed point of f and g, that is, x is a solution for (.). □ Competing interests Competing interests The authors declare that they have no competing interests. Acknowledgements h l f d d c o edge e ts This article was funded by the Deanship of Scientiﬁc Research (DSR), King Abdulaziz University, Jeddah. Therefore, the ﬁrst author acknowledges with thanks DSR, KAU for ﬁnancial support. Received: 6 July 2013 Accepted: 13 September 2013 Published: 07 Nov 2013 10.1186/1029-242X-2013-486 Cite this article as: Hussain et al.: Common ﬁxed point results for weak contractive mappings in ordered b-dislocated metric spaces with applications. Journal of Inequalities and Applications 2013, 2013:486 Hussain et al. Journal of Inequalities and Applications 2013, 2013:486 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 References References 1. Aghajani, A, Abbas, M, Roshan, JR: Common ﬁxed point of generalized weak contractive mappings in partially ordered b-metric spaces. Math. Slovaca (2012, in press) 2. Kelle, JL: General Topology. Van Nostrand, New York (1960) 3. Jungck, G: Compatible mappings and common ﬁxed points. Int. J. Math. Math. Sci. 9(4), 771-779 (1986) 4. Nieto, JJ, Rodríguez-López, R: Existence and uniqueness of ﬁxed points in partially ordered sets and applications to ordinary diﬀerential equations. Acta Math. Sin. Engl. Ser. 23, 2205-2212 (2007) Page 21 of 21 http://www.journalofinequalitiesandapplications.com/content/2013/1/486 5. Abbas, M, Nemeth, SZ: Finding solutions of implicit complementarity problems by isotonicity of metric projection. Nonlinear Anal. 75, 2349-2361 (2012) 6. Hussain, N, Kadelburg, Z, Radenovi´c, S, Al-Solamy, FR: Com 6. Hussain, N, Kadelburg, Z, Radenovi´c, S, Al-Solamy, FR: Comparison functions and ﬁxed point results in partial metric spaces. Abstr. Appl. Anal. 2012, Article ID 605781 (2012) spaces. Abstr. Appl. Anal. 2012, Article ID 605781 (2012) p pp 7. Hussain, N, Ðori´c, D, Kadelburg, Z, Radenovi´c, S: Suzuki-type ﬁxed point results in metric type spaces. Fixed Point Th A l 2012 126 (2012) 7. Hussain, N, Ðori´c, D, Kadelburg, Z, Radenovi´c, S: Suzuki-type ﬁxed point results in metric type spaces. Fixed Point Theory Appl. 2012, 126 (2012) 7. Hussain, N, Ðori´c, D, Kadelburg, Z, Radenovi´c, S: Suzuki-ty 7. Hussain, N, Ðori´c, D, Kadelburg, Theory Appl. 2012, 126 (2012) y 8. Shah, MH, Simic, S, Hussain, N, Sretenovic, A, Radenovic, S: Common ﬁxed points theorems for occasionally weakly compatible pairs on cone metric type spaces J Comput Anal Appl 14 290-297 (2012) 9. Hussain, N, Shah, MH: KKM mappings in cone b-metric spaces. Comput. Math. Appl. 62, 1677-1684 ( 10. Khamsi, MA, Hussain, N: KKM mappings in metric type spaces. Nonlinear Anal. 73, 3123-3129 (2010) 1. Hitzler, P, Seda, AK: Dislocated topologies. J. Electr. Eng. 51(12), 3- 12. Hitzler, P: Generalized metrics and topology in logic programming semantics. PhD thesis, National University of Ireland (University College, Cork) (2001) 12. Hitzler, P: Generalized metrics and topology in logic programming seman Ireland (University College, Cork) (2001) Ireland (University College, Cork) (2001) Ireland (University College, Cork) (2001) 13. Aage, CT, Salunke, JN: The results on ﬁxed points in dislocated and dislocated quasi-metric space. Appl. Math. Sci. 2(59), 2941-2948 (2008) 13. Aage, CT, Salunke, JN: The results on ﬁxed points in dislocated and dislocated quasi-metric space. Appl. Math. Sc 2(59), 2941-2948 (2008) 14. Isufati, A: Fixed point theorems in dislocated quasi-metric space. Appl. Math. Sci. 4(5), 217-233 (2010) ari, PS: On dislocated metric spaces. Int. J. Math. Arch. 3(1), 72-77 (20 15. Sarma, IR, Kumari, PS: On dislocated metric spaces. Int. J. Math. 16. Shrivastava, R, Ansari, ZK, Sharma, M: Some results on ﬁxed points in dislocated and dislocated quasi-metric spaces J. Adv. Stud. Topol. 3(1), 25-31 (2012) 17. Shrivastava, M, Qureshi, K, Singh, AD: A ﬁxed poi spaces. Int. J. Theor. Appl. Sci. 4(1), 39-40 (2012) 17. Shrivastava, M, Qureshi, K, Singh, AD: A ﬁxed point theorem for continuous mapping in dislocated quasi-metric spaces. Int. J. Theor. Appl. Sci. 4(1), 39-40 (2012) spaces. Int. J. Theor. Appl. Sci. 4(1), 39-40 (2012) 18. Kumari, PS, Kumar, VV, Sarma, IR: Common ﬁxed point theorems on weakly compatible maps on dislocated metric spaces. Math. Sci. 6, 71 (2012) 18. Kumari, PS, Kumar, VV, Sarma, IR: Common ﬁxed 18. Kumari, PS, Kumar, VV, Sarma, spaces. Math. Sci. 6, 71 (2012) 19. Kumari, PS: Some ﬁxed point theorems in generalized dislocated metric spaces. Math. Theory Model. 1(4), 16-22 (2011) 20. Daheriya, RD, Jain, R, Ughade, M: Some ﬁxed point theorem for expansive type mapping in dislocated metric space. ISRN Math. Anal. 2012, Article ID 376832 (2012) ISRN Math. Anal. 2012, Article ID 376832 (2012) 21. George, R: Cyclic contractions and ﬁxed points in dislocated metric spaces. Int. J. Math. Anal. 7(9), 403-411 (2013) h ﬁ d h d l d l h S ( ) ( ) 22. Jha, K: A common ﬁxed point theorem in dislocated metric space. Appl. Math. Sci. 6(91), 4497-4503 (2012) 23 Z t K S lt i di l t d d di l t d i t i A l M th S i 6(71) 3519 3526 (2012) 23. Zoto, K: Some new results in dislocated and dislocated quasi-metric spaces. Appl. Math. Sci. 6(71), 3519-3526 (2012) 23. Zoto, K: Some new results in dislocated and dislocated quasi-metric spaces. Appl. Math. Sci. q pp http://www.journalofinequalitiesandapplications.com/content/2013/1/486 Ireland (University College, Cork) (2001) 6(71), 3519-3526 (2012) 10.1186/1029-242X-2013-486 Cite this article as: Hussain et al.: Common ﬁxed point results for weak contractive mappings in ordered b-dislocated metric spaces with applications. Journal of Inequalities and Applications 2013, 2013:486
https://openalex.org/W4285062180	https://ojs.sites.ufsc.br/index.php/mixsustentavel/article/download/4697/3810	Portuguese	null	PROPOSTA DE SISTEMA MODULAR EM BAMBU PARA HABITAÇÃO EMERGENCIAL	Mix Sustentável	2,021	cc-by	8,726	RESUMO RESUMO O crescente número de refugiados e desabrigados no Brasil tem direcionado debates em relação ao direito de moradia e humanização no processo de integração no país. A arquitetura tem um papel fundamental na caracterização social, principalmente na sua capacidade de prover condições de moradia. Aos refugiados, representa mais que uma edificação, mas também o sentimento de pertencer, o reestabelecimento da dignidade individual, além da melhora da qualidade de vida. Assim, este trabalho buscou desenvolver uma metodologia projetual através da revisão bibliográfica, tendo como estratégias os princípios de coordenação modular, associados ao uso dos colmos de bambu, como principal sistema construtivo, para a viabilização da proposta arquitetônica de baixo custo e baixo impacto ambiental – características importantes para a implantação de habitações emergenciais. 102 PALAVRAS CHAVE: Sistemas construtivos sustentáveis; Modularidade; Refugiados; Metodologia de Projeto RAS CHAVE: Sistemas construtivos sustentáveis; Modularidade; Refugiados; Metodologia de PROPOSTA DE SISTEMA MODULAR EM BAMBU PARA HABITAÇÃO EMERGENCIAL PROPOSAL OF BAMBOO MODULAR SYSTEM FOR EMERGENCY HOUSING ARTIGOS ARTIGOS 1. INTRODUÇÃO Apesar da existência de diversos abrigos temporários para apoio aos refugiados, estes encontram-se em sua capacidade máxima, fazendo com que muitos imigrantes vivam como moradores de ruas. Zuker (2019), observa que no início do ano de 2019 cerca de 300 imigrantes venezuelanos viviam ao redor da rodoviária de Manaus, dormindo sob lonas de caminhão, em barracas improvisadas ou em colchões no chão, em situação de extrema vulnerabilidade. O relatório do Alto Comissariado das Nações Unidas para os Refugiados (ACNUR), descreve que o deslocamento forçado afeta uma em cada 97 pessoas. Até o fim de 2019, 79,5 milhões de pessoas em todo o mundo foram forçadas a deixar suas casas (ACNUR, 2019). Os deslocamentos advêm por diversos fatores. Desastres naturais como, secas, inundações e terremotos, tais como o ocorrido na capital de Porto Príncipe no Haiti em 2010. Annoni e Lima Manzi (2016), em seu discurso, disseram que foi o terremoto mais dramático, provando a morte de quase 300 mil pessoas e deixando pelo menos um milhão de desabrigados. Silva (2016), diz que o processo de inserção social é marcado muitas vezes pela desonestidade de empregadores sem escrúpulos, que querem transformá-los em possíveis escravos. Esses eventos causam danos humanos, prejuízos econômicos e sociais, agravando a desigualdade social no país. Outra possível causa de deslocamentos são os desastres causados pela ação do homem, como por exemplo, os rompimentos das barragens ocorrido no estado de Minas Gerais, no município de Mariana em 2015, que causou o maior impacto ambiental da história brasileira e o maior do mundo envolvendo barragens de rejeitos (IBGE, 2017) e no município de Brumadinho em 2019, que deixou 270 pessoas mortas e outras 11 desaparecidas, número que corresponde a quase 1% de toda a população local (CONECTAS, 2020). Uma das formas mais imediatas e básicas de se encaminhar a questão é o provimento de abrigo. A moradia é um marco de estabilidade. Desta forma, o objetivo desta pesquisa é apresentar uma proposta de habitação emergencial em bambu adotando sistema construtivo modular, a fim de racionalizar o processo de construção, possibilitando o acesso a moradia a essas famílias vítimas de desastres e refugiados. Observa-se ainda a ocorrência de crises humanitárias, tais como conflitos armados, colapsos econômicos e políticos, situação enfrentada atualmente por muitos grupos humanos e, em particular na América do Sul, os venezuelanos. 103 ABSTRACT ABSTRACT The elevated number of refugees and homeless people in Brazil has driven to debates regarding the right to housing and humanization in the integration process in the country. Architecture has a fundamental role in social welfare, mainly in its ability to provide housing conditions. For refugees, it represents not only a building but also a feeling of belonging while restoring individual dignity and life quality. Thus, this work sought to develop a design methodology through the literature review, using as strategies the principles of modular coordination, associated with the use of bamboo culms, as the main constructive system, for the feasibility of the architectural proposal of low-cost and low environmental impact - since this material is highly renewable, witch is an important feature for emergency housing. KEY WORDS: Sustainable building systems; Modularity; Refugees; Design Methodology. Mix Sust \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 http://dx.doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 ISSN: 2447-0899 (IMPRESSA) \| 2447-3073 (ONLINE) Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 2.1 Refugiados na Região Norte do Brasil O Brasil foi o primeiro país do Cone Sul a ratificar a Convenção relativa ao Estatuto dos Refugiados de 1951, no ano de 1960 e a realizar um Comitê Executivo do ACNUR (2018), responsável pela aprovação dos programas e orçamentos anuais da agência. Santana (2018) afirma que: O Brasil é reconhecido como país seguro para refugiados, abriga a maior população de refugiados da América do Sul, vinda de 80 diferentes países. (SANTANA, Carmen. v. 34, n. 11, 2018) 104 Segundo Mamed (2016), a região Norte começou a receber os refugiados haitianos vindo das fronteiras para os estados brasileiros do Amazonas e do Acre em 2010, quando a capital de Porto Príncipe no Haiti foi duramente afetada pelo terremoto. A construção de um lar seguro e privativo é o primeiro passo para que uma pessoa tenha sua dignidade restituída. Outra forma de conectar os refugiados com sua moradia é o princípio de Co- design que segundo Mandola (2020) suscita criatividade coletiva através das experiências compartilhadas, logo os usuários podem se tornar parte do grupo nas fases do planejamento do projeto desse abrigo, como também na sua montagem e futuras ampliações, permitindo que essas famílias se conectem ao espaço, gerando o sentimento de pertencimento. A cidade acreana de Assis Brasil, região da tríplice fronteira entre Brasil, Peru e Bolívia, foi o palco do fluxo mais constante e consolidado, sendo a principal porta de entrada destes refugiados no período de 2010 a 2015. Segundo os dados da SEJUDH (2015) em 2012 os imigrantes haitianos somavam 2.225 mil pessoas. Esse número foi acentuado nos anos de 2013 e 2014, quando chegaram até a região cerca de 10.779 mil e 14.333 mil, respectivamente. Shigueru Ban (2013), arquiteto japonês reconhecido pelo seu trabalho com abrigos emergenciais, construções sustentáveis e, principalmente pelo caráter humanitário de seu trabalho, afirma: A tendência de crescimento continuou até meados de 2015, mas desde então os números de entrada diminuíram e o ano fechou com o registro da passagem de 9.315 mil pessoas pela fronteira acreana. Somando-se os imigrantes de outras nacionalidades que chegam pela mesma rota, o número total é de 43.558 mil. Eu me senti muito decepcionado com a minha profissão enquanto arquiteto, porque percebi que não estávamos trabalhando em prol da sociedade. Estávamos trabalhando para pessoas privilegiadas: pessoas ricas, políticos, empresários desenvolvedores. Todos eles possuem dinheiro e poder. 2. REVISÃO BIBLIOGRÁFICA Segundo Anders (2007), é possível definir o conceito de abrigos emergenciais como uma edificação construída de maneira rápida com o objetivo de abrigar pessoas e/ou famílias após a passagem destas por um evento desfavorável – como é a situação dos refugiados a ser discutida. O Cômite Nacional para os Refugiados (CONARE) na 4ª edição de 2019 no relatório “Refúgio em Números”, divulgou que o Brasil reconheceu, apenas em 2018, um total de 1.086 refugiados de diversas nacionalidades. Com isso, o país atinge a marca de 11.231 mil pessoas reconhecidas como refugiadas pelo Estado Brasileiro. Neste caso, as famílias chegam em um país novo com a necessidade de auxílio externo, forçados a improvisar moradias durante ao menos um período inicial de adaptação e reconhecimento de sua existência. Em alguns casos, este abrigo antes de emergência, acaba por se tornar permanente quando não há apoio do Estado. O relatório ainda apontou um aumento exponencial do deslocamento de venezuelanos. No total, foram mais de 80 mil solicitações de refúgio no Brasil, sendo 61.681 desta nacionalidade. Outro número importante na contextualização deste problema foi apontado pela agência da ONU para refugiados, que mostra que os estados com mais solicitações em 2018 foram Roraima com 50.770 mil, Amazonas com 10.500 mil e São Paulo com 9.977 mil. Quarantelli (1991), um dos primeiros autores a discutir definições em relação ao assunto, classifica os abrigos de quatro formas: • Emergency sheltering: ou abrigo emergencial para vítimas de desastres que buscam alojamentos fora da sua residência atual por períodos curtos. Esses imigrantes vêm ao Brasil em busca de refúgio e oportunidade, diante dos eventos em seu país. No entanto, as dificuldades encontradas vão além do processo de inserção sociocultural e dos desafios socioeconômicos enfrentados. • Temporary sheltering: ou abrigo temporário, em que as vítimas buscam alojamento por um período um pouco maior do que o mencionado anteriormente, porém não reestabelecem seus hábitos e rotinas familiares até que retornem a uma nova casa permanente. Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. 2. REVISÃO BIBLIOGRÁFICA Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 • Temporary housing: ou residência temporária, em que a vítima sabe que, apesar de não ser permanente, ficará por muito tempo nesta habitação (meses, anos). Por isso, trata aquele espaço como se já fosse seu lar: estabelece rotina, personaliza a decoração etc. usuário, enquanto ser humano para que seja possível desenvolver um projeto funcional e formalmente aceitável. Deve, portanto, projetar uma edificação que não só abrigue, como também dê a oportunidade de acolhimento e traga integridade ao usuário e em particular a este, que já vive uma situação dramática. • Permanent housing: residência permanente. Trata-se do lugar onde há o sentido de lar, de sentimento, de identificação do indivíduo e sua personalidade dentro daquele espaço. Sendo assim, para uma melhor abordagem, este trabalho foi dividido em dois principais subtópicos, visando a apresentação do público alvo da proposta projetual apresentada e então, a discussão dos conceitos iniciais para desenvolvimento do estudo de projeto arquitetônico. Estes conceitos são importantes para que seja compreendida a tipologia a ser projetada, visto que à medida em que período de utilização deste espaço aumenta, é preciso que a construção ofereça um caráter crescente de robustez. Uma habitação pode ser classificada em mais de uma das categorias apresentadas. A sua classificação está relacionada ao tempo em que o usuário vai necessitar deste abrigo e na projeção de durabilidade desta edificação, o que já deve estar previsto anteriormente ao projeto. Desta forma, visto que a intenção é a de que o estudo aqui apresentado não se torne a residência permanente deste público, estão sendo utilizados para esta proposta a ideia de um Temporary sheltering e, caso haja uma necessidade real, poderá até se tornar um Temporary housing caso o tempo de uso precise ser um pouco estendido. Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 2.1.1 Características arquitetônicas da Região A região Amazônica em sua morfologia natural, possui grande presença de rios e igarapés. Com a grande quantidade de água corrente na cidade de Manaus a tornando-se conhecida na década de 1920, como a Veneza dos Trópicos. Nogueira (2016), desenvolveu na Amazônia um estudo para conhecer o modo de vida do povo que habita as margens dos rios da região, mais conhecidos como ‘ribeirinhos’. A arquitetura vernacular ribeirinha corresponde a construção que está relacionada aos costumes e tradições de um lugar, bem como sua economia e os materiais tradicionais encontrados na região. Figura 2. Casa flutuante em 1960 à margem do Rio Amazonas - AM. Fonte: IBGE, 2021. A pequena casa de madeira, com boa inércia térmica, simétrica, com o pé-direito alto, telhado de duas águas, varanda na frente e ventilação na parte inferior na casa, aparentemente isolada em meio a floresta, característica correspondente ao modo de vida distinta de subsistência de seus habitantes. Esses são os aspectos correspondentes aos sistemas constitutivos na cultura da região Amazônica, segundo Nogueira (2016). De forma mais poética, a maneira ribeirinha de habitar a terra pode ser entendida como o encontro entre homem e natureza, revelado a partir de sua materialização na paisagem na forma das casas que estes constroem e na maneira em como constroem materialmente e simbolicamente o lugar no qual habitam, construindo uma cultura material a partir de uma relação simbiótica com o ambiente que os cercam. 2.1 Refugiados na Região Norte do Brasil (BAN, 2013) Silva (2016) diz que a cidade de Manaus até maio de 2014 havia passado pela capital amazonense mais de 8 mil haitianos, dos quais mais de mil continuam na cidade. Em 2017 a região Norte passou por mais dificuldades para atender o alto Logo, é papel e responsabilidade do arquiteto: propor projeto bem desenvolvido, associado à humanização no caráter de tomadas de decisões sobre o usuário a ser atendido; pensar no Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 fluxo migratório de venezuelanos. A imigração tem implicações econômicas, sociais e culturais, tanto no local de partida quanto no local de chegada ou de passagem. construtiva é o resultado do conjunto de fenômenos de seca e cheia dos rios, atrelado ao período de chuvas da região. Se adaptando muito bem ao clima da região, exemplificadas nas figuras 1 e 2. Segundo a CONARE (2019), até dezembro de 2018, a região Norte recebeu cerca de 85.438 mil solicitações de reconhecimento de condição de refugiados da Venezuela. Destas, 81% das solicitações foram apresentadas no estado de Roraima que faz fronteira com a Venezuela. Figura 1. Casa palafita em 1958 perto da Avenida Silves em Manaus – AM. Fonte: IBGE, 2021. Em agosto 2018 o Estado de Roraima, encontrava-se com oito abrigos temporários em funcionamento e dois abrigos transitórios. Nesse período, Boa Vista estava com aproximadamente 5.000 pessoas abrigadas. Além destes locais, estava sendo finalizado o abrigo transitório para servir de espaço de passagem. Nesse período os venezuelanos encontravam-se em centros coletivos com ginásios, galpões e em acampamentos planejados (CARBONARI e LIBRELOTTO, 2020). Dados do ACNUR (2018) mostraram que em agosto de 2018 esses abrigos estavam a capacidade superior à planejada, com um total de 754 residentes, sendo 400 do sexo masculino e 354 do sexo feminino. Logo, estes números mostram-se altos e difíceis de serem ignorados pela população e poder público. Figura 1. Casa palafita em 1958 perto da Avenida Silves em Manaus – AM. Fonte: IBGE, 2021. Figura 2. Casa flutuante em 1960 à margem do Rio Amazonas - AM. Fonte: IBGE, 2021. 105 2.1.1 Características arquitetônicas dos abrigos na região Logo, as palafitas e flutuantes são construções usadas a décadas na região Amazônica. A técnica O modo de vida dos imigrantes venezuelanos difere do modo de vida da população da região Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Segundo Pasello (2020), a existência de princípios modulares que orientam estrategicamente o processo de projeto de um determinado produto trazem grandes benefícios, como a flexibilidade em adaptar-se com o mínimo de esforço, de maneira mais econômica possível. Amazônica. No entanto, o público encontrado nos abrigos da região é em grande parte de indígenas venezuelanos. A população indígena vinda da Venezuela mostrou particularidades relacionadas principalmente com suas atividades do cotidiano, como o modo de comer, dormir e vestir, que lembra ao modo de habitar a terra dos ribeirinhos. Gosling et al. (2016) afirmam que as abordagens modulares podem ajudar a resolver problemas persistentes, a volatilidade do fluxo de trabalho, pontualidade e alterações. Segundo Carbonari e Librelotto (2020), os abrigos que recebiam os imigrantes venezuelanos na região Amazônica, careciam de algumas infraestruturas, como instalações sanitárias próximos ao abrigo, com melhor iluminação e visibilidade, áreas verdes com arborização para projetar sombra, necessidade básica em locais de clima quente como a região Amazônica, espaços para o preparo das refeições, pois muitos dos imigrantes não se adaptaram com as marmitas fornecidas pelo exército brasileiro e área de redário, que caracteriza o modo de dormir dos indígenas venezuelanos. No Projeto Chacras (2020), localizado no Equador, o conceito de modularidade foi aplicado através da proposta de módulo de dimensões similares aos do pallet de madeira, de aproximadamente 1,20m X 1,20m. Através deste elemento, a fachada e as medidas do edifício tornaram possível a construção da moradia, conforme fica evidenciado na figura 3 (NAKASATO, 2020). Figura 3. Fachada e Área externa Projeto Chacras. Fonte: Nakasato, 2020. 2.1.1 Características arquitetônicas dos abrigos na região Outra particularidade encontrada nos abrigos da região norte do Brasil, estão relacionados aos meios de subsistência dos imigrantes, pois grande parte dos homens eram trabalhadores agrícolas e grande parte das mulheres artesãs, deste modo, houve a necessidade de desenvolver um espaço específico para as mulheres realizarem a venda de seus artesanatos e uma horta comunitária para que os homens pudessem desenvolver suas atividades agrícolas. 106 Os refugiados passam por vários obstáculos na sua integração no país. Diferenças culturais, étnicas e econômicas, dificuldades com o idioma, perda de relações familiares e sociais, restrições ao reconhecimento da formação acadêmica, violência relacionada às circunstâncias que forçaram o deslocamento. Adicionalmente, sofrem problemas sociais que afetam os brasileiros: dificuldades de inserção profissional, no acesso à educação superior, no acesso à moradia e à saúde (SANTANA, 2018). Figura 3. Fachada e Área externa Projeto Chacras. Fonte: Nakasato, 2020. É possível pois, se manter um padrão nas dimensões da edificação e combinar estes elementos facilitando tanto a sua construção quanto a compreensão de sua lógica construtiva pelos próprios usuários do espaço. Isso permite que os mesmos tenham a capacidade de construir o espaço no qual vão morar, numa dinâmica de transferência do conhecimento profissional-usuário. Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 2.2.2 Bambu O bambu é um recurso natural renovável, de fácil manuseio, versatilidade e que apresenta grande resistência mecânica e flexibilidade. Além de sustentável, permite a instalação seca, podendo ser montado in loco, sem geração de resíduos e com rápida e efetiva execução (CARBONARI, 2017). Benavides e Ceolin (2015) afirmam que vários são os fatores que interferem na redução do preço final de um projeto arquitetônico, entre eles estão o aproveitamento dos materiais, de modo a não gerar desperdícios, facilitando a montagem e consequentemente reduzindo o tempo no canteiro de obras. Os princípios de coordenação modular podem facilitar o processo de construção, através da pré-fabricação, permitindo rapidez na montagem e substituição de elementos nos processos de manutenção. A ABNT NBR 16.828 (2020) é a primeira legislação brasileira que regulamenta os projetos de estruturas em bambu. Segundo a norma, o amadurecimento dos colmos de bambu ocorre Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 geralmente com idade entre 3 anos e 7 anos, dependendo da espécie, sendo considerado uma planta de rápido crescimento. bambu é fortemente empregado desde os tempos antigos por sua força natural e flexibilidade e hoje em dia seu emprego na construção combina saberes populares com técnicas contemporâneas, resultando em edificações que respondem às mais diversas necessidades (BENAVIDES, 2019). Seus colmos são tubulares com diâmetros e espessuras de paredes variáveis e, na própria peça os diâmetros das suas extremidades são diferentes entre si. Necessitando e uma caracterização para avaliar as variações de diâmetro, as irregularidades e as espessuras dos colmos, que podem ser relativamente finas, o que ocasionaria fragilidade nas ligações. Figura 4. Esquema de conexão da Casa de bambu em Parque Estadual de Niterói – RJ. Fonte: Adaptada do https://sustentarqui.com.br/uma-casa-de-bambu-em- parque-estadual. Apesar da pouca exploração do material no Brasil, a partir da criação da chamada Lei do bambu - Lei nº 12.484/2011, que institui a Política Nacional de Incentivo ao Manejo Sustentado e ao Cultivo do Bambu (PNMCB), se iniciou um grande estímulo ao cultivo e à comercialização deste produto não apenas na indústria têxtil e de papel, e ainda na construção civil. 2.2.2 Bambu O governo federal procura, com a lei, incentivar a transformação dos bambus brasileiros em cultivo capaz de gerar emprego, renda e até créditos de carbono (DRUMOND E WIEDMAN, 2017). Figura 4. Esquema de conexão da Casa de bambu em Parque Estadual de Niterói – RJ. Fonte: Adaptada do https://sustentarqui.com.br/uma-casa-de-bambu-em- parque-estadual. Figura 4. Esquema de conexão da Casa de bambu em Parque Estadual de Niterói – RJ. Fonte: Adaptada do https://sustentarqui.com.br/uma-casa-de-bambu-em- parque-estadual. Com a publicação da ABNT NBR 16.828 (2020) de bambu, é possível afirmar que o bambu já é um material estrutural de importância reconhecida tanto à nível nacional, quanto internacionalmente. Benavides e Ceolin (2015), propôs uma habitação de interesse social com 36m², a partir de estruturas e vedações feitas em bambu, aplicando os princípios de coordenação modular. Apesar do bambu possui desvantagens a serem consideradas, a compreensão a respeito do material foi essencial para a organização espaços propostos. 4. RESULTADOS Neu et al. (2016) afirmam que a região Amazônica sofre grande contaminação biológica das águas, pela precariedade das infraestruturas de saneamento básico, principalmente em áreas rurais e ribeirinhas, áreas sujeitas a alagamento. A proposta de sistema modular em bambu para habitação emergencial a ser apresentada, é uma simulação do exercício de projeto que contempla todo o referencial bibliográfico abordado anteriormente. Para propor uma solução mais adequada à realidade, foram avaliadas as condicionantes climáticas, as características arquitetônicas vernaculares da região Amazônica, as características culturais dos imigrantes indígenas venezuelanos, assim como, as soluções construtivas e estruturais, impostas pelos materiais construtivos utilizados. Os autores sugerem o uso de banheiro ecológico ribeirinho (BER), onde os dejetos humanos são isolados em recipiente impermeável (bombona) e são misturados a cascas, papeis, folhas e serragens, convertendo-se lentamente em adubo, através do processo de compostagem. Figueiredo et al. (2019) falam que a bacia de evapotranspiração é uma das possíveis soluções para tratamento de esgoto em comunidades rurais e em áreas isoladas. Esse sistema conhecido popularmente como fossa de bananeiras, se destaca por aliar plantas com grande potencial de evapotranspiração ao tratamento biológico realizado por bactérias anaeróbias produzindo pouca quantidade de lodo e gerando pouco ou nenhum efluente final. 4.1. Da Modularidade Aplicada Na busca por funcionalidade e adequação a um sistema construtivo rápido, o projeto buscou desenvolver uma proposta de habitação emergencial a partir da matéria-prima principal do Bambu, empregando ao conceito da modularidade. Através desta estratégia de projeto foi possível desenvolver módulos que podem ser combinados entre si para construção de tipologias diferentes. Segundo os autores, o sistema faz com que a água percole do fundo para o topo lentamente, atravessando camadas de entulho, brita, areia e solo, até alcançar a zona de raízes. Os patógenos ficam presos no sistema, uma vez que a água não extravasa. Dessa forma, é possível realizar o aproveitamento da água e de grande parte dos nutrientes contidos no esgoto doméstico pelas plantas do sistema, sem necessidade de realizar pós-tratamento. Para a concepção desta proposta, buscou-se um elemento comum que fosse capaz de determinar toda a modulação do projeto baseado na literatura consultada. A ideia partiu do interesse de se obter um elemento específico que fizesse parte da base e/ou piso da habitação permitindo modificação da planta conforme as necessidades de cada família. Optou-se então pela utilização do conceito de modularidade a partir das dimensões das placas de Oriented Strand Board (OSB) com dimensões de 122cm x 244cm, utilizadas no piso, como mostra a figura 5. Com as dimensões da mesma, foi possível determinar a planta baixa e a área necessária para cada módulo da edificação (Figura 6). 108 2.2.3 Pneus reutilizáveis dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 2.2.3 Pneus reutilizáveis 107 Floriani, Funaretto e Sehnem (2016), afirmam que a partir dos anos de 1990 com o aumento da fabricação e utilização de automóveis inicia-se o grande acúmulo de pneus descartados e, consequentemente, um grande impacto ambiental visto que após o seu desgaste completo ele não possui mais nenhuma utilidade. Além de suas características anatômicas, o bambu é suscetível à degradação e a ataques de insetos xilófagos (caruncho), necessitando de tratamento adequado para protegê-lo, preferencialmente, de menor impacto ambiental, assim como o uso de elementos arquitetônicos que evitem os efeitos do clima sob o material. Raiol et al. (2019) realizaram o levantamento da geração de pneus inservíveis na cidade de Capanema, na Amazônia Oriental no estado do Pará em 2018. Segundo os questionários recebidos das empresas da região, cerca de 50 pneus eram descartados mensalmente, onde muitos deles eram encaminhados ao lixão municipal. Rashmi Manandhar et al. (2019) fala sobre os avanços no tratamento de pragas e fungos, aumentando a vida útil do material até 30-40 anos permitindo o emprego do bambu no desenvolvimento de belas e ousadas arquiteturas, assim como, conexões inovadoras, como mostra a figura 4. Os autores ressaltam a necessidade da ampliação de parcerias entre empresas do ramo de pneus com os demais órgãos ambientais, além de incentivos para a criação de associações que permitam a reciclagem ou o reaproveitamento desses materiais. Os autores ainda apontam alguns tipos de tratamentos preservativos a serem realizados e ressaltam a importância de regulamentações rigorosas sobre o uso de produtos químicos no tratamento do bambu. Com a disseminação dos perigos e precauções sobre o uso, o manuseio e o descarte, a fim de minimizar esses perigos. Segundo o Serviço Social do Transporte (SESNAT, 2021), 450 mil toneladas de pneus são descartadas todos os anos no Brasil. Considerando que os mesmos levam em média 600 anos para se decompor é possível imaginar o tamanho do impacto ambiental negativo que este produto causa no país. No âmbito social, a população mundial está familiarizada com o bambu. Em países asiáticos o Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 3. METODOLOGIA E ESTRATÉGIAS Como ponto de partida para a elaboração da proposta de sistema modular em bambu para habitação emergencial, foram estabelecidas algumas diretrizes, tais como, público alvo e região para a implantação do abrigo, assim como o estudo das características das mesmas. Figura 5. Configuração do piso em OSB. Fonte: elaborada pelos autores. Para o processo de concepção arquitetônica, a principal estratégia utilizada foi a modularidade e a sustentabilidade. A partir da aplicação deste conceito na elaboração da planta baixa, todas as demais decisões arquitetônicas foram sendo tomadas. Optou-se pela utilização do bambu, placas de Oriented Strand Board (OSB) e pneus reutilizáveis como principais materiais construtivos para a habitação proposta, conforme o referencial teórico. Figura 5. Configuração do piso em OSB. Fonte: elaborada pelos autores. Finalmente, foi desenvolvido a modulação das estruturas, painéis de vedação e instalações propostas, de forma sintetizada para que possam serem utilizadas em propostas futuras. De forma esquemática, a figura 6 mostra as possibilidades de combinações pretendidas com a criação do módulo base, conferindo flexibilidade Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 bu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 (Figura 9). O segundo Kit acrescenta um quarto extra, estruturado como um mezanino e acesso por uma escada marinheiro, desta forma permitindo atender uma família de até 6 pessoas (Figura 10). Figura 8. Kit de fechamento dos módulos. Fonte: elaborada pelos autores. Figura 9. Planta baixa da configuração térrea. Fonte: elaborada pelos autores. Figura 10. Planta baixa mostrando a adição de mezanino. Fonte: elaborada pelos autores. Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 (Figura 9). O segundo Kit acrescenta um quarto extra, estruturado como um mezanino e acesso por uma escada marinheiro, desta forma permitindo atender uma família de até 6 pessoas (Figura 10). projetual a habitação emergencial. Além disso, o módulo permite que sejam feitas ampliações, extensões e o acoplamento da estrutura em diferentes níveis de topografias, através da adaptação da altura da cobertura. Figura 8. Kit de fechamento dos módulos. Fonte: elaborada pelos autores. Figura 8. Kit de fechamento dos módulos. Fonte: elaborada pelos autores. Figura 6. Modelo esquemático de composições modulares. Fonte: elaborada pelos autores. Figura 6. Modelo esquemático de composições modulares. Fonte: elaborada pelos autores. Figura 8. 3. METODOLOGIA E ESTRATÉGIAS Kit de fechamento dos módulos. Fonte: elaborada pelos autores. Figura 9. Planta baixa da configuração térrea. Fonte: elaborada pelos autores. Figura 10. Planta baixa mostrando a adição de mezanino. Fonte: elaborada pelos autores. Figura 8. Kit de fechamento dos módulos. Fonte: elaborada pelos autores. Figura 6. Modelo esquemático de composições modulares. Fonte: elaborada pelos autores. As placas de OSB de 15mm nesta proposta estão instaladas sobre uma base de vigas simples de bambu, por sua vez, fixadas por parafusos em vigas duplas do mesmo material, conectadas com as fundações feitas com dois pneus reutilizados, preenchidos com terra compactada e no centro com concreto, buscando rigidez em todo o subsistema de piso, exemplificado no esquema da figura 7. Figura 8. Kit de fechamento dos módulos. Fonte: elaborada pelos autores. Figura 8. Kit de fechamento dos módulos. Fonte: elaborada pelos autores. Figura 9. Planta baixa da configuração térrea. Fonte: elaborada pelos autores. Figura 7. Corte esquemático da fundação. Fonte: elaborada pelos autores. Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 4.2. Da Humanização Aplicada Baseado na reflexão de Shigueru Ban (2013) mencionada anteriormente, é preciso não apenas entender o usuário a ser atendido, bem como pensar nas habitações no sentido de restaurar o conceito de “Lar” àqueles que já o perderam, conforme observa-se na figura 13 a intenção da apropriação dos espaços de varandas criadas através das esquadrias. Criam-se então janelas venezianas feitas em bambu e esteira de cana de bambu, que permitem a iluminação e a ventilação natural controlada dos espaços internos. A função do lar não é referente apenas ao espaço de abrigo, mas também se relaciona as emoções e vivências de “estar se sentindo em casa. Desta forma entende-se que o ambiente construído influencia no comportamento e nas emoções, principalmente em crianças e adolescentes. A Corte esquemático do abrigo de 25m². Fonte: elaborada pelos autores. As janelas tipo basculante, concedem a interação proposta pela humanização, servindo também como extensão do ambiente, projetando uma varanda coberta, característica relevante para os refugiados indígenas. É possível que este espaço se caracterize como uma área de permanência, possibilitando a apreciação do entorno e o próprio acesso a edificação, principalmente pelas crianças (Figura 14). Figura 14. Perspectiva interna do quarto. Fonte: elaborada pelos autores. 110 Figura 12. Corte esquemático do abrigo de 25m². Fonte: elaborada pelos autores. O objetivo destas duas propostas de abrigo é mostrar as possibilidades de combinação do conjunto em relação às necessidades de seus usuários, além de exemplificar como seria uma possível adaptação/ampliação do abrigo. Figura 14. Perspectiva interna do quarto. Fonte: elaborada pelos autores. Outra ferramenta utilizada com o objetivo de humanização é a iluminação e ventilação natural. Sendo assim, as aberturas aqui apresentadas foram desenvolvidas para o melhor aproveitamento da ventilação, iluminação e características naturais da região Amazônica. Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 109 Figura 7. Corte esquemático da fundação. Fonte: elaborada pelos autores. Figura 9. Planta baixa da configuração térrea. Fonte: elaborada pelos autores. Figura 9. Planta baixa da configuração térrea. Fonte: elaborada pelos autores. A solução proposta para a fundação, conforme apresentado na figura 5, além de possuir baixo custo, evitar a exposição do material vegetal com o solo (as vigas de bambu neste caso) garantindo uma maior durabilidade e resistência, remete a característica regional de palafitas, usadas na região norte do Brasil. Figura 10. Planta baixa mostrando a adição de mezanino. Fonte: elaborada pelos autores. Na proposta também se observa a modularidade na sistematização no processo de montagem dos painéis de fechamento, com base no modelo esquemático de composições modulares da figura 5. A partir deste esquema, foi desenvolvido um sistema básico, nomeado nesse projeto de Kit, como mostra a figura 8. Figura 10. Planta baixa mostrando a adição de mezanino. Fonte: elaborada pelos autores. Essa solução permitiu a combinação de módulos formados pelo dimensionamento das chapas, admitindo ao projeto flexibilidade no momento da construção, segundo a necessidade de cada O primeiro Kit configura uma habitação, do modelo mais básico, tem 20m² de área interna e possui uma configuração térrea, sem mezaninos e abriga uma família de no máximo 4 pessoas Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 família, como mostra os cortes esquemáticos dos abrigos nas figuras 11 e 12. família, como mostra os cortes esquemáticos dos abrigos nas figuras 11 e 12. partir disso, o projeto buscar estimular o sentimento de pertencimento através da apropriação do espaço com a utilização de aberturas, que permitem interações sociais e futuras ampliações (Figura 13). abrigos nas figuras 11 e 12. Figura 11. Corte esquemático do abrigo de 20m². Fonte: elaborada pelos autores. Figura 12. Corte esquemático do abrigo de 25m². Fonte: elaborada pelos autores. Figura 11. Corte esquemático do abrigo de 20m². Fonte: elaborada pelos autores. Figura 13. Perspectiva da janela tipo basculante aberta. Fonte: elaborada pelos autores. Figura 11. Corte esquemático do abrigo de 20m². Fonte: elaborada pelos autores. Figura 13. Perspectiva da janela tipo basculante aberta. Fonte: elaborada pelos autores. Figura 12. Corte esquemático do abrigo de 25m². Fonte: elaborada pelos autores. Figura 12. 4.3.1 Implantação Para a proposta modular de habitação emergencial, sugere-se uma implantação com grupos de 4 unidades, conectadas através de passarelas externas. Acredita-se que a implantação proposta na figura 15 estimula o sentimento de comunidade entre as diferentes famílias e proporciona a vigilância natural de uma área determinada. Figura 16. Visão interna da passarela de conexão. Fonte: elaborada pelos autores. No entanto há outras possibilidades decorrentes ao uso da modularidade enquanto estratégia de projeto, trazendo flexibilidade na planta baixa e na cobertura. Os módulos podem ser inseridos em lotes padrões de quadras com ruas, dispostos um do lado do outro ou ainda com parede germinada, em lotes com aclives, decliveis, conforme a necessidade e a disponibilidade de terreno a ser cedido. Figura 15. Sugestão de implantação. Fonte: elaborada pelos autores. 4.3. Da Composição Arquitetônica Além das características discutidas anteriormente sobre a modularidade e a humanização aplicada ao projeto, algumas Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Figura 16. Visão interna da passarela de conexão. Fonte: elaborada pelos autores. definições, características físicas e qualitativas não foram abordadas nas soluções apresentadas. A seguir essas definições serão apresentadas separadamente cada qual com seu conjunto de elementos e justificativas para um melhor entendimento do processo construtivo da habitação emergencial proposta. 4.3.4 Instalações O projeto adotou um módulo sanitário com sistema de banheiro seco, comum ao conjunto de 4 unidades, conectadas através das passarelas externas. O módulo foi instalado intencionalmente afastado das demais edificações para evitar que qualquer problema de mau cheiro seja contido antes de alcançar as habitações. 4.3.2 Cobertura Como foi dito anteriormente a proposta modular de habitação emergencial contempla as características regionais da região Amazônica. Essa região possui dois períodos que marcam as condições climáticas, com chuvas abundantes nos meses de dezembro a fevereiro, onde a temperatura média varia de 24°C a 30°C e o período de grande calor, nos meses de setembro a outubro onde a temperatura média varia de 25°C a 33°C. 111 Figura 15. Sugestão de implantação. Fonte: elaborada pelos autores. Além disso, a proposta de implantação permite a criação de espaços para atividades recreativas, descanso, interação social e horta comunitária sob a vigilância das famílias. Apesar de não ter sido incialmente previsto, as passarelas de conexão poderão ser cobertas para que não haja problemas com as intempéries agindo sobre as placas de OSB e as estruturas de bambu (Figura 16). Sendo assim, o projeto buscou soluções para minimizar os efeitos climáticos na estrutura de bambu, com a utilização de beirais longos e testeiras do telhado para a proteção da estrutura de bambu, evidenciado de forma mais clara nas figuras 17 e 21. Figura 17. Perspectiva das coberturas. Fonte: elaborada pelos autores. Figura 17. Perspectiva das coberturas. Fonte: elaborada pelos autores. Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Abaixo da estrutura da cobertura, foram projetadas aberturas para ventilação natural inspiradas na ideia de brise-soleil permitindo a ventilação cruzada dentro da edificação, com bambu e cana de bambu. Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 4.3.3 Mobiliário Os pallets de madeira são materiais facilmente disponíveis no mercado e possuem um custo relativamente baixo. Além disso, os pallets são fáceis de montar e podem ser adaptados para diversos tipos de uso, permitindo o aumento do seu ciclo de vida. No processo de compostagem com sistema de banheiro seco é importante que o tambor não seja totalmente vedado, logo faz-se necessário a instalação de um “suspiro” na tampa para a saída de gases e a aeração para que o processo da decomposição ocorra. (Figura 20). Como sugestão para a composição do mobiliário da habitação emergencial, foi utilizado pallet de madeira de 120cm x 120cm (Figura 18). Figura 20. Perspectiva do módulo sanitário. Fonte: elaborada pelos autores. Figura 18. Proposta de mobiliário. Fonte: elaborada pelos autores. 112 Figura 18. Proposta de mobiliário. Fonte: elaborada pelos autores. Figura 20. Perspectiva do módulo sanitário. Fonte: elaborada pelos autores. Conforme o número de módulos utilizados nas habitações a quantidade necessária de móveis pode ser reconfigurada para atender a todos as possíveis combinações de modulações, conforme ilustrado na figura 5. A perspectiva interna da figura 19 sugere uma possível disposição do mobiliário nos módulos das habitações. Para as águas residuais do banho e lavatório, foi empregado o sistema de tratamento com bacia de evapotranspiração, conhecida popularmente como fossa de bananeiras. Neste sistema, o retorno da água para a natureza se dá através do vapor da transpiração das folhas de bananeira, evitando a contaminação do solo e do lençol freático. Figura 19. Perspectiva interna da sala e cozinha. Fonte: elaborada pelos autores. Além das instalações de água e esgoto, o projeto prevê o reaproveitamento da água das chuvas com a instalação de mini cisternas em cada unidade, onde a água coletada pela calha será armazenada em bombonas, podendo ser reaproveitada para diversas funções exceto o consumo, evidenciadas nos cortes esquemáticos das figuras 11 e 12. A figura 21 apresenta a visão geral externa da proposta arquitetônica. Figura 19. Perspectiva interna da sala e cozinha. Fonte: elaborada pelos autores. Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Figura 21. Visão geral externa da proposta arquitetônica. Fonte: elaborada pelos autores. Figura 21. Visão geral externa da proposta arquitetônica. Fonte: elaborada pelos autores. 4. CONSIDERAÇÕES FINAIS construtivo (pelo menos no plano Nacional), atreladas ao processo de humanização e socialização do refugiado e estrangeiro. Levando em consideração o cenário enfrentado por vários grupos refugiados no Brasil e o número crescente de desabrigados, considera-se suma importância a reflexão sobre a questão de seu acolhimento e integração, sendo que a moradia, é uma das primícias a ser abordada. REFERÊNCIAS Disponível em: <https://repositorio.ufsc.br/handle/123456789/204499> IBGE \| Biblioteca \| Detalhes \| Casa palafita perto da Avenida Silves em Manaus (AM). Disponível em: <https://biblioteca.ibge.gov.br/ biblioteca-catalogo.html?id=47005&view=detalhes>. Acesso em: 25 de fevereiro de 2021. IBGE \| Biblioteca \| Detalhes \| Casa palafita perto da Avenida Silves em Manaus (AM). Disponível em: <https://biblioteca.ibge.gov.br/ biblioteca-catalogo.html?id=47005&view=detalhes>. Acesso em: 25 de fevereiro de 2021. CARBONARI, L. T., & LIBRELOTTO, L. I. (2020). Abrigo Temporário para Refugiados Venezuelanos Indígenas. ENSUS, Revista Gestão e Sustentabilidade Ambiental, Florianópolis, v. 9, n. esp, p. 372-391, jul. 2020. IBGE \| Biblioteca \| Detalhes \| Casa flutuante à margem do Rio Amazonas (AM). Disponível em: <https://biblioteca.ibge.gov.br/index.php/biblio teca-catalogo?view=detalhes&id=47554>. Acesso em: 25 de fevereiro de 2021. CARBONARI, G. et al. Bambu – O Aço Vegetal. MIX Sustentável, v. 3, n. 1, p. 17, 2017. DOI: https://doi.org/10.29183/2447- 3073.MIX2017.v3.n1.17-25. MAMED, L. H. (2016). Haitianos na Amazônia: a morfologia da imigração haitiana pelo Acre e o horizonte de inserção precarizada no Brasil. v. 10 n. 1 (2016). RURIS - Revista Do Centro De Estudos Rurais - UNICAMP, 10(1). Disponível em: <https://www.ifch.unicamp.br/ojs/index.php/ruris/article/v iew/2636>. Acesso em 25 de fevereiro de 2021. MAMED, L. H. (2016). Haitianos na Amazônia: a morfologia da imigração haitiana pelo Acre e o horizonte de inserção precarizada no Brasil. v. 10 n. 1 (2016). RURIS - Revista Do Centro De Estudos Rurais - UNICAMP, 10(1). Disponível em: <https://www.ifch.unicamp.br/ojs/index.php/ruris/article/v iew/2636>. Acesso em 25 de fevereiro de 2021. Dados sobre refúgio no Brasil - ACNUR Brasil. Disponível em: <https://www.acnur.org/portugues/dados- sobre-refugio/dados-sobre-refugio-no-brasil/>. Acesso em: 25 de fevereiro de 2021. MANDOLA, Juliana Bambini; IMAI, Cesar. Instrumentos Utilizados no Processo de Co- Design. In: ENCONTRO NACIONAL DE TECNOLOGIA DO AMBIENTE CONSTRUÍDO, 18., 2020, Porto Alegre. Anais... Porto Alegre: ANTAC, 2020. DADOS: Os Números Da Tragédia De Brumadinho. A Conectas selecionou alguns dados que sobre a maior tragédia socioambiental do país. Disponível em: <https://www.conectas.org/ noticias/fact-sheets-o-numeros-da-tragedia-de-brumadinho>. Acesso em: 03 de março de 2021. NAKASATO, M. S. Abrigos temporários: Recomendações de projetos de sistemas construtivos em madeira no Paraná. [s.l.] Universidade Estadual de Maringá, 2020. DRUMOND, Patrícia Maria, WIEDMAN, Guilherme. Bambus no Brasil: da biologia à tecnologia / organização Patrícia Maria Drumond, Guilherme Wiedman. - 1. ed. - Rio de Janeiro: ICH, 2017. 655 p. Disponível em: <http://www.infoteca. cnptia.embrapa.br/infoteca/handle/doc/1078373> NEU, V.; SOUZA Dos Santos, M. A.; FERRAZ Meyer, L. F. Banheiro ecológico ribeirinho: saneamento descentralizado para comunidades de várzea na Amazônia. Revista Em Extensão, v. 15, n. 1, p. 28-44, 25 jul. 2016. DOI: https://doi.org/10.14393/REE-v15n12016_art02. REFERÊNCIAS ANDERS, G. C. Abrigos Temporários de Caráter Emergencial. [s.l.] Universidade de São Paulo, 2007 Disponível em: <https://teses.usp.br/teses/disponiveis/16/16134/tde- 19092007-102644/publico/Dissertacao.pdf.> A proposta de habitação emergencial em sistema construtivo modular apresentada neste artigo, ainda que baseada nas indicações da literatura, é um protótipo virtual de caráter preliminar e, por conseguinte, incidindo numa série de incertezas. 113 <https://teses.usp.br/teses/disponiveis/16/16134/tde- 19092007-102644/publico/Dissertacao.pdf.> ANNONI, Danielle; LIMA MANZI, Maria Júlia. Política migratória Brasileira e seus reflexos para os estados da UNASUL: Um estudo a partir do tratamento dado pelo Brasil ao caso dos haitianos \| Boletín Mexicano de Derecho Comparado (elsevier.es). DOI: 10.22201/iij.24484873e.2016.146.10506. O mesmo apresenta características e elementos construtivos/estruturais que são passiveis de revisão, reprojeto e redesenho de acordo com as constatações que a construção de um protótipo físico poderia fornecer. Sobretudo em relação ao desempenho estrutural do bambu e as decisões estruturais tomadas, bem como o as soluções construtivas adotadas para o enfrentamento das condições climáticas regionais, paira ainda uma grande interrogação. ASSOCIAÇÃO BRASILEIRA DE NORMAS TÉCNICAS. NBR 16.828-1: Estruturas de Bambu – Parte 1: Projeto. 2020. BAN, S. Emergency shelters made from paper. 2013. Disponível em: https://www.ted.com/talks/shigeru_ban_emergency_shelter s_made_from_paper#t-686748. Acesso em: 02 de março de 2021. Além destes aspectos, não se sabe como o usuário (alvo) se comportará na utilização do edifício, sendo necessário a utilização de uma análise de pós-ocupação que poderia fechar um ciclo mais abrangente de desempenho e aceitação de tal proposta. BENAVIDES, Andrea Salomé Jaramillo; CEOLIN, Eliká Deboni. (2015). Coordenação Modular aplicada à habitação de interesse social com estrutura de bambu Guadua roliço. Disponível em: 2015_IMED_Passo_Fundo_Coordenacao_Modular_ aplicado_a_Habitacao_de_Interesse_Social-with-cover-page- v2.pdf (d1wqtxts1xzle7.cloudfront.net) Apesar da complexidade da questão, e da multiplicidade de aspectos a serem tratados, acredita-se que a proposta possa auxiliar no desenvolvimento de projetos e estudos de caráter semelhante. BENAVIDES, Andrea Salomé Jaramillo. (2019). Manifestações Patológicas e Decisões Projetuais que incidem na durabilidade do Bambu em Edificações no Sul do Brasil / Andrea Aponta para novas perspectivas formais e construtivas no uso do bambu como material Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Salomé Jaramillo Benavides; orientador, Lisiane Librelotto, coorientador, Ângela do Valle, 2019. 282 p. AGRADECIMENTOS Ao Programa de Pós-graduação em Metodologia de Projeto da Universidade Estadual de Londrina, em especial à disciplina de Sistemas Construtivos. RASHMI Manandhar, JIN-HEE Kim & JUN-TAE Kim (2019) Environmental, social and economic sustainability of bamboo and bamboo-based construction materials in buildings, Journal of Asian Architecture and Building Engineering, 18:2, 49-59, DOI: 10.1080/13467581 .2019.1595629. AUTORES ORCID: https: 0000-0001-8175-9091 ANDRESSA MIKHAELLA DOS SANTOS BRITO OLIVEIRA (AMSBO) \| Universidade Estadual de Londrina \| Arquitetura e Urbanismo \| Londrina, PR - Brasil \| Correspondência para: Rua Leonora Armstrong, 55 - Champagnat, Londrina - PR, 86062-520 \| e-mail: andressa.oliveira@uel.br SANTANA, Carmen. Humildade cultural: conceito estratégico para abordar a saúde dos refugiados no Brasil. Cad. Saúde Pública, Rio de Janeiro, v. 34, n. 11, e00098818, maio 2018. DOI: https://doi.org/10.1590/0102-311X00098818. SAVI, A. E. Contribuições da Arquitetura sobre o acolhimento de crianças e adolescentes em abrigos institucionais. Tese (Doutorado em Arquitetura e Urbanismo) – Universidade Federal de Santa Catarina. Florianópolis, p. 303. 2018. ORCID: https:// 0000-0003-4264-0682 BRUNA ALESSANDRA GAFFURI (BAG) \| Universidade Estadual de Londrina \| Arquitetura e Urbanismo \| Londrina, PR - Brasil \| Correspondência para: Rua Gen. Alcides Etchegoyen, 1047 - Jardim La Salle, Toledo - PR, 85903-010 \| e-mail: bruna.alessandra@uel.br SECRETARIA NACIONAL DE JUSTIÇA. Refúgio em Números. Ministério da Justiça e Segurança Pública, 4a edição, p. 46, 2019. Disponível em: <https://portaldeimigracao.mj.gov.br/images/d ados/Ref%C3%BAgio%20em%20n%C3%BAmeros/RE F%C3%9AGIO%20EM%20N%C3%9AMEROS.pdf>. SEMINÁRIO DE METODOLOGIA DO IBGE, 2017. Disponível em: <https://eventos.ibge.gov.br /downloads/smi2017/apresentacoes/minicursos/ MC4%20_Parte1_Pedro%20Silva_Ricardo%20Cardo so_Sonia%20Oliveira.pdf>. Acesso em: 03 de março de 2021. ORCID: https:// 0000-0001-9778-8524 JORGE DANIEL DE MELO MOURA (JDMM), Dr. \| Universidade Estadual de Londrina \| Arquitetura e Urbanismo \| Londrina, PR - Brasil \| Correspondência para: Rodovia Celso Garcia Cid, PR 445 Km 380 - UEL-CAMPUS UNIVERSITÁRIO- CTU, Londrina - PR, 86057-970 \| e-mail: jordan@uel.br 115 SEST. Serviço Social do Transporte, 2021. Cerca de 450 mil toneladas de pneus são descartadas por ano no Brasil. Disponível em: <https://www.sestsenat.org.br/imprensa/noticia /cerca-de-450-mil-toneladas-de-pneus-sao- descartados-por-ano-no-brasil>. Acesso em: 04 de março de 2021. SEST. Serviço Social do Transporte, 2021. Cerca de 450 mil toneladas de pneus são descartadas por ano no Brasil. Disponível em: <https://www.sestsenat.org.br/imprensa/noticia /cerca-de-450-mil-toneladas-de-pneus-sao- SEST. Serviço Social do Transporte, 2021. Cerca de 450 mil toneladas de pneus são descartadas por ano no Brasil. Disponível em: <https://www.sestsenat.org.br/imprensa/noticia /cerca-de-450-mil-toneladas-de-pneus-sao- p descartados-por-ano-no-brasil>. Acesso em: 04 de março de 2021. REFERÊNCIAS 114 FIGUEIREDO, Isabel Campos Salles; BARBOSA, Ariane Corrêa; MIYAZAKI, Caroline Kimie; SCHNEIDER, Jerusa; COASACA, Raúl Lima; MAGALHÃES, Taína Martins; TONETTI, Adriano Luiz. Bacia de Evapotranspiração (BET): uma forma segura e ecológica de tratar o esgoto de vaso sanitário. Revista DAE \| núm. 220 \| vol. 67 \| São Paulo \| Edição Especial - Novembro 2019. DOI:https://doi.org/10.4322/dae.2019.059 NOGUEIRA, L. R. B.. Arquitetura Vernacular e Paisagem Amazônica: um Caminho na Busca pelo Habitar Poético. Revista da Abordagem Gestáltica (Impresso), v. xxii, p. 171-180, 2016. PASELLO, Bruno José Olivari. (2020). A Modularidade na indústria da construção brasileira: O Desenvolvimento de um Sistema Modular para o Pantanal / Bruno José Olivari Pasello; orientador, Jorge Daniel de Melo Moura, coorientador, Rovenir Bertola Duarte, 2020. FLORIANI, M. A.; FURLANETTO, V. C.; SEHNEM, S. Descarte sustentável de pneus inservíveis. Navus - Revista de Gestão e Tecnologia, v. 6, n. 2, p. 37–51, 14 abr. 2016. DOI: https://doi.org/10.22279/navus.2019.v9n1.p165- 181.806. Projeto Chacras (El Oro, Ecuador) - Natura Futura Arquitectura + Colectivo Cronopios. Disponível em: <https://www.archdaily.com.br/br/ 792993/projeto-chacras-natura-futura-arquitectura-plus- colectivo-cronopios>. Acesso em: 11 novembro de 2020. GREVEN, H. A.; BALDAUF, A. S. F. Introdução a Coordenação Modular da Construção no Brasil: Uma abordagem atualizada. Coleção Habitare, 9. Porto Alegre: ANTAC, 2007. 72 p. ISBN 978-85- 89478-23-6. QUARANTELLI, E. L. Patterns of sheltering and housing in US disasters. Disaster Prevention and Management: An International Journal, v. 4, n. 3, p. 43–53, 1991. DOI: https://doi.org/10.1108/09653569510088069. GOSLING, J.; PERO, M.; SCHOENWITZ, M.; TOWILL, D.; CIGOLINI, R. Defining and Categorizing Modules in Building Projects: An International Perspective. American Society of Civil Engineers. J. Constr. Eng. Manage., 04016062, DOI: 10.1061/(ASCE)CO.1943- 7862.0001181. 2016. g , , 3, p. 43–53, 1991. DOI: https://doi.org/10.1108/09653569510088069. RAIOL, L. L., SANTOS, D. C. R., dos SANTOS SILVA, K. W., da SILVA, L. P., FERREIRA, M. T., GOMES, J. A., ... & da SILVA JÚNIOR, J. F. (2019). Gerenciamento de pneus inservíveis na Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 cidade de Capanema, Pará. Brazilian Journal of Development, 5(12), 30989-31004. DOI:10.34117/bjdv5n12-204 HOW TO CITE THIS ARTICLE OLIVEIRA, Andressa Mikhaella Dos Santos Brito; GAFFURI, Bruna Alessandra; MOURA, Jorge Daniel de Melo. Proposta de Sistema Modular em Bambu para Habitação Emergencial. MIX Sustentável, v. 8, n. 1, p. 102-116, jan. 2022. ISSN 24473073.Disponível em: <http://www.nexos.ufsc.br/index.php/mixsuste ntavel>. https://doi.org/10.29183/2447- 3073.MIX2022.v8.n1.102-116 SILVA, Sidney A. Entre o Caribe e a Amazônia: haitianos em Manaus e os desafios da inserção sociocultural. Estud. av., São Paulo, v. 30, n. 88, p. 139-152, dezembro, 2016. OLIVEIRA, Andressa Mikhaella Dos Santos Brito; GAFFURI, Bruna Alessandra; MOURA, Jorge Daniel de Melo. Proposta de Sistema Modular em Bambu para Habitação Emergencial. MIX Sustentável, v. 8, n. 1, p. 102-116, jan. 2022. ISSN 24473073.Disponível em: TECHIO, L. M.; ZAMBONATO, B.; GRIGOLETTI, G. de C.; CLARO, A. Iluminação natural em habitação multifamiliar: o caso do conjunto residencial videiras, Santa Maria, RS. PARC Pesq. em Arquit. e Constr., Campinas, SP, v. 12, 2020. DOI: https://doi.org/10.20396/parc.v12i00.8659780. ZUKER, Fabio. Abrigos lotados, acampamentos improvisados e viadutos – A dura vida dos venezuelanos em Manaus \| National Geographic (nationalgeographicbrasil.com), 2019. ZUKER, Fabio. Abrigos lotados, acampamentos improvisados e viadutos – A dura vida dos venezuelanos em Manaus \| National Geographic (nationalgeographicbrasil.com), 2019. Submitted: 30/03/2021 Approved: 12/08/2021 Published: 01/12/2021 Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Proposta de Sistema Modular em Bambu para Habitação Emergencial \|A. M. dos S. B. Oliveira; B. A. Gaffuri & J. D. de M. Moura https://doi.org/10.29183/2447-3073.MIX2022.v8.n1.102-116 Mix Sustentável \| Florianópolis \| v.8 \| n.1 \| p.102-116 \| JAN \| 2022 Editora Responsável: Lisiane Ilha Librelotto Registro da contribuição de autoria: Taxonomia CRediT (http://credit.niso.org/) AMSBO, BAG, JDMM: conceituação, metodologia, recursos, administração do projeto, visualização, escrita -rascunho original, escrita -revisão e edição. JDMM i ã Registro da contribuição de autoria: JDMM: supervisão Declaração de conflito: nada foi declarado. Declaração de conflito: nada foi declarado. 116
W4383620101.txt	https://journals.qucosa.de/kbzf/article/download/52/39	de		Möglichkeiten kultureller Identifikation in Howard Shores Soundtrack zu Peter Jacksons THE LORD OF THE RINGS	Kieler Beiträge zur Filmmusikforschung	2,023	cc-by	7,404	Möglichkeiten kultureller Identifikation in Howard Shores Soundtrack zu Peter Jacksons THE LORD OF THE RINGS Martin Schröder (Rostock) Im Masterseminar Musikwissenschaft an der Hochschule für Musik und Theater Rostock führte ich im Jahr 2012 folgenden spielerischen Versuch durch: Ich zeigte den Studierenden das Musikvideo zu Enigmas »Return to Innocence« (Emimusic 2009) aus dem Jahr 1993. Im Refrain wurde offensichtlich das Sample eines traditionellen Gesangs verwendet. Ich fragte die Anwesenden daher, in welchem Teil der Welt sie diesen Gesang verorten würden. Die Mehrheit der Studierenden entschied sich für Nordamerika, offensichtlich in der Annahme, bei dem verwendeten Sample handele es sich um den Gesang nordamerikanischer Natives. Die Wahl Nordamerikas hatte ich vorausgesehen und verstärkte diesen Eindruck zum einen durch den Hinweis, dass der Song für die musikalische Untermalung von TVWerbespots für die olympischen Spiele 1996 in Atlanta genutzt wurde, und zum anderen mit einer vorbereiteten Karte der Stammesgebiete der indigenen Stämme Nordamerikas. Daraufhin spielte ich den Studierenden zwei Aufnahmen der CD Authentic Music of the American Indian (Various 2008) vor: einen »Kiowa Slow War Dance« (Track 10) und einen »Shanee Stomp Dance« (Track 18) verbunden mit der Frage, wessen Gesang nun in Enigmas Song Verwendung gefunden hat. Nachdem sich eine Mehrheit auf den Shanee Stomp Dance geeinigt hatte, eröffnete ich den Teilnehmern, dass ihre Vermutung falsch sei, woraufhin ich das Video einer Live-Performance eines Joiks präsentierte. Joiks sind der traditionelle Gesang der Sami, einer transnationalen Ethnie im Norden Skandinaviens. Das Konzept des Joiks spiegelt sich hervorragend in der Ästhetik des Enigma-Videos wider. Joiks haben keinen Anfang und kein Ende. Passend dazu läuft das Musikvideo Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 41 rückwärts ab. Plötzlich befanden wir uns also geografisch in Nordeuropa. Die Verwendung eines gesampelten Joiks war ebenso plausibel wie die der kurz zuvor gehörten Gesänge der amerikanischen Natives. Viele Teilnehmer waren nun überzeugt, einen Sami-Joik in »Return to Innocence« zu hören. Allein, auch diese Gewissheit stellte sich als falsch heraus. Das von Michael Cretu, Produzent und kreativer Kopf hinter Enigma, genutzte Sample lässt sich sehr präzise zurückverfolgen – nicht nur auf eine bestimmte Ethnie, sondern sogar auf zwei konkrete Personen. Die Aufnahme entstammt dem Jubilant Drinking Song von Kuo Ying-nan (geb. Difang Duana) und Kuo Hsiu-chu, zwei Mitglieder der Ami People, einem indigenen Stamm in Taiwan. Eine Verortung der Musik in Asien hätte keiner der Studierenden vermutet, ebensowenig wie ich selbst, ging ich doch jahrelang davon aus, bei dem Sample handele es sich um einen Sami-Joik. Infolge dieser Begebenheit drängte sich die Frage auf: Kann Musik überhaupt eine Identität besitzen, Musik als eine kulturelle Praxis? Die obigen Ausführungen zeigen sehr deutlich: Die Wahrnehmung von Musik ist ein Konstrukt des Rezipienten und kulturell beeinflusst. Die Musik selbst ist das Produkt eines schöpferischen Geistes. Der Komponist hat eine bestimmte Vorstellung von seiner Kreation, die umgesetzt, zum Leben erweckt, gespielt werden muss. Eventuell wird diese im Zuge eines Aufnahmeprozesses auch materiell fixiert. Dann aber erfolgt ein wichtiger Schritt: die Rezeption. Étienne Souriau zufolge ist die Existenz eines Musikwerkes in der konkreten Wirklichkeit trotz seiner klanglichen Realisation solange nicht wirklich vollkommen, bis es zu einem gemeinsamen Empfang des Werkes kommt, der sein Dasein konstituiert (zitiert nach: Lissa 1975, 111). Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 42 Die Veränderlichkeit von Musik und ihre Wahrnehmung hängen also maßgeblich vom Rezipienten ab. Die Charakteristik und Eigentümlichkeit eines Musikstücks ist letztlich nichts anderes als eine prozessuale Verhandlung zwischen der Idee von Komponist (und Regisseur) und dem Hörer und seiner Wahrnehmung. Diese Wahrnehmung des Hörers ist wiederum beeinflusst von individueller musikalischer Sozialisation und kultureller Prägung. So wie die personale Identität des Menschen als ein dynamisches, ständig neu verhandeltes Konstrukt erscheint, so ist auch die vermeintliche ›Identität‹ eines Musikstücks das Ergebnis eines Konstruktionsprozesses der beteiligten Personen, insbesondere des Hörers. Darüber hinaus ist die Musikwahrnehmung heutzutage stark medial beeinflusst. Schließlich sind alle möglichen Musiken heutzutage über das Internet abruf- und konsumierbar. Nur diesem Umstand ist es zu verdanken, dass wir mit Musiken aus aller Welt in Kontakt kommen und diese kennenlernen, was letztlich die Voraussetzung dafür ist, vermeintliche Ähnlichkeiten ausmachen und unbekannte Musik geographisch und kulturell verorten zu können. Die einleitenden Schilderungen aus dem Seminar zeigen jedoch, dass solche kulturellen Verortungen in Bezug auf Musik in die Irre und leicht zu essentialistischen Zuschreibungen führen können. Howard Shores Soundtrack zu Peter Jacksons Literaturverfilmung THE LORD OF THE RINGS (USA/NZ 2001–2003) stellt dabei keine Ausnahme dar. So schreibt David Bratman in seiner Kritik: So Tolkien meant us to think of the hobbits as English country folk singing English folk songs. Which is why it was so disconcerting to go to Peter Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 43 Jackson’s FELLOWSHIP OF THE RING movie and hear the hobbits depicted by music that is not English at all but neo-Celtic: composer Howard Shore’s attempt at copies of Irish whistle and fiddleand-accordion tunes in lush orchestral arrangements. Tolkien spent a lifetime differentiating English civilization from Celtic. He would have been deeply unhappy, to say the least, to have heard Irish music, and imitation Irish music at that, applied to his hobbits. (Bratman 2010, 142) Inwieweit Bratmans Einschätzung zum Film und dessen Musik zutreffend ist, soll im Folgenden erörtert werden. Jacksons Film-Trilogie wurde von mir nicht nur aus persönlichem Interesse, sondern auch aufgrund ihres hohen Bekanntheitsgrades als Resultat einer weltweiten medialen Vermarktung ausgewählt, was sich auch im enormen Erfolg der Filme (17 Oscar-Auszeichnungen) widerspiegelt. Dies gilt ebenso für die Musik (Academy Awards für die beste Filmmusik in den Jahren 2001 und 2003) von Howard Shore. Wie die Filme selbst dürfte auch der Soundtrack einer breiten Masse bekannt sein – auch weil dieser losgelöst vom Film und in adaptierter Form als 6-sätzige Sinfonie existiert. Für Jacksons Film-Trilogie erschuf Shore ab Oktober 2000 musikalisches Material von insgesamt 11 Stunden. Für beinahe jeden Charakter und jede Kultur komponierte er musikalische Themen, die er – teilweise über alle drei Filme hinweg – zum Teil leitmotivartig verwendet. Shores Score sollte nach Jacksons Vorstellungen eine »cultural significance« besitzen und »eine Menge über die Kulturen dieser Welt zu erzählen haben«1 [Übers. des 1 »[…] it’s doing it in such a way that it’s also telling you a lot about the cultures of this world«. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 44 Autors] (Jackson 2002, 00:02:43–00:03:13).2 Es war offensichtlich von Beginn an eine Identifikation des Scores mit bestimmten Kulturregionen in Tolkiens literarischem und Jacksons filmischem Raum intendiert. Musik ist soziokulturell geprägt (Merriam 1964, 27–32). Ja, Musik ist Kultur (Herndon/McLeod 1982, 9f.), und kann daher per definitionem kulturelle Spezifika reflektieren. Wie vielschichtig und ambivalent diese Reflexion sein kann, soll anhand der Phänomene ›Englishness‹ und ›Celticness‹ verdeutlicht werden. Die leitende Fragestellung dabei lautet daher: Inwiefern bietet Shores Score Möglichkeiten der Reflexion von englisch bzw. keltisch geprägter kultureller Identität? Welche Anknüpfungspunkte lassen sich in der Musik finden, die eine solche kulturelle Identifizierung des Hörers erst ermöglichen? Vergleichend dazu sollen auch Tolkiens eigene Vorstellungen berücksichtigt werden, die letztlich Eingang in sein literarisches Werk gefunden haben. Diese Fragen sollen durch die Untersuchung des Auenland-Themas beantwortet werden. Das »Shire Theme« war nicht nur das erste musikalische Thema, das Shore für die Filme komponierte (Adams 2010, 392), es dürfte auch das am besten bekannte sein.3 2 Die Time-Code-Angabe bezieht sich auf die Warner Special Extended DVD Edition (2002). 3 Denkbar und reizvoll wäre auch eine Untersuchung der anderen Themen, wie etwa des Rohan-Themas, das dem Hörer aufgrund der Verwendung der norwegischen Hardanger-Geige in Verbindung mit skandinavisch anmutender Architektur eine Identifikation mit der normannischen Kultur nahelegt, jedoch den Ansichten und Einstellungen Tolkiens zuwiderläuft, der zeitlebens die nahezu vollständige Auslöschung der angelsächsischen Kultur durch die Invasion der Normannen beklagte. Aufgrund der Fülle an Themen und Motiven und deren leitmotivartiger Verwendung wäre auch eine Untersuchung ihrer kompositorischen Variation interessant, gerade auch vor dem Hintergrund der neu entstandenen HOBBIT-Trilogie (USA/NZ 2012–2014, Peter Jackson). Eine umfassende Betrachtung des vielschichtigen Werkes kann an dieser Stelle selbstverständlich nicht erfolgen. Vielmehr soll der Frage nach kulturellen Zuschreibungen durch die Musikschaffenden und Rezipienten genauer nachgegangen werden, was eine Beschränkung des Untersuchungsmaterials erforderlich macht. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 45 Das Auenland-Thema und ›Celticness‹ Laut Doug Adams wollte Shore das Auenland durch eine »Celtic influenced peasant melody« repräsentieren (Adams 2010, 26). Der Begriff »Celtic« ist ein relativ unscharfer Terminus. Vor allem im Zuge des World-MusicBooms der 1990er Jahre in Mode gekommen4 (Melhuish 1998, 7, 29; Reiss 2003, 145), bezieht er sich im Wesentlichen auf die Formen traditioneller Musik der Länder mit keltischen Sprachen wie etwa Irland und Schottland, aber auch die Bretagne sowie Wales und andere britische Regionen wie die Isle of Man oder Cornwall, wie auch die Region Nova Scotia in Kanada. Beim Hören des Auenland-Themas wird schnell deutlich, dass Shore vor allem eine musikalische Verbindung mit der irischen respektive schottischen Musik intendierte. Ein Blick in die Partitur (Adams 2010, 25) bestätigt diesen Höreindruck. Die bei Adams reproduzierten Takte 9–16 des Shire Theme (›Rural Setting‹) weisen mit Mandoline, Fiddle und Bòdhran in der Instrumentierung charakteristische Instrumente der irischen und schottischen traditionellen Musik auf. Nun ist es so, dass internationale Filme und insbesondere Blockbuster wie THE LORD OF THE RINGS, wollen sie kommerziell erfolgreich sein, die Verständigung mit einem globalen Publikum suchen müssen. Das gilt ebenso für die Filmmusik. Die Instrumentierung ist ein Charakteristikum, über das am einfachsten eine solche Verständigung möglich ist, was in der vorliegenden Untersuchung eine kulturelle Identifizierungsmöglichkeit bedeutet. Nicht umsonst werden 4 »Celtic« ist in diesem Kontext vor allem auch als ein medial repräsentiertes Label und als Kategorie der Musikindustrie zu sehen. Doch bereits zu Beginn des 20. Jahrhunderts wurde von gälischen Aktivisten wie Malcolm MacFarlane der inflationäre Gebrauch des Begriffs beklagt, vor allem, wenn damit eigentlich der Terminus »Gaelic« gemeint war (Scott 2013, 15f). Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 46 in Filmen, die mit bestimmten Ländern oder Kulturregionen verbunden sind, typische Instrumente in der Filmmusik verwendet – unabhängig davon, ob nun kulturelle Anknüpfungspunkte oder einfach nur Lokalkolorit erzeugt werden sollen. Beispiele hierfür wären die Balalaika, die typischerweise mit Russland verbunden wird (z. B. in DOCTOR ZHIVAGO [USA/I/F 1965, David Lean, Musik: Maurice Jarre]), oder aber auch der Dudelsack5, der für gewöhnlich mit Schottland identifiziert wird und daher als kultureller Marker in der Filmmusik dient (z. B. in BRAVE [USA 2012, Mark Andrews/Brenda Chapman/Steve Purcell, Musik: Patrick Doyle]), ebenso wie der Song »Tha Mo Ghaol Air Aird A ˈChuain« der schottisch-gälischen Sängerin Julie Fowlis im Trailer zum obengenannten Film, in dem die Musik eine Art ›Scottishness‹ transportieren soll und somit paraphrasierend zu dem im Bild Gezeigten wirkt. Doch nicht nur die Instrumentierung vermag als kultureller Marker für Celticness zu fungieren. Betrachtet man das in Abb. 1 dargestellte Auenland-Thema im ›Pensive Setting‹, so fällt zunächst eine pentatonisch gefärbte Umgebung auf, wobei aufgrund der Töne cis² und g1 (mit einem X markiert) keine reine Pentatonik vorliegt: 5 Dass auch eine solch vermeintlich eindeutige und einfache kulturelle Zuordnung wie Dudelsack–Schottland nicht unproblematisch ist, zeigt die Tatsache, dass der Dudelsack bzw. die Sackpfeife ein pan-europäisches Instrument ist, das in unterschiedlicher Ausführung auf dem Balkan, in Nordeuropa, in Italien, Spanien und England gespielt wird. Als Hümmelchen und Dudey war er zur Zeit der Renaissance auch in Deutschland in Gebrauch. Ironischerweise kam der Dudelsack in den schottischen Highlands, mit denen er gemeinhin verbunden wird, zeitlich gesehen zuletzt an (Collinson 1966, 160; Cannon 1995, 131). Auch in BRAVEHEART (USA 1995, Mel Gibson), dem wohl bekanntesten Film, der schottische Kultur und Geschichte (wenn auch in vielen Punkten historisch nicht korrekt) zum Gegenstand hat, erklingt im Soundtrack von James Horner nicht etwa eine schottische Great Highland Bagpipe, sondern eine Uilleann Pipe, eines der Nationalinstrumente Irlands. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 47 Abb. 1: Auenland-Thema im ›Pensive Setting‹ Allerdings kann die zugrunde liegende Skala auch als hexatonisch bezeichnet werden, da das cis² relativ exponiert als Teil der Melodie angesehen werden kann, das g1 jedoch lediglich Teil einer Verzierung ist. Schottischen und irischen Tunes und Songs liegen häufig pentatonische wie auch hexatonische Skalen zugrunde. Beispiele hierfür sind unter anderem Robert Burns‘ Song »Auld Lang Syne« oder auch die Fiddle Tune »Niel Gow’s Lament for the Death of his Second Wife«. Die als Scots Snap (und nicht etwa Scotch Snap, wie häufig zu lesen ist) bekannte eingerahmte Kombination aus betonter Sechzehntelnote mit nachfolgend unbetonter punktierter Achtelnote ist eine wichtige rhythmische Figur in der schottischen Musik und laut Collinson (1966, 29) »the very life-blood of Scots musical rhythm«. Der Scots Snap ist gebräuchlich sowohl in Scots als auch in Gaelic Songs. Besonders prominent ist er in Strathspeys, einer Art von Tanz und dazugehöriger Tune, die als spezifisch schottisch angesehen werden kann. Er findet sich als (sprach)rhythmische Figur besonders häufig im Gälischen wie auch im Scots, da in diesen Sprachen überdurchschnittlich viele Worte mit einer kurzen betonten Silbe gefolgt von einer langen unbetonten Silbe existieren – im Gegensatz etwa zum Italienischen oder Französischen (Temperley/Temperley 2011, 53). 6 6 Der Lombard-Rhythmus, der ebenfalls aus einer Sechzehntelnote mit nachfolgend punktierter Achtelnote besteht, war bei italienischen Komponisten des 17. und 18. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 48 Abb. 2: »Miss Gordon of Fochabers«, Strathspey von William Marshall Abb. 3: Gälisches Lament »Cumha do Mhinisteir a Chaidh a Bhàthadh« Jahrhunderts überaus beliebt, existierte in der Vokalmusik jedoch zumeist als Melisma auf einer einzigen Silbe, in der Instrumentalmusik häufig gebunden als Ornamentierung bzw. Vorhalt (Temperley 2011, 53f). Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 49 Letztlich finden wir im Auenland-Thema auch eine große Anzahl von grace notes, welche als einfachste Form der Ornamentierung in traditioneller schottischer und irischer Musik überaus gebräuchlich ist. Besonders exponiert finden sich diese Verzierungen in der Variation ›Rural Setting‹. Abb. 4: Auenland-Thema im ›Rural Setting‹ Obgleich das Auenland-Thema auf den ersten Blick eine Menge Ansatzpunkte und Marker bietet, die eine kulturelle Identifizierung mit der Celtic Culture durch das Publikum erlauben, stellt sich die Sachlage bei näherer Betrachtung als deutlich vielschichtiger heraus. Pentatonik wie auch hexatonische Skalen sowie Ornamentierung in Tunes sind in der schottisch-irischen, natürlich aber auch in der englischen traditionellen Musik verbreitet. Ohnehin gibt es Überschneidungen im schottischen, irischen und englischen Repertoire, wie es natürlicherweise aufgrund von kulturellen flows und Interdependenzen auch keine reinen puren Musikkulturen geben kann, in diesem Fall eine distinkt ›englische‹ bzw. ›schottisch-irisch/keltische‹. Auch der Scots Snap ist selbstverständlich aus dem gleichen Grund in der englischen Sprache bzw. in der englischen traditionellen Musik prominent vertreten wie er es in der schottischen oder speziell gälischen Musik ist. Immerhin hat sich das Scots aus dem Northumbrian Old English entwickelt. Im Barock galt der Scots Snap laut Philip Tagg (2011, 0:56–0:58) gar als Indikator für ›Englishness‹. In seinem Video »Scotch Snaps – The Big Picture« zitiert Tagg aus dem Harvard Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 50 Dictionary of Music (1958). Dort heißt es in den Worten des Autors Willi Apel: This rhythm figures prominently in English music of the seventeenth century (John Blow, Henry Purcell), in which it is used effectively in order to bring out the short, but accented, first syllables which occur in so many English dissylabics. Philip Taggs Musematic Analysis als Analyse-Tool Im vorangegangenen Abschnitt wurden musikimmanente Charakteristika aufgezeigt, die dem Zuhörer bzw. Zuschauer eine kulturelle Verortung der Musik in Schottland bzw. Irland ermöglichen, gleichzeitig ist deren Ambivalenz aufgezeigt worden, sodass bestimmte kulturelle Identifikationen durch den Hörer denkbar erscheinen, jedoch nicht zwangsläufig genau so erfolgen müssen. Erhellend kann an dieser Stelle ein Wechsel der Untersuchungsperspektive sein. Diesen anderen Blickwinkel auf einen potenziellen Bedeutungsgehalt und die Wirkung von Filmmusik auf den Rezipienten bietet Philip Taggs aus der Semiotik stammender Ansatz der »Musematic Analysis« (Abb. 5). Demnach wird eine zu analysierende Musik A in musikalische Bausteine, sogenannte Musemes, zerlegt. Dieser von Charles Seeger geprägte Begriff bezeichnet kleinste Träger musikalischer Bedeutung, letztlich also Zeichen, die wiederum interpretiert werden müssen.7 7 Tagg beschreibt das musical morpheme oder museme folgendermaßen: »These component parts […] are nevertheless the smallest units or basic building bricks in the construction of musical phrases and […] can be combined (like the morphemes of spoken language) in different ways to communicate different messages […].« (Tagg 2000, 61). Seeger definiert die Mindestlänge eines Musemes mit drei Tönen, Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 51 Abb. 5: Schematische Darstellung von Taggs Musematic Analysis Musemes können ineinander übergehen, zeitlich nacheinander erklingen (museme strings) oder auch zeitgleich auftreten (museme stacks) (108–110). Die Musemes ihrerseits sind durch basale Elemente (musical phonemes) musikalischen Ausdrucks charakterisiert. Dies können einzelne Töne, aber auch Parameter wie Dynamik und Tempo sein (108).8 Nach der Zerlegung des Analyseobjekts in seine einzelnen Musemes wird dieses mittels »interobjective comparison« mit den Referenzmusiken B, C, D usw. verglichen. Diese müssen aus der gleichen Kulturgemeinschaft stammen wie Tagg hingegen erachtet die Stille vor dem ersten Ton als ebenso konstitutiv, sodass ein einfaches Intervall aus zwei Tönen (bzw. drei Ereignissen, wenn die Stille vor dem entsprechenden Ton mitbedacht wird) bereits ein vollständiges Museme darstellen kann (106f). Man denke in diesem Zusammenhang beispielsweise an die Signalwirkung von aufsteigenden Quarten oder Quinten. 8 Die Änderung des Bedeutungsgehalts der von Shore komponierten Themen und Motive bzw. der möglichen Affekte bei den Rezipienten aufgrund der Modifikation dieser musikalischen Phoneme wäre, wie in Anmerkung drei bereits angedacht, eine lohnende Untersuchungsperspektive. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 52 das Analyseobjekt A und eine soziomusikalische Bedeutung besitzen, das heißt, bei den Mitgliedern einer Gesellschaft aufgrund von Tradierungsprozessen bestimmte Assoziationen wecken oder Emotionen evozieren. Lassen sich nun sowohl in der Analysemusik als auch in den Referenzmusiken die gleichen musikalischen Bausteine finden, ist anzunehmen, dass die Analysemusik mit den gleichen Bedeutungsinhalten belegt sein und gleiche oder ähnliche Affekte beim Rezipienten hervorrufen könnte (Tagg 2000, 112–114; Tagg 2004, 2). Zwei Beispiele sollen das Prinzip Musematic Analysis in aller gebotenen Kürze verdeutlichen. Beispiel a) Museme des Lamentobasses: Diese diatonisch oder chromatisch absteigende Bassfigur vom Grundton in den Quintton einer Tonart ist als Topos seit dem frühen 17. Jahrhundert von vielen Komponisten verwendet worden, um Klage, Trauer, Schmerz und Leid auszudrücken.9 So zum Beispiel im zweiten Satz der Kantate »Weinen, Klagen, Sorgen, Zagen« BWV 12 von Johann Sebastian Bach und auch in dessen Parodie, dem »Crucifixus« aus der Messe in h-Moll (Abb. 6). In Mozarts »Qui tollis« aus der Großen Messe in c-Moll lässt der Lamentobass die Schwere der Sünden beinahe körperlich spürbar werden (Abb. 7). Auch in Filmmusiken ist der Lamentobass effektvoll als Symbol für Trauer eingesetzt worden. Als wohl bekanntestes Beispiel darf sicherlich das klagende »Harmonica Theme« in Ennio Morricones Soundrack zu C'ERA UNA VOLTA IL WEST (I/USA/E 1968, Sergio Leone) gelten (Abb. 8). 9 Dies geschah zumeist in Form eines absteigenden Moll-Tetrachords. In den frühen Opern des 17. Jahrhunderts wurde dieser beinahe ausschließlich in Verbindung mit Klagegesängen verwandt. Durch diese enge Verbindung erhielt der Lamentobass auch nach dessen Eingang in die Instrumentalmusik seine Bedeutung als Zeichen von Trauer und Schmerz (Vgl. Rosand 1979, 358f). Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 53 Abb. 6: Johann Sebastian Bachs »Crucifixus«, h-Moll Messe (T. 13–19) Abb. 7: Wolfgang Amadeus Mozarts »Qui tollis«, Große Messe in c-Moll (T. 4–7) Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 54 Abb. 8: »Harmonica Theme« aus C'ERA UNA VOLTA IL WEST (T. 9–16) Die Bedeutung des Lamentobasses als Zeichen von Leid, Schmerz und Trauer wurde von Monteverdi über Bach bis Morricone, bis in unsere heutige Zeit tradiert. Eine dem Zuhörer unbekannte Musik, die jedoch die Lamentobass-Figur enthält, wird daher von diesem womöglich mit den gleichen Affekten und Assoziationen verknüpft, so zumindest die Logik Taggs. Beispiel b) Museme des absteigenden Terzfalls: Als besonders eindrückliches Beispiel für einen absteigenden Terzfall soll an dieser Stelle die Dies Irae-Sequenz angeführt werden. Sie markiert den Beginn des Hymnus vom Jüngsten Gericht und ist Teil der lateinischen Totenmesse. Sie steht als Symbol für das Gottesgericht oder auch allgemein für den Tod. Abb. 9: »Dies Irae«-Sequenz Mit dieser Konnotation wurde sie auch in Filmmusiken verwendet, wie etwa in ESCALATION (I 1968, Roberto Faenza, Musik: Ennio Morricone) oder aber in THE SHINING (USA/UK 1980, Stanley Kubrick, Musik: Wendy Carlos/Rachel Elkind). Im Soundtrack zu THE LORD OF THE RINGS verarbeitet Howard Shore die Dies Irae-Sequenz passenderweise in der Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 55 Begleitfigur im Motiv für Sauron/Mordor. Auch in diesem Fall soll die Musik dem Zuhörer eine Verbindung von Tod und Verderben mit der Figur des dunklen Herrschers Sauron ermöglichen, eine Verbindung, die durch die Bilder ohnehin erfolgt, von der unterlegten Musik jedoch gekonnt unterstützt wird, wenngleich die Dies Irae-Sequenz nicht als vollständiges Zitat vorliegt: Abb. 10: »The Power of Mordor«-Thema mit absteigendem Terzfall In seiner Masterarbeit Projecting Tolkien’s Musical Worlds: A Study of Affect in Howard Shore’s Soundtrack to LORD OF THE RINGS (2007) versucht Mathew David Young, mit Hilfe von Taggs Musematic Analysis mögliche Wirkungen auf das Publikum bzw. dessen potenzielle Assoziationsspielräume aufzuzeigen. Dabei untersucht er systematisch die Hauptthemen des Scores. Durch den Vergleich des Auenland-Themas mit Musiken, die schottisch–irische Musik nachbilden bzw. dem schottischirischen Kulturraum entstammen, ist Young bestrebt, diesem eine ›Celticness‹ zuzuschreiben, die so auch vom Publikum wahrgenommen werden kann. Hierzu vergleicht er das Auenland-Thema mit dem Thema »The Secret Wedding« aus James Horners Soundtrack zu Mel Gibsons BRAVEHEART (1995) (Young 2007, 43). Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 56 Wie THE LORD OF THE RINGS ist auch BRAVEHEART überaus erfolgreich gewesen und dürfte einem breiten Publikum bekannt sein. Mit dem Freiheitskampf von William Wallace bearbeitet Gibson ein zentrales Thema schottischer Geschichte und Kultur, wenngleich sein Film diverse historische Ungenauigkeiten und fiktive Elemente beinhaltet, die an dieser Stelle jedoch nicht weiter ausgeführt werden sollen. Nimmt man die Flötenvariation des Auenland-Themas10 als Referenz, so gleicht Shores Orchestrierung der des ›Wedding Theme‹ von James Horner (0:33:10– 0:33:50)11 auf verblüffende Weise. In beiden Fällen liegt eine einzelne Flötenmelodie über ausgehaltenen diatonischen Streicherakkorden vor (im Horner-Score übernehmen ab T. 5 die oktavierten Streicher für ein kurzes Zwischenspiel die Melodie). Am Ende des in Abb. 11a dargestellten Ausschnitts ist sogar der für schottische Musik scheinbar so typischen Scots Snap in den Streichern zu hören. 10 Die Variation des Auenland-Themas im ›Pensive-Setting‹ für Tin Whistle bildet in instrumentaler Hinsicht einen weiteren kulturellen Marker und möglichen Anknüpfungspunkt für eine kulturelle Identifikation des Hörers mit der schottischirisch geprägten Musikkultur. Wie Young überzeugend darlegt, bleibt die Whistle über alle drei Filme hinweg nicht nur mit dem Auenland allgemein, sondern im Speziellen mit der Figur Frodos verbunden (Young 2007, 50). 11 Die Time-Code-Angabe bezieht sich auf die Twentieth Century Fox 2-Disc Limited Edition DVD (2009). Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 57 Abb. 11a & Abb. 11b: »The Secret Wedding«-Thema und AuenlandThema im Vergleich Eine weitere Übereinstimmung sieht Young im Museme der aufsteigenden Auftaktnoten12, in teilweise ähnlichen oder gar identischen Intervallfolgen sowie in einem gleichartigen Melodieverlauf. Schrittweise Auftaktachtel lassen sich zudem, so Young, in vielen Beispielen irischer und schottischer Tunes finden (43–45). Folgt man der Argumentation Youngs, hat das Auenland-Thema aufgrund der Ähnlichkeiten zu Melodien aus dem 12 Problematisch ist bei Youngs Argumentation die Tatsache, dass das AuenlandThema, legt man die bei Adams abgedruckte Transkription für einen Vergleich zugrunde, nicht mit einem Auftakt sondern einem Volltakt beginnt. Erst der Nachsatz der 8-taktigen Periode wird mit zwei auftaktigen Sechzehntelnoten eröffnet. Da Adams in direktem Kontakt mit dem Komponisten stand, ist in diesem Fall der ›offiziellen‹ Transkription zu folgen. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 58 schottisch-irischen Kulturraum (Traditional Tunes) bzw. zu anderen Musiken, die diese nachbilden (wie denen des BRAVEHEART-Soundtracks), das Potential, eine ›keltisch‹ geprägte Atmosphäre zu schaffen. Der Vorteil von Taggs Musematic Analysis liegt auf der Hand: Bei einer guten Auswahl der Referenzmusiken können Ähnlichkeiten zwischen den Stücken auch von Nichtmusikern wahrgenommen werden. Allerdings birgt die Methode auch gewisse Probleme, die nicht unerwähnt bleiben sollen: 1.) Die Musematic Analysis zur Untersuchung möglicher Assoziationen und Affekte beim Hörer bleibt hochgradig subjektiv. Zum einen müssen dem Hörer die Referenzmusiken bzw. -themen natürlich bekannt sein, um Ähnlichkeiten ausmachen zu können, und selbst dann bleibt die Reaktion auf bestimmte Musiken ein sehr individuelles Phänomen. 2.) Die Wahrnehmung bestimmter Ähnlichkeiten hängt stark von der individuellen Einbildungskraft und dem Abstraktions- und Kategorisierungsvermögen des Rezipienten ab. Selbst, wenn diesem die Referenzmusiken bekannt sind, muss er das fragliche Museme – etwa den Lamentobass – während des Rezeptionsvorgangs erst einmal erkennen.13 Das heißt, der Hörer muss eine Vorstellung davon haben, welche Eigenschaften das Museme als solches konstituieren. 3.) Auch wenn gewisse Konsistenzen in den Hörerreaktionen auf diese Musiken konstatiert werden (»intersubjective comparison«) (Tagg 2000, 112), wird in China eine bestimmte Musik womöglich ganz anders wahrgenommen als etwa in Europa. Ein Museme oder eine Museme-Kette können gewisse Bedeutungszuschreibungen durch Hörergruppen plausibel erscheinen lassen und durch die weltweite Mediensphäre und Warenströme 13 Dabei soll nicht außer Acht gelassen werden, dass das Hören von Musik in Teilen auch ein unbewusster Prozess ist. Der Rezipient kann beispielsweise auf affektiver Ebene beeinflusst werden, ohne dass dies ein bewusstes Wahrnehmen erfordern würde. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 59 kommt es selbstverständlich zu Homogenisierungstendenzen in der Wahrnehmung von Musik. Dennoch spielen spezifische, kulturell geprägte Hörerfahrungen natürlich weiterhin eine große Rolle. 4.) Die Referenz muss wahrnehmbar bzw. hörbar sein. In dem angeführten Beispiel des Mordor/Sauron-Themas ist die angedeutete Dies Irae-Sequenz in den tiefen Bläsern nur schwer herauszuhören.14 Beim einmaligen Sehen des Films wird dem Hörer daher diese Anspielung womöglich gar nicht auffallen, sofern die Sequenz überhaupt bekannt ist.15 5.) Taggs Methode wirkt zum Teil arg konstruiert, zumindest in der Anwendung durch Mathew Young. Selbstverständlich findet sich sowohl im Auenland-Thema als auch bei James Horners ›Wedding Theme‹ und vielen irischen und schottischen Tunes das Museme der aufsteigenden Auftaktnoten. Dass es dem Rezipienten jedoch gelingt, diese subtile Verbindung zu erkennen und somit eine Konnotation von Celtic Culture auf das Auenland-Thema zu übertragen, erscheint zumindest fragwürdig, zumal es für dieses Argument genug Gegenbeispiele gibt. Das heißt, je reduzierter die Auswahl und Gestalt der zu betrachtenden musikalisch bedeutsamen Einheiten ist, desto schwieriger und spekulativer ist es, mögliche musikalische Bedeutungszuschreibungen oder affektive Beeinflussungen beim Hörer auszumachen. 14 Z. B. bei 0:32:13–0:32:30 in der Extended Version von THE FELLOWSHIP OF THE RING. 15 Die Sequenz ist beispielsweise in der Abmischung und Orchestrierung der 6sätzigen Sinfonie, die ebenfalls als CD-Aufnahme erhältlich ist, wesentlich besser herauszuhören (Shore 2011, CD1, Track 01, 09:04–09:20). Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 60 Das Auenland und ›Englishness‹ Eine weitere Problematik, die bei Young keine Erwähnung findet, ist die Intention Tolkiens. Tolkien wollte mit seiner Romanvorlage eine ›Mythologie für England‹ schaffen. So schreibt er in einem Brief aus dem Jahre 1956: Having set myself a task, the arrogance of which I fully recognized and trembled at: being precisely to restore to the English an epic tradition and present them with a mythology of their own: it is a wonderful thing to be told that I have succeeded, at least with those who still have the undarkened heart and mind. (Tolkien 2000, 230f). Dieses Bedürfnis Tolkiens erklärt sich aus der Tatsache, dass nach der Schlacht von Hastings im Jahr 1066 und im Zuge der darauf folgenden normannischen Invasion beinahe alle Artefakte aus der angelsächsischen Zeit vernichtet wurden und außer dem Beowulf keine angelsächsischen Epen oder Fragmente vergleichbarer Texte die Zeit überdauert haben. Mythologische Textsammlungen wie etwa die auf Altisländisch verfasste Edda oder die von Elias Lönnrot nach mündlichen Überlieferungen kompilierte finnische Kalevala16 finden sich im angelsächsischen Raum nicht. Dass trotz der beabsichtigten »Mythologie für England« Tolkiens Romanvorlage wie auch die Schilderungen der Begebenheiten des Ersten 16 Die Edda und Kalevala sind vom Philologen Tolkien intensiv studiert und Elemente dieser Erzählungen in sein »Legendarium« miteinbezogen worden. Das ›Quenya‹ (Hochelbisch) basiert auf der finnischen Sprache, die ›Silmaril‹ oder auch die zwei Bäume ›Laurelin‹ und ›Telperion‹ Valinors weisen Ähnlichkeiten zum magischen Objekt ›Sampo‹ aus der Kalevala auf (Vgl. Kahlas-Tarkka 2014, 264, 267f). Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 61 und Zweiten Zeitalters im Silmarillion durchaus auch keltische Motive aufweisen und diese daher ebenso ambivalent sind wie die Musik Howard Shores zu Peter Jacksons Verfilmung, überrascht nicht.17 Das Auenland jedoch war für Tolkien englisch geprägt. Schon der Begriff ›Shire‹ stammt aus dem Altenglischen und bezeichnete in angelsächsischer Zeit ein Verwaltungsgebiet. Tolkien selbst äußert sich in einem Brief an seinen Verleger mit folgenden Worten: »The Shire […] is in fact more or less a Warwickshire village of about the period of the Diamond Jubilee« (Tolkien 2000, 230). Mit dieser Aussage bezieht sich der Autor auf das diamantene Kronjubiläum Königin Victorias im Jahre 1897. Zu diesem Zeitpunkt war Tolkien 5 Jahre alt. Erst ein Jahr zuvor war er mit seiner Mutter in das kleine Städtchen Sarehole, einen Vorort von Birmingham, gezogen. Dieser Ort und seine Umgebung dienten als Inspiration für Tolkiens Vision vom Auenland. Am eindrücklichsten zeigt sich dies im Landschaftsmarker der Sarehole Mill, die ihre literarische Entsprechung – wenn auch in idealisierter Weise – in der Mühle von Hobbingen hat.18 Eine englische Prägung des Auenlandes in der Vorstellung Tolkiens zeigt sich jedoch nicht nur in geografischer oder materieller Hinsicht. Auch in musikalischer Hinsicht lässt sich eine Verbindung des Auenlandes im Roman bzw. des Auenland-Themas von Howard Shore mit dem schottisch17 So basiert das von den Grauelben gesprochene Sindarin auf dem Walisischen, die Akallabêth, die Geschichte vom Untergang Númenors aus dem Silmarillion, weist eine große motivische Ähnlichkeit zur sagenhaften versunkenen Insel Hy-Brasil aus der keltischen Mythologie auf und Earendil der Seefahrer erinnert an den irischen Heiligen Brendan den Reisenden (Vgl. Lyman-Thomas 2014, 276f.; Vgl. Flieger 2005, 121–136). 18 Besonders deutlich wird dieser Zusammenhang durch Tolkiens eigenhändige Illustration von Hobbingen in der originalen Ausgabe von The Hobbit, die im Vordergrund exponiert eine Mühle zeigt. Die Ähnlichkeit zur Sarehole Mill ist dabei nicht zu übersehen. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 62 irischen Kulturraum, wie von Mathew David Young unter Zuhilfenahme der Analysemethode Taggs konstruiert, zumindest relativieren. Das betrifft a) die nahegelegte Ähnlichkeit zum »Secret Wedding«-Thema aus BRAVEHEART sowie b) das Buch und c) den Score, die für den Leser und Hörer jeweils alternative Identifikationsangebote bieten. Zu a): Hat Mathew David Young durch Taggs Musematic Analysis scheinbar plausibel eine Ähnlichkeit zwischen Shores und Horners BRAVEHEARTScore nachgewiesen, soll an dieser Stelle noch einmal auf den Konstruktcharakter dieser Methode hingewiesen werden unter Verwendung eines ebenso plausiblen Gegenbeispiels. Dieses finden wir in der Hymne »I Vow to Thee, My Country«. Im Jahr 1921 versah Komponist Gustav Holst den 1908 von Cecil Spring Rice verfassten und 1918 umgeschriebenen und umbenannten Text (ursprünglich mit dem Namen »Urbs Dei« bzw. »The Two Fatherlands« betitelt) mit einer Melodie, die er dem Mittelteil des »Jupiter«-Satzes seiner Suite The Planets op. 32 (1914–1916) entlehnt hatte. Holst gilt neben Ralph Vaughan Williams als einer der führenden Komponisten des First British Folk Revival, indem er unter anderem englische Folksongs als Grundlage für seine Kompositionen nutzte. Zu nennen wären in diesem Zusammenhang beispielsweise die Cecil Sharp19 gewidmete A Somerset Rhapsody20 (1907) oder auch Three Folk Tunes (March on Three Folk Tunes) (1905). Die als unisono-Melodie mit Orchesterbegleitung harmonisiert und vorgesehene unter der Komposition Nummer 188 wurde als von Hymne mit Holst der 19 Sharp gilt als bekanntester Sammler und Protagonist des First British Folk Revival. 20 Diese Komposition basiert auf drei Folksongs, die sich in den Sammlungen Sharps finden lassen. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 63 Melodiebezeichnung »Thaxted«21 in die 1925 edierte Sammlung Songs of Praise (Dearmer et al.) übernommen. Abb. 12a: Beginn von »I Vow to Thee, My Country« Abb. 12b: »The Secret Wedding«-Thema (ohne Vorschlagnoten) »I Vow to Thee, My Country« (transponiert nach C-Dur) zeigt zu Beginn eine deutliche Übereinstimmung mit dem »The Secret Wedding«-Thema, auch was den weiteren an- und absteigenden Verlauf der Melodie betrifft, die in beiden Fällen, nach vier bzw. zwei Takten zum Ausgangston zurückkehrt. Wenn also der Beginn des Auenland-Themas den Hörer an das »The Secret Wedding«-Thema aus BRAVEHEART denken lassen soll, um daraufhin dessen schottisch-keltische Konnotation auf Shores Komposition zu übertragen, ist es ebenso denkbar, dass der möglicherweise musikalisch anders sozialisierte (möglicherweise sogar englische) Hörer eben nicht an BRAVEHEART und Schottland, sondern an Holst und England denkt, denn die Hymnen-Sammlung Songs of Praise wurde vorwiegend für den Schulgebrauch konzipiert, war in englischen Schulen weit verbreitet und ist auch sonst eng mit der anglikanischen Kirche verbunden. Unabhängig davon ist »I Vow to Thee, My Country« ein patriotisches britisches Lied, 21 Benannt nach dem Ort Thaxted in Essex, in dem Holst zwischen 1917 und 1925 lebte. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 64 welches sowohl zur Beerdigung von Winston Churchill und Margaret Thatcher als auch während der Hochzeitszeremonie von Charles und Diana gespielt worden ist. Es ist also ganz sicher kein distinktiv schottisches. An dieser Stelle wird eine weitere Problematik offensichtlich, dass nämlich auch die Vergleichsmusiken, die herangezogen wurden, um eine vermeintliche ›Englishness‹ bzw. ›Celticness‹ des Auenland-Themas plausibel erscheinen zu lassen, auch ihrerseits das Produkt eines Aushandlungsprozesses sind. Auch sie erhalten ihre scheinbare ›Identität‹ letztlich aufgrund von kulturellen Zuschreibungen durch den Hörer.22 Zu b): In der Buchvorlage (The Fellowship of the Ring, Kapitel 12 ›Flight to the Ford‹) stimmt Frodos treuer Begleiter Samweis Gamdschie auf Bitten der Gemeinschaft ein Lied an. Da sich die Gefährten just an dem Platz befinden, an dem sich im Hobbit, der Vorgeschichte zu The Lord of the Rings, drei Trolle im Sonnenlicht zu Stein verwandeln, intoniert Sam folgendes Lied: Troll sat alone on his seat of stone, And munched and mumbled a bare old bone; For many a year he had gnawed it near, For meat was hard to come by. Done by! Gum by! In a case in the hills he dwelt alone, And meat was hard to come by. (Tolkien 1994, 201) Dieses Gedicht wurde von Tolkien zeitlich weit vor der Veröffentlichung des Romans geschrieben. Die früheste Version stammt aus dem Jahr 1926, damals noch unter dem Titel ›Pero & Podex‹. Ein erster Abdruck erfolgte 22 Wenn in der vorliegenden Publikation also von schottischen, irischen oder englischen Melodien gesprochen wird, ist diese Relativierung immer mit zu bedenken. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 65 1936 auf private Initiative von Studenten der Leeds University in der Sammlung Songs for the Philologists. Diese enthielt unter anderem 13 Gedichte Tolkiens, so auch das Lied vom Steintroll unter dem Namen ›The Root and the Boot‹. Interessanterweise hat Tolkien schon in der damaligen ersten Druckversion eine Melodie vorgeschlagen, zu der die Verse gesungen werden sollten – nicht etwa eine irische oder schottische, sondern die eines englischen Folksongs mit dem Titel »The fox and the goose« bzw. »The fox went out on a winter’s night«23, dessen textliche Ursprünge bis in das 15. Jahrhundert reichen (Scull/Hammond 2006, 967f). Wie sehr sich die Melodie dieses Folksongs in der Vorstellung Tolkiens mit dem Gedicht vom Steintroll verbunden hat, zeigt eine Audioaufnahme aus dem Jahr 1952. Bereits zwei Jahre vor der Erstveröffentlichung von The Lord of the Rings liest Tolkien im Hause seines Freundes George Sayer Auszüge aus seinem Roman und dem bereits 1937 erschienenen The Hobbit, darunter auch »Sam‘s Rhyme of the Troll« (Track 13) (Tolkien 2001). 24 Dass es sich bei dem von Tolkien eingesungenen Lied tatsächlich um den genannten Folksong handelt, kann ein Vergleich der Melodien klären. Als Referenz dazu dient eine Transkription von Cecil Sharp aus dem Jahr 1913 mit dem Titel »(Daddy) Fox & the Goose« aus der Cecil Sharp Manuscript Collection in der Vaughan Williams Memorial Library (Sharp 1913). 25 Für eine bessere Vergleichbarkeit der Versionen sind die Melodie von Tolkiens 23 Wie bei mündlich tradierten Folksongs üblich, gibt es für dieses Lied textliche Varianten und verschiedene Titelbezeichnungen. So ist der Song auch unter dem Namen »The fox went out on a chilly night« bekannt (Roud No. 131). 24 Die Erstveröffentlichung dieser Aufnahmen erfolgte durch Caedmon im Jahr 1975. 25 Dass Sharp diesen Song 1913 transkribieren konnte, zeigt, dass das Lied während Tolkiens Kindheit und Jugend tatsächlich gesungen wurde (zumindest in Südengland) und er damit wahrscheinlich schon in jungen Jahren in Kontakt gekommen ist. Vgl. auch Bratman 2010, 141f. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 66 Gesang von D-Dur nach G-Dur transponiert und die Notenwerte augmentiert worden. Abb. 13a: »Sam’s Rhyme of the Troll« Abb. 13b: »The fox and the goose« Die markierten Abschnitte weisen deutliche Übereinstimmungen entweder in der Melodik oder Rhythmik auf. Tolkiens Wahl des englischen Folksongs Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 67 belegt neben seinen eigenen Aussagen, dass seine Vision vom Auenland eine im Wesentlichen englisch geprägte ist. Es an dieser Stelle bei einer simplen dichotomen Gegenüberstellung: Shore/celtic vs. Tolkien/englisch zu belassen, würde dennoch zu kurz greifen, denn auch Howard Shores offenkundig schottisch-irisch anmutendes Auenland-Thema basiert – über Umwege – auf einer englischen Folk Tune. Zu c): In einer der zentralen Szenen von THE FELLOWSHIP OF THE RING überwindet Samweis Gamdschie seine Furcht vor dem Unbekannten und entfernt sich so weit von zu Hause wie noch niemals zuvor (0:43:12– 0:43:29). Passenderweise wandelt sich in diesem Moment das AuenlandThema sowohl in der Instrumentierung als auch in seiner Melodik. Im sogenannten ›Hymn-Setting‹ übernimmt das Horn (zusammen mit den hohen Streichern) die Melodie, wobei Shore auf Material des ›PensiveSettings‹ zurückgreift, der Gestus jedoch ein nobler und erhabener ist. Dieser Moment markiert zum einen die beginnende Transformation des Charakters Samweis Gamdschie, der sich letztlich als der ›eigentliche‹ Held der Geschichte herausstellt, zum anderen beginnt sich in diesem Augenblick die Gemeinschaft des Ringes zu formieren (Adams 2010, 149). Abb. 14: ›Hymn-Setting‹ des Auenland-Themas Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 68 Interessant für die vorliegende Untersuchung ist die Melodik des in dieser Szene etablierten ›Hymn-Setting‹. Den Namen trägt diese Variation des Auenland-Themas nicht zufällig. Sie ähnelt sehr der bekannten Hymnenmelodie »Terra Beata« von Franklin L. Sheppard aus dem Jahre 1915. Damit vertonte Sheppard den vom New Yorker Pfarrer Maltbie Davenport Babcock im Jahr 1901 verfassten Text »This is my Father’s World«. Die Melodie erschien mit diesem Text erstmals im von Sheppard selbst herausgegebenen Hymnal Alleluia unter der Nummer 180 (Sheppard 1915). In dieser Kombination ist das Lied noch heute eine beliebte christliche Hymne in Amerika. Im Vergleich der beiden Themen wird deutlich, dass der Beginn des Vordersatzes bezüglich der Melodik identisch ist und sich auch die beiden folgenden Takte im Melodieverlauf stark ähneln. Abb. 15: Beginn von »This is My Father’s World« Laut Ethel K. Porter ließ sich Sheppard bei seiner Komposition jedoch von einer noch älteren Melodie inspirieren, die er als kleiner Junge von seiner Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 69 Mutter gelernt hatte (Porter 1966, 366). Diese sei die Tune »Rusper«, als 4stimmiger Satz zuerst veröffentlicht im Jahre 1906 in The English Hymnal (Dearmer et al.) unter der Nummer 379. Auch dieser Satz verweist auf eine englische Melodie als Grundlage, wobei sich Rusper auf den gleichnamigen Ort in West Sussex zwischen London und Brighton beziehen muss, in dem die Tune offenbar gesammelt worden ist. Abb. 16: Beginn von »Rusper« »Rusper« zeigt in seiner Melodik (insbesondere in den ersten beiden der abgebildeten Takte, jedoch auch im weiteren Verlauf) eine deutliche Ähnlichkeit zu »Terra Beata«, was die Vermutung Porters zumindest plausibel erscheinen lässt, zumal auch Sheppard anmerkt, dass »Terra Beata« lediglich ein Arrangement sei. Die beinahe identische melodische Ausformung in »Terra Beata« und dem Auenland-Thema in der Hornvariation lässt keinen anderen Schluss zu, als dass Shore diese Hymnenmelodie gekannt haben muss, was auch die Bezeichnung ›Hymn-Setting‹ nahelegt. Auch wenn dem Komponisten möglicherweise die englischen Wurzeln dieser Melodie nicht bekannt waren, ist doch festzuhalten, dass die Komposition Shores nicht so eindeutig mit dem Label ›Celtic‹ oder ›Englisch‹ belegt werden kann. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 70 Es ist hoffentlich auf den vorangegangenen Seiten zweierlei deutlich geworden. Erstens: Eine vermeintlich einfache Zuschreibung des Auenlandmotivs zum schottisch-irischen Kulturraum unter alleiniger Berücksichtigung der Instrumentierung ist aus musikwissenschaftlicher Perspektive haltlos. Dass es keine reinen Musikkulturen gibt, ist an anderer Stelle bereits festgestellt worden. Sowohl im Song- und Tune-Repertoire als auch bezüglich musikalischer Charakteristika und in der Instrumentierung gibt es zwischen der schottisch-irischen und der englischen Folk Music Überschneidungen. So ist selbstverständlich die Fiddle Teil der englischen ebenso wie der irischen oder schottischen Musik, was seit dem 19. Jahrhundert auch auf das Akkordeon oder die Concertina zutrifft. Dass das Auenland-Thema, wie Bratman eingangs konstatiert, »not English at all« sei, konnte, so hoffe ich, auch schlüssig widerlegt werden. Zweitens: Die Bedeutung und das Wesen von Musik im Allgemeinen und von Filmmusik im Speziellen, ihre ›Identität‹, ist im Wesentlichen ein Konstrukt und liegt in den kulturell bestimmten Hörerfahrungen des Rezipienten begründet. Abhängig vom Blickwinkel kann das Auenland-Thema vermeintlich irischschottisch oder aber auch englisch geprägt sein bzw. ist es in der Lage eine solcherart geprägte kulturelle Identität zu reflektieren. Letztlich sind essentielle ethnische Zuschreibungen und statische Kategorisierungen immer problematisch. Mit diesem Fazit wäre sicher auch Tolkien einverstanden. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 71 Literatur Adams, Doug (2010) The Music of the Lord of the Rings Films. Van Nuys, CA: Carpentier. Bratman, David (2010) Liquid Tolkien: Music, Tolkien, Middle-earth, and More Music. In: Middle-earth Minstrel. Essays on Music in Tolkien. Hrsg. v. Bradford Lee Eden. Jefferson, NC u. a.: McFarland, S. 140–170. Cannon, Roderick D. (1995) The Highland Bagpipe and its Music. Edinburgh: John Donald. Collinson, Francis (1966) The Traditional and National Music of Scotland. London: Routledge and Kegan Paul. Dearmer, Percy/Shaw, Martin/Williams, Ralph V. (1925) Songs of Praise. London: Oxford University Press. Dearmer, Percy/Williams, Ralph V./Birkbeck, William J. et al. (1933) The English Hymnal. 2nd Edition. London: Oxford University Press. Flieger, Verlyn (2005) Interrupted Music. The Making of Tolkien’s Mythology. Kent, OH: The Kent State University Press. Herndon, Marcia/McLeod, Norma (1982) Music as Culture. Second Edition. Darby, PA: Norwood Editions. Kahlas-Tarkka, Leena (2014) Finnish: The Land and Language of Heroes. In: A Companion to J. R. R. Tolkien. Hrsg. v. Stuart D. Lee. Malden, MA/Oxford/Chichester: Wiley Blackwell. S. 259–271. Lissa, Zofia (1975) Heinrichshofen. Neue Aufsätze zur Musikästhetik. Wilhelmshaven: Lyman-Thomas, J. S. (2014) Celtic: »Celtic Things« and »Things Celtic« – Identity, Language and Mythology. In: A Companion to J. R. R. Tolkien. Hrsg. v. Stuart D. Lee. Malden, MA/Oxford/Chichester: Wiley Blackwell. S. 272–285. Melhuish, Martin (1998) Celtic Tides. Traditional Music in a new Age. Kingston, Ontario: Quarry Press. Merriam, Alan P. (1964) The Anthropology of Music. Chicago: Northern University Press. Porter, Ethel K. (1966) Guide to the Pilgrim Hymnal. Philadelphia/Boston: United Church Press. Reiss, Scott (2003) Tradition and Imaginary: Irish Traditional Music and the Celtic Phenomenon. In: Celtic Modern. Music at the Global Fringe, Hrsg. v. Martin Stokes/Philip V. Bohlman, Lanham, MD/Oxford: The Scarecrow Press. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 72 Rosand, Ellen (1979) The Descending Tetrachord: An Emblem of Lament. In: The Musical Quarterly 65, S. 346–359. Scott, Priscilla (2013) ›With heart and voice ever devoted to the cause‹: Women in the Gaelic Movement, 1886–1914. PhD Thesis, University of Edinburgh. Scull, Christina/Hammond, Wayne G. (2006) The J. R. R. Tolkien Companion and Guide, Bd. 2: Reader’s Guide, London: HarperCollins. Sharp, Cecil (1913) »(Daddy) Fox & the Goose«, Cecil Sharp Manuscript Collection, VWML, http://www.vwml.org/record/CJS2/10/2824 (Stand: 02.09.2017). Sheppard, Franklin L. (1915) Alleluia. Philadelphia: The Westminster Press. Tagg, Philip (2000) Kojak: 50 Seconds of Television Music. Toward the Analysis of Affect in Popular Music. 2. Aufl. New York: Mass Media Music Scholars' Press. Tagg, Philip (2004) Film music, anti-depressants and anguish management, http://www.tagg.org/articles/xpdfs/jochen0411.pdf (Stand: 31.08.2017). Temperley, Nicholas/Temperley, David (2011) Music-Language Correlations and the ›Scotch Snap‹. In: Music Perception 29, S. 51–63. Tolkien, J. R. R. (1994 [1954]) The Fellowship of the Ring. Being the First Part of the Lord of the Rings. New York: Houghton Mifflin. Tolkien, J. R. R. (2000 [1981]) The Letters of J. R. R. Tolkien. Hrsg. v. Humphrey Carpenter. New York: Houghton Mifflin. Young, Mathew David (2007) Projecting Tolkien’s Musical Worlds: A Study of Affect in Howard Shore’s Soundrack to Lord of the Rings. Saarbrücken: VDM Verlag Dr. Müller. Filmografie/Videos: BRAVEHEART (USA 1995, Mel Gibson). Emimusic (10.03.2009) Enigma – Return to Innocence, http://www.youtube.com/watch?v=jyLXjPdBKEA (Stand: 31.08.2017). Jackson, Peter (2002) Interview, THE LORD OF THE RINGS: THE FELLOWSHIP OF THE RING – Special Extended Edition, The Appendices: Part 2 – From Vision to Reality, »Music for Middle-Earth«. Tagg, Philip (11.05.2011) Scotch Snaps – The Big Picture, http://tagg.org/Clips/ScotchSnap/ScotchSnap.mp4 (Stand: 31.08.2017). Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 73 THE LORD OF THE RINGS: THE FELLOWSHIP OF THE RING – Special Extended DVD Edition (USA/NZ 2002, Peter Jackson). Diskografie: Shore, Howard (2011) The Lord of the Rings Symphony, Howe Records. Tolkien, J. R. R. (2001) The J. R. R. Audio Collection, Caedmon. Various (2008) Authentic Music of the American Indian, Essential Media Group. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 74 Empfohlene Zitierweise Schröder, Martin: Möglichkeiten kultureller Identifikation in Howard Shores Soundtrack zu Peter Jacksons THE LORD OF THE RINGS. In: Kieler Beiträge zur Filmmusikforschung 13 (2017), S. 41–75, DOI: https://doi.org/10.59056/kbzf.2017.13.p41-75. Kieler Beiträge zur Filmmusikforschung (ISSN 1866-4768) Die Inhalte dieses Werks werden unter der Lizenz CC BY 4.0 Creative Commons Namensnennung 4.0 zur Verfügung gestellt. Hiervon ausgenommen ist das Bildmaterial, das abweichenden, in den Bildlegenden spezifizierten Bestimmungen unterliegt. Kieler Beiträge zur Filmmusikforschung, 13, 2017 // 75
https://openalex.org/W2090946017	https://bmcresnotes.biomedcentral.com/counter/pdf/10.1186/1756-0500-4-193	English	null	Prevalence and risk factors of Hepatitis C among individuals presenting to HIV testing centers, Hawassa city, Southern Ethiopia	BMC research notes	2,011	cc-by	4,036	* Correspondence: techalew03@yahoo.com 1Department of Medical Laboratory Science, Hawassa University, P. O. Box: 1560 Hawassa, Ethiopia Full list of author information is available at the end of the article SHORT REPORT Open Access Alemayehu et al. BMC Research Notes 2011, 4:193 http://www.biomedcentral.com/1756-0500/4/193 Alemayehu et al. BMC Research Notes 2011, 4:193 http://www.biomedcentral.com/1756-0500/4/193 © 2011 Shimelis et al; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Background: Hepatitis C virus (HCV), either alone or in combination with Human Immunodeficiency virus (HIV), constitutes a major public health concern. This study was conducted to describe the prevalence and risk factors for HCV infection in people with and without HIV infection. Methods: Blood samples and data on socio-demographic and risk factors for HCV infection were collected from consecutive 400 HIV- positive and 400 HIV- negative individuals attending HIV testing centers in Hawassa city, from October to December, 2008. All sera were tested for antibody to HCV infection (anti-HCV) using enzyme linked immunosorbent assay (ELISA). Sera positive for anti-HCV were further tested for viral ribonucleic acid (RNA) levels using real-time polymerase chain reaction. Results: The rate of anti-HCV positivity was 10.5% in the HIV- infected individuals compared with 6% in the HIV negative group (p = 0.002). HCV-RNA was detected in 9.1% of anti-HCV positive samples and rates were comparable between HIV- infected and HIV- non-infected individuals. There was no significant difference in odds of HCV infection in participants with and without HCV risk factors in either HIV sero-group. Conclusion: HIV infected individuals had significantly higher rate of anti-HCV although most of them showed no evidence of viraemia. Hence, while priority should be given for HIV infected patients, testing those with anti-HCV for HCV-RNA remains important. Keywords: Prevalence, hepatitis C virus, human immunodeficiency virus, co- infection Keywords: Prevalence, hepatitis C virus, human immunodeficiency virus, co- infection estimated 40 million individuals infected with HIV worldwide, about 4-5 million are chronically infected with HCV [4]. Prevalence and mode of transmission of each of the two viral infections are the main determi- nants of co-infection rate, and vary according to the geographic region. For instance, co-infection is high in patients that acquire the viruses through parenteral routes compared with those infected through sexual route [5]. Addisu Alemayehu1, Yayehyirad Tassachew1, Zufan Sisay2 and Techalew Shimelis1* Addisu Alemayehu1, Yayehyirad Tassachew1, Zufan Sisay2 and Techalew Shimelis1* Addisu Alemayehu1, Yayehyirad Tassachew1, Zufan Sisay2 and Techalew Shimelis1* Methods where testing for viral ribonucleic acid (RNA) using real-time polymerase chain reaction (PCR) (Abbott Real- Time™HCV) was performed. According to the manu- facturer’s guidelines, the real-time PCR test has detection sensitivity of a viral load as low as 12 IU/ml of serum. All samples were tested in accordance with the manufacturer’s instructions and in duplicates to ensure reproducibility. Despite the possibility that the ELISA test introduces false positive result, we interpreted a positive anti-HCV test as HCV-exposure. Presence of HCV-RNA indicates active viral replication. A sero-prevalence study of HCV was conducted in peo- ple with and without HIV infection attending HIV test- ing centers (Hawassa Referral Hospital, Hawassa Health Centre, Organization for Social Services for AIDS (OSSA), and Beth-Zeta Hospital) from October to December, 2008. These centers are located in Hawassa, the capital city of the Southern Nations, Nationalities and Peoples’ Regional State and one of the administra- tive regions in Ethiopia. The centers provide voluntary HIV counseling and testing services using two approaches: a client-initiated approach to serve people seeking to know their HIV status, and a provider- initiated approach which enables health care provider offer targeted medical services. Clients who are tested HIV positive are usually referred to the Hawassa Refer- ral Hospital. Clinical and immunological assessments (CD4+ T cell count) at enrollment and at six-monthly follow-up visits identify those who are eligible for anti- retroviral therapy (ART). Those receiving ART monitor their disease status on a regular basis. Services including HIV counselling and testing, clinical and immunological assessments as well as ART are provided free of charge. Data entry and analysis were performed using SPSS Version-16. Descriptive summary was presented in terms of mean, range, and proportions depending on the scale of the variable. Pearson’s Chi-square test was used to assess differences in proportion of anti-HCV marker between HIV-infected and HIV-non-infected individuals. The strength of association between HCV status and risk factors was measured by calculating odds ratios. A given statistical test was reported significant whenever it resulted in a p-value < 0.05. The study was approved by the Ethics Committees of Aklilu Lemma Institute of Pathobiology, Addis Ababa Uni- versity and the College of Medicine and Health Sciences, Hawassa University. All participants gave informed written consent, and doctors managed HCV-positive cases. Introduction Hepatitis C virus (HCV) represents a major public health burden in both industrialized and developing countries. The prevalence of HCV infection is esti- mated to be 2.2-3.0% (130-170 million people) world- wide [1]. Region-specific estimates range from < 1.0% in Northern Europe to > 2.9% in Northern Africa, with the highest prevalence (15%-20%) reported in Egypt [2]. About 15 to 20% of HCV infections progress to potentially serious cirrhosis and end-stage liver disease [3]. Due to shared mode of transmission, co-infection of HCV and human immunodeficiency virus (HIV) has become a growing public health concern. Among the In Ethiopia, a previous population-based survey had reported a moderate prevalence (2%) of HCV infection [6]. However, information is scarce regarding the rate of HIV-HCV co-infection in our setting where HIV is pre- valent. This study was conducted to describe the preva- lence and risk factors for HCV infection in people with and without HIV infection. * Correspondence: techalew03@yahoo.com 1Department of Medical Laboratory Science, Hawassa University, P. O. Box: 1560 Hawassa, Ethiopia Full list of author information is available at the end of the article Alemayehu et al. BMC Research Notes 2011, 4:193 http://www.biomedcentral.com/1756-0500/4/193 Alemayehu et al. BMC Research Notes 2011, 4:193 http://www.biomedcentral.com/1756-0500/4/193 Alemayehu et al. BMC Research Notes 2011, 4:193 http://www.biomedcentral.com/1756-0500/4/193 Page 2 of 5 Methods p g In this study, consecutive HIV-positive individuals who were at their first enrollment at the ART clinic as well as those who had been receiving ART for at least six months were recruited prospectively. Consecutive clients, who were tested HIV negative at any of the cen- ters during the study period, were recruited for com- parative purposes. The sample size was estimated using HCV prevalence rates of 4.5% and 0.8% in HIV-positive and HIV-negative subjects, respectively [7]. With 95% level of confidence and 80% power of the study, the required sample size would be 800 (400 individuals in each HIV sero-group). In both HIV sero-groups, only individuals aged 15 years and above were included. Among HIV-positive clients, those found with a CD4+ T cell counts below 50 cells/mm3 were excluded from the study due to the unreliability of HCV serological tests in a state of severe immunosuppression [8]. Results Of the 402 HIV-positive and 404 HIV-negative indivi- duals approached during the study period, 2 HIV-posi- tive and 4 HIV-negative participants were excluded due to insufficient blood volume and refusal, respectively. The majority (94.7%) of HIV-infected participants were urban dwellers and married (48%) (Table 1). The mean age of HIV-infected participants was 31.7 years (range 18-70 years; SD 8.4) compared with 26.6 years (range 16-62 years; SD 7.3) in HIV- non-infected clients. The male to female ratio was 0.6:1 in HIV-positive and 0.9:1 in HIV-negative clients. Two-hundred (50%) HIV- infected participants were receiving ART. The prevalence of anti-HCV in HIV-positive indivi- duals was 10.5% compared to 6% in HIV-negative indivi- duals (p = 0.002) (Table 2). A comparable rate of anti- HCV marker was observed in HIV-infected individuals with or without ART. HCV- RNA was detected in 6 (9.1%) of all 66 anti-HCV positive samples and viral load ranged from 31 IU/ml - 4.294 million IU/ml of serum. Rate of HCV- RNA was similar among anti- HCV- positive participants with or without HIV infec- tion (7.1% versus 12.5%, respectively; p = 0.39), and in those HIV-infected with or without ART (5.3% versus 8.7%, respectively; p = 0.43). Counsellor nurses interviewed the study subjects using structured questionnaires on socio-demographic charac- teristics and risk factors of HCV infection such as his- tory of traditional and modern medical practices, sexual practices, history of sexual transmitted infections (STIs), liver disease, and family history of liver disease. Remain- ing sera left from HIV testing and CD4+ T cell mea- surement were stored at -80°C for the purpose of HCV testing. All sera were screened at the Hawassa Referral Hospital laboratory for HCV antibody (anti-HCV) using Enzyme Linked Immunosorbent Assay (ELISA) (HUMAN anti-HCV ELISA, Germany). Sera found to be positive for anti-HCV were transferred in cold boxes to the International Clinical Laboratories in Addis Ababa The exposure of HIV-infected and HIV-non-infected participants to different risk factors for HCV infection is Alemayehu et al. BMC Research Notes 2011, 4:193 http://www.biomedcentral.com/1756-0500/4/193 Page 3 of 5 Page 3 of 5 Page 3 of 5 summarised in Table 3. Histories of ear piercing and tooth extraction were the most frequently reported exposures in both HIV sero-groups. Among females with HIV infection, 22.5% reported having ahistory of abortion. Discussion The sero-prevalence of anti-HCV was 10.5% in HIV- positive and 6% in HIV-negative individuals. None of the risk factors included in the analysis significantly increased the odds of HCV exposure. The rate of HIV- HCV co-infection in the present study was higher com- pared to findings in an earlier report which investigated residents of Addis Ababa (4.5%) [7]. Although we found a lower co-infection rate compared with developed nations [9,10], the magnitude of HCV in Ethiopia would become a concern as HIV is more prevalent in our settings. Despite a weak association, we found a significant dif- ference in rates of anti-HCV positivity between HIV- infected and HIV-non-infected individuals. In agree- ment, a previous report showed that HIV positive inha- bitants in Addis Ababa had about five-fold rate of anti- HCV compared with HIV negative individuals [7]. How- ever, associations between HCV and HIV infections in our settings were not as strong as those reported from industrialized countries where HIV can occur among HCV infected people exclusively [10,11]. Unlike most western countries, risk factors such as repeated transfusions, haemodialysis, or injection drug use are very rare in Ethiopia where transmission of HIV is primarily through heterosexual exposure [12]. While it remains to be fully established, rising evidences point to implications of sexual activity in HCV transmission [13]. Sexual transmission of HCV that might be facilitated by HIV and other STIs has also been reported [14,15]. Therefore, the role of sexual transmission should not be neglected as it may contribute to the higher rate of anti- HCV among HIV-infected individuals. In the current study, the odds of HCV infection were not shown to be significantly affected by the exposure status of participants to various medical practices. This result contrasts with afinding from Egypt where risk fac- tors including blood transfusion, invasive medical proce- dures, and frequent injections significantly increased the odds of HCV infection [16]. This may be due to the strength of risk factors varying according to HCV epide- miology in a given geographical area. Moreover, unlike a report from Brazil [13], our investigation as well as a previous study from Egypt [16] showed that the odds of HCV infection was not significantly higher in summarised in Table 3. Histories of ear piercing and tooth extraction were the most frequently reported exposures in both HIV sero-groups. Among females with HIV infection, 22.5% reported having ahistory of abortion. Results Sero- prevalence rate of anti-HCV was higher in HIV-infected participants who reported having multi- Table 1 Socio-demographic characteristics of HIV-positive and HIV-negative individuals at Hawassa HIV testing centers, 2008 Characteristics HIV positive (%) HIV negative (%) Total (%) Residence Rural 21 (5.3) 17 (4.2) 38(4.8) Urban 379 (94.7) 383(95.8) 762(95.2) Sex Male 151(37.8) 191(47.8) 342 (42.8) Female 249 (62.2) 209(52.2) 458 (57.2) Age (years) <20 6(12.5) 42(87.5) 48 (6) 20-29 164(40.5) 241(59.5) 405 (50.6) 30-39 161(63.6) 92(36.4) 253 (31.6) 40-49 55(75.3) 18(24.7) 73 (9.1) ≥50 14(66.7) 7(33.3) 21 (2.6) Marital status Married 192(48) 97(24.2) 289 (36.1) Single 60(15) 274(68.5) 334 (41.8) Divorced 60(15) 20(5.0) 80 (10 ) Widowed 88(22) 9(2.2) 97 (12.1) Religion Orthodox 250(62.5) 163(40.8) 413 (51.6) Protestant 112(28.0) 199(49.8) 311 (38.9) Muslim 31(7.8) 37(9.2) 68 (8.5) Other 7(1.8) 1(0.2) 8(1.0) Educational status Educational stat Illiterate 57(14.2) 19(4.8) 76 (9.5) Primary school 169(42.2) 72(18.0) 241 (30.1) Secondary school 150(37.5) 139(34.8) 289 (36.1) Diploma 16(4.0) 119(29.8) 135 (16.9) Degree 8(2) 51(12.8) 59 (7.4) Table 1 Socio-demographic characteristics of HIV-positive and HIV-negative individuals at Hawassa HIV testing centers, 2008 Discussion Sero- prevalence rate of anti-HCV was higher in HIV-infected participants who reported having multi- ple sexual partners. There was no significant difference in odds of HCV infection in participants with and with- out HCV risk factors in either HIV sero-group (Table 3). Table 2 Distribution of anti-HCV and HCV-RNA in HIV-positive and HIV-negative individuals at Hawassa HIV testing centers, 2008 HCV Marker Total tested Number (%) positive HIV-positive HIV-negative On ART ART naives Total Tested +ve (%) Tested +ve (%) Tested +ve (%) Tested +ve (%) Anti-HCV 800 66 (8.2) 200 19 (9.5) 200 23 (11.5) 400 42 (10.5) 400 24 (6) HCV-RNA 66 6 (9.1) 19 1 (5.3) 23 2 (8.7) 42 3 (7.1) 24 3 (12.5) f anti-HCV and HCV-RNA in HIV-positive and HIV-negative individuals at Hawassa HIV testing Alemayehu et al. * having more than one sexual partner; have ever been diagnosed for sexual transmitted infections (excluding HIV); * ever had minor or major surgical operation; **** have ever been diagnosed for liver disease References 8. Toro C, Simon A, Amor A, Jimenez V, Rios P, Soriano V: HCV seronegative infections in HIV patients with profound immunosuppresion. Fourth International Workshop on HIV and Hepatitis Co-infection Madrid, Spain; 2008. 9. Buxton AJ, Yu A, Kim HP, Spinelli JJ, Kuo M, Alvarez M, Gilbert M, Krajden M: HCV co-infection in HIV positive population in British Columbia, Canada. BMC Public Health 2010, 10:225-236. 9. Buxton AJ, Yu A, Kim HP, Spinelli JJ, Kuo M, Alvarez M, Gilbert M, Krajden M: HCV co-infection in HIV positive population in British Columbia, Canada. BMC Public Health 2010, 10:225-236. 10. Rockstroh JK, Mocroft A, Soriano V, Tural C, Losso HM, Horban A, Kirk O, Phillips A, Ledergerber B, Lundgren J, for the EuroSIDA Study Group: Influence of hepatitis C virus infection on HIV-1 disease progression and response to highly active antiretroviral therapy. J Infect Dis 2005, 192:992-1002. 11. Updated US Public Health Service guidelines for the management of occupational exposures to HBV, HCV and HIV and recommendations for postexposure prophylaxis. MMWR 2003, 50:1-42. In conclusion, HIV-infected individuals had signifi- cantly higher positivity rate of anti-HCV although most of them had no sign of viraemia. Hence, with priority given for HIV- infected patients, testing those with anti- HCV for HCV-RNA is important in order to manage HIV-HCV co-infected patients appropriately. Further studies on the epidemiology of HCV infection in our settings are imperative to plan intervention methods systematically. 12. Ethiopian Public Health Association (EPHA). Identifying HIV/AIDS, sexually transmitted infections and tuberculosis research gaps and priority setting agenda in Ethiopia; 2005. g p 13. Brandão BMA, Fuchs CS: Risk factors for hepatitis C virus infection among blood donors in southern Brazil: a case-control study. BMC Gastroenterol 2002, 2:18-26. 13. Brandão BMA, Fuchs CS: Risk factors for hepatitis C virus infection among blood donors in southern Brazil: a case-control study. BMC Gastroenterol 2002, 2:18-26. 14. Shev S, Widell A, Bergström T, Hermodsson S, Lindholm A, Norkrans G: Herpes simplex virus-2 may increase susceptibility of the sexual transmission of hepatitis C. Sex Transm Dis 1995, 22:210-216. 14. Shev S, Widell A, Bergström T, Hermodsson S, Lindholm A, Norkrans G: Herpes simplex virus-2 may increase susceptibility of the sexual transmission of hepatitis C. Sex Transm Dis 1995, 22:210-216. 15. Mohsen AH, Easterbrook P, Taylor CB, Norris S: Hepatitis C and HIV-1 coinfection. Gut 2002, 51:601-608. 16. Discussion BMC Research Notes 2011, 4:193 http://www.biomedcentral.com/1756-0500/4/193 Page 4 of 5 Page 4 of 5 Table 3 Anti-HCV positivity in relation to HCV risk factors in HIV-positive and HIV- negative individuals at Hawassa HIV testing centers, 2008 Characteristics HIV-positive HIV-negative No (%) tested No (%) positive for anti-HCV Crude odds ratio (95% CI) No (%) tested No (%) positive for anti-HCV Crude odds ratio (95% CI) Hospital admission No 300(75.0) 35(11.7) 1 342 (85.5) 21(6.1) 1 Yes 100(25.0) 7(7.0) 0.57(0.25-1.33) 58 (14.5) 3(5.2) .83(0.24-2.89) Blood transfusion No 388(97.0) 41(10.6) 1 398(99.5) 24(6.0) Yes 12(3.0) 1(8.3) 0.77(0.10-6.11) 2(0.5) 0 - Unsafe injection No 375(93.8) 40(10.7) 1 384(96.0) 24(6.2) Yes 25(6.2) 2(8.0) 0.73(0.17-3.20) 16(4.0) 0 - Multiple sexual partners * No 315(78.8) 31(9.8) 1 344(86.0) 18(5.2) 1 Yes 85(21.2) 11(12.9) 1.3(0.65-2.84) 56(14.0) 6(10.7) 2.1(0.82-5.74) Sexual transmitted infections No 282(70.5) 26(9.2) 1 380(95.0) 23(6.1) 1 Yes 118(29.5) 16(13.6) 1.54(0.80-3.0) 20(5.0) 1(5.0) 0.82(0.11-6.38) Tooth extraction No 278(69.5) 27(9.7) 1 286(71.5) 18(6.3) 1 Yes 122(30.5) 15(12.3) 1.30(0.67-2.55) 114(28.5) 6(5.3) 0.83(0.32-2.14) Catheterization No 393(98.2) 41(10.4) 1 399(99.8) 24(6.0) - Yes 7(1.8) 1(14.3) 1.43(0.17-12.18) 1(0.2) 0 Scarification No 373(93.2) 39(10.5) 1 367(91.8) 23(6.3) 1 Yes 27(6.8) 3(11.1) 1.07(0.31-3.72) 33(8.2) 1(3.0) 0.47(0.06-3.58) Abortion No 193(77.5) 18(9.3) 1 195(93.3) 11(5.6) 1 Yes 56(22.5) 7(12.5) 1.39(0.55-3.51) 14(6.7) 1(7.1) 1.29(0.15-10.8) Ear piercing No 240(60.0) 22(9.2) 1 292(73.0) 16(5.5) 1 Yes 160(40.0) 20(12.5) 1.42(0.74-2.69) 108(27.0) 8(7.4) 1.38(0.57-3.32) Surgery * No 376(94.0) 40(10.6) 1 388(97.0) 24(6.2) Yes 24(6.0) 2(8.3) 0.76(0.17-3.37) 12(3.0) 0 - Blood letting No 390(97.5) 41(10.5) 1 398(99.5) 24(6.0) Yes 10(2.5) 1(10.0) 0.95(0.12-7.66) 2(0.5) 0 - Tattooing No 365(91.2) 38(10.4) 1 376(94.0) 23(6.1) 1 Yes 35(8.8) 4(11.4) 1.11(0.37-3.32) 24(6.0) 1(4.2) 0.67(0.09-5.16) History of liver disease **** No 373(93.2) 41(11) 1 387(96.8) 23(5.9) 1 Yes 27(6.8) 1(3.7) 0.31(0.04-2.36) 13(3.2) 1(7.7) 1.32(0.16-10.6) Contact with a person having liver disease No 372(93.0) 39(10.5) 1 379(94.8) 24(6.3) Yes 28(7.0) 3(10.7) 1.03(0.30-3.55) 21(5.2) 0 - * having more than one sexual partner; have ever been diagnosed for sexual transmitted infections (excluding HIV); * ever had minor or major surgical operation; **** have ever been diagnosed for liver disease Table 3 Anti-HCV positivity in relation to HCV risk factors in HIV-positive and HIV- negative individuals at Hawassa HIV testing centers, 2008 Alemayehu et al. BMC Research Notes 2011, 4:193 http://www.biomedcentral.com/1756-0500/4/193 Page 5 of 5 Page 5 of 5 Author details 1D f 1Department of Medical Laboratory Science, Hawassa University, P. O. Box: 1560 Hawassa, Ethiopia. 2Aklilu Lemma Institute of Pathobiology, Addis Ababa University, P. O. Box: 1176, Addis Ababa, Ethiopia. 1Department of Medical Laboratory Science, Hawassa University, P. O. Box: 1560 Hawassa, Ethiopia. 2Aklilu Lemma Institute of Pathobiology, Addis Ababa University, P. O. Box: 1176, Addis Ababa, Ethiopia. 20. Shores NJ, Maida I, Soriano V, Núňez M: Sexual transmission is associated with spontaneous HCV clearance in HIV-infected patients. J hepatol 2008, 49:323-328. doi:10.1186/1756-0500-4-193 Cite this article as: Alemayehu et al.: Prevalence and risk factors of Hepatitis C among individuals presenting to HIV testing centers, Hawassa city, Southern Ethiopia. BMC Research Notes 2011 4:193. References Medhat A, Shehata M, Magder LS, Mickhali N, Abdel-Baki L, Nafeh M, Abdel- Hamid M, Strckland GT, Fix AD: Hepatitis C in a community in Upper Egypt: risk factors for infection. Am J Trop Med Hyg 2002, 66:633-638. Acknowledgements Th h ld lik The authors would like to thank the counsellor nurses and laboratory personnel in Hawassa Referral Hospital for their kind assistance during data collection. We extend our appreciation to Dr. Tekalign Debalke for valuable comments and Mrs Askale Tesfaye for laboratory works. We are also most grateful to the study subjects who demonstrated willingness to participate in the study. 17. Wolff FH, Fuchs SC, Barcellos NNT, Ricardo de Alencastro P, Ikeda R ML, Brandão BMA, Falavigna M, Fuchs DF: Co-infection by hepatitis C virus in HIV-infected patients in southern Brazil: genotype distribution and clinical correlates. PLoS ONE 2010, 5:e10494. 18. Orland J, Murphy LE, Wright JD, Kleinman S, for the NHLBI Retrovirus Epidemiology Donor Study (REDS) Group: Correlates of hepatitis C virus (HCV) RNA negativity among HCV-seropositive blood donors. Transfusion 2006, 46:469-475. We acknowledge the Government of Netherlands (Maastricht University) for considerable financial support; and the Aklilu Lemma Institute of Pathobiology, Addis Ababa University, and the International Clinical Laboratories for supplementary backing. 19. Thomas DL, Astemborski J, Rai RM, Anania FA, Schaeffer M, Galai N, Nolt K, Nelson KE, Strathdee SA, Johnson L, Laeyendecker O, Boitnott J, Wilson LE, Vlahov D: The natural history of hepatitis C virus infection: host, viral, and environmental factors. JAMA 2000, 284:450-456. Received: 14 February 2011 Accepted: 15 June 2011 Published: 15 June 2011 g p HCV infection commonly takes achronic course and studies have reported viraemia persisting in over 80% of infected individuals [17,18]. In the present study, how- ever, HCV-RNA was detected in 9.1% of anti-HCV posi- tive individuals. Such a lower rate of HCV-RNA is unusual and points to a need for further investigation in our settings to confirm or challenge this finding. Factors such as intact immune status [19] and heterosexual HCV transmission [20] were suggested to contribute to varying rates of spontaneous HCV clearance in different geographical regions. In contrast, our study lacked enough power to assess whether factors such as demo- graphic characteristics, routes of transmission, HIV and ART status were influencing rate of HCV clearance. References 1. Lavanchy D: The global burden of hepatitis C. Liver Int 2009, 29:74-81. 2. Alter MJ: Epidemiology of hepatitis C virus infection. Gastroenterol 2007, 13:2436-2441. 3. Liang TJ, Rehermann B, Seeff LB, Hoofnagle JH: Pathogenesis, natural history, treatment, and prevention of hepatitis C. Ann Intern Med 2000, 132:296-305. 4. Alter MJ: Epidemiology of viral hepatitis and HIV co-infection. J Hepatol 2006, 44:6-9. 4. Alter MJ: Epidemiology of viral hepatitis and HIV co-infection. J Hepatol 2006, 44:6-9. 5. Thimme R, Spangenberg H, Blum H: Hepatitis B or hepatitis C and human immunodeficiency virus infection. J Hepatol 2005, 42:37-44. 5. Thimme R, Spangenberg H, Blum H: Hepatitis B or hepatitis C and human immunodeficiency virus infection. J Hepatol 2005, 42:37-44. y p 6. Frommel D, Tekle-Haimanot R, Berhe N, Aussel L, Verdier M, Preux P, 6. Frommel D, Tekle-Haimanot R, Berhe N, Aussel L, Verdier M, Preux P, 6. Frommel D, Tekle-Haimanot R, Berhe N, Aussel L, Verdier M, Preux P, Denis F: A survey of antibodies to hepatitis C virus in Ethiopia. Am J Trop Hyg 1993, 49:435-439. Denis F: A survey of antibodies to hepatitis C virus in Ethiopia. Am J Trop Hyg 1993, 49:435-439. 7. Ayele W, Nokes DJ, Abebe A, Messele T, Dejene A, Enquselassie F, Rinke de Wit FT, Fontanet LA: Higher prevalence of anti-HCV antibodies among HIV positive compared to HIV negative inhabitants of Addis Ababa, Ethiopia. J Med Virol 2002, 68:12-17. 7. Ayele W, Nokes DJ, Abebe A, Messele T, Dejene A, Enquselassie F, Rinke de Wit FT, Fontanet LA: Higher prevalence of anti-HCV antibodies among HIV positive compared to HIV negative inhabitants of Addis Ababa, Ethiopia. J Med Virol 2002, 68:12-17. Findings in this study need to be interpreted in light of its methodological limitations. First, lost participants, who tested HIV positive at the centers and did not regis- ter at the ART clinic, may introduce selection bias. Sec- ond, there could be a survival bias as a result of deaths occurring among patients with viral hepatic C within the six months following the ART initiation. Third, lack of confirming ELISA positive test results using an immuno- blot test might lead to misinterpretation of anti-HCV false positive results as HCV exposures. 8. Toro C, Simon A, Amor A, Jimenez V, Rios P, Soriano V: HCV seronegative infections in HIV patients with profound immunosuppresion. Fourth International Workshop on HIV and Hepatitis Co-infection Madrid, Spain; 2008. Competing interests participants with tattoos or ear piercing compared with those without the respective risk factors in either HIV sero-groups. Competing interests The authors declare that they have no competing interests. Received: 14 February 2011 Accepted: 15 June 2011 Published: 15 June 2011 Authors’ contributions All authors contributed to the study design, AA carried out data collection and laboratory works; AA and TS performed data analysis and interpretation; all authors contributed to final write up and approved the manuscript.
https://openalex.org/W4226029121	https://cyberleninka.ru/article/n/methods-for-accuracy-increasing-of-solid-brittle-materials-fracture-toughness-determining/pdf	English	null	Methods for Accuracy Increasing of Solid Brittle Materials Fracture Toughness Determining	Pribory i metody izmerenij	2,022	cc-by	6,842	Received 10.11.2021 Accepted for publication 18.01.2022 Received 10.11.2021 Accepted for publication 18.01.2022 Abstract Method for determining of the fracture toughness of brittle materials by indentation is described. The critical stress intensity factor KIC quantifies the fracture toughness. Methods were developed and applied to improve the accuracy of KIC determination due to atomic force microscopy and na- noindentation. It is necessary to accurately determine parameters and dimensions of the indentations and cracks formed around them in order to determine the KIC . Instead of classical optical and scanning electron microscopy an alternative high-resolution method of atomic force microscopy was proposed as an imaging method. g g Three methods of visualization were compared. Two types of crack opening were considered: along the width without vertical displacement of the material and along the height without opening along the width. Due to lack of contact with the surface of the samples under study, the methods of optical and scan- ning electron microscopy do not detect cracks with a height opening of less than 100 nm (for optical) and less than 40–50 nm (for scanning electron microscopy). Cracks with opening in width are determined within their resolution. Optical and scanning electron microscopy cannot provide accurate visualization of the deformation area and emerging cracks when applying small loads (less than 1.0 N). The use of ato- mic force microscopy leads to an increase in accuracy of determining of the length of the indent diagonal up to 9.0 % and of determining of the crack length up to 100 % compared to optical microscopy and up to 67 % compared to scanning electron microscopy. The method of atomic force microscopy due to spatial three-dimensional visualization and high accuracy (XY ± 0.2 nm, Z ± 0.03 nm) expands the possibili- ties of using indentation with low loads. A method was proposed for accuracy increasing of KIC determination by measuring of microhard- ness from a nanoindenter. It was established that nanoindentation leads to an increase in the accuracy of KIC determination by 16–23 % and eliminates the formation of microcracks in the indentation. Keywords: fracture toughness, accuracy, indentation method, atomic force microscopy, nanoindentation. DOI: 10.21122/2220-9506-2022-13-1-40-49 DOI: 10.21122/2220-9506-2022-13-1-40-49 Адрес для переписки: Лапицкая В.А. Институт тепло- и массообмена имени А.В. Лыкова НАН Беларуси, ул. П. Бровки, 15, г. Минск 220072, Беларусь e-mail: vasilinka.92@mail.ru Address for correspondence: Lapitskaya V.A. A.V. Luikov Heat and Mass Transfer Institute of NAS of Belarus, P. Brovki str., 15, Minsk 220072, Belarus e-mail: vasilinka.92@mail.ru Для цитирования: V.A. Lapitskaya, T.A. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Methods for Accuracy Increasing of Solid Brittle Materials Fracture Toughness Determining 1A.V. Luikov Heat and Mass Transfer Institute of NAS of Belarus, P. Brovki str., 15, Minsk 220072, Belarus 2Belarusian National Technical University, Nezavisimosty Ave., 65, Мinsk 220013, Belarus 3Koszalin University of Technology, Sniadeckich str., 2, Koszalin 75-453, Poland 1A.V. Luikov Heat and Mass Transfer Institute of NAS of Belarus, P. Brovki str., 15, Minsk 220072, Belarus 2Belarusian National Technical University, Nezavisimosty Ave., 65, Мinsk 220013, Belarus 3Koszalin University of Technology, Sniadeckich str., 2, Koszalin 75-453, Poland 1A.V. Luikov Heat and Mass Transfer Institute of NAS of Belarus, P. Brovki str., 15, Minsk 220072, Belarus Received 10.11.2021 Accepted for publication 18.01.2022 Received 10.11.2021 Accepted for publication 18.01.2022 Abstract Kuznetsova, S.A. Chizhik, B. Warcholinski. Methods for Accuracy Increasing of Solid Brittle Materials Fracture Toughness Determining. Приборы и методы измерений. 2022. – Т. 13, № 1. – С. 40–49. DOI: 10.21122/2220-9506-2022-13-1-40-49 For citation: V.A. Lapitskaya, T.A. Kuznetsova, S.A. Chizhik, B. Warcholinski. Methods for Accuracy Increasing of Solid Brittle Materials Fracture Toughness Determining. Devices and Methods of Measurements. 2022, vol. 13, no. 1, рр. 40–49. DOI: 10.21122/2220-9506-2022-13-1-40-49 DOI: 10.21122/2220-9506-2022-13-1-40-49 40 Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Поступила 10.11.2021 Принята к печати 18.01.2022 Приведено описание метода определения вязкости разрушения хрупких материалов индентирова- нием. Количественно вязкость разрушения характеризуется критическим коэффициентом интенсивности разрушения KIC . Использование атомно-силовой микроскопии и наноиндентирования позволило разработать и применить способы повышения точности определения KIC . Для определения KIC необходимо точно определять параметры и размеры отпечатков индентирования и образованных вокруг них трещин. В качестве метода визуализации вместо классических оптической и сканирующей электронной микроскопий предложен альтернативный высокоразрешающий метод атомно-силовой микроскопии. Проведено сравнение трёх методов визуализации. Рассмотрено два типа раскрытия трещин: по ширине без смещения материала по вертикали и по высоте без раскрытия по ширине. Методы оптической и сканирующей электронной микроскопий из-за отсутствия контакта с поверхностью исследуемых образцов не определяют трещины с раскрытием по высоте менее 100 нм (для оптической) и менее 40–50 нм (для сканирующей электронной микроскопии). Трещины с раскрытием по ширине определяют в рамках своей разрешающей способности. Оптическая и сканирующая электронная микроскопии не могут обеспечить точную визуализацию области деформации и формирующихся трещин при применении малых нагрузок (меньше 1,0 Н). Применение атомно-силовой микроскопии приводит к повышению точности определения длины диагонали отпечатка до 9,0 % и определения длины трещины до 100 % по сравнению с оптической микроскопией и до 67 % по сравнению со сканирующей электронной микроскопией. Метод атомно-силовой микроскопии благодаря пространственной трёхмерной визуализации и высокой точности (по XY ± 0,2 нм, по Z ± 0,03 нм) расширяет возможности применения индентирования с применением низких нагрузок. Предложен способ повышения точности определения KIC за счёт измерения микротвёрдости с наноиндентора. Установлено, что наноиндентирование приводит к повышению точности определе- ния KIC на 16–23 % и исключает образование микротрещин в отпечатке. Ключевые слова: вязкость разрушения, точность, метод индентирования, атомно-силовая микро- скопия, наноиндентирование. Адрес для переписки: Лапицкая В.А. Институт тепло- и массообмена имени А.В. Лыкова НАН Беларуси, ул. П. Бровки, 15, г. Минск 220072, Беларусь e-mail: vasilinka.92@mail.ru Address for correspondence: Lapitskaya V.A. A.V. Luikov Heat and Mass Transfer Institute of NAS of Belarus, P. Brovki str., 15, Minsk 220072, Belarus e-mail: vasilinka.92@mail.ru Для цитирования: V.A. Lapitskaya, T.A. Kuznetsova, S.A. Chizhik, B. Warcholinski. Methods for Accuracy Increasing of Solid Brittle Materials Fracture Toughness Determining. Приборы и методы измерений. 2022. – Т. 13, № 1. – С. 40–49. DOI: 10.21122/2220-9506-2022-13-1-40-49 For citation: V.A. Lapitskaya, T.A. Kuznetsova, S.A. Chizhik, B. Warcholinski. Methods for Accuracy Increasing of Solid Brittle Materials Fracture Toughness Determining. Devices and Methods of Measurements. 2022, vol. 13, no. 1, рр. 40–49. DOI: 10.21122/2220-9506-2022-13-1-40-49 DOI: 10.21122/2220-9506-2022-13-1-40-49 Способы повышения точности определения вязкости разрушения твёрдых хрупких материалов при индентировании В.А. Лапицкая1,2, Т.А. Кузнецова1,2, С.А. Чижик1,2, Б. Вархолински3 1Институт тепло- и массообмена имени А.В. Лыкова НАН Беларуси, ул. П. Бровки, 15, г. Минск 220072, Беларусь 2Белорусский национальный технический университет, пр-т Независимости, 65, г. Минск 220013, Беларусь 3Кошалинский технологический университет, ул. Снядецких, 2, г. Кошалин 75-453, Польша В.А. Лапицкая1,2, Т.А. Кузнецова1,2, С.А. Чижик1,2, Б. Вархолински3 1Институт тепло- и массообмена имени А.В. Лыкова НАН Беларуси, ул. П. Бровки, 15, г. Минск 220072, Беларусь В.А. Лапицкая1,2, Т.А. Кузнецова1,2, С.А. Чижик1,2, Б. Вархолински3 1И б А В Л НАН Б В.А. Лапицкая1,2, Т.А. Кузнецова1,2, С.А. Чижик1,2, Б. Вархолински3 2Белорусский национальный технический университет, пр-т Независимости, 65, г. Минск 220013, Беларусь 3Кошалинский технологический университет, ул. Снядецких, 2, г. Кошалин 75-453, Польша Поступила 10.11.2021 Принята к печати 18.01.2022 Introduction The purpose of the work was to improve the ac- curacy of determining the critical stress intensity fac- tor KIC by using the methods of atomic force micros- copy and nanoindentation, to establish the influence of the visualization method of the deformation area, the calculation model, and microhardness values on the accuracy of determining KIC . Reducing of the measurement error of any physical quantity is always an urgent task. Deve- lopment of new methods, use of other physical techniques allows them to be applied to existing measurement methods to improve accuracy and re- duce the error in determining of physical quantities. In this work, we will present a method for measu- ring the physical and mechanical properties of hard brittle materials, namely, fracture toughness (or crack resistance). This parameter characterizes the limiting state of any solid material and the ability to resist crack development [1–6]. It is extremely important for various types of ceramics (carbide, oxide, nitride) [7, 8] and coatings based on them, which combine special physical-mechanical, ther- mophysical, bioinert, antioxidant and wear-resistant properties [2, 7]. One of the main tasks of ceramic production technology is to increase their strength, to prevent the likelihood of sudden brittle fracture, appearance of chips on the surface, or even the de- struction of the part. DOI: 10.21122/2220-9506-2022-13-1-40-49 41 Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Analysis of the method for determining fracture toughness by indentation Determination of fracture toughness KIC by in- dentation is based on the introduction of a diamond indenter in the form of a tetrahedral Vickers pyramid into the surface of the test sample (Figure 1) under a selected load depending on the material and size of the sample [3–6]. Performed at least three inden- tations at each load. The prints are visualized in an optical microscope (OM) or a scanning electron mi- croscope (SEM) after indentation, the length of the print diagonals d1 and d2 is determined, and the print half-diagonal length a = (d1+d2) / 4 is calculated. Measured the length of the cracks (l is the length of the crack near the indent, c is the length of the crack measured from the center of the indent) near each indent, and then determine the average values of the crack lengths for the sample. Determined the physi- cal and mechanical properties of the material (micro- hardness HV and elasticity modulus E ). On the ratio c/a determines the type of cracks (Palmquist cracks or median cracks) [6] around the prints after deter- mining the values of a, l, c, HV and E. A mathemati- cal calculation model is selected depending on the type of cracks and the critical stress intensity factor KIC is determined. The critical stress intensity factor KIC quanti- fies the fracture toughness [1, 3–6]. There are many methods to determine this characteristic (bending, torsion, rupture, etc.). However, all of them are of limited use due to the complexity or impossibility of preparing test samples with the required notch ge- ometry and are economically unprofitable. The in- dentation method [4–6, 9] does not require samples of complex shape. It uses thin sections and consists in the study of the deformation area on the material surface after indentation, followed by the calculation of KIC . Figure 1 – Method and indentation imprint with defined parameters Figure 1 – Method and indentation imprint with defined parameters s Figure 1 – Method and indentation imprint with defined parameters To accurately determine KIC , it is important both to use mathematical calculation models (adequate to the range not only of high loads from 1.0 N and above, but also to low loads of 0.01–0.5 N) and accu- rate experimental characteristics of the material and the deformation region. Analysis of the method for determining fracture toughness by indentation There are factors that affect Development of methods for reducing the error in determination of fracture toughness by the indentation method Development of methods for reducing the error in determination of fracture toughness by the indentation method Determination of fracture toughness by inden- tation is a computational and experimental method. 42 Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. the accuracy of determining the fracture toughness of a material by indentation. These include the fol- lowing factors: the accuracy of determining the length of the diagonals of the indentation mark and the length of the cracks around it; method for deter- mining the microhardness of a material. three-dimensional visualization (which neither OM nor SEM has), it becomes possible to determine the depth of the indentation imprint and cracks, as well as the height of their opening [11]. In AFM the resolution depends on the radius of curvature of the tip of the used probe and the nature of its interac- tion with the surface of the object (contact or semi- contact) [12]. The AFM resolution in XY is limited by the capabilities of the device and is 0.2–0.5 nm.i g Mathematical Model of Calculation. There are many mathematical models for determining the critical stress intensity factor KIC [10]. In most of the models used earlier in the literature, loads of more than 1 N were used. In cases where small loads (0.25–0.75 N) were used, the values of a, l and c were determined inaccurately, and because of this, at low loads, the difference in a, l and c was not detected. Justification of the choice and deter- mination of the correctness of mathematical models for calculating KIC and their ability to maintain the stability of KIC values in a wide range of loads, pri- marily at low loads (from 0.01 to 0.5 N), will expand the capabilities of the indentation method. Microhardness Definition. To determine the fracture toughness KIC , it is necessary to determine the microhardness and elastic modulus of the mate- rial under study. Indentations are carried out at vari- ous loads, three to five (minimum number) of inden- tations at each load to determine KIC . Development of methods for reducing the error in determination of fracture toughness by the indentation method The classical method for determining fracture toughness KIC uses the values of microhardness HV determined by the Vickers method (GOST 9377–81). Microhardness HV according to the Vickers method is determined by the formula (1) [13]: Visualization of the Deformation Area. OM and SEM are used in the classical approach to visualize indentation prints, determine the length of the diago- nals of the indentation print and the length of cracks around it on the surface of the material under study. Each of these methods allows, within its resolution, to determine the linear geometric dimensions of the indentation imprint and the formed cracks. H P d V mean = 1 854 2 . , (1) (1) where dmean is arithmetic mean of the diagonals of the imprint of a tetrahedral Vickers pyramid after in- dentation, m; P is indenter load, N. During indentation, all the energy consists of the elastic and plastic components of the de- formation and is spent on the formation of an im- print on the surface sample. The presence of cracks around the indentation indicates that part of the energy was spent on their formation in the sample, as well as the inability to reliably determine the micro- hardness of the sample material. Unreliable values of Vickers microhardness HV lead to incorrect deter- mination of KIC . The resolution of an optical microscope is cha- racterized by the smallest distance between two points that are visible separately in the microscope. This distance is directly proportional to the wave- length of the light radiation incident on the object under study. The minimum dimensions of a distin- guishable object are approximately equal to half the wavelength of the incident light. Standard op- tical microscopes use visible light. Objects around 0.25∙10–6 m or 250 nm in size can be seen under a microscope. An alternative method for determining micro- hardness can be the method of nanoindentation (NI). A Berkovich-type diamond indenter and loads of no more than 5 mN are used when measuring micro- hardness on a nanoindenter (ISO 14577-1:2015). The microhardness HBer by the NI method is deter- mined by the formula [14]: The resolution of the SEM is determined by the electron wavelengths and the numerical aperture of the system. The wavelength of electron radiation depends on its energy, which is affected by the ac- celerating voltage. Samples and equipment To compare the visualization methods of the deformation area, diagonals of the length of the in- dentation print and length cracks, we used: an opti- cal microscope MICRO-200 (JSC Planar, Republic of Belarus) and a lens with a magnification of 400×, SEM – JSM-7001F (JEOL, Japan) with resolu- tion in secondary electrons 1.2 nm (at an accelera- ting voltage of 30 kV) and AFM – Dimension Fast- Scan (Bruker, USA) with XY resolution ± 0.2 nm, Z ± 0.03 nm. A section of silicon carbide ceramic was used as a sample [7, 8]. Microhardness imprints on the sample were made using a PMT-3M microhardness tester (LOMO, Russia) with a Vickers tetrahedral dia- mond pyramid at a constant load of 1.0 and 2.0 N.l K l a H E H a IC V V =     ⋅    ⋅           − − 0 048 1 2 2 5 1 2 . ; Φ Φ K c a H E H a IC V V =     ⋅    ⋅           − − 0 129 3 2 2 5 1 2 . , Φ Φ (4) (5) (4) Experimental determination of the influence of the mathematical calculation model and the method for determining microhardness on the value of frac- ture toughness KIC was carried out on several materi- als – single-crystal silicon wafers Si of three orien- tations (100), (110) and (111) (JSC “INTEGRAL”, Belarus) with a size of Ø100 mm and a thickness of 0.5 mm, AT-cut quartz plates with a diameter of 12 mm and a thickness of 3 mm after chemical-me- chanical (CMP) and magnetorheological (MRF) po- lishing, slide and cover glass. Indentation prints were made using a PMT-3M microhardness tester with a load from 0.01 to 5.0 N (5) where l is crack length near the indent, m; a is half- diagonal length, m; E is elasticity modulus, GPa; Ф is bond reaction index in the crystal lattice (Ф ≈ 3); HV is Vickers hardness, GPa; c is crack length from the center of the indent, m. In works [5, 6] in addition to models, described the conditions for their selection depending on the type of cracks (Figure 1). Development of methods for reducing the error in determination of fracture toughness by the indentation method The electron energy is E = V·e, where V – potential difference, traversable by elec- tron; e – electron charge. Thus, objects larger than 1.0–1.2 nm in size can be distinguished in the SEM. If you change the OM and SEM measure- ment method to a micromechanical method – atomic force microscopy (AFM) with a vertical resolution of at least 0.03 nm, then due to spatial The resolution of the SEM is determined by the electron wavelengths and the numerical aperture of the system. The wavelength of electron radiation depends on its energy, which is affected by the ac- celerating voltage. The electron energy is E = V·e, where V – potential difference, traversable by elec- tron; e – electron charge. Thus, objects larger than 1.0–1.2 nm in size can be distinguished in the SEM. H P A Ber = max , (2) (2) where Pmax is maximum indentation force of the Berkovich pyramid, N; A is resulting contact area under this load, m2. If you change the OM and SEM measure- ment method to a micromechanical method – atomic force microscopy (AFM) with a vertical resolution of at least 0.03 nm, then due to spatial Contact area A was determined by the formu- la [14]: 43 Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. The contribution of the mathematical model of calculation to the accuracy of determining the fracture toughness A S Er =     π 4 2 , (3) (3) where S is unload curve stiffness; Er is reduced mo- dule, GPa. With the existing set of models for calculating KIC using the indentation method, it has been estab- lished that it is impossible to correctly determine the fracture toughness of the material under study using most models, especially at low indentation loads (0.01–0.5 N). The models used for compari- son are given in [15]. KIC for each sample was de- termined at loads of 0.01–5.5 N using six formulas from [15]. It was found that the mathematical mo- dels of calculation (4) and (5) [3, 5, 6, 15] given below show the correctness and stability of the KIC values in the entire range of loads (from 0.01 to 5.5 N). These models are designed and are given in [5, 6]. Unlike others, they include the parameters of the indentation imprint (diagonals length d ) and the formed cracks length l and c. Also include pa- rameters characterizing the material (microhard- ness H and elasticity modulus E ): The use of a high-precision NI method and low loads makes it possible to exclude the formation of cracks in the material. Accordingly, the microhard- ness values determined by this method are correct and accurate. Samples and equipment The condition is as fol- lows: if c /a ≤ 2.5, then Palmquist cracks form in the sample (Figure 1) and the calculation is carried out according to model (4), and if c /a > 2.5, then median cracks form in the sample (Figure 1) and the calcula- tion is carried out by model (5) [5, 6]. Five indentations were performed for each load. Then visualization was carried out using AFM, the indentation parameters (d1 , d2 ) and crack length (l, c) were determined. Microhardness was determined by two methods: by Vickers using formula (1), by the NI method (load 5 mN). Microhardness HV by the Vickers method was determined on PMT-3M. The NI method was determined using a Hysitron 750 Ubi nanoindenter (Bruker, United States) with a Berkovich-type diamond tip with a curvature radius of 60 nm [14]. Then KIC was determined. According to the obtained values for each load, the average value was determined using six formu- las from [15] (Table 1). The values according to for- mulas (4) and (5) depending on the c /a ratio were taken as the actual KIC values for the test material with a standard deviation of less than 10 % [15]. Bold type in Table 1 indicates the actual KIC values for each sample. 44 Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. It has been established that deviations from the actual values of fracture toughness of 12–74 % (Table 2) give mathematical models [15] that depend directly on the load P (Figure 3). It has been established that deviations from the actual values of fracture toughness of 12–74 % (Table 2) give mathematical models [15] that depend directly on the load P (Figure 3). Samples and equipment Table 1 Fracture toughness and deviation from the actual value for silicon, quartz and glass Silicon KIC , (MPa·m1/2 ) Deviation of the KIC value in % from the actual value (100) (110) (111) (100) (110) (111) 0.59±0.11 1.10 ± 0.55 0.74 ± 0.42 50.9 12.4 37.8 0.74±0.22 1.06 ± 0.54 0.86 ± 0.15 38.0 15.7 27.5 1.20±0.05 1.26 ± 0.09 1.19 ± 0.10 0.0 0.0 0.0 0.38±0.07 0.55 ± 0.32 0.46 ± 0.33 68.1 56.4 61.5 1.23±0.07 1.24 ± 0.08 1.20 ± 0.08 2.5 1.6 0.8 0.31±0.13 0.45 ± 0.21 0.39 ± 0.22 74.3 64.3 67.2 Quartz CMP MRF CMP MRF 1.89 ± 0.89 1.15 ± 0.51 29.8 17.5 1.78 ± 0.77 1.57 ± 0.69 22.0 11.9 1.46 ± 0.12 1.40 ± 0.17 0.0 0.0 0.95 ± 0.38 0.88 ± 0.38 34.7 37.3 1.54 ± 0.13 1.33 ± 0.15 5.3 4.7 0.58 ± 0.23 0.46 ± 0.21 60.0 67.2 Glass Slide Сover Slide Сover 1.97 ± 0.47 1.67 ± 0.08 37.48 51.49 2.15 ± 0.04 1.88 ± 0.09 50.25 70.76 1.42 ± 0.03 1.10 ± 0.05 0.0 0.0 1.06 ± 0.30 0.92 ± 0.04 25.88 16.03 1.39 ± 0.01 1.11 ± 0.05 2.46 0.91 0.60 ± 0.17 0.49 ± 0.02 57.70 55.67 Fracture toughness and deviation from the actual value for silicon, quartz and glass Comparison of visualization methods and the accuracy of determining the diagonals of the imprint of length and cracks These imprints were then visualized using SEM and AFM (Figure 2). It is very difficult to accu- rately determine the presence of cracks around the imprint and their length from optical images. The length d1 and d2 of the imprints is determined ap- proximately (Table 2). SEM images do not always clearly show the borders of the imprint (Figure 2а ). Diagonal length cannot be determined exactly. When comparing the determination of the length of the in- dent diagonals (d1 and d2 ), it was found that AFM Visualization of the deformation area (or inden- tation imprint) was carried out on OM, SEM and AFM after applying marks (Figure 2). Comparison of the quality and accuracy of imaging compared on a specific selected imprint. On optical images (Fi- gure 2a, b ), the selected indentations are marked with a red square. 45 Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. allows increasing the accuracy of determining the length of the diagonal up to 9 % compared to OM and up to 2.3 % compared to SEM, if scanning fields of 10 × 10 μm2–50 × 50 μm2 are used. and up to 2.3 % compared to SEM, if scanning fields of 10 × 10 μm2–50 × 50 μm2 are used. а b c d Figure 2 – Optical (400×), images from scanning electron and atomic force microscopy of imprints (a, b ), features of crack opening (c) and influence on the accuracy of determining their length (d ): a – 1.0 N; b – 2.0 N; c – two types of crack opening; d – determination of l crack (using the example of crack No. 2 in Figure 2b) T bl 2 а b а b b а c d d c Figure 2 – Optical (400×), images from scanning electron and atomic force microscopy of imprints (a, b ), features of crack opening (c) and influence on the accuracy of determining their length (d ): a – 1.0 N; b – 2.0 N; c – two types of crack opening; d – determination of l crack (using the example of crack No. Comparison of visualization methods and the accuracy of determining the diagonals of the imprint of length and cracks of the second type (with opening) are either not de- tected by OM and SEM, or they are determined, but not the entire length: the height difference should be at least 100 nm for OM and at least 40–50 nm for SEM. The study of the deformation area around the indentations using OM showed good visualization of cracks, the opening width of which is greater than its resolution – 250 nm (Figure 2a, b). After apply- ing a load of 1.0 N, it was not possible to determine the presence of cracks around the indentation imprint using OM (Figure 2a). After applying a load of 2.0 N near the studied imprint, it was possible to determine the presence and length of three cracks (Figure 2b, Table 3). It was also found that during visualization, the crack is visible with a large opening. Closer to the crack tip, the opening decreases and becomes in- visible in an optical microscope. For this reason, the crack length is incorrectly determined. scope shows a 100 % error compared to AFM, i. e. OM reveals nothing compared to AFM or SEM. The accuracy of determining the crack tip on AFM is due to the surface profile. You can see how a crack of the second type with vertical opening is visualized using OM, SEM and AFM using the example of crack No. 2 in Figure 2b. Schematically, for comparison, this is shown in Fi- gure 2d. On OM this crack could not be determined. The SEM shows only a part of a crack of length l1 with a large difference (56–73 nm) in height without opening. AFM showed that the crack has a length of l1 + l2 and is almost twice as large (Table 2) as com- pared to the SEM value. Detection of cracks after 2.0 N is due to the larger width of crack opening compared to AFM and makes it possible to detect only 37 % of all cracks. SEM makes it possible to detect from 33 to 75 % of all cracks, depending on the applied load, i. e. the higher the load, the larger the cracks and better visible in the SEM. When examining the same prints in the SEM, from two (Figure 2a, marked with white arrows) to five cracks around the imprint (Figure 2b, marked with white arrows) were detected. Comparison of visualization methods and the accuracy of determining the diagonals of the imprint of length and cracks 2 in Figure 2b) Indentation diagonal length and percentage difference compared to optical microscopy OM SEM AFM P, N 1.0 N 2.0 N 1.0 N 2.0 N % 1.0 N 2.0 N % 1.0 N 2.0 N 1.0 N 2.0 N d1 , μm 9.97 10.61 9.00 11.25 9.7 6.0 9.19 11.52 7.8 8.6 d2 , μm 10.07 11.54 9.20 11.35 8.6 1.6 9.24 11.37 8.2 1.5 During indentation, two types of crack opening are formed: the first type is widthwise opening, the second type is vertical opening (Figure 2c). Cracks of the first type open in the XY plane without ver- tical displacement of the material. Cracks of the second type open along the Z axis with little or no width opening (Figure 2c). OM and SEM detect cracks of the first type (in width) only within their resolution. OM reveals cracks with a width opening of at least 250 nm, SEM – at least 1–1.2 nm. Cracks Indentation diagonal length and percentage difference compared to optical microscopy OM SEM AFM P, N 1.0 N 2.0 N 1.0 N 2.0 N % 1.0 N 2.0 N % 1.0 N 2.0 N 1.0 N 2.0 N d1 , μm 9.97 10.61 9.00 11.25 9.7 6.0 9.19 11.52 7.8 8.6 d2 , μm 10.07 11.54 9.20 11.35 8.6 1.6 9.24 11.37 8.2 1.5 Indentation diagonal length and percentage difference compared to optical microscopy second type open along the Z axis with little or no width opening (Figure 2c). OM and SEM detect cracks of the first type (in width) only within their resolution. OM reveals cracks with a width opening of at least 250 nm, SEM – at least 1–1.2 nm. Cracks During indentation, two types of crack opening are formed: the first type is widthwise opening, the second type is vertical opening (Figure 2c). Cracks of the first type open in the XY plane without ver- tical displacement of the material. Cracks of the 46 Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Comparison of visualization methods and the accuracy of determining the diagonals of the imprint of length and cracks The SEM per- fectly visualizes cracks with an opening or a height difference of more than 40–50 nm. Now letʼs compare the cracks, the length of which was determined by all three methods, as well as the error obtained by incorrect determina- tion of the crack length (Table 3). These cracks cor- respond to numbers No. 1, 3 and 4 in Table 2 after a load of 2.0 N. As a result, the use of AFM makes it possible to increase the determination accuracy up to 100 % (Table 4). AFM made it possible to identify from six (Fi- gure 2a, marked with yellow arrows) to eight cracks around the indentation imprint (Figure 2b, marked with yellow arrows). The absence of cracks after a load of 1.0 N when visualized with an optical micro- Table 3 Crack length around indentation imprint and length errors compared to atomic force microscopy Type OM SEM AFM Load, N 1.0 % 2.0 % 1.0 % 2.0 % 1.0 2.0 Crack No, μm 1 no cracks found 100 14.01 30 15.80 3 16.40 18 16.29 20.09 2 – 100 6.30 9 6.94 39 6.39 11.39 3 10.53 46 – 100 19.80 1 6.78 19.64 4 10.08 24 – 100 10.28 23 4.52 13.31 5 – 100 – 100 16.50 15 5.97 19.41 6 – 100 – 100 13.04 0 10.06 13.02 7 – 100 – – – 100 – 10.79 8 – 100 – – – 100 – 9.57 Crack length around indentation imprint and length errors compared to atomic force microscopy gth around indentation imprint and length errors compared to atomic force microscopy 47 Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. Resolution and errors of visualization methods Type Resolution, nm Error in determining, % d c OM at least 250 up to 9.0 up to 100 SEM at least 1.0 up to 2.3 up to 67.0 AFM at least 0.2 (by XY ), at least 0.03 (by Z ) up to 2.0 Resolution and errors of visualization methods After determining a, l, c, E and H, one of the formulas (4) and (5) was selected with respect to c/a. Comparison of visualization methods and the accuracy of determining the diagonals of the imprint of length and cracks Then KIC was calculated. It has been estab- lished that the values of microhardness with NI compared with Vickers differ by 1.2–1.6 times or by 25–38 % (Table 5). This leads to an error in deter- mining the KIC of 16–23 % (Table 5). Comparison of methods for determining microhardness The results of determining the microhardness are shown in Table 5. The values of the elasti- city modulus E of the samples were measured on the NI. Values of microhardness, fracture toughness and errors of their determination by two me Values of microhardness, fracture toughness and errors of their determination by two methods Sample Microhardness H, GPa KIC , MPa·m1/2 Vickers NI % Vickers NI % Si (100) 8.6 ± 0.9 13.8 ± 0.6 37.7 0.97 ± 0.05 1.20 ± 0.05 19.2 Si (110) 8.8 ± 0.4 13.6 ± 0.7 35.3 1.00 ± 0.05 1.26 ± 0.09 20.6 Si (111) 8.4 ± 0.3 13.4 ± 0.7 37.6 0.99 ± 0.06 1.19 ± 0.10 16.8 Quartz MRF 10.0 ± 0.7 13.5 ± 0.1 25.9 1.17 ± 0.17 1.40 ± 0.21 16.4 Glass slide 10.1 ± 0.6 6.7 ± 0.1 33.6 1.82 ± 0.04 1.42 ± 0.03 22.0 Cover glass 10.1 ± 0.2 6.4 ± 2.4 36.6 1.45 ± 0.07 1.10 ± 0.05 23.9 rohardness, fracture toughness and errors of their determination by two methods Acknowledgments This research was supported by the grant of Be- larusian Republican Foundation for Fundamental Research BRFFR No. F20M-083 and Т21МS-029, as part of the assignment No. 2.3 SPSR “Energy and nuclear processes and technologies”. 9. Scholz T., Schneider G.A., Muñoz-Saldaña J., Swain M.V. Fracture toughness from submicron derived indentation cracks. Applied Physics Letters, 2004, vol. 84, pp. 3055. DOI: 10.1063/1.1711164 10. Sergejev F. Comparative study on indentation fracture toughness measurements of cermets and hard- metals. Estonian Journal of Engineering, 2006, vol. 12(4), pp. 43–48. DOI: 10.3176/eng.2006.4.07 Conclusion mechanical (atomic force microscopy) leads to a decrease of the error in determining of the indent diagonal length by 2.3–9.0 %. This also leads to a decrease of the error in determining of the crack length by 46–100 % compared to optical microscopy and 24–67 % compared to scanning electron micros- copy. The method of atomic force microscopy pro- posed in this work for visualizing of the deformation region due to spatial three-dimensional visualiza- tion, atomic forces and high accuracy (XY ± 0.2 nm, Z ± 0.03 nm) can significantly expand the possi- bilities of using the indentation method. It becomes possible to determine the fracture toughness Three ways to improve the accuracy of determin- ing of the critical stress intensity factor KIC , which quantitatively characterizes the fracture toughness, were considered: the choice of a mathematical calcu- lation model, the use of atomic force microscopy to visualize the deformation region, and the nanoinden- tation method to determine the microhardness and elasticity modulus of the material. It was established that changing the physi- cal principle of the visualization method for cracks and indentation parameters from optical to micro- 48 Devices and Methods of Measurements 2022, vol. 13, no. 1, pp. 40–49 V.A. Lapitskaya et al. Приборы и методы измерений 2022. – Т. 13, № 1. – С. 40–49 V.A. Lapitskaya et al. of individual phases in a material and individual elements of microelectromechanical systems through the use of small loads. densification of reaction bonded Si/SiC ceramics. Jour- nal of the European Ceramic Society, 2018, vol. 38, pp. 4815–4823. DOI: 10.1016/j.jeurceramsoc.2018.07.014 DOI: 10.1016/j.jeurceramsoc.2018.07.014 The use of the nanoindentation method in- stead of the Vickers method made it possible to in- crease the accuracy of determining of the materialʼs microhardness up to 38 %, as well as the critical stress intensity factor KIC up to 23 %. 8. Grinchuk P.S., Kiyashko M.V., Abuhimd H.M., Alshahrani M.S., Solovei D.V., Stepkin M.O., Aku- lich A.V., Shashkov M.D., Kuznetsova T.A., Danilova- Tretiak S.M., Evseeva L.E., Nikolaeva K.V. Advanced technology for fabrication of reaction-bonded SiC with controlled composition and properties. Journal of the European Ceramic Society, 2021, vol. 41, pp. 5813–5824. DOI: 10.1016/j.jeurceramsoc.2021.05.017 References 1. Gdoutos E.E. Fracture Mechanics. Cham: Sprin- ger, 2020, vol. XIX, 477 p. DOI: 10.1007/1-4020-3153-X 11. Lapitskaya V.A., Kuznetsova T.A., Khabara- va A.V., Chizhik S.A., Aizikovich S.M., Sadyrin E.V., Mitrin B.I., Sun Weifu. The use of AFM in assessing the crack resistance of silicon wafers of various orienta- tions. Engineering Fracture Mechanics, 2022, vol. 259, pp. 107926. 2. Ritchie R., Liu D. Introduction to Fracture Me- chanics. 1st Edition. Elsevier, 2021, 160 p. DOI: 10.1016/C2020-0-03038-0 DOI: 10.1016/C2020-0-03038-0 DOI: 10.1016/j.engfracmech.2021.107926 3. Moradkhani A., Baharvandi H., Tajdari M., Lati- fi H., Martikainen J. Determination of fracture toughness using the area of micro-crack tracks left in brittle mate- rials by Vickers indentation test. Journal of Advanced Ceramics, 2013, vol. 2(1), pp. 87–102. 12. Pogotskaia I.V., Kuznetsova T.A., Chizhik S.A. Elastic Modulus Determination of Nanolayers by the Atomic-Force Microscopy Method. Mekhanika ma- shin, mekhanizmov i materialov [Mechanics of Machines, Mechanisms and Materials], 2011, no. 3(16), pp. 43–48 (in Russian). DOI: 10.1007/s40145-013-0047-z 4. Evans A.G. Fracture Toughness Determinations by Indentation. Journal of the American Ceramic Society, 1976, vol. 59, pp. 371–372. 13. Oreshko E.I., Utkin D.A., Erasov V.S., Lyak- hov A.A. Methods of measurement of hardness of mate- rials (review). Trudy VIAM [Proceedings of VIAM], 2020, no. 1(85), pp. 101–117 (in Russian). DOI: 10.18577/2307-6046-2020-0-1-101-117 DOI: 10.1111/j.1151-2916.1976.tb10991.x 5. Niihara K., Morena R., Hasselman D.P.H. Evalua- tion of KIC of brittle solids by the indentation method with low crack-to-indent ratios. Journal of Materials Science Letters, 1982, vol. 1, pp. 13–16. DOI: 10.1007/BF00724706 DOI: 10.18577/2307-6046-2020-0-1-101-117 14. Oliver W.C., Pharr G.M. An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments. Jour- nal of Materials Research, 1992, vol. 7, pp. 1564–1583. DOI: 10.1557/JMR.1992.1564 6. Niihara K. A fracture mechanics analysis of in- dentation-induced Palmqvist crack in ceramics. Journal of Materials Science Letters, 1983, vol. 2, pp. 221–223. DOI: 10.1007/BF00725625 DOI: 10.1557/JMR.1992.1564 15. Lapitskaya V.A., Kuznetsova T.A., Khudo- ley A.L., Khabarava A.V., Chizhik S.A., Aizikovich S.M., Sadyrin E.V. Influence of polishing technique on crack resistance of quartz plates. International Journal of Frac- ture, 2021, vol. 231, no. 1, pp. 61–77. 7. Grinchuk P.S., Kiyashko M.V., Abuhimd H.M., Alshahrani M.S., Stepkin M.O., Toropov V.V., Khort A.A., Solovei D.V., Akulich A.V., Shashkov M.D., Liakh M.Yu. Effect of technological parameters on DOI: 10.1007/s10704-021-00564-5 49
https://openalex.org/W4313127324	https://hal.science/hal-03884294/document	English	null	Physics for understanding life	Europhysics news	2,022	cc-by	2,899	To cite this version: Felix Ritort, Bart van Tiggelen. Physics for Understanding Life. Europhysics News, 2022, EPN, 53 (5), pp.24-27. 10.1051/epn/2022503. hal-03884294 Physics for Understanding Life Felix Ritort, Bart van Tiggelen To cite this version: Felix Ritort, Bart van Tiggelen. Physics for Understanding Life. Europhysics News, 2022, EPN, 53 (5), pp.24-27. 10.1051/epn/2022503. hal-03884294 To cite this version: Distributed under a Creative Commons Attribution 4.0 International License Physics for Understanding Life Dyson argued that metabolism and replication could have originated separately and that "life actually began twice". Physics is not enough for understanding life, but life may reveal new physics. To understand life, we need interdisciplinary collaborations between biologists, computer-scientists, chemists, physicists, astrophysicists and engineers. Can science discover what conditions have facilitated the origin of life on Earth? Does some kind of “life” exist on extra-terrestrial planets? Can we create “artificial life” that self-replicates or has intelligence? Can we design and implement “artificial intelligence”? These challenges are hopefully within reach on the Horizon 2050. Is life nothing but “vital dust”, i.e. a natural consequence of non-equilibrium physics and chemistry [3], or was it rather a “magnificent accident”? Famous physicists like Erwin Schrödinger [4] and Freeman Dyson [5] struggled with the role of physics in the origin of life. Schrödinger wondered how life manages to stay out of equilibrium while respecting the Second Law. Dyson argued that metabolism and replication could have originated separately and that "life actually began twice". Physics is not enough for understanding life, but life may reveal new physics. To understand life, we need interdisciplinary collaborations between biologists, computer-scientists, chemists, physicists, astrophysicists and engineers. HAL Id: hal-03884294 https://hal.science/hal-03884294v1 Submitted on 16 Dec 2022 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License Physics for Understanding Life Felix Ritorta and Bart van Tiggelenb aSmall Biosystems Laboratory, Condensed Matter Physics Department, Facultat de Física, Universitat de Barcelona, Spain bUniversity Grenoble Alpes, CNRS, LPMMC, Grenoble, France “Living” matter distinguishes itself from “ordinary” matter by its capacity to grow, to reproduce or to multiply, and most of all by its autonomous functional activity, baptized “intelligence”, to sense its environment to adapt or to survive. This last feature truly is one of the miracles of life. Our current knowledge of life is relatively recent. The “molecule of life”, Deoxyribose Nucleic Acid or DNA, was first discovered by Miescher in 1860. The discovery of the double helix structure by Watson, and Crick in 1951 [1], and the determination of its sequence of nucleic acids by Sanger and Coulson in 1975 [2] have been major breakthroughs of the last century. Can science discover what conditions have facilitated the origin of life on Earth? Does some kind of “life” exist on extra-terrestrial planets? Can we create “artificial life” that self-replicates or has intelligence? Can we design and implement “artificial intelligence”? These challenges are hopefully within reach on the Horizon 2050. Is life nothing but “vital dust”, i.e. a natural consequence of non-equilibrium physics and chemistry [3], or was it rather a “magnificent accident”? Famous physicists like Erwin Schrödinger [4] and Freeman Dyson [5] struggled with the role of physics in the origin of life. Schrödinger wondered how life manages to stay out of equilibrium while respecting the Second Law. Dyson argued that metabolism and replication could have originated separately and that "life actually began twice". Physics is not enough for understanding life, but life may reveal new physics. To understand life, we need interdisciplinary collaborations between biologists, computer-scientists, chemists, physicists, astrophysicists and engineers. Can science discover what conditions have facilitated the origin of life on Earth? Does some kind of “life” exist on extra-terrestrial planets? Can we create “artificial life” that self-replicates or has intelligence? Can we design and implement “artificial intelligence”? These challenges are hopefully within reach on the Horizon 2050. Is life nothing but “vital dust”, i.e. a natural consequence of non-equilibrium physics and chemistry [3], or was it rather a “magnificent accident”? Famous physicists like Erwin Schrödinger [4] and Freeman Dyson [5] struggled with the role of physics in the origin of life. Schrödinger wondered how life manages to stay out of equilibrium while respecting the Second Law. Where do we come from? How life on Earth began remains an unsolved, complex scientific problem [6]. After the brilliant experiments by Pasteur and Tyndall in the nineteenth century, the concept of “spontaneous generation” of life from non-living matter was rejected for once and for all. Life may have emerged only 400 million years after the “Hadean” Earth was created (Hades was the king of the ancient Greek underworld), only 200 million years after liquid water had first appeared. The absence of a well-preserved crust older than 3.3 billion years makes it difficult to get a precise information about the chemical composition of the Hadean Earth. absence of a well-preserved crust older than 3.3 billion years makes it difficult to get a precise information about the chemical composition of the Hadean Earth. The time variations on all scales of the Earth-Sun connection must have played a major role in the chemistry that created life. Stellar evolution models tell us that the young Sun was 70 % fainter than today. The “faint young Sun paradox” states that this low flux implies water on Earth to be frozen until well beyond the Hadean age, which we know is not true. Greenhouse gases such as C02 must have been present, heating up the atmosphere much like they do today [7]. The rapid rotation of the young Sun in only a few days led to strong magnetic activity and large space weather, so that energetic X-ray fluxes must have been up to 100 times larger than today. The habitable zone predicted on the basis of liquid water present at the surface of the planet. (source: Wikepedia) The habitable zone predicted on the basis of liquid water present at the surface of the planet. (source: Wikepedia) The presence of liquid water is a major condition for life to form, but other habitability conditions exist, such as the presence of organic molecules and energy sources – providing at least 150 kJ/mol - to initiate prebiotic chemistry. Energetic solar photons must have provided a lot of this energy. Organic molecules such as HCN and CH2O could also have been created by the lightning-triggered dissociation of carbon-dioxide and abiotic methane. A mostly reducing (oxygen-poor) early atmosphere with energetic sparks could have been a play-ground for prebiotic chemistry and a world-famous “Miller-Urey” experiment supporting this vision was done in 1952 [8,9]. Where do we come from? Carbonaceous meteorites impacted the Earth and brought also many organic compounds, including amino acids. Where exactly prebiotic chemistry took place is still subject to a large debate. Energetic radiation does not penetrate more than 1 cm in water. What is the nature of the human mind? Artificial intelligence (AI) covers all techniques that enable machines to solve tasks like humans do. AI applications are immense, from medicine to robotics and basic science. AI has seen fast growth since the fifties when cybernetics and feedback, the branch of science introduced by Wiener in 1948 [11], took off as an effort to understand information processes in machines and living beings. The Greek term “cybernetics” (kybernetiké), the art of guiding, was introduced by Ampère in his classification of sciences published in 1834, subsequently taken up by Maxwell [12]. Subsequent work by Bertalanffy on the general theory of biological systems, Maturana and Varela on autopoiesis (the idea of self-sustaining cycles), Von Neumann, Shannon, and Gabor on automatic systems have set the basis of modern AI. Good old-fashioned AI (GOFAI), as it is called nowadays [13], embraces the power of using basic symbols (words, pictures, actions, etc..) to represent physical patterns to build up high- order symbolic structures to be manipulated with programs and algorithms. A key limitation of GOFAI is that programs can neither adapt themselves nor acquire new knowledge on their own, which led to the "AI winter" in the eighties without significant progress. The subsequent development of connectionism in artificial neural networks and statistics-based inference methods (machine learning) that allow data internal structure to evolve, has led to new functionalities. In the current view, intelligence stands for the ability to avert the ever-present threat of the exponential explosion of search options. What about consciousness? If brains can be considered as probabilistic prediction machines they must somehow deal with uncertainty. Is this related to consciousness? Are we going to understand consciousness one day? Life for understanding new physics In his 1944's beautiful essay Erwin Schrödinger wrote [4]: “The large and important and very much discussed question is: how can the events in space and time which take place within the spatial boundary of a living organism be accounted for by physics and chemistry?”. The marvellous complexity of living matter raises the question of whether biology hides new, emerging physical laws awaiting discovery. From the Brownian motion of pollen grains to cell motility, animal motion, the many behaviours of active matter, and the evolution of bacterial populations, the interface between biology and physics offers an unprecedented wealth of new phenomena that can be investigated at an exquisite level of detail. For example, single-molecule techniques permit us to monitor how a protein folds and measurements of energies with 1 kT ( 10-21J) accuracy [10] reveal how a molecular motor moves one step at a time. By increasing the bar of spatial, temporal, and energy measurement accuracy, we hope to uncover vital discrepancies between our current theoretical knowledge and experiments on biological systems. Teleonomy is the continuous “purposefulness” of living beings to sleep, feed, trade, play, laugh, etc. They are not found in non-living matter where just the laws of physics and chemistry prevail. Does it escape our comprehension? What is the role of physical information in biology? Does information set a new paradigm for emerging complexity that connects biology to quantum physics? All these are relevant questions waiting for an answer. Can we build artificial life? Artificial life (ALIFE) aims at building sustainable self-replicating systems resembling living beings. ALIFE comes in three forms: computer-designed AI programs and algorithms (soft ALIFE), hardware-based robots (hard ALIFE), and chemical and biochemical systems with life-like behaviour (wet ALIFE). The basic tenet of ALIFE is that if Nature has found one way of organizing living matter, other (unknown) ways should exist. Can we build synthetic cells endowed with homeostasis, self-reproduction, and evolution? Man-designed cell-like droplets exhibit limited properties and functionalities, such as self-propulsion, artificial chemotaxis, organism-like multi-droplets, collective-like behaviours, etc. Building synthetic ALIFE systems represents a daunting task ahead. Besides possessing basic properties such as robustness, autonomy, efficiency, recycling, intelligence, self-repair, adaptation, self- replication, etc., they must show open-ended evolution and teleonomy [14]. Are we ever going to leap this gap? Is there anybody out there? The question whether extra-terrestrial life exists is as old as mankind. No matter how small the probability is for life to emerge, its existence “somewhere, sometime” in the huge Universe is largely accepted by scientists. The Drake equation [15], actually proposed as an agenda item for a meeting on alien communications in 1961 but today argued to be one of the most important equations in science, gives estimates of up to many millions of detectable alien civilisations in Is there anybody out there? Is there anybody out there? The question whether extra-terrestrial life exists is as old as mankind. No matter how small the probability is for life to emerge, its existence “somewhere, sometime” in the huge Universe is largely accepted by scientists. The Drake equation [15], actually proposed as an agenda item for a meeting on alien communications in 1961 but today argued to be one of the most important equations in science, gives estimates of up to many millions of detectable alien civilisations in our own Galaxy. The famous paradox raised by Enrico Fermi during an after-dinner talk in 1950 states that if there are really so many, then “where are they?”. One possible answer is that intelligent civilisations tend to self-destruct. With the recent observation of exoplanets, the first proof of extra-terrestrial life may be within reach. Will it have the same biochemistry that created life on Earth? The establishment of a list of “biosignatures” - molecules whose origin we believe requires a biological agent - has top- priority. Can we build artificial life? Molecular oxygen is a clear biosignature, and is hardly found elsewhere in our solar system. A molecular biosignature must be stable with respect to the local planetary environment, though its detection may be false alarm due to abiotic reactions. Within our solar system several bodies such as Mars, Jupiter’s moon Europe and Saturn’s moon Titan have been identified as possible candidates to host or to have hosted life. Sample-return missions and in- situ search for “biosignatures” started with the Viking lander missions to Mars in 1976 and continue today with the rover Perseverance. Indications exist that Mars might host methane- producing microorganisms at its subsurface. The complication on Mars is that many identified biosignatures were possibly destroyed by (per)chlolates. Our knowledge about exoplanets outside the solar system entirely depends on remote sensing. Today, more than 4000 exoplanets have been identified. The recently launched James Webb telescope will facilitate the study of infrared absorption lines in the spectra of exoplanet atmospheres, observed while transiting in front of the star, and possibly due to biosignatures. The perfect exoplanet to observe “Earthian” life would be an Earth-sized planet in an Earth- like orbit around a Sun-like star. Unfortunately, due to observational selections effects, few such perfect exoplanets are currently known. A recently observed candidate is one of the three exoplanets of our nearest star Proxima Centauri [16]. It is Earth-like and estimated to be in the habitable zone of liquid water. Dragonfly mission concept of entry, descent, landing, surface operations, and flight at Titan. Figure credit NASA onfly mission concept of entry, descent, landing, surface operations, and flight at Titan. Figure credit NAS Several next-generation proposals exist to overcome the problem detecting exoplanets of fainter, more distant stars, such as the “Seventy Billion Mile Space Telescope” FOCAL that uses the Sun as a gravitational lens [17]. We might even launch relativistic “spacechips” to Proxima Centauri. Despite huge technological progress in the decennia to come, no watertight proof of life on exoplanets may ever come. A fundamental question is whether we can live with that. The authors would like to thank professors Patricia Bassereau, Angelo Cangelosi, Jitka Cejkova, Carlos Gerkenson, Raymond Goldstein, Zita Martin, Sarah Matthews, Sara Seager, David Vernon and Francis Westall for their precious contributions to the Grands Challenges. 2. F. Sanger and A.R. Coulson, A rapid method for determining sequences in DNA by primed synthesis with DNA polymerase, J. Mol. Biol. 94 (3): 441 (1975). Can we build artificial life? The authors would like to thank professors Patricia Bassereau, Angelo Cangelosi, Jitka Cejkova, Carlos Gerkenson, Raymond Goldstein, Zita Martin, Sarah Matthews, Sara Seager, David Vernon and Francis Westall for their precious contributions to the Grands Challenges. , y , ( ) 2. F. Sanger and A.R. Coulson, A rapid method for determining sequences in DNA by primed synthesis with DNA polymerase, J. Mol. Biol. 94 (3): 441 (1975). 2. F. Sanger and A.R. Coulson, A rapid method for determining sequences in DNA by primed synthesis with DNA polymerase, J. Mol. Biol. 94 (3): 441 (1975). 3. C. De Duve, Vital Dust (Basic Books, 1995). 4. E. Schrödinger, What is life?, Cambridge University Press (1944). 5. F. Dyson, Origins of Life, Cambridge University Press (1992). 6. H.J. Cleaves, Prebiotic Chemistry: What We Know, What We Don't, Evolution: Education and Outreach 5, 342 (2012). 7. C. Sagan and G. Mullen, Earth and Mars: Evolution of Atmospheres and Surface Temperatures, Science. 177 (4043): 52 (1972). 8. H.C. Urey, On the early chemical history of the Earth and the origin of life, P.N.A.S. 38, 351 (1952 9. S.L. Miller, Production of Amino Acids Under Possible Primitive Earth Conditions, Science 117, 528 (1953). 10. F. Ritort, The Noisy and Marvellous Molecular World of Biology, Inventions 4.2, 24 (2019). 10. F. Ritort, The Noisy and Marvellous Molecular World of Biology, Inventions 4.2, 24 (2019). 11. N. Wiener, Cybernetics or Control and Communication in the Animal and the Machine, MIT pre 11. N. Wiener, Cybernetics or Control and Communication in the Animal and the Machine, MI (2019). 12. J.C. Maxwell, On Governors. Proceedings of the Royal Society of London 16, 270-283 (1868). 13. O.A. Newell and A.S. Herbert, Computer science as empirical inquiry: Symbols and search, ACM Turing award lectures, 2007 (1975). 14. A. Pross, What is life?: How chemistry becomes biology. Oxford University Press, 2016. 15. F. Drake and D. Sobel, The Origin of the Drake Equation, Astronomy Beat 46, 1 (2010) 16. J.P. Faria etal., A candidate short-period sub-Earth orbiting Proxima Centauri, Astronomy & Astrophysics 658, A115 (2022). 17. V.R. Eshleman, Gravitational Lens of the Sun: Its Potential for Observations and Communications over Interstellar Distances, Science, 205 (4411), 1133 (1979).
https://openalex.org/W2933782819	https://espace.curtin.edu.au/bitstream/20.500.11937/79839/3/79941.pdf	English	null	Effect of Ionic Liquid Structure on the Oxygen Reduction Reaction Under Humidified Conditions	Journal of physical chemistry. C./Journal of physical chemistry. C	2,019	cc-by	11,023	Abstract The cation was found to have a large effect on the reduction current of oxygen even at low humidity levels (< 40 RH %), whereas the anion mainly influenced the current at higher humidity levels (> 65 RH%). Consequently, the choice of cation needs to be carefully considered when selecting a suitable RTIL solvent for oxygen reduction in humidified environments. The size, structure and hydrophobicity of the ions was found to dictate the degree at which the RTIL is susceptible to changes in humidity. The physical characteristics of the RTIL electric double layer on platinum electrode surfaces were further investigated by atomic force microscopy force-curve studies in three selected RTILs. The results suggest that there is a significant amount of water incorporated at the electrode-RTIL interface in [C2mim][NTf2] and [N4,1,1,1][NTf2], but not in the more hydrophobic [P14,6,6,6][NTf2]. The presence of moisture has a significant impact on ORR currents in [C2mim][NTf2] even at extremely low humidity levels, which was verified by the higher level of water incorporation in [C2mim][NTf2] compared to [N4,1,1,1][NTf2] and [P14,6,6,6][NTf2]. Hydrophobic and large RTIL cations and anions (e.g. [P14,6,6,6]+ and [FAP]-) are recommended for applications where a stable ORR current response is required under humidified conditions. Simon DoblingerA, Junqiao LeeA, Debbie S. SilvesterA,B Simon DoblingerA, Junqiao LeeA, Debbie S. SilvesterA,B A Curtin Institute for Functional Molecules and Interfaces, and School of Molecular and Life Sciences, Curtin University, GPO Box U1987, Perth, WA 6845, Australia B Corresponding author. Email: d.silvester-dean@curtin.edu.au Abstract The oxygen reduction reaction (ORR) is widely studied in room temperature ionic liquids (RTILs), but typically in dry environments. Since water is known to affect diffusion coefficients and reaction outcomes, the influence of water on the ORR is expected to be significant. We have therefore studied the effect of RTIL structure on the ORR at different relative humidity (RH) levels using cyclic voltammetry. A broad range of cations including imidazolium-, ammonium-, pyrrolidinium-, pyridinium, sulfonium- and phosphonium-, and anions such as [BF4]-, [PF6]-, [NTf2]- and [FAP]- were employed. The cation was found to have a large effect on the reduction current of oxygen even at low humidity levels (< 40 RH %), whereas the anion mainly influenced the current at higher humidity levels (> 65 RH%). Consequently, the choice of cation needs to be carefully considered when selecting a suitable RTIL solvent for oxygen reduction in humidified environments. The size, structure and hydrophobicity of the ions was found to dictate the degree at which the RTIL is susceptible to changes in humidity. The physical characteristics of the RTIL electric double layer on platinum electrode surfaces were further investigated by atomic force microscopy force-curve studies in three selected RTILs. The results suggest that there is a significant amount of water incorporated at the electrode-RTIL interface in [C2mim][NTf2] and [N4,1,1,1][NTf2], but not in the more hydrophobic [P14,6,6,6][NTf2]. The presence of moisture has a significant impact on ORR currents in [C2mim][NTf2] even at extremely low humidity levels, which was verified by the higher level of water incorporation in [C2mim][NTf2] compared to [N4,1,1,1][NTf2] and [P14,6,6,6][NTf2]. Hydrophobic and large RTIL cations and anions (e.g. [P14,6,6,6]+ and [FAP]-) are recommended for applications where a stable ORR current response is required under humidified conditions. The oxygen reduction reaction (ORR) is widely studied in room temperature ionic liquids (RTILs), but typically in dry environments. Since water is known to affect diffusion coefficients and reaction outcomes, the influence of water on the ORR is expected to be significant. We have therefore studied the effect of RTIL structure on the ORR at different relative humidity (RH) levels using cyclic voltammetry. A broad range of cations including imidazolium-, ammonium-, pyrrolidinium-, pyridinium, sulfonium- and phosphonium-, and anions such as [BF4]-, [PF6]-, [NTf2]- and [FAP]- were employed. 1. Introduction The reliable monitoring of gases under extreme conditions is an important criteria for sensors in industry, and is especially important for health and safety requirements.1 Commercially available sensors usually consist of a three- electrode setup, with an aqueous based electrolyte (usually water/H2SO4), and a gas permeable membrane that prevents solvent evaporation at high temperatures.2 However, this membrane also reduces the diffusion rate of gases towards the electrode.3 An alternative, proposed by Buzzeo et al., is the use of room-temperature ionic liquids (RTILs) as a non-volatile electrolyte, which removes the need for a membrane layer.4 RTILs are composed of cations and anions that are liquid at room temperature due to asymmetry in at least one of the ions5 and weaker ion-ion interaction forces.6 RTILs are increasingly used in synthesis,7 as well as in electrochemistry applications,8-11 due to their promising characteristics, such as negligible volatility, high stability up to certain temperatures without decomposition, and high conductivity.12 Due to their intrinsic conductivity, they have been extensively explored as alternative electrolyte materials.13 The wide electrochemical windows (ca. 4–6 V)14 of these solvents make them suitable candidates for electrochemical sensing applications; with the detection of gaseous analytes like hydrogen, oxygen or ammonia already reported in the literature.15 However, the hygroscopic nature of RTILs is an issue for their use in practical environments that currently hinders their implementation in commercial membrane-free sensors. The tendency for RTILs to absorb moisture can lead to instability and unreliability in the measurements when the gas-sensor is subjected to real-atmospheres with changing humidity levels.16-18 It is hence clear that water is a significant impurity in RTIL-based gas-sensors that need to be investigated; this can be indirectly monitored by studying the oxygen reduction reaction (ORR). The ORR has been widely reported in aprotic RTILs for the purposes of oxygen gas detection.4, 8, 19-20 Katayama et al. Keywords Oxygen reduction reaction; humidity; room temperature ionic liquids; cyclic voltammetry; atomic force microscopy; electrical double layer; oxygen sensing. 1 1. Introduction published an almost reversible redox couple in ammonium-based ionic liquids with a current density ratio of 0.97.21 It has also been shown that the reduction of oxygen can undergo different mechanistic pathways, depending on the presence and concentration of protic species such as water, which could cause a shift from a 1e– (equation 1) to a 2e– (equation 2) or even a 4e– (equation 3) reduction reaction.22-24 This significantly complicates the reliable calibration of a sensor when exposed to different humidity levels, since water can act as an effective proton source,25 opening up electrochemical reaction pathways that are dependent upon the concentration of water. 𝑒 O2 + 𝑒𝑒−→O2 −● (1) O2 + 2H+ + 2𝑒𝑒−→H2O2 (2) O2 + 4H+ + 4𝑒𝑒−→2H2O (3) (1) (2) (3) Therefore, several groups (including ours), have used hydrophobic RTILs,26-28 or mixed the electrolyte with a hydrophobic polymer to reduce the water uptake.29-30 However, this did not eliminate the water absorption problem entirely. Therefore, several groups (including ours), have used hydrophobic RTILs,26-28 or mixed the electrolyte with a hydrophobic polymer to reduce the water uptake.29-30 However, this did not eliminate the water absorption problem entirely. When examining the impact of absorbed moisture on the electrochemical processes at the electrode-RTIL interface, the structure of the RTIL at the interface must be taken into account.17 Research performed by other groups has shown 2 that as with conventional solvents, ionic liquids form electrical double layers (EDLs) at the electrode/solution interface.31-32 The EDL structure at the electrode interface could have a severe impact on the transport of analytes such as gases or moisture towards the surface, which is required to induce an electrochemical reaction. Several methods including atomic force microscopy (AFM),33-34 surface force apparatus (SFA),35 neutron reflectometry,36 sum frequency generation vibrational spectroscopy (SFG-VS)37 and surface-enhanced infrared absorption spectroscopy (SEIRAS)38 have been used to try to elucidate this behaviour. For RTILs, the thickness and number of layers at the electrode is strongly dependent on the charge density at the surface and the nature of the cation and anion. 1. Introduction Depending on the cation and anion structure, the EDL that is formed when a potential is applied can be either more or less pronounced.34, 39-41 Single as well as ion-pair layers have been detected at the interface using AFM.33, 42 With regards to layer thickness, theoretical simulations of these EDL structures are in good accordance with experimental data.43 It is already well known that when the electrode is highly charged, more than one cation or anion layer can be formed near the electrode44 which means that either the cation-analyte or anion-analyte interactions may influence the concentration profile of various dissolved species. Hence, knowledge of the EDL structure is crucial to explain changes in the response in electrochemical experiments. However, the structuring of RTILs on surfaces in the presence of other species (e.g. oxygen or water) is not very well understood. Computational studies have been carried out to investigate the influence of adsorption of water45 or ferrocene46 in RTILs. However, despite the extensive literature published on the structure of the EDL of RTILs, experimental studies on the influence of species, such as absorbed gases or water, are still lacking.23 Characteristics of the RTIL, such as the charge carrying atom, as well as the alkyl chain length, number, and position(s) on both the cation and the anion, appear to be key determining factors of water sorption properties of RTILs. COSMO-RS predictions have shown that there is a strong interaction between water and ionic liquid anion in the bulk phase.47 This means that for different RTILs that have the same anion, the mutual solubility of water only varies slightly, while for RTILs with same cation and different anions, the effect is more pronounced.47 It should also be noted that the bulk-phase behaviour of water in RTILs can be very different to the behaviour near a charged electrode surface. For example, Bi et al.48 showed that the electrosorption of water on electrodes in humid, hydrophobic ILs was higher than in hydrophilic RTILs even though the bulk phase solubility of water showed the opposite trend. Therefore, the effect of water on electrode processes may be different to the trends expected based on the hydrophobicity/hydrophilicity of the RTILs. In this work, we demonstrate the influence of the RTIL structure at the EDL – varying both the cation and the anion – and show its effect on the ORR in humidified environments using cyclic voltammetry (CV). 1. Introduction Six different cation types (imidazolium-, pyrrolidinium-, pyridinium-, ammonium-, sulfonium- and phosphonium-) with different alkyl chain lengths and several different anions are analysed. Atomic force microscopy (AFM) is then performed under unbiased and biased conditions to provide insights into the impact of water on the EDL structure of RTILs on a charged surface, and to give an explanation for the electrochemistry results. 3 3 2. Experimental 2.1 Chemical reagents Figure 1. Chemical structures and abbreviations of the room temperature ionic liquid cations and anions used in this study. N N H2n+1Cn [Cnmim]+ N CnH2n+1 H2n+1Cn CxH2x+1 CnH2n+1 [Nx,n,n,n]+ P CnH2n+1 H2n+1Cn CxH2x+1 CnH2n+1 [Px,n,n,n]+ S [S2,2,1]+ N [C4mpyrr]+ B F F F F P F F F F F F P C2F5 F C2F5 C2F5 F F N S S O O O O F3C CF3 [NTf2]- [BF4]- [PF6]- [FAP]- Cation Anion n, x = 2-14 N [C4py]+ N N H2n+1Cn [Cnmim]+ N CnH2n+1 H2n+1Cn CxH2x+1 CnH2n+1 [Nx,n,n,n]+ P CnH2n+1 H2n+1Cn CxH2x+1 CnH2n+1 [Px,n,n,n]+ S [S2,2,1]+ N [C4mpyrr]+ Cation n, x = 2-14 N [C4py]+ N CnH2n+1 H2n+1Cn CxH2x+1 CnH2n+1 [Nx,n,n,n]+ P CnH2n+1 H2n+1Cn CxH2x+1 CnH2n+1 [Px,n,n,n]+ Cation Cation [Cnmim]+ [Cnmim]+ [Px,n,n,n]+ [Nx,n,n,n]+ n, x = 2-14 S [S2,2,1]+ N [C4mpyrr]+ 14 N [C4py]+ N [C4mpyrr]+ N [C4py]+ S [S2,2,1]+ [C4py]+ [S2,2,1]+ [C4mpyrr]+ P F F F F F F P C2F5 F C2F5 C2F5 F F [PF6]- [FAP]- Anion Anion N S S O O O O F3C CF3 [NTf2]- [NTf2]- Figure 1. Chemical structures and abbreviations of the room temperature ionic liquid cations and anions used in this study. Figure 1. Chemical structures and abbreviations of the room temperature ionic liquid cations and anions used in this study. All RTILs were obtained at the highest purities possible. Diethylmethylsulfonium bis(trifluoromethylsulfonyl)imide ([S2,2,1][NTf2], 99 %, IoLiTec-Ionic Liquids Technologies GmbH, Heilbronn, Germany), butyltrimethylammonium bis(trifluoromethylsulfonyl)imide ([N4,1,1,1][NTf2], 99,5 %, IoLiTec), butylpyridinium bis(trifluoromethylsulfonyl)imide ([C4py][NTf2], 99 %, IoLiTec), trihexyltetradecylphosphonium tris(pentafluoroethyl)trifluorophosphate ([P14,6,6,6][[FAP], >98 %, Merck Pty. Ltd. Kilsyth, Victoria, Australia), butylmethylpyrrolidinium bis(trifluoromethylsulfonium)imide ([C4mpyrr][NTf2], 99.5 %, IoLiTec), 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([C2mim][NTf2], 99.5 %, IoLiTec), 1-ethyl-3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate ([C2mim][FAP], for synthesis >98 %, Merck), 1-butyl-methylimidazolium bis(trifluoromethylsulfonyl)imide ([C4mim][NTf2], 99.5 %, IoLiTec), 1-butyl- 3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate ([C4mim][FAP], high purity >99 %, Merck), 1-butyl-3- methylimidazolium tetrafluoroborate ([C4mim][BF4], 99 %, IoLiTec), 1-butyl-3-methylimidazolium hexafluorophosphate ([C4mim][PF6], high purity >99 %, Merck), 1-hexyl-3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate ([C6mim][FAP], high purity, Merck) were used as received. Trihexyltetradecylphosphonium bis(trifluoromethylsulfonyl)imide ([P14,6,6,6][[NTf2]) was kindly donated by the group of Professor Chris Hardacre (now at the University of Manchester, UK) when he was located at Queens University Belfast, UK. 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([C6mim][NTf2], 99 %, IoLiTec) was kindly donated by Professor Rob Atkin (University of Western Australia). The RTIL heptyltriethylammonium 4 bis(trifluoromethylsulfonyl)imide ([N2,2,2,7][NTf2]) was synthesized according to standard literature procedures49 (see electronic supplementary information (ESI) Figure S1 for the NMR data). 2.2 Electrochemical experiments CV measurements were performed with a PGSTAT101 Autolab potentiostat (Metrohm Autolab, Gladesville, NSW, Australia) interfaced to a computer with NOVA 1.11 software. The electrochemical cell was housed in an aluminium Faraday cage to reduce electromagnetic interferences. The working electrode (WE), counter electrode (CE) and reference electrode (RE) of the thin-film electrodes (TFEs) (ED-SE1 Pt, MicruX Technologies, Oviedo, Spain) are composed of platinum (Pt) deposited on a Pyrex substrate, with a WE diameter of 1 mm. Wires were soldered onto the connecting pads of the chip. The WE was electrochemically activated prior to each experiment by CV cycling (> 300 times) at 1 Vs-1 in nitrogen purged 1 M H2SO4(aq) between -0.75 and +0.7 V. The activated chips were then rinsed twice with ultrapure water and acetone before drying under a nitrogen stream. To account for the unstable Pt quasi- RE on the TFE, modification of the RE with lithium iron phosphate (LFP) (previously developed by our group50) was attempted. However, the interaction of superoxide with the LFP material led to a change in ORR peak shape and current over consecutive scans, making measurements unreliable. Therefore, to investigate the influence of humidity on CV peak potentials for selected RTILs, ferrocene was added in-situ, and potentials were shifted such that the midpoint of the ferrocene/ferrocenium (Fc/Fc+) redox couple was at 0 V. Seven µL of the ionic liquid was drop-cast to cover all three electrodes on the TFE and purged for at least 45 min in a high purity nitrogen stream at a flow rate of 500 mL min-1 to remove dissolved gases and impurities such as oxygen and carbon dioxide. The relatively high volume of electrolyte on the small chip (compared to our previous experiments) ensures that the electrogenerated products from each scan can be effectively diluted, so that their accumulation is negligible. For all oxygen measurements, the integrated CE and WE were used. After purging the electrolyte with nitrogen gas for at least 45 min, oxygen was introduced into one arm of a modified glass T-cell.51 The oxygen line was additionally connected to a humidity generator (Owlstone Humidity Generator OHG-4, Owlstone, Cambridge, UK) that separates the gas flow into ‘dry’ and ‘wet’ streams by bubbling the gas through a container of 400 mL of ultrapure water at a flow rate of 500 mL min-1. 2. Experimental The chemical structures of the ionic liquid cations and anions used in this study are summarized in Figure 1. Ultrapure water with a resistivity of 18.2 MΩ cm prepared by a Milli-Q laboratory water purification system (Millipore Pty Ltd., North Ryde, NSW, Australia) was used for humidifying the oxygen gas stream. Acetone (CHROMASOLV®, for HPLC, ≥99.0 %, Sigma-Aldrich) and ferrocene (98 %, Sigma-Aldrich) were used as received. A 1 M stock solution of H2SO4(aq) (prepared with ultrapure water from a 95-98 wt% H2SO4 solution, Ajax Finechem, WA, Australia) was used for the activation of the thin-film electrodes. High purity oxygen gas (>99.5 %) and high purity nitrogen gas (99.99 %) cylinders were purchased from BOC Gases (Welshpool, WA, Australia). 2.2 Electrochemical experiments Adjusting the dry to wet flow ratio results in different humidity levels and the measured dew point (Td) was converted to obtain the relative humidity percentage (RH%), using equation (4)52: 𝑅𝑅𝑚𝑇𝑑𝑇𝑎𝑎𝑎 𝑅𝑅𝑅𝑅% = 100% ∗10 𝑚𝑚൬ 𝑇𝑇𝑑𝑑 𝑇𝑇𝑑𝑑+𝑇𝑇𝑛𝑛− 𝑇𝑇𝑎𝑎𝑎𝑎𝑎𝑎 𝑇𝑇𝑎𝑎𝑎𝑎𝑎𝑎+𝑇𝑇𝑛𝑛൰ (4) (4) 5 with the constants m = 7.59 and Tn = 240.73 °C, the ambient temperature Tamb and the dew point Td in °C (Humidity Conversion Formulas, Vaisala).53 The dew point was measured at a location close to the electrochemical cell to ensure that the actual humidity introduced into the T-cell is not affected by condensation or adsorption along the tubing (see Figure 2). Figure 2. Schematic of the gas humidifying system used for voltammetric measurements on thin-film electrodes (TFEs). Figure 2. Schematic of the gas humidifying system used for voltammetric measurements on thin-film elec Since the humidity generator is not able to detect humidity levels lower than 1 RH%, measurements were conducted at a starting humidity level of ~1 RH%. It is also known in the literature that even with prolonged vacuum purging, a significant amount of water is retained (in the ~10–100 ppm range)54 in the RTIL. An equilibration time of 45 min was employed before performing CV scans at each humidity level to ensure that the gas was fully saturated, particularly into the most viscous ionic liquid. Repeated CV scans for O2 reduction were carried out under constant humidity conditions to identify the duration for a stable peak current to be achieved (see Figures S3-S6 in the supporting information). 45 mins was found to be more than sufficient for equilibrium to be attained across all RTILs and humidities. Hence, before the commencement of experiments, the RTILs were also first purged under N2 gas for more than 45 mins. It is noted that water uptake times of up to 180 min were observed for RTILs at different humidity environments,55 but that study employed a larger volume of ionic liquid and used a static set-up, in contrast to the large surface-to-volume (seven microliter RTIL droplet spread out over a thin-film electrode) and the constant flow system used in our experiments. All experiments were carried out at least in triplicate and all data points were used to establish the trendline that was fitted by a logistic regression analysis. 2.2 Electrochemical experiments In order to remove the contributions caused by different solubilities and diffusion coefficients in the RTILs, the currents were normalised, where the measured currents at different humidities were divided by those measured at 1 RH%. 2.3 AFM measurements Atomic force microscopy (AFM) experiments were carried out with a Bruker Dimension Icon interfaced to a computer with NanoScope 9.4 software in contact mode using activated Pt TFEs. Force curves were collected continuously with a scan rate of 0.16 Hz while the ramp size was between 100 and 160 nm. A silicone barrier (Selleys Silicone Sealant, Padstow, NSW, Australia) was constructed around the electrode cell to confine the ionic liquid during 6 the AFM measurements. All experiments were performed with a silicon nitride AFM tip (SNL-10, spring constant 0.35 Nm-1). The tips were carefully cleaned in sulfuric acid (1 M) for 60 s, washed with ultrapure water and acetone, and finally dried under a nitrogen stream. Measurements were conducted on an uncharged and a negatively charged electrode surface under mildly humidified conditions (ca. 40 RH%). When the electrode was negatively polarised, rapid formation and growth of H2 bubbles were observed at < -1 V in [C2mim][NTf2] (see ESI for a detailed discussion). Hence, a less negative biasing potential of -0.8 V was chosen for the studies. 2.4 Computational estimation of ion dimensions and water-affinity The structure of several RTIL cations and the [NTf2]- anion was computationally optimised (Spartan ’14, Wavefunction, Inc., California, USA) with a semi-empirical approach, using the parameterized model number 3 (PM3) method for single isolated ions in vacuum. For the cations with significant degree of freedom (e.g. [N4,1,1,1]+ and [P14,6,6,6]+) the alkyl chains were pre-oriented before the structures were minimised. The dimensions were estimated from the optimised structures based on their expected orientations. Re-positioning of the RTIL side-chains, due to interactions with the electrode surface (unbiased/negatively-biased), and neighbouring species (e.g. cation, anion, water) can affect the layer thickness created by the cations/anions. It is noted that the computationally obtained values are simply to aid with the analysis of AFM force-curve measurements; a more thorough computational investigation of layer thicknesses is beyond the scope of this work. The value estimated for [C2mim]+ agrees well with literature.56 The computationally estimated orthogonal dimensions, corresponding space-filling (CPK) volumes for each of the ions and the water molecule, as well as their solvation energies (SEs) are summarised in Table S1. The CPK volumes follow the trend: [P14,6,6,6]+ >>> [N7,2,2,2]+ > [FAP]– > [C6mim]+ > [C4mpyrr]+ > [C4mim]+ ≈ [C4py]+ > [NTf2]– ≈ [N4,1,1,1]+ > [C2mim]+ ≈ [S2,2,1]+ >> [PF6]– > [BF4]– > H2O (5) [P14,6,6,6]+ >>> [N7,2,2,2]+ > [FAP]– > [C6mim]+ > [C4mpyrr]+ > [C4mim]+ ≈ [C4py]+ > [NTf2]– ≈ [N4,1,1,1]+ > [C2mim]+ ≈ [S2,2,1]+ >> [PF6]– > [BF4]– > H2O (5) ] (5) The PM3 calculated SEs of selected single isolated RTIL molecules are also listed in Table S1 to provide the local affinity of each ion towards surrounding water molecules; the greater the SE value, the more hydrophobic the molecule is likely to be. It is noted that these values are relevant on the inter-molecular scale but may not apply for the bulk RTIL as they do not account for RTIL-RTIL and bulk interactions, and hence may not coincide with bulk properties of hygroscopicity and hydrophobicity. The SEs (listed in Table S1) follow the trend: The PM3 calculated SEs of selected single isolated RTIL molecules are also listed in Table S1 to provide the local affinity of each ion towards surrounding water molecules; the greater the SE value, the more hydrophobic the molecule is likely to be. 3.1 Cyclic voltammetry of RTILs in dry vs wet environments CV was carried out in different RTILs under a pure oxygen atmosphere at a range of humidity levels from < 1 RH% up to ~95 RH%. Representative cyclic voltammograms in [N4,1,1,1][NTf2] at four different humidity levels are presented in Figure 3. A reversible one-electron reduction to superoxide (see equation 1) was observed in a dry environment (< 1 RH%, black line). The CV shape is typical for an electrochemically quasi-reversible process, which is well reported for the oxygen/superoxide redox couple in RTILs.16 At increased humidity levels, the reductive peak current becomes larger and the reduction process became more irreversible, resulting in a decreased superoxide oxidation peak. Ferrocene was added to the solution as an internal reference due to significant potential shifting occurring at the Pt quasi-RE in the presence of the humidified gas. The voltammograms were shifted so that the midpoint of the ferrocene/ferrocenium redox process was at 0 V, and a substantial shift (~600 mV) of the ORR potential, from ca. -1.3 V (vs. Fc/Fc+) at < 1 RH% (dry) to ca. -0.7 V at 95 RH% (wet) was observed. This is likely due to the increased presence of proton-donors that catalyse the reduction reaction (see equations 2 and 3), which is further supported by the significant increase in the ORR peak current. From < 1 RH% to 55 RH%, a current increase of approximately 2.5 times (from -6 µA to -15 µA) was observed, indicating a mix of 2 and 4 electron processes, which is supported by the absence of the superoxide oxidation peak. At 95 RH %, the current increased by a factor of three (to -19 ca. µA), which suggests closer to a 4-electron process is dominating (equation 3). A change in the oxygen diffusion coefficient as the result of viscosity changes in the water-saturated RTIL may also contribute to the increased currents, but a detailed study of this effect is well beyond the scope of this work. In contrast with [N4,1,1,1][NTf2], the peak position vs. the Pt quasi-RE in [P14,6,6,6][FAP] was substantially less affected by humidity (see ESI Figure S2b), and the ORR peak current only showed a ~20 % increase, likely owing to the much higher hydrophobicity of the [P14,6,6,6]+ cation compared to [N4,1,1,1]+ (see trend in equation 6). Figure 3. 2.4 Computational estimation of ion dimensions and water-affinity It is noted that these values are relevant on the inter-molecular scale but may not apply for the bulk RTIL as they do not account for RTIL-RTIL and bulk interactions, and hence may not coincide with bulk properties of hygroscopicity and hydrophobicity. The SEs (listed in Table S1) follow the trend: [P14,6,6,6]+ >>> [FAP]- > [C6mim]+ > [N7,2,2,2]+ > [C4mim]+ > [C2mim]+ ≈ [C4mpyrr]+ ≈ [C4py]+ > [N4,1,1,1]+ > [S2,2,1]+ >> [NTf2]- >> [PF6]- >> H2O > [BF4]- (6) [P14,6,6,6]+ >>> [FAP]- > [C6mim]+ > [N7,2,2,2]+ > [C4mim]+ > [C2mim]+ ≈ [C4mpyrr]+ ≈ [C4py]+ > [N4,1,1,1]+ > [S2,2,1]+ >> [NTf2]- >> [PF6]- >> H2O > [BF4]- (6) (6) 7 3.1.1 Effect of Anion The type of anion is known to have a significant impact on the hygroscopic behaviour of RTILs.54, 57 The bulk solubility of water in ionic liquids containing the [FAP]– anion is much lower compared to [NTf2]–, [PF6]–, or [BF4]– – see trend in equation 6, and also in reference [57]. It was previously shown that electro-absorption of water in the EDL is more pronounced for hydrophilic anions at negative potentials.58 In an ionic liquid, absorbed water molecules are known to preferentially interact with the anion.6, 47, 58-59 Hence, it is important to investigate how different anionic species may affect the ORR reaction at different humidity levels. Figure 4 shows the normalised ORR current as a function of increasing humidity levels in four different [C4mim]+-based RTILs. The current is mostly unaffected at low humidity levels (< 15 RH%) suggesting that the cation is the determining factor at low humidity levels. This is consistent with AFM32-33, 39, 60 and SEIRAS38 studies, where it is well known that the innermost layer at a negatively biased surface is mainly composed by cations. The normalised ORR current then begins to increase above 15 RH%. The current at high humidity levels (> 95 RH%) follows the trend: [FAP]– < [NTf2]– < [PF6]– < [BF4]– (7) (7) which is consistent with the computationally predicted molecular-hydrophobicity and volume trends for individual anionic molecules (see trends 5 and 6), and by COSMO-RS calculations47 from a separate study. For the [BF4]– RTIL, there is a lower than expected current trend between 20 and 80 RH%. It was suspected that this may be due to insufficient equilibration time. However, waiting 130 min instead of 45 min before recording the first scan did not affect the current response (see ESI Figure S3). Instead, repeated CV scanning caused an increase in the current, suggesting a possible side-reaction and build-up of electrogenerated products contributing to the current. This is not surprising considering that [BF4]-, as well as [PF6]-, can undergo hydrolysation.61 At 90 RH%, the electrogenerated products are likely saturated at the EDL from the initial scan; consecutive scans at this humidity level can cause severe changes in the ORR current due chemical reactions with previously electrogenerated species, such as H2O2. Overall, the results in Figure 4 suggest that bulky, hydrolytically stable and hydrophobic anions such as [FAP]– are recommended for applications where a stable ORR response in humidified environments is required, e.g. 3.1 Cyclic voltammetry of RTILs in dry vs wet environments Cyclic voltammograms at 100 mV s-1, in [N4,1,1,1][NTf2] on a Pt TFE, with 7 mM ferrocene as internal reference in an O2- saturated flow with RH% of < 1 RH% (black), 13 RH% (green), 55 RH% (blue) and approximately 95 RH% (red). -1.6 -1.2 -0.8 -0.4 0.0 -20 -16 -12 -8 -4 0 4 < 1 RH% 13 RH% 55 RH% ~ 95 RH% I / µA E vs Fc/Fc+ / V Figure 3. Cyclic voltammograms at 100 mV s-1, in [N4,1,1,1][NTf2] on a Pt TFE, with 7 mM ferrocene as internal reference in an O2- saturated flow with RH% of < 1 RH% (black), 13 RH% (green), 55 RH% (blue) and approximately 95 RH% (red). 8 3.1.2 Effect of cation with a hydrophobic anion, [NTf2]– A negatively polarised electrode during oxygen reduction voltammetry implies that the EDL structure at the electrode-RTIL interface will be dominated by the RTIL cation. In contrast with anions, cations generally only have a slight influence on the solubility of water in RTILs.45, 47 Properties such as the molecular affinity to water (predicted by SE values), coupled with cation size (based on CPK volumes) – which may sterically hinder the adsorption of other species at the electrode surface (see Table S1 in the ESI) – need to be considered. In this study, only RTILs displaying quasi-reversible ORRs were included, i.e. the pyridinium ionic liquid was omitted due to an irreversible ORR process (see ESI, Figure S7). The presence of an obvious superoxide oxidation peak for the remaining RTILs implies that any impurities in the RTILs (e.g. left over from the synthesis procedure) are minimal, and do not affect the electrochemical processes occurring during the oxygen reduction reaction. Figure 5 presents normalised ORR peak currents vs. humidity for eight different RTILs sharing the same (moderately hydrophobic) [NTf2]– anion. Compared to the previous study of different anions (section 3.1.1), the type of cation appears to have a more significant influence on the ORR current at lower humidity levels (i.e. < 40 RH%). There is a clear difference between the humidity level at which the ORR current starts to sharply increase – referred to as the “onset-humidity” – typically occurring at below 30 RH%, and follows the trend: [P14,6,6,6]+ > [C6mim]+ > [C4mim]+ ≈ [N7,2,2,2]+ ≈ [N4,1,1,1]+ >> [C4mpyrr]+ ≈ [S2,2,1]+ >> [C2mim]+ (8) [P14,6,6,6]+ > [C6mim]+ > [C4mim]+ ≈ [N7,2,2,2]+ ≈ [N4,1,1,1]+ >> [C4mpyrr]+ ≈ [S2,2,1]+ >> [C2mim]+ (8) This is in contrast with the different anions discussed in Section 3.1.1, where the onset-humidity appears to be almost unaffected by the type of cation present. Significant variability in the data points towards higher humidity levels (see discussions in section 3.2.3) prevents a confident evaluation of the 95 RH% current trends, although it is clear that the ORR current in the [P14,6,6,6]+ RTIL is much less affected by water compared to the other cations. The variability in current could be due several factors including increased follow-up chemical reactions and even bubble formation at the electrode-RTIL interface due to water-splitting reactions (see Figure S11 in the ESI). 3.1.1 Effect of Anion for oxygen sensing. 9 Figure 4. Effect of humidified environments on the normalised current for oxygen reduction in [C4mim][BF4] (●), [C4mim][PF6] (), [C4mim][NTf2] () and [C4mim][FAP] () with increasing relative humidity percentages (RH%). CVs (at 100 mVs-1) were obtained on Pt-TFEs, with each scan was taken after 45 min equilibration time. All experiments were performed at least in triplicate. 0 20 40 60 80 100 0 1 2 3 4 5 6 [C4mim][BF4] [C4mim][PF6] [C4mim][NTf2] [C4mim][FAP] Inorm Humidity / RH% 0 20 40 60 80 100 0 1 2 3 4 5 6 [C4mim][BF4] [C4mim][PF6] [C4mim][NTf2] [C4mim][FAP] Inorm Humidity / RH% [C4mim][BF4] [C4mim][PF6] [C4mim][NTf2] [C4mim][FAP] 9 Figure 4. Effect of humidified environments on the normalised current for oxygen reduction in [C4mim][BF4] (●), [C4mim][PF6] (), [C4mim][NTf2] () and [C4mim][FAP] () with increasing relative humidity percentages (RH%). CVs (at 100 mVs-1) were obtained on Pt-TFEs, with each scan was taken after 45 min equilibration time. All experiments were performed at least in triplicate. 9 3.1.2 Effect of cation with a hydrophobic anion, [NTf2]– (8) Zhong et al.60 reported that at low water content, a stable structuring of the RTIL at the electrode exists and is dominated by the cation. As increasing humidity levels, water molecules begin to disrupt and weaken the structuring of the RTILs at an electrode. Hence, the intermediate RH% trends are dominated by the tendency of the EDL structure to be affected by the presence of water. The trends at high RH% are thus a better reflection of the bulk hydrophobicity of the RTILs. 10 Figure 5. Effect of humidified environments on the normalised current for oxygen reduction in [C2mim][NTf2] (), [S2,2,1][NTf2] (─), [C4mpyrr][NTf2] (), [C4mim][NTf2] (●), [N4,1,1,1][NTf2] (), [N7,2,2,2][NTf2] (), [C6mim][NTf2] () and [P14,6,6,6][NTf2] (I) with increasing relative humidity percentage (RH%). CVs at 100 mV s-1 were carried out on Pt-TFEs and each scan was taken after 45 min equilibration time. The scans were taken at 100 mVs-1 and with a step potential of -2.5 mV. 0 20 40 60 80 100 0 1 2 3 4 [C2mim][NTf2] [S2,2,1][NTf2] [C4pyrr][NTf2] [N4,1,1,1][NTf2] [N7,2,2,2][NTf2] [C4mim][NTf2]) [C6mim][NTf2] [P14,6,6,6][NTf2] Inorm Humidity / RH% 0 20 40 60 80 100 0 1 2 3 4 [C2mim][NTf2] [S2,2,1][NTf2] [C4pyrr][NTf2] [N4,1,1,1][NTf2] [N7,2,2,2][NTf2] [C4mim][NTf2]) [C6mim][NTf2] [P14,6,6,6][NTf2] Inorm Humidity/ RH% Figure 5. Effect of humidified environments on the normalised current for oxygen reduction in [C2mim][NTf2] (), [S2,2,1][NTf2] (─), [C4mpyrr][NTf2] (), [C4mim][NTf2] (●), [N4,1,1,1][NTf2] (), [N7,2,2,2][NTf2] (), [C6mim][NTf2] () and [P14,6,6,6][NTf2] (I) with increasing relative humidity percentage (RH%). CVs at 100 mV s-1 were carried out on Pt-TFEs and each scan was taken after 45 min equilibration time. The scans were taken at 100 mVs-1 and with a step potential of -2.5 mV. Between the three imidazolium-based RTILs, the trends are in order of increasing alkyl chain length, as expected in terms of increased cation hydrophobicity. The ORR in [C2mim][NTf2] appears to be the most affected by the presence of moisture at extremely low humidity levels, where a sharp ORR current increase from 0 to 3 RH% is present. This is in contrast with the other RTILs where a relatively more stable region of low ORR current vs. humidity is present. The significant dependency of the onset-humidity on the type of cations is a factor to be considered when designing humidity sensors where sensitivity within a particular humidity range is desired; or for the case of an oxygen sensor, where sensitivity to moisture should be avoided. 3.1.2 Effect of cation with a hydrophobic anion, [NTf2]– Surface spectroscopy measurements37 have shown that imidazolium cations with small alkyl chains (e.g. [C2mim]+) tend to align with the aromatic ring structure flat against platinum surfaces. Applying a negative potential stabilizes this conformation, while increasing the alkyl chain length causes a tilting of the planar structure. However, it is surprising that there is a substantial difference in the ORR CV response current behaviour between [C2mim]+ (black squares), [C4mim]+ (red circles) and [C6mim]+ (purple hollow circles). Nevertheless, the longer alkyl chains of [C4mim]+ and [C6mim]+ allow the formation of a more hydrophobic innermost layer that prevents accumulation of water near the electrode. Interestingly, the two ammonium ionic liquids, [N4,1,1,1][NTf2] and [N7,2,2,2][NTf2], with significantly different cation alkyl chain lengths, produced similar ORR-current behaviour in Figure 5. A layer of ammonium cations will form at a negatively charged surface. At sufficiently high negative surface polarisation, the positively charged nitrogen group will be pulled towards the negative surface charge, inducing a reorientation of the short alkyl chains. Ultimately, this results in a dense structure of cations at the electrode interface, where longer alkyl groups are expected to point away from the charged surface, orientating in the most stable energetic state to maximise interaction with the anion layer above. However, the three shorter alkyl chains of the ammonium RTILs studied are too short and less flexible to effectively cover the electrode surface area between repelling cations, and do not form a highly dense hydrophobic layer at the electrode interface. This can leave cavities and pathways for water transportation to the electrode surface. 11 Furthermore, the positively charged nitrogen atom is less shielded by the shorter alkyl chains. This can allow water molecules to interact via the electronegative oxygen, and with the two aligning hydrogens more strongly attracted to the negatively polarised surface. However, in the case of [P14,6,6,6][NTf2], the reorientation of the four long alkyl chains forms a dense hydrophobic layer that prevents the absorption of water in the EDL close to the electrode surface. Hence it is recommended for bulky and highly hydrophobic cations (e.g. [P14,6,6,6]+) to be employed for stable responses with less interaction from water. 3.1.3 Effect of cation with a highly hydrophobic anion, [FAP]– Figure 6 shows ORR peak currents vs. humidity for different RTILs with the most hydrophobic [FAP]– anion used in this study, and different cations, namely [P14,6,6,6][FAP], [C2mim][FAP], [C4mim][FAP], and [C6mim][FAP]. As expected, normalised currents at > 30 RH% for the [FAP]–-based RTILs are lower than the corresponding [NTf2]–-based RTILs (Figure 5), because of the higher hydrophobicity of [FAP]–. Additionally, the [FAP]– anion has an onset-humidity of 10 RH% for [C2mim][FAP] (Figure 6) which is a significant improvement compared to [C2mim][NTf2] (Figure 4). The onset-humidity follows the trend: [P ]+ [C i ]+ [C i ]+ [C i ]+ (9) (9) [P14,6,6,6]+ >> [C4mim]+ ≈ [C6mim]+ > [C2mim]+ (9) Interestingly, the onset-humidity trend for the three imidazolium cations with the [FAP]- anion is not as expected, and is different to that observed for [NTf2]–. It is not entirely obvious as to why it is the case for the [FAP]--based RTILs, but this unusual behaviour is consistent with a previous AFM study under dry conditions that revealed differences in the packing density of the three imidazolium ionic liquid cations that did not follow the expected trend based of increasing alkyl chain length.33 Figure 6. Effect of humidified environments on the normalised current for oxygen reduction of [C2mim][FAP] (●), [C4mim][FAP] (), [C6mim][FAP] () and [P14,6,6,6][FAP] () with increasing relative humidity percentages (RH%). CVs were carried out on Pt- TFEs. Each scan was taken after 45 min equilibration time and all experiments were performed at least in triplicate. The scans were taken at 100 mVs-2 and with a step potential of -2.5 mV. 0 20 40 60 80 100 0 1 2 3 [C2mim][FAP] [C4mim][FAP] [C6mim][FAP] [P14,6,6,6][FAP] Inorm Humidity / RH% 0 20 40 60 80 100 0 1 2 3 [C2mim][FAP] [C4mim][FAP] [C6mim][FAP] [P14,6,6,6][FAP] Inorm Humidity/ RH% [C2mim][FAP] [C6mim][FAP] [C4mim][FAP] [P14,6,6,6][FAP] Figure 6. Effect of humidified environments on the normalised current for oxygen reduction of [C2mim][FAP] (●), [C4mim][FAP] (), [C6mim][FAP] () and [P14,6,6,6][FAP] () with increasing relative humidity percentages (RH%). CVs were carried out on Pt- TFEs. Each scan was taken after 45 min equilibration time and all experiments were performed at least in triplicate. The scans were taken at 100 mVs-2 and with a step potential of -2.5 mV. 12 3.2. Physical characteristics of the ionic liquid electrical double layer on a Pt-TFE. Fedorov and Kornyshev’s62 study of the ionic liquid (IL) EDL at the electrified solid-IL interfaces reveals a “generally more complicated than expected” structure, due to the tendency of ILs to self-assemble. Previous work suggests an EDL structure that is comparable to aqueous electrolytes, consisting of an adsorbed ion layer succeeded by an electrostatically bound diffuse layer.63-65 However, it was found from AFM studies that a multilayer of both single and cation-anion pair layers can form at the interface, and the number of layers increased and were more tightly organised at higher potentials.34, 66 Here, three RTILs: [C2mim][NTf2], [N4,1,1,1][NTf2], and [P14,6,6,6][NTf2] were selected to perform AFM force curve measurements under humidified conditions at ca. 40 RH% (Figure 7), and will be contrasted with the CV results in the previous sections. These RTILs represent three distinct behaviours observed in the electrochemical studies above: (case 1) [C2mim][NTf2] shows a significant change in ORR current at low humidities, (case 2) [N4,1,1,1][NTf2] gives a stable ORR current at low RH%, and (case 3) [P14,6,6,6][NTf2] where the ORR current remains stable up to high moisture contents. The electrode was held at the open circuit potential (OCP) and biased at -0.8 V (vs. the quasi-Pt-RE) to compare effect of a negative surface biasing during ORR on the EDL. Several force-distance curves were collected and one representative curve per sample and condition is used for the discussion. It is noted that the studies were conducted on rough platinum thin-film electrodes which complicates the interpretation of results, but were performed to represent the real conditions used for electrochemical sensing. Three overlayed force-distance curves per sample can be found in the ESI (Figure S10) to show the reproducibility at different areas on the electrode. We note that the layering of RTILs such as [C2mim][NTf2],55 [C4mpyrr][NTf2]60 or ammonium RTILs38 on a flat electrode under ‘dry’ conditions has been previously reported, but these conclusions may not be valid for a significantly rougher surface. Figure 7. Representative force vs distance curves for a silicon nitride cantilever approaching a Pt-TFE surface at the OCP (grey) and -0.8 V (vs. pseudo Pt RE) (orange) in (A) [C2mim][NTf2], in (B) [N4,1,1,1][NTf2] and in (C) [P14,6,6,6][NTf2] in 40 RH% (in air). An overlay of three force-curves for each experimental case can be found in the ESI (Figure S10). Figure 7. 3.2.1 General trends in AFM force-curves Overall, the force-curves in Figure 7 for both charged and uncharged surfaces show step-like features, superimposed on a broad curve with the force gradually increasing towards 0 nm separation. The broad curve is related to the compression of stable EDL structures by the tip as well as the interaction of the tip with the rough electrode surface. The steps correspond to tip-interaction with structured layers of molecules (e.g. cations, anions, water, etc.) orientated and assembled over the electrode surface. Subtle step-features are also present in between the more pronounced ones, which are attributed to either different conformations of ions, the presence of interstitial water or may be caused by roughness of the electrode surface. In the following discussions, the more pronounced layers will be referred to as layers 1, 2, 3, etc., with layer 1 being the surface-adsorbed innermost ion layer. The thickness of the EDL can be observed by the occurrence of a “snap-on” event. This is defined as the distance at which the descending tip experiences the first significant attractive force towards the diffuse layer, and is observed as an abrupt drop in the measured force coinciding with the start of the compression behaviour. Within the EDL, the pronounced layer-thicknesses can be estimated by measuring the width between each step of the force-curves, and are expected to be a function of the ion or ion-pair sizes. The force required for the AFM tip to puncture through a layer is referred to as the “push-through” force. Zhong et al.60 reported that the EDL of [C4mpyrr][NTf2] is structurally weakened in the presence of water, even at very low water contents around 66 ppm. Based on the results shown in figure 5, [C2mim][NTf2], [S2,2,1][NTf2] and [C4mpyrr][NTf2] show an early onset humidity, meaning that there is a significant amount of water present in the EDL even at low humidity levels. This is in contrast to the research reported by Cheng et al., where AFM studies were carried out in dry and humidified [C2mim][NTf2] on mica surfaces, with a silica-terminated silicon AFM tip (of 5–10 nm radius).55 The steps appeared more pronounced under humidified conditions (43 RH%), and a negligible compression behaviour was observed. This is probably because of the atomically smooth and uniform surface of Mica as well as the fact that water is required to charge the Mica surface. 3.2. Physical characteristics of the ionic liquid electrical double layer on a Pt-TFE. Representative force vs distance curves for a silicon nitride cantilever approaching a Pt-TFE surface at the OCP (grey) and -0.8 V (vs. pseudo Pt RE) (orange) in (A) [C2mim][NTf2], in (B) [N4,1,1,1][NTf2] and in (C) [P14,6,6,6][NTf2] in 40 RH% (in air). An overlay of three force-curves for each experimental case can be found in the ESI (Figure S10). 13 3.2.1 General trends in AFM force-curves The uniformly charged and smooth surface of Mica allows highly ordered establishment of the first layer, which can then allow more ordered subsequent layers to form, whereas more disordered and less defined structures are expected on the rougher Pt-TFE surface used in our study (see discussions in ESI). Also, an interface-selective vibrational spectroscopy study on the behaviour of ammonium ILs at Au interfaces showed that alkyl chains form a loose structure that reaches into the anion layer, resulting in a more interdigitated structure rather than a clearly separated alternating cation/anion layer.38 This behaviour could further add to the contribution of the superimposed compression-like feature, especially for the longer chain RTILs (e.g. [P14,6,6,6][NTf2]). For this work, we have chosen to examine the EDL on the non-ideal Pt-TFE (also used for the CV experiments) since this is a more typical representation of an electrode surface found in real sensors. Case 1: Unbiased electrode (at OCP) In [C2mim][NTf2] and [N4,1,1,1][NTf2] several pronounced layers (Figures 7A and 7B grey arrows) are visible in the force-curves. However, in [P14,6,6,6][NTf2] no pronounced layers are discernible. During tip-engagement, the initial interaction with the EDL is with the diffuse layer, which typically occurs at a separation between 3.5 and 4.5 nm. For [C2mim][NTf2] the initial “snap-on” behaviour occurs from ca. 4.5 nm to ca. 3.8 nm, which could indicate a higher degree of water-incorporation into the EDL structure compared to the bulk RTIL. In [C2mim][NTf2], layer thicknesses are approximately 0.3 nm (layer 1), 0.63 nm (layer 2), and ~0.51 nm each for layers 3, 4 and 5. [C2mim]+ is known to rest relatively flat and well-ordered on the surface.37 The stability of this orientation will determine the effectiveness of the cation to block moisture adsorption at the electrode surface. Comparing the layer thicknesses with Figure 7, it can be assumed that the innermost layer almost equals the size of a cation (≈0.25 nm), followed by an anion layer (≈0.6 nm), as expected. Further layers are significantly thicker (around 0.5nm) that would match a water + cation layer, or an anion layer. So the layer thickness suggests water incorporation between the ion layers, which was also concluded by Zhong et al. for [C4mpyrr][NTf2].60 [N4,1,1,1][NTf2] has approximately the following layer thicknesses: 0.5 nm (layer 1), 1.1 nm (layer 2), and the subsequent three layers are ca. 0.7–0.9 nm thick. Again, it can be assumed based on previous research that the innermost layer is formed by cations, followed by an anion layer. However, in this case no separate cation and anion layers can be detected and therefore it cannot be determined whether there is water incorporation or simply a less well separated cation–anion layer due to the presence of the longer alkyl chain. It may even be possible that the innermost layer is formed of a cation bilayer followed by the expected cation-anion double layers as predicted by Kornyshev et al. for “tricky quaternary structures in combination with more compact anions”.67 Examining the structure of [N4,1,1,1]+, the three short methyl groups (believed to be facing the electrode surface) would not be effective at shielding the localised positive charge on the nitrogen atom, allowing interaction with the water molecules and preventing adsorption on the Pt-surface. 3.2.2 Comparison of the EDLs of individual RTILs To aid with the discussion, estimations of the layer thicknesses were obtained from computationally optimised molecular structures of each ion, as shown in Figure S13 in the ESI. Orientations of the ions were based on suggestions 14 from other research.37-38. Since only the first few layers are expected to have a strong influence on the electrochemistry at the Pt-RTIL interface, they will be focussed on in the following sections. Case 1: Unbiased electrode (at OCP) The absence of a “snap-on” behaviour, indicates that there is insignificant water incorporation within the EDL of this RTIL, and is consistent with the results in Figure 5. However it can be seen that the innermost layer is thicker for [N4,1,1,1]+ compared to [C2mim]+. In contrast, [P14,6,6,6][NTf2] does not form very prominent layers. Most force curves in this RTIL were dominated by a very obvious compression-like behaviour without obvious structuring. These features could be associated with conformational changes of the long (14-carbon) alkyl chains coupled with re-organisation of anions sandwiched between the cation layers that are held in an interdigitated pattern. Like with [N4,1,1,1][NTf2], no initial “snap-on” behaviour was present in [P14,6,6,6][NTf2], which suggests an insignificant amount of water-incorporation into the EDL, consistent with the electrochemistry results. In summary, even on the rough Pt surface, ionic liquid structuring in the EDL can be observed in humid environments for the smaller cations, but no conclusions can be drawn for the larger 15 phosphonium one – characterization on a flat electrode or SEIRAS measurements are required to determine whether structuring is present or not. Case 2: Biased Pt electrode (at -0.8 V) At a negatively biased Pt surface, the force-curves for [C2mim][NTf2] and [N4,1,1,1][NTf2] (Figures 7A and 7B, orange) reveal greater forces on the cantilever below a separation of ~1.5 nm compared to the unbiased case. In general, higher forces are required to push through the innermost layers which are composed of more ordered cations that are attracted more strongly to the negatively charged surface, as previously reported by Rivera-Rubero et al.37 However, for [P14,6,6,6][NTf2], the long hydrophobic alkyl chains are able to significantly shield the influence of the surface electric field, thus reducing the stabilising effect of a negative potential bias and therefore no obvious changes in the structuring can be seen. Our experimental results agree well with a molecular dynamics simulation study,48 and also support the CV observations for the [P14,6,6,6]+-based RTILs in Figures 5 and 6. Under negative potential biasing and above a separation of 0.5 nm, less pronounced steps are observed in humidified [C2mim][NTf2] (Figure 7A orange) as a result of water molecules competing for space with ions in the layers and therefore weakening the layering compared to the unbiased surface (Figure 7A grey) or dry environment.34, 60, 66 In contrast, a higher number of pronounced layers are observed on the charged Pt electrode in [N4,1,1,1][NTf2], which may hint at less disruption from an increased number of water molecules incorporating in the EDL and competing for space with the RTIL ions under humidified conditions. For [P14,6,6,6][NTf2], the much longer alkyl group of the cation prevents condensed and stable structures from forming, but shielding of the positive charge coupled with the high hydrophobicity inhibits the incorporation of polar water molecules into the EDL. Furthermore, at distances > 1.5 nm, the layers remain relatively unaffected by negative biasing, which further substantiates this interpretation. In conclusion, potential biasing results in an increased force required to push through the innermost layers, and for the small RTIL cations, obvious structural layers begin to vanish due to the competition between water and IL ions. 16 4. Conclusions The ORR mechanism was studied by CV on Pt-TFEs in several different RTILs under humidified conditions. Peak potentials and currents were found to vary with humidity levels, showing a gradual shift in the reduction mechanism with increasing humidity (from < 1 to 95 RH%). The degree by which a given RTIL is susceptible to humidity changes was found to be a function of ion size, structure, and water affinity. Generally speaking, the cation influences the ORR current at humidity levels below 40 RH% due to the more pronounced structuring of the ions near the negatively charged electrode, while the anion has a more significant impact on the ORR reaction above 65 RH% where water molecules begin to disintegrate the EDL structures. AFM force-curve results indicated that water is present at the electrode-RTIL interface in [C2mim][NTf2] and [N4,1,1,1][NTf2], but is not detected in the hydrophobic RTIL [P14,6,6,6][NTf2], which agrees well with the CV results. A much greater level of water incorporation in [C2mim][NTf2] was detected, which explains why the ORR current in this RTIL was highly affected by the presence of moisture even at extremely low humidity levels (e.g. < 5 RH %). The innermost EDL layers are stabilised at the negatively polarised electrode, and the water layer near the electrode-RTIL interface becomes more apparent in [C2mim][NTf2] and [N4,1,1,1][NTf2] due to the presence of ions and/or water molecules whereas the outer layers vanish due to the water incorporation. However, in [P14,6,6,6][NTf2], the longer alkyl chain of the cation reduces the layering effect of the negative surface bias, and even at an electrode bias of -0.8 V, no indication of a water-layer at the electrode-RTIL interface was observed. These results suggest that for applications where a stable ORR current response is necessary, the use of highly hydrophobic and large RTIL cations and anions (e.g. [P14,6,6,6]+ and [FAP]-) is recommended. Alternatively, if a large change with humidity is desired (e.g. for humidity sensors), the choice of hydrophillic ionic liquids could be favourably exploited. References 1. Stetter, J. R.; Li, J., Amperometric Gas Sensors - a Review. Chem. Rev. 2008, 108, 352-366. 2. Xiong, L.; Compton, R. G., Amperometric Gas Detection: A Review. Int. J. Electrochem. Sci. 2014, 9, 7152-7181. 3. Freeman, B. D., Basis of Permeability/Selectivity Tradeoff Relations in Polymeric Gas Separation Membranes. Macromolecules 1999, 32. 4. Buzzeo, M. C.; Hardacre, C.; Compton, R. G., Use of Room Temperature Ionic Liquids in Gas Sensor Design. Anal. Chem. 2004, 76, 4583-4588. 5. Malvaldi, M.; Chiappe, C., From Molten Salts to Ionic Liquids: Effect of Ion Asymmetry and Charge Distribution. J. Phys.: Condens. Matter 2008, 20, 035108. 6. Freire, M. G.; Santos, L. M. N. B. F.; Fernandes, A. M.; Coutinho, J. A. P.; Marrucho, I. M., An Overview of the Mutual Solubilities of Water-Imidazolium-Based Ionic Liquid Systems. Fluid Ph. Equilibria 2007, 261, 449-454. 7. Welton, T., Room-Temperature Ionic Liquids. Solvents for Synthesis and Catalysis. Chemical Reviews 1999, 99, 2071-2084. 8. Buzzeo, M. C.; Evans, R. G.; Compton, R. G., Non‐Haloaluminate Room‐Temperature Ionic Liquids in Electrochemistry—a Review. ChemPhysChem 2004, 5, 1106-1120. 9. Shiddiky, M. J. A.; Torriero, A. A. J., Application of Ionic Liquids in Electrochemical Sensing Systems. Biosens Bioelectron 2011, 26, 1775-1787. 10. Shvedene, N. V.; Chernyshov, D. V.; Pletnev, I. V., Ionic Liquids in Electrochemical Sensors. Russian Journal of General Chemistry 2008, 78, 2507-2520. 11. Silvester, D. S., Recent Advances in the Use of Ionic Liquids for Electrochemical Sensing. Analyst 2011, 136, 4871-4882. 12. Marsh, K. N.; Boxall, J. A.; Lichtenthaler, R., Room Temperature Ionic Liquids and Their Mixtures- a Review. Fluid Ph. Equilibria 2004, 219, 93-98. 13. Barrosse-Antle, L. E.; Bond, A. M.; Compton, R. G.; O'Mahony, A. M.; Rogers, E. I.; Silvester, D. S., Voltammetry in Room Temperature Ionic Liquids: Comparisons and Contrasts with Conventional Electrochemical Solvents. Chemistry – An Asian Journal 2010, 5, 202-230. 14. Hayyan, M.; Mjalli, F. S.; Hashim, M. A.; AlNashef, I. M.; Mei, T. X., Investigating the Electrochemical Windows of Ionic Liquids. J. Ind. Eng. Chem. 2013, 19, 106-112. 15. Rehman, A.; Zeng, X., Methods and Approaches of Utilizing Ionic Liquids as Gas Sensing Materials. RSC Adv. 2015, 5, 58371 - 58392. 16. Khan, A.; Gunawan, C. A.; Zhao, C., Oxygen Reduction Reaction in Ionic Liquids: Fundamentals and Applications in Energy and Sensors. ACS Sustain. Chem. Eng. 2017, 5, 3698-3715. 17. 5. Acknowledgements This work was supported by an Australian Research Council (ARC) Future Fellowship Award (FT170100315) for DSS. The authors thank both Professor Christopher Hardacre and Professor Rob Atkin for kind donation of two of the RTILs used in this work. The authors acknowledge the use of equipment of the instruments of the Scanning Probe Microscopy facility of the CIFMI/Discipline of Chemistry at Curtin University, funded by ARC LIEF grant number LE130100121. 17 References Chi, X.; Tang, Y.; Zeng, X., Electrode Reactions Coupled with Chemical Reactions of Oxygen, Water and Acetaldehyde in an Ionic Liquid: New Approaches for Sensing Volatile Organic Compounds. Electrochimica Acta 2016, 216, 171-180. 18. Lee, J.; Silvester, D. S., Electrochemical Detection of Explosive Compounds in an Ionic Liquid in Mixed Environments: Influence of Oxygen, Moisture, and Other Nitroaromatics on the Sensing Response. Australian Journal of Chemistry 2018. 19. Barnes, A. S.; Rogers, E. I.; Streeter, I.; Aldous, L.; Hardacre, C.; Wildgoose, G. G.; Compton, R. G., Unusual Voltammetry of the Reduction of O2 in [C4dmim][N(Tf)2] Reveals a Strong Interaction of O2•− with the [C4dmim]+ Cation. J. Phys. Chem. C 2008, 112, 13709-13715. 20. Hayyan, M.; Mjalli, F. S.; Hashim, M. A.; AlNashef, I. M.; Tan, X. M., Electrochemical Reduction of Dioxygen in Bis(Trifluoromethylsulfonyl)Imide Based Ionic Liquids. J. Electroanal. Chem. 2011, 657, 150-157. 21. Katayama, Y.; Onodera, H.; Yamagata, M.; Miura, T., Electrochemical Reduction of Oxygen in Some Hydrophobic Room-Temperature Molten Systems. J. Electrochem. Soc. 2004, 151, A59-A63. 22. Switzer, E. E.; Zeller, R.; Chen, Q.; Sieradzki, K.; Buttry, D. A.; Friesen, C., Oxygen Reduction Reaction in Ionic Liquids: The Addition of Protic Species. J. Phys. Chem. C 2013, 117, 8683-8690. 23. Zhang, D.; Okajima, T.; Matsumoto, F.; Ohsaka, T., Electroreduction of Dioxygen in 1-N-Alkyl-3- Methylimidazolium Tetrafluoroborate Room-Temperature Ionic Liquids. J. Electrochem. Soc. 2004, 151, D31-D37. 23. Zhang, D.; Okajima, T.; Matsumoto, F.; Ohsaka, T., Electroreduction of Dioxygen in 1 N Alkyl 3 Methylimidazolium Tetrafluoroborate Room-Temperature Ionic Liquids. J. Electrochem. Soc. 2004, 151, D31-D37. 24. Yuan, X.-Z.; Alzate, V.; Xie, Z.; Ivey, D. G.; Qu, W., Oxygen Reduction Reaction in 1-Butyl-1-Methyl-Pyrrolidinium Bis(Trifluoromethanesulfonyl)Imide: Addition of Water as a Proton Species. J. Electrochem. Soc. 2014, 161, A451-A457. 25. Xiong, L.; Barnes, E. O.; Compton, R. G., Amperometric Detection of Oxygen under Humid Conditions: The Use of a Chemically Reactive Room Temperature Ionic Liquid to 'Trap' Superoxide Ions and Ensure a Simple One Electron Methylimidazolium Tetrafluoroborate Room-Temperature Ionic Liquids. J. Electrochem. Soc. 2004, 151, D31-D37. 24. Yuan, X.-Z.; Alzate, V.; Xie, Z.; Ivey, D. G.; Qu, W., Oxygen Reduction Reaction in 1-Butyl-1-Methyl-Pyrrolidinium Bis(Trifluoromethanesulfonyl)Imide: Addition of Water as a Proton Species. J. Electrochem. Soc. 2014, 161, A451-A457. 25 Xiong L ; Barnes E O ; Compton R G Amperometric Detection of Oxygen under Humid Conditions: The Use 24. Yuan, X.-Z.; Alzate, V.; Xie, Z.; Ivey, D. References Rivera-Rubero, S.; Baldelli, S., Surface Spectroscopy of Room-Temperature Ionic Liquids on a Platinum Electrode: A Sum Frequency Generation Study. J. Phys. Chem. B 2004, 108, 15133-15140. 38. Nishi, N.; Minami, K.; Motobayashi, K.; Osawa, M.; Sakka, T., Interfacial Structure at the Quaternary Ammonium-Based Ionic Liquid/Gold Electrode Interface Probed by Surface-Enhanced Infrared Absorption Spectroscopy: Anion Dependence of the Cationic Behavior. J. Phys. Chem. C 2017, 121, 1658-1666. 39. Hoth, J.; Hausen, F.; Mueser, M. H.; Bennewitz, R., Force Microscopy of Layering and Friction in an Ionic Liquid. J. Phys. Condens. Matter 2014, 26, 284110. 40. McLean, B.; Li, H.; Stefanovic, R.; Wood, R. J.; Webber, G. B.; Ueno, K.; Watanabe, M.; Warr, G. G.; Page, A.; Atkin, R., Nanostructure of [Li(G4)] Tfsi and [Li(G4)] No3 Solvate Ionic Liquids at Hopg and Au(111) Electrode Interfaces as a Function of Potential. Phys. Chem. Chem. Phys. 2015, 17, 325-333. 41. Zhong, Y.-X.; Yan, J.-W.; Li, M.-G.; Zhang, X.; He, D.-W.; Mao, B.-W., Resolving Fine Structures of the Electric Double Layer of Electrochemical Interfaces in Ionic Liquids with an Afm Tip Modification Strategy. J. Am. Chem. Soc. 2015, 136, 14682-14685. 42. Carstens, T.; Hayes, R.; El Abedin, S. Z.; Corr, B.; Webber, G. B.; Borisenko, N.; Atkin, R.; Endres, F., In Situ Stm, Afm and Dts Study of the Interface 1-Hexyl-3-Methylimidazolium Tris(Pentafluoroethyl)Trifluorophosphate/Au(111). Electrochim. Acta 2012, 82, 48-59. 43. Ivanistsev, V.; O'Connor, S.; Fedorov, M. V., Poly(a)Morphic Portrait of the Electrical Double Layer Ionic Liquids. Electrochem. Commun. 2014, 48, 61-64. 44. Fedorov, M. V.; Kornyshev, A. A., Towards Understanding the Structure and Capacitance of Electrical Double Layer in Ionic Liquids. Electrochim. Acta 2008, 53, 6835-6840. 45. Feng, G.; Jiang, X.; Qiao, R.; Kornyshev, A. A., Water in Ionic Liquids at Electrified Interfaces: The Anatomy of Electrosorption. ACS Nano 2014, 8, 11685-11694. 46. Nikitina, V. A.; Kislenko, S. A.; Nazmutdinov, R. R.; Bronshtein, M. D.; Tsirlina, G. A., Ferrocene/Ferrocenium Redox Couple at Au(111)/Ionic Liquid and Au(111)/Acetonitrile Interfaces: A Molecular-Level View at the Elementary Act. J. Phys. Chem. C 2014, 118, 6151-6164. y 47. Zhou, T.; Chen, L.; Ye, Y.; Chen, L.; Qi, Z.; Freund, H.; Sundmacher, K., An Overview of Mutual Solubility of Ionic Liquids and Water Predicted by Cosmo-Rs. Ind. Eng. Chem. Res. 2012, 51, 6256-6264. 48. Bi, S.; Wang, R.; Liu, S.; Yan, J.; Mao, B.; Kornyshev, A. A.; Feng, G., Minimizing the Electrosorption of Water from Humid Ionic Liquids on Electrodes. Nat. Commun. References G.; Qu, W., Oxygen Reduction Reaction in 1-Butyl-1-Methyl-Pyrrolidinium Bis(Trifluoromethanesulfonyl)Imide: Addition of Water as a Proton Species. J. Electrochem. Soc. 2014, 161, A451-A457. 25. Xiong, L.; Barnes, E. O.; Compton, R. G., Amperometric Detection of Oxygen under Humid Conditions: The Use of a Chemically Reactive Room Temperature Ionic Liquid to 'Trap' Superoxide Ions and Ensure a Simple One Electron Reduction. Sens. Actuators B 2014, 200, 157-166. 18 26. Tian, S.; Hou, Y.; Wu, W.; Ren, S.; Qian, J., Hydrophobic Task-Specific Ionic Liquids: Synthesis, Properties and Application for the Capture of So2. J. Hazard. Mat. 2014, 278, 409-416. 27. Fukay, Y.; Ohno, H., Hydrophobic and Polar Ionic Liquids. Phys. Chem. Chem. Phys. 2013, 15, 4066-4072. 28. Messadi, A.; Mohamadou, A.; Boudesocque, S.; Dupont, L.; Fricoteaux, P.; Nguyen-Van-Nhien, A.; Courty, M., Syntheses and Characterisation of Hydrophobic Ionic Liquids Containing Trialkyl(2-Ethoxy-2-Oxoethyl)Ammonium or N-(1-Methylpyrrolidyl-2-Ethoxy-2-Oxoethyl)Ammonium Cations. J. Mol. Liq. 2013, 184, 68-72. 29. Fuller, J.; Breda, A. C.; Carlin, R. T., Ionic Liquid-Polymer Gel Electrolytes from Hydrophilic and Hydrophobic Ionic Liquids. J. Electroanal. Chem. 1998, 459, 29-34. 30. Guan, J.; Li, Y.; Li, J., Stretchable Ionic-Liquid-Based Gel Polymer Electrolytes for Lithium-Ion Batteries. Ind. Eng. Chem. Res. 2017, 56, 12456-12463. 31. Ruzanov, A.; Lembinen, M.; Jakovits, P.; Srirama, S. N.; Voroshylova, I. V.; Cordeiro, M. N. D. S.; Pereira, C. M.; Rossmeisl, J.; Ivaništšev, V. B., On the Thickness of the Double Layer in Ionic Liquids. Phys. Chem. Chem. Phys. 2018, 20, 10275-10285. , 32. Hayes, R.; Warr, G. G.; Atkin, R., Structure and Nanostructure in Ionic Liquids. Chem. Rev. 2015, 115, 6357- 6426. 33. Li, H.; Endres, F.; Atkin, R., Effect of Alkyl Chain Length and Anion Species on the Interfacial Nanostructure of Ionic Liquids at the Au(111)-Ionic Liquid Interface as a Function of Potential. Phys. Chem. Chem. Phys. 2013, 15, 14624- 14633. 34. Hayes, R.; Borisenko, N.; Tam, M. K.; Howlett, P. C.; Endres, F.; Atkin, R., Double Layer Structure of Ionic Liquids at the Au(111) Electrode Interface: An Atomic Force Microscopy Investigation. J. Phys. Chem. C 2011, 115, 6855-6863. 35. Perkin, S., Ionic Liquids in Confined Geometries. Phys. Chem. Chem. Phys. 2012, 14, 5052-5062. 35. Perkin, S., Ionic Liquids in Confined Geometries. Phys. Chem. Chem. Phys. 2012, 14, 5 36. Lauw, Y.; Rodopoulos, T.; Gross, M.; Nelson, A.; Gardner, R.; Horne, D., Electrochemical Cell for Neutron Reflectometry Studies of the Structure of Ionic Liquids at Electrified Interface. Rev. Sci. Instrum. 2010, 81, 074101. 37. References 2018, 9, 5222. 49. Sun, J.; MacFarlane, D. R.; Forsyth, M., Synthesis and Properties of Ambient Temperature Molten Salts Based on the Quaternary Ammonium Ion. Ionics 1997, 3, 356-362. 19 50. Wandt, J.; Lee, J.; Arrigan, D.; Silvester, D. S., A Lithium Iron Phosphate Reference Electrode for Ionic Liquid Electrolytes. Electrochem. Commun. 2018, 93, 148-151. 51. Lee, J.; Du Plessis, G.; Arrigan, D. W. M.; Silvester, D. S., Towards Imroving the Robustness of Electrochemical Gas Sensors: Impact of Pmma Addition on the Sensing of Oxygen in an Ionic Liquid. Anal. Methods 2015, 7, 7327-7335. 51. Lee, J.; Du Plessis, G.; Arrigan, D. W. M.; Silvester, D. S., Towards Imroving the Robustness of Electrochemical Gas Sensors: Impact of Pmma Addition on the Sensing of Oxygen in an Ionic Liquid. Anal. Methods 2015, 7, 7327-7335. 52. Wagner, W.; Pruss, A., The Iapws Formulation 1995 for the Thermodynamic Properties of Ordinary Water Substance for General Scientific Use. J. Phys. Chem. Ref. Data 2002, 31, 387535. 52. Wagner, W.; Pruss, A., The Iapws Formulation 1995 for the Thermodynamic Properties of Ordinary Water Substance for General Scientific Use. J. Phys. Chem. Ref. Data 2002, 31, 387535. 53. Oyi, V. Humidity Conversion Formulas; B210973EN-F; 2013. 53. Oyi, V. Humidity Conversion Formulas; B210973EN-F; 2013. 54. O’Mahony, A. M.; Silvester, D. S.; Aldous, L.; Hardacre, C.; Compton, R. G., Effect of Water on the Electrochemical Window and Potential Limits of Room-Temperature Ionic Liquids. J. Chem. Eng. Data 2008, 53, 2884- 2891. 55. Cheng, H. W.; Dienemann, J. N.; Stock, P.; Merola, C.; Chen, Y. J.; Valtiner, M., The Effect of Water and Confinement on Self-Assembly of Imidazolium Based Ionic Liquids at Mica Interfaces. Sci. Rep. 2016, 6, 30058. 56. Cheng, H.-W.; Stock, P.; Moeremans, B.; Baimpos, T.; Banquy, X.; Renner, F. U.; Valtiner, M., Characterizing the Influence of Water on Charging and Layering at Electrified Ionic-Liquid/Solid Interfaces. Adv. Mater. Interfaces 2015, 2, 1500159. 57. Ignat'ev, N. V.; Welz-Biermann, U.; Kucheryna, A.; Bissky, G.; Willner, H., New Ionic Liquids with Tris(Perfluoroalkyl)Trifluorophosphate (Fap) Anions. J. Fluor. Chem. 2005, 126, 1150-1159. 58. Budkov, Y. A.; Kolesnikov, A. L.; Goodwin, Z. A. H.; Kiselev, M. G.; Kornyshev, A. A., Theory of Electrosorption of Water from Ionic Liquids. Electrochim. Acta 2018, 284, 346-354. 59. Khan, I.; Kurnia, K. A.; Mutelet, F.; Pinho, S. P.; Coutinho, J. A. P., Probing the Interactions between Ionic Liquids and Water: Experimental and Quantum Chemical Approach. J. Phys. References Chem. B 2014, 118, 1848-1860. 60. Zhong, Y.; Yan, J.; Li, M.; Chen, L.; Mao, B., The Electric Double Layer in an Ionic Liquid Incorporated with Water Molecules: Atomic Force Microscopy Force Curve Study. ChemElectroChem 2016, 3, 2221-2226. 61. Freire, M. G.; Neves, C. M. S. S.; Marrucho, I. M.; Coutinho, J. A. P.; Fernandes, A. M., Hydrolysis of Tetrafluoroborate and Hexafluorophosphate Counter Ions in Imidazolium-Based Ionic Liquids. J. Phys. Chem. A 2010, 114, 3744-3749. 62. Fedorov, M. V.; Konrnyshev, A. A., Ionic Liquids at Electrified Interfaces. A. Chem. Rev. 2 62. Fedorov, M. V.; Konrnyshev, A. A., Ionic Liquids at Electrified Interfaces. A. Chem. Rev. 2014, 114, 2978-3036. 63. Feng, G.; Zhang, J. S.; Qiao, R., Microstructure and Capacitance of the Electrical Double Layers at the Interface of Ionic Liquids and Planar Electrodes. J. Phys. Chem. C 2009, 113, 4549-4559. , ; y , , q , , 63. Feng, G.; Zhang, J. S.; Qiao, R., Microstructure and Capacitance of the Electrical Double Layers at the Interface of Ionic Liquids and Planar Electrodes. J. Phys. Chem. C 2009, 113, 4549-4559. 64. Lockett, V.; Horne, M.; Sedev, R.; Rodopoulos, T.; Ralston, J., Differential Capacitance of the Double Layer at the Electrode/Ionic Liquids Interface. Phys. Chem. Chem. Phys. 2010, 12, 12499-12512. 65. Lockett, V.; Sedev, R.; Harmer, S.; Ralston, J.; Horne, M.; Rodopoulos, T., Orientation and Mutual Location of Ions at the Surface of Ionic Liquids. Phys. Chem. Chem. Phys. 2010, 12, 13816-13827. 66. Hayes, R. Structure in Ionic Liquids. The University of Newcastle, Australia, 2014. 67. Kornyshev, A. A., Double-Layer in Ionic Liquids: Paradigm Change? J. Phys. Chem. B 2 20
https://openalex.org/W2977583882	http://izdat.istu.ru/index.php/ISM/article/download/3394/2187	English	null	Estimation of Module for the Function Which Is Analytic at Half-Plane	Intellektualʹnye sistemy v proizvodstve	2,017	cc-by	8,200	?- ? , ?, @A f z @ ? C A - A y f iy e ෶෶ ෶ ෶ , ? D E E f z ෶ ෶ , z (2) (3) ! Rei z : ! Rei z : 2 2 Re Re Re Re Re Re 1. Re Re i i k k i i k k i i i k k i i i k k R z R z R z R z e R z R z e R z R z (4) (4) ! cos sin z t i , 0, , z t , t R : ISSN 1813-7911. . 2016. 4 (31) ISSN 1813-7911. . 2016. 4 (31) 8 517.547(045) . . , . . , . . ! "#$% &"'* +%+'/:$;"< !"'!'";";/ =%# , - . ! , " # - $. . % . . . : ? @A , , B , ? , @A . . ! . 2 2 2 2 2 2 2 2 0 ln ln 1 ln Re 2 Re Re k R k k z Q k k i i i R z z R z z f z R z z R z z R z R z f d z z 2 2 2 2 2 2 2 2 1 sin sin ln 1 sin sin . R R R R r R r f t dt t z R zt r R r d t t z R zt (2) "A 2 2 ln k R k k z Q k k R z z R z z R z z R z z . (3) . ., 2016 ?- ? , ?, @A f z @ ? C A - A y f iy e ෶෶ ෶ ෶ , ? D E E f z ෶ ෶ , z E D E? ?. / ? ? - C = – ' @ , , , E ? @A A B . !E A- ?B , ? B - ? B @A ? F – / . G ? E H. . & E D «& ? E- » 1966 ., @A f z , ? CJ A, ? - A, C? E? C ?, ? [1]. " , ? D K . , ?, @A f z D . . ! [2], D & . / - CJ A : z f z Ke ෶෶ ෶ ෶ , 1 0 K e K A ? [3]. : 3 3 4 3 2 2 2 2 2 2 2 2 0 4 sin sin 4 sin sin ln Re 1 ln 2 2 4 cos cos 4 sin sin cos cos i R r Rr f d f z R r R Rr r R r Rr R 4 2 3 2 2 2 3 2 2 2 2 4 2 2 2 2 2 sin sin sin sin ln 1 . cos sin cos sin 2 cos R R R r r t R rt R r f t dt t r r R t r r r t R tr (14) 2 2 2 2 2 2 2 2 2 2 2 2 sin 1 ln cos sin cos sin R R r R r R t f t dt t r r R rt r t 4 2 3 2 2 2 3 2 2 2 2 4 2 2 2 2 2 1 sin sin sin sin ln cos sin cos sin 2 cos R R R r r t R rt R r f t dt t r r R t r r r t R tr . 2 2 2 2 1 k k k k k k k k R t z R z t t z R z t R z t t z R z t R t z . (5) 1. f z " ! : 0 x i y y ; &' : (5) ! A , ? (4) (5) R Q , ? 2 2 0 ln 1 ln 2 lim ln Re , i R f t y f z r f d dt R t x y 2 2 0 ln 1 ln 2 lim ln Re , i R f t y f z r f d dt R t x y 2 2 ln ln 1 0 k R k R k k z Q z Q k k R z z R z z R z z R z z . (6) (1) (1) ! z (6) . & f z '& ' ( " R Q [4] \| R R , i R z re Q : F D 9 2 2 2 2 2 2 2 2 sin sin 1 sin r R r R t z t z R zt R zt , (7) i z x i y re . 2 2 2 2 2 2 2 2 2 1 sin cos sin cos sin R r t r r R rt r t 4 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 cos 2 cos sin cos sin cos sin R r t R rt R t r R R rt r t r r R rt r t 2 2 2 2 2 2 2 2 2 2 2 2 2 2 sin cos sin cos sin R R r t R r r t r r R rt r t 2 2 2 2 2 2 2 2 2 2 2 2 sin 0 cos sin cos sin r R r R t t r r R rt r t , (8) , ? 2 2 2 2 1 sin sin 0 R R r R r d t t z R zt , (9) 0. d t F 2 2 2 2 1 sin sin ln R R r R r f t dt t z R zt , (10) ? (8), ? (11) (11) , : 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 cos cos sin sin cos cos sin sin cos cos sin sin cos cos sin sin Re Re i i i i R r R r R r R r R r R r R r R r R r R r R r re re 2 2 2 2 2 2 2 2 2 2 2 2 2 4 sin sin 4 cos cos 4 sin sin cos cos R r Rr R r Rr R r R r 3 3 4 3 2 2 2 2 2 2 2 2 1 4 sin sin 4 sin sin 2 4 cos cos 4 sin sin cos cos R r Rr R r R Rr r R r Rr R . (12) (12) / D 3 3 2 2 2 2 2 2 4 3 2 2 2 2 2 2 2 2 0 0 4 sin sin 4 sin sin ln Re 1 1 ln Re . 2 2 Re Re 2 4 cos cos 4 sin cos n c s si o i i i i R r Rr f d R z R z f d R r z z R Rr r R r Rr R 3 3 2 2 2 2 2 2 4 3 2 2 2 2 2 2 2 2 0 0 4 sin sin 4 sin sin ln Re 1 1 ln Re . 2 2 Re Re 2 4 cos cos 4 sin cos n c s si o i i i i R r Rr f d R z R z f d R r z z R Rr r R r Rr R R (13) ? (6), (9), (11) (13), ? : (13) ? (6), (9), (11) (13), ? : ? (6), (9), (11) (13), ? : ? (6), (9), (11) (13), ? : 3 3 4 3 2 2 2 2 2 2 2 2 0 4 sin sin 4 sin sin ln Re 1 ln 2 2 4 cos cos 4 sin sin cos cos i R r Rr f d f z R r R Rr r R r Rr R 4 2 3 2 2 2 3 2 2 2 2 4 2 2 2 2 2 sin sin sin sin ln 1 . , ? K D - CJ E : 2 2 2 2 2 2 2 2 2 2 2 2 sin 1 ln cos sin cos sin R R r R r R t f t dt t r r R rt r t 4 2 3 2 2 2 3 2 2 2 2 4 2 2 2 2 2 1 sin sin sin sin ln cos sin cos sin 2 cos R R R r r t R rt R r f t dt t r r R t r r r t R tr . (11) , ? (7) 2 2 2 2 1 sin sin 0 R R r R r d t t z R zt , (9) (9) 0. d t F 0. d t F 2 2 2 2 1 sin sin ln R R r R r f t dt t z R zt , (10) (10) ? (8), ? , ? K D - CJ E : ? (8), ? , ? K D - CJ E : 2 2 2 2 2 2 2 2 2 2 2 2 sin 1 ln cos sin cos sin R R r R r R t f t dt t r r R rt r t 4 2 3 2 2 2 3 2 2 2 2 4 2 2 2 2 2 1 sin sin sin sin ln cos sin cos sin 2 cos R R R r r t R rt R r f t dt t r r R t r r r t R tr . (11) , : 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 cos cos sin sin cos cos sin sin cos cos sin sin cos cos sin sin Re Re i i i i R r R r R r R r R r R r R r R r R r R r R r re re 2 2 2 2 2 2 2 2 2 2 2 2 2 4 sin sin 4 cos cos 4 sin sin cos cos R r Rr R r Rr R r R r 3 3 4 3 2 2 2 2 2 2 2 2 1 4 sin sin 4 sin sin 2 4 cos cos 4 sin sin cos cos R r Rr R r R Rr r R r Rr R . (12) / D 3 3 2 2 2 2 2 2 4 3 2 2 2 2 2 2 2 2 0 0 4 sin sin 4 sin sin ln Re 1 1 ln Re . 2 2 Re Re 2 4 cos cos 4 sin cos n c s si o i i i i R r Rr f d R z R z f d R r z z R Rr r R r Rr R (13) ? (6), (9), (11) (13), ? / @A C . . ! . 3. * ! p f z N , 3. * ! p f z N , ln ln a a . (18) (18) 2 2 ln ln f it x f z dt t y x (24) ? (18), (15) ? (18), (15) (24) 2 2 0 ln 1 ln 2 lim ln Re . i R f t y f z r f d dt R t x y " ! : 0 z x i y x . " ! : 0 z x i y x . " ! : 0 z x i y x . . ! 2. =A ln f z A . ! 2. =A ln f z A ) . 2. f z ! : 0 x i y x ; &' : 2 2 2 2 ln 1 ln 2 lim ln Rei R f it x f z r f d dt R t y x . 2 2 2 2 ln 1 ln 2 lim ln Rei R f it x f z r f d dt R t y x . / ! p f z N , 2 2 2 2 ln 1 ln 2 lim ln , . 2 2 2 sin ln 1 , cos sin r f t dt t r r ) . "E ? ? ! p N ? B ! { : 0} x iy x @A f z , CJ B : ? , ? cos x r , sin y r , D 0 sin 1 0, 0 sin 1 0, , ? ? , ? cos x r , sin y r , D 0 sin 1 0, 0 sin 1 0, , ? 1) 2 0 _ ln sup 1 p x f x iy dy y # $ $ ! % $ $ " & , 0, p 2 2 0 ln 1 ln 2 lim ln Re . i R f t dt y f z r f d R t x y ln , 0; ln 0, 0 1. a a a a ! " (15) ( ) ! @A C ln a CJ E : ( ) ! @A C ln a CJ E : ln ln ln a a a , (16) ln ln ln a a a , (16) 2) 2 2 1 lim ln Re 0. i R f d R ln , 1 ln 0, 0 1 a a a a ! " 0, 1 ln ln , 0 1 a a a a ! " (17) ln ln a a (18) / @A C . . ! . cos sin cos sin 2 cos R R R r r t R rt R r f t dt t r r R t r r r t R tr (14) 3 3 4 3 2 2 2 2 2 2 2 2 0 4 sin sin 4 sin sin ln Re 1 ln 2 2 4 cos cos 4 sin sin cos cos i R r Rr f d f z R r R Rr r R r Rr R 4 2 3 2 2 2 3 2 2 2 2 4 2 2 2 2 2 sin sin sin sin ln 1 . cos sin cos sin 2 cos R R R r r t R rt R r f t dt t r r R t r r r t R tr (14) (14) ISSN 1813-7911. . 2016. 4 (31) 10 ! R : ! R : ! R : ? (22) (23), (20) ? (22) (23), (20) 0 2 2 2 1 ln 2 lim sin sin ln Re sin ln 1 , cos sin i R f z r f d R r f t dt t r r 0 1 ln 2 lim sin sin ln Rei R f z r f d R 2 2 2 2 ln 1 ln 2 lim ln Rei R f it x f z r f d dt R t y x . i R f it x f z r f Re d dt z R t y x (19) ! p f z N , / (19) 2 2 1 lim ln Re 0 i R f d R , C- 2 2 1 lim ln Re 0 i R f d R , C- . ! 1 @A f w - ? B ! : 0 u iv v A: 2 2 0 ln 1 ln 2 lim ln Re , . ! i R f t v f w r f d dt w R t u v . ! 1 @A f w - ? B ! : 0 u iv v A: . ! 1 @A f w - ? B ? B ! : 0 u iv v A: ? B ! : 0 u iv v A: J : 2 2 ln ln f it x f z dt t y x . 2 2 0 ln 1 ln 2 lim ln Re , . ! i R f t v f w r f d dt w R t u v ) . "E ? ? A D B - CJ B B @A t [0, ) , CJ B : (20) (20) @ E ! ! J @A i w z . (21) w i z , u y , v x , (22) ln ln , 2 2 f w f i z f z ! ! . (23) i w z . (21) w i z , u y , v x , (22) (21) (21) a) te t 0; a) te t 0; b) 2 0 1 t dt t ! ; c) lim t t e t . ? ln ln ln , 2 2 f w f i z f z ! ! . (23) 11 1. 1. , @A m s , - @A C –": , @A m s , - @A C –": ! p f z N , 1 p , (25) (25) (25) 0 max s s t T t m s . (33) ) - 0 max s s t T t m s . (33) (33) ! & - & =A T t J CE 1 t , - J =A T t J CE 1 t , - ( ) y f iy e y ! ! ෶෶ , (26) (26) J 0 ln max ln ln s T t s t m s . (34) t A . + & ! (39) (39) G D, ? @A m s – C- J 0 s "A , J ? - (29). 1 0 m e . (32) (32) (37)–(39) 2 2 2 2 0 0 ln ln ln t t x e x e f z dt dt t y x t y x 1 1 2 2 2 2 0 0 ln ln( ) K T t K T t x x dt dt t y x t y x 0 0 1 0 1 0 2 2 2 2 0 0 (ln ln ln ) (ln ln ln ) s s K m s t K m s t x x dt dt t y x t y x 2 2 2 2 0 0 ln ln ln t t x e x e f z dt dt t y x t y x 1 1 2 2 2 2 0 0 ln ln( ) K T t K T t x x dt dt t y x t y x 0 0 1 0 1 0 2 2 2 2 0 0 (ln ln ln ) (ln ln ln ) s s K m s t K m s t x x dt dt t y x t y x 1 1 0 0 0 0 2 2 2 2 2 2 2 2 0 0 0 0 ln ln ln ln . (35) (35) F @A C 1 max s t t t m s e . (30) ? , 1 max s t t t m s e . (30) 1 max s t t t m s e . (30) ? , ? e' , 1 max s t t t m s e . (30) ? , ? e' , (30) (30) (30) ln , 0 e T e ' ' ( ' , (36) t e T t ( , 1 t , (37) 0 1 1 1 , , t t e K T t K e e T t . (38) ln , 0 e T e ' ' ( ' , (36) (36) =A m s B 0 s , B s , ? , ? t A , J- (37) 1 0 ln max ln max t m s s t t s e' ' ' . (31) 0 1 1 1 , , t t e K T t K e e T t . (38) ln m s @A , H @A e' . ln m s @A , H @A e' . ! 0s – A ? , , ? ! 0s – A ? , , ? 0 0 sr T r m s . (34) z f z Ke ෶෶ ෶ ෶ , 1 0 K e . (27) (27) =A lnT e' – H @A ln m s , , lnT t J- B 1 t . . 3 (26) - , ? F ? , 0 1 t ! . ! K B - ? B @A T t D , K ? s @A st m s E 2 2 2 2 ln ln . f it t x x f z dt dt t y x t y x ෶෶ (28) K ? s @A st m s E ! K CJ E : 2 2 2 2 0 0 ln ln ln t t x e x e f z dt dt t y x t y x . (29) K 1 (0 1) 0 T t t m ! (32) ? 2 2 2 2 0 0 ln ln ln t t x e x e f z dt dt t y x t y x . (29) F @A C K 1 (0 1) 0 T t t m ! . (32) ? 1 T t e , 0 1 t ! . (35) (32) ? 1 T t e , 0 1 t ! . 1. % $. . "E E @A , ?- B // & ? E . – 1966. – /. 69, 4. – ;. 658–662. 2. . . ? ? ?B - ? B @A . – &. ; '. : B , 1950. – 336 . 3. / . ., 0! *. . " B A B ? B @A // & D ?- . ; B ?. – D, 2003. – ;. 161–162. % A ! @A 2 2 1 , ln u x y x y I. V. Dzogij, Senior Teacher, Bryansk State University n.a. Academician I. G. Petrovski (43) (43) (41) (42) (41) (42) , z f z Ke ෶෶ ෶ ෶ 1 0 1 1 . K K e (49) (41) (42) , z f z Ke ෶෶ ෶ ෶ 1 0 1 1 . K K e (49) M ? @A , ? 0,0 , (49) + . 2 2 2 2 ln 2 x y x x x y ) ) ; 2 2 2 2 2 2 2 2 2 ln 2 2 x y y x x x y ) ) ; 2 2 2 2 ln 2 x y y y x y ) ) ; 2 2 2 2 2 2 2 2 2 ln 2 2 x y x y y x y ) ) ; 2 2 2 2 2 2 1 1 2 2 2 2 2 2 2 2 , , 2 2 2 2 0 u x y u x y y x x y x y x y x y ) ) ) ) . (44) K K m s m s s t s t x x x x dt dt dt dt t y x t y x t y x t y x (40) (40) ISSN 1813-7911. . 2016. 4 (31) 12 ISSN 1813-7911. 12 ? , J ? (40): 1 1 0 0 2 2 2 0 0 ln ln 1 K K t y d m s m s x x dt t y x t y x 1 1 0 0 0 ln ln arctg arctg 2 K K m s m s t y y x x ; (41) 1 1 0 0 2 2 2 0 0 ln ln 1 K K t y d m s m s x x dt t y x t y x 1 1 0 0 0 ln ln arctg arctg 2 K K m s m s t y y x x . (42) F @A C ? , J ? (40): 2 ln 0 2 t ln t . (45) ? , J ? (40): 2 ln 0 2 t ln t . (45) 2 ln 0 2 t ln t . (45) ? , J ? (40): 2 ln 0 2 t ln t . (45) ! @A C 2 2 1 , ln u x y x y 2ln r ? ! Re 0 z : 1 1 0 0 2 2 2 0 0 ln ln 1 K K t y d m s m s x x dt t y x t y x 2 2 2 2ln ln t x r dt t y x . (46) (46) D E ? K 0 2 s , - ? ? 0 0 2 2 ln s s ln t x r dt t y x 0 0 2 2 0 0 2 2 2 2 0 0 ln ln ln . s s ln t x r dt t y x s t s t x x dt dt t y x t y x (47) 0 1 1 0 0 2 2 2 0 0 ln ln 1 K K t y d m s m s x x dt t y x t y x (47) ! ? (47) D ? (40), E , ? 0 0 0 1 1 0 ln ln ln ln ln . s s ln f z r m s K r K m s (48) (48) 0 F @A C 0 F @A C 2 2 1 , ln u x y x y . Estimation of Module for the Function Which Is Analytic at Half-Plane The paper considers the estimation of module for the function which is analytic at the right half-plane provided that there is some estimation for the decrease in the function module at an imaginary axis or at its part. The extension of the function class to the class of Privalov I. I. has been obtained, the statement of the theorem by Maslyakov Y. I. being valid for this extension. Keywords: module of the analytic function, right half-plane, upper half-plane, unit circle, function class by Privalov I. I. !?: 02.11.16
https://openalex.org/W4300597678	https://repository.law.umich.edu/context/articles/article/1213/viewcontent/25BerkeleyTechLJ.pdf	English	null	The Invention of Common Law Play Right	Deep Blue (University of Michigan)	2,010	public-domain	35,332	The Invention of Common Law Play Right Available at: https://repository.law.umich.edu/articles/214 Follow this and additional works at: https://repository.law.umich.edu/articles Part of the Common Law Commons, Entertainment, Arts, and Sports Law Commons, Intellectual Property Law Commons, and the Legal History Commons Follow this and additional works at: https://repository.law.umich.edu/articles Part of the Common Law Commons, Entertainment, Arts, and Sports Law Commons, Intellectual Property Law Commons, and the Legal History Commons Follow this and additional works at: https://repository.law.umich.edu/articles University of Michigan Law School University of Michigan Law School Scholarship Repository Articles Faculty Scholarship 2010 The Invention of Common Law Play Right Jessica D. Litman University of Michigan Law School, jdlitman@umich.edu Available at: https://repository.law.umich.edu/articles/214 Follow this and additional works at: https://repository.law.umich.edu/articles Part of the Common Law Commons, Entertainment, Arts, and Sports Law Commons, Intellectual Property Law Commons, and the Legal History Commons This Article is brought to you for free and open access by the Faculty Scholarship at University of Michigan Law School Scholarship Repository. It has been accepted for inclusion in Articles by an authorized administrator of University of Michigan Law School Scholarship Repository For more Recommended Citation Litman, Jessica D. "The Invention of Common Law Play Right." Berkeley Tech. L. J. vol. 25, no. 3 (2010): 1381-425. ty of Michigan Law School rsity of Michigan Law School Scholarship Repository Faculty Scholarship 0 2010 Jessica Litman. t John F. Nickoll Professor of Law, University of Michigan. I'm grateful to Pam Samuelson for thinking up this conference and inviting me to do this research, and to Jon Weinberg, Jody Kraus, Jane Ginsburg, Becky Eisenberg, and H. Tomis G6mez-Arostegui for their helpful comments on earlier drafts. 1. EATON S. DRONE, TREATISE ON THE LAW OF PROPERTY IN INTELLECTUAL PRODUCTIONS IN GREAT BRITAIN AND THE UNITED STATES 553 (1879). Copyright in both Britain and the United States was initially limited to the right to print and sell copies. The Recommended Citation Recommended Citation Litman, Jessica D. "The Invention of Common Law Play Right." Berkeley Tech. L. J. vol. 25, no. 3 (2010): 1381-425. This Article is brought to you for free and open access by the Faculty Scholarship at University of Michigan Law School Scholarship Repository. It has been accepted for inclusion in Articles by an authorized administrator of University of Michigan Law School Scholarship Repository. For more information, please contact mlaw.repository@umich.edu. THE INVENTION OF COMMON LAW PLAY RIGHT Jessica Lztmant I. INTRODUCTION............................. ........... 1382 II. PLAY RIGHT IN ENGLAND ........................... 1387 A. ENGLISH THEATRE BEFORE THE STATUTE OF ANNE.................... 1389 B. ENGLISH THEATRE FROM THE STATUTE OF ANNE TO THE 1833 DRAMATIC LITERARY PROPERTY ACT......... ............... 1396 C. STATUTORY PLAY RIGHT IN ENGLAND ......................... 1397 III. PLAY RIGHT IN AMERICA ............................ 1400 A. EARLY AMERICAN THEATRE.................................. ....... 1401 B. THE 1856 DRAMATIC COPYRIGHT ACT.............................................. 1402 C. COMMON LAW RIGHTS ................................................ 1403 D. EATON DRONE'S NATURAL LAW PLAY RIGHT ................................ 1410 E. WHOSE PLAY RIGHT? ......................................... 1416 IV CONCLUSION 1424 IV. CONCLUSION A dramatic composition is capable of two distinct public uses. It may be printed as a book and represented as a drama... . The exclusive right of multiplying copies is called copyright. But this does not embrace the right of representation.... The sole liberty of publicly performing a dramatic composition might more properly be called dramatic right or acting right... . I have adopted playright as being, in my judgment, the best name for the purpose. It is a convenient euphonious word, and its formation is analogous to that of copyright. As the latter word literally means the right to copy a work, or the right to the copy, so playright means the right to play a drama, or the right to the play. Eaton Sylvester Drone' BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1382 -Eaton Sylvester Drone' -Eaton Sylvester Drone' HeinOnline -- 25 Berkeley Tech. L.J. 1381 2010 British Parliament and United States Congress later expanded copyright to encompass other rights. See infra notes 106-18, 134-38, 213-15, 261 and accompanying text. I. INTRODUCTION American copyright law nominally vests exclusive rights in "authors." 2 "Authors" in America, as often as not, are the employers of the individuals who actually create copyrightable works.' Even when the law vests copyright in creators, the architecture of the system encourages them to assign their copyrights to intermediaries, who are motivated by potential profits to disseminate the works to the public.4 Employers and assignees stand in the authors' shoes and may control a work's exploitation to the exclusion of the individuals who created it.' That's the essence of a copyright system that is fundamentally utilitarian in its design. With the narrow exceptions of painters and sculptors, American authors have no enforceable attribution or integrity rights.' The originality standard for "meriting" copyright protection is low, and follow-on creators who add even a little creativity to authorized adaptations of copyrighted works are entitled to exclusive rights in their versions of those works. This is the American version of copyright law in a nutshell.' In stark contrast to that model, consider the American playwright: the playwright keeps her copyright, rather than assigning it. In the United States, playwrights license public exploitation, and pay the intermediary exploiters with a share of the proceeds rather than ownership of the copyright.' Playwrights assert strong, apparently enforceable rights to attribution and integrity.' Playwrights, finally, insist that other creators who contribute 4. See Jessica Litman, Real Copyright Reform, 96 IOWA L. REV. (forthcoming 2010) (on file with author). 5. See, e.g., Okla. Natural Gas Co. v. Larue, Copyright L. Rep. (CCH) 27,814 (10th Cir. 1998); Gross v. Seligman, 212 Fed. 930 (2d Cir. 1914). 6. See ROBERTA KWALL, THE SOUL OF CREATIVITY: FORGING A MORAL RIGHT LAW FOR THE UNITED STATES 23-35 (2009). 7. See ARTHUR R. MILLER & MICHAEL H. DAVID, INTELLECTUAL PROPERTY: PATENTS, TRADEMARKS, AND COPYRIGHT IN A NUTSHELL 296-304, 383-404 (4th ed. 2007). ) 8. See, e.g., DONALD C. FARBER, PRODUCING THEATRE 1-33 (3d rev. ed. 2006). 9. See, e.g., DONALD C. FARBER, FROM OPTION TO OPENING: A GUIDE TO PRODUCING PLAYS OFF BROADWAY 21 (5th ed. rev. 2005); Samuel G. Freedman, Who's to Say Whether a Playwnght is Wronged, N.Y. TIMES, Dec. 23, 1984, at E6; United Press Int'l, Albee Seeking to Close All-Male "Woof "N.Y. TIMES, Aug. 3, 1984, at CS; Dramatists Guild of Am., Dramaists Bill of Rights, http://www.dramatistsguild.com/billofrights/ (last visited Aug. 2. See 17 U.S.C. § 201(a) (2006). 3. See U.S. Copyright Office, Works Made for Hire Under the 1976 CopyrightAct, Circular 9 (revised Apr. 2010). , g , 9. See, e.g., DONALD C. FARBER, FROM OPTION TO OPENING: A GUIDE TO PRODUCING PLAYS OFF BROADWAY 21 (5th ed. rev. 2005); Samuel G. Freedman, Who's to Say Whether a Playwnght is Wronged, N.Y. TIMES, Dec. 23, 1984, at E6; United Press Int'l, Albee Seeking to Close All-Male "Woof "N.Y. TIMES, Aug. 3, 1984, at CS; Dramatists Guild of Am., Dramaists Bill of Rights, http://www.dramatistsguild.com/billofrights/ (last visited Aug. 3, 2010); Dramatists Play Service, Frequenty Asked Quesions (FAQs), g f , , 2. See 17 U.S.C. § 201(a) (2006). 4. See Jessica Litman, Real Copyright Reform, 96 IOWA L. REV. (forthcoming 2010) (on file with author). British Parliament and United States Congress later expanded copyright to encompass other rights. See infra notes 106-18, 134-38, 213-15, 261 and accompanying text. 2. See 17 U.S.C. § 201(a) (2006). 3. See U.S. Copyright Office, Works Made for Hire Under the 1976 CopyrightAct, Circular 9 (revised Apr. 2010). 4. See Jessica Litman, Real Copyright Reform, 96 IOWA L. REV. (forthcoming 2010) (on file with author). 5. See, e.g., Okla. Natural Gas Co. v. Larue, Copyright L. Rep. (CCH) 27,814 (10th Cir. 1998); Gross v. Seligman, 212 Fed. 930 (2d Cir. 1914). 6. See ROBERTA KWALL, THE SOUL OF CREATIVITY: FORGING A MORAL RIGHTS LAW FOR THE UNITED STATES 23-35 (2009). 7. See ARTHUR R. MILLER & MICHAEL H. DAVID, INTELLECTUAL PROPERTY: PATENTS, TRADEMARKS, AND COPYRIGHT IN A NUTSHELL 296-304, 383-404 (4th ed. 2007). 8. See, e.g., DONALD C. FARBER, PRODUCING THEATRE 1-33 (3d rev. ed. 2006). 9. See, e.g., DONALD C. FARBER, FROM OPTION TO OPENING: A GUIDE TO PRODUCING PLAYS OFF BROADWAY 21 (5th ed. rev. 2005); Samuel G. Freedman, Who's to Say Whether a Playwnght is Wronged, N.Y. TIMES, Dec. 23, 1984, at E6; United Press Int'l, Albee Seeking to Close All-Male "Woof "N.Y. TIMES, Aug. 3, 1984, at CS; Dramatists Guild of Am., Dramaists Bill of Rights, http://www.dramatistsguild.com/billofrights/ (last visited Aug. 3, 2010); Dramatists Play Service, Frequenty Asked Quesions (FAQs), 7. See ARTHUR R. MILLER & MICHAEL H. DAVID, INTELLECTUAL PROPERTY: PATENTS, TRADEMARKS, AND COPYRIGHT IN A NUTSHELL 296-304, 383-404 (4th ed. 2007). 10. See, e.g., John Weidman, Protecting the American Playwrght: The Seventh Annual Media and Society Lecture, 72 BROOK. L. REV. 639, 641-42 (2007); Ralph Sevush, The Urinetown Papers: The U.S. Copyright Ofice Debunks the Notion of a 'Directors Copyright," THE DRAMATIST, Mar.-Apr. 2009, at 30; Dramatists Guild of Am., Dramatists' Copyright & Intellectual Property Rights, Apr. 2000, http://www.dramatistsguild.com/about-statements-copyright.aspx (last visited Feb. 2, 2010); see also Thomson v. Larsen, 147 F.3d 195 (2d Cir. 1998) (holding dramaturg was not a joint author of musical because composer/dramatist did not intend to share authorship credit); Erickson v. Trinity Theatre, 13 F.3d 1061 (7th Cir. 1994) (holding actors were not joint authors of three plays because their contributions were not independently copyrightable); Childress v. Taylor, 945 F.2d 500 (2d Cir. 1991) (holding actress not a joint author of one-woman play because dramatist did not intend to share authorship credit). But see, e.g., Carrie Ryan Gallia, Note, To Fix or Not to Fix, Copynght's Fixation Requirement and the Rights of Theatrical Collaborators, 92 MINN. L. REV. 231, 264 (2007) ("By transforming a stack of pages into a live performance, a director demonstrates sufficient originality."); Margit Livingston, Inspiration or Imitation? Copynrght Protection for Stage Directions, 50 B.C. L. REV. 427, 486 (2009) ("[A] straightforward application of traditional copyright law would dictate that stage directions are subject to copyright protection."). I. INTRODUCTION 3, 2010); Dramatists Play Service, Frequenty Asked Quesions (FAQs), HeinOnline -- 25 Berkeley Tech. L.J. 1382 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 2010] 1383 significant creative expression to licensed productions of their scripts have added no authorship and should receive no copyright protection for their additions. What accounts for the differences? There's no obvious language in the copyright statute suggesting that dramatists or their plays be treated as unique legal beasts. The current copyright statute includes several relatively minor provisions that single dramatic works out for more favorable treatment in particular narrow contexts." In general, the category of dramatic works (which Congress doesn't even bother to define) is treated the same way as other subject matter categories. Federal copyright cases involving dramatic works are surprisingly scant. For most of the 18th, 19th, and 20th centuries, the contours of federal statutory copyright and the formal prerequisites for perfecting it fit dramatic works poorly. Until 1978, American copyright law focused chiefly on works that had been published-for most works, publication was the quid pro quo 10. See, e.g., John Weidman, Protecting the American Playwrght: The Seventh Annual Media and Society Lecture, 72 BROOK. L. REV. 639, 641-42 (2007); Ralph Sevush, The Urinetown Papers: The U.S. Copyright Ofice Debunks the Notion of a 'Directors Copyright," THE DRAMATIST, Mar.-Apr. 2009, at 30; Dramatists Guild of Am., Dramatists' Copyright & Intellectual Property Rights, Apr. 2000, http://www.dramatistsguild.com/about-statements-copyright.aspx (last visited Feb. 2, 2010); see also Thomson v. Larsen, 147 F.3d 195 (2d Cir. 1998) (holding dramaturg was not a joint author of musical because composer/dramatist did not intend to share authorship credit); Erickson v. Trinity Theatre, 13 F.3d 1061 (7th Cir. 1994) (holding actors were not joint authors of three plays because their contributions were not independently copyrightable); Childress v. Taylor, 945 F.2d 500 (2d Cir. 1991) (holding actress not a joint author of one-woman play because dramatist did not intend to share authorship credit). But see, e.g., Carrie Ryan Gallia, Note, To Fix or Not to Fix, Copynght's Fixation Requirement and the Rights of Theatrical Collaborators, 92 MINN. L. REV. 231, 264 (2007) ("By transforming a stack of pages into a live performance, a director demonstrates sufficient originality."); Margit Livingston, Inspiration or Imitation? Copynrght Protection for Stage Directions, 50 B.C. L. REV. 427, 486 (2009) ("[A] straightforward application of traditional copyright law would dictate that stage directions are subject to copyright protection."). 11. http://www.dramatists.com/faqsmanager/applications/faqsmanager/ (last visited Aug. 3, 2010). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1384 for statutory copyright protection. 12 The vast majority of plays, even those that were exploited commercially, were never published as U.S. copyright law defined it.' 3 Playwrights in America until 1909 faced a choice of arranging to publish their scripts to secure federal statutory protection or relying on whatever copyright protection state courts might afford unpublished works.14 Th t dil d d C i 1909 t it i ht i t ti for statutory copyright protection. 12 The vast majority of plays, even those that were exploited commercially, were never published as U.S. copyright law defined it.' 3 Playwrights in America until 1909 faced a choice of arranging to publish their scripts to secure federal statutory protection or relying on whatever copyright protection state courts might afford unpublished works.14 That dilemma persuaded Congress in 1909 to permit copyright registration for unpublished plays." While thousands of unpublished play scripts were registered," very few of them became involved in federal litigation. 7 Important questions about how the copyright statute applied to works registered as unpublished, such as the duration of rights for unpublished works," remained unresolved for years." for statutory copyright protection. 12 The vast majority of plays, even those that were exploited commercially, were never published as U.S. copyright law defined it.' 3 Playwrights in America until 1909 faced a choice of arranging to publish their scripts to secure federal statutory protection or relying on whatever copyright protection state courts might afford unpublished works.14 That dilemma persuaded Congress in 1909 to permit copyright registration for statutory copyright protection. 12 The vast majority of plays, even those that were exploited commercially, were never published as U.S. copyright law defined it.' 3 Playwrights in America until 1909 faced a choice of arranging to publish their scripts to secure federal statutory protection or relying on whatever copyright protection state courts might afford unpublished works.14 That dilemma persuaded Congress in 1909 to permit copyright registration for unpublished plays." While thousands of unpublished play scripts were registered," very few of them became involved in federal litigation. 17. See, e.g., Rosen v. Loew's, Inc., 162 F.2d 785 (2d Cir. 1947) (holding the authors of "The Mortal Storm" did not copy "The Mad Dog of Europe"); Marx v. United States, 96 F.2d 204 (9th Cir. 1938) (affirming criminal conviction for infringement of radio script); Davis v. E.I DuPont De Nemours & Co., 240 F. Supp. 612 (S.D.N.Y. 1965) (holding play based on novel "Ethan Frome" to be infringed by TV movie). 15. Copyright Act of Mar. 4, 1909 5 11, 35 Stat. 1075, 1078 (later § 12). 16. See LIBRARY OF CONG., FiFrY-FOURTH ANNUAL REPORT OF THE REGISTER OF COPYRIGHTS FOR THE FISCAL YEAR ENDING JUNE 30, 1951, at 2 (1952); LIBRARY OF CONG., REPORT OF THE REGISTER OF COPYRIGHTS FOR THE FISCAL YEAR 1924-25, at 196- 97 (1925); LIBRARY OF CONG., REPORT OF THE REGISTER OF COPYRIGHTS FOR THE FISCAL YEAR 1918-19, at 138-39 (1919). I. INTRODUCTION Section 110 expressly limits some of its exemptions for noncommercial performances or displays to "non-dramatic literary or musical works," excluding dramatic works, choreographic works, pictorial graphic and sculptural works, motion pictures and other audiovisual works and architectural works. Section 118 limits the scope of the statutory noncommercial broadcasting license to "published nondramatic musical works and published pictorial, graphic and sculptural works." Section 121 limits the scope of the exemption for reproduction and distribution of copies or phonorecords in formats accessible for people with disabilities to "previously published, non-dramatic literary work." Those distinctions, though, don't begin to explain the more fundamental differences in copyright ownership and control described in the text accompanying notes 8-10. 11. Section 110 expressly limits some of its exemptions for noncommercial performances or displays to "non-dramatic literary or musical works," excluding dramatic works, choreographic works, pictorial graphic and sculptural works, motion pictures and other audiovisual works and architectural works. Section 118 limits the scope of the statutory noncommercial broadcasting license to "published nondramatic musical works and published pictorial, graphic and sculptural works." Section 121 limits the scope of the exemption for reproduction and distribution of copies or phonorecords in formats accessible for people with disabilities to "previously published, non-dramatic literary work." Those distinctions, though, don't begin to explain the more fundamental differences in copyright ownership and control described in the text accompanying notes 8-10. HeinOnline -- 25 Berkeley Tech. L.J. 1383 2010 18. See Marx v. United States, 96 F.2d 204 (9th Cir. 1938) (holding that twenty-eight year duration for published works should also apply to registered unpublished work, but the date of deposit should be deemed the date of publication). , ( 15. Copyright Act of Mar. 4, 1909 5 11, 35 Stat. 1075, 1078 (later § 12). 14. See Revision of the Copyright Laws: Hearings on Pending Bills to Amend and Consolidate the Acts Respecting Copynght Before the House and Senate Comm. on Patents, 60th Cong. 21-41 (1908) (testimony of witnesses representing the American Dramatists Club and the National Association of Theatrical Managers); 1 LIBRARY OF CONG. COPYRIGHT OFFICE, DRAMATIC COMPOSITIONS COPYRIGHTED IN THE UNITED STATES 1870 TO 1916, at 1 (1918). 12. L. Ray Patterson, Free Speech, Copyright and Fair Use, 40 VAND. L. REV. 1, 2-12, 33- 34 (1987); Pamela Samuelson, CONTU Revisited: The Case Against Copynght Protecion for Computer Programs in Machine-Readable Form, 1984 DUKE L.J. 663, 705-06. 13. See, e.g., BARRETT H. CLARK, 1 AMERICA'S LOST PLAYS v-vi (1940). 14. See Revision of the Copyright Laws: Hearings on Pending Bills to Amend and Consolidate the Acts Respecting Copynght Before the House and Senate Comm. on Patents, 60th Cong. 21-41 (1908) (testimony of witnesses representing the American Dramatists Club and the National Association of Theatrical Managers); 1 LIBRARY OF CONG. COPYRIGHT OFFICE, DRAMATIC COMPOSITIONS COPYRIGHTED IN THE UNITED STATES 1870 TO 1916, at 1 (1918). 15. Copyright Act of Mar. 4, 1909 5 11, 35 Stat. 1075, 1078 (later § 12). 16. See LIBRARY OF CONG., FiFrY-FOURTH ANNUAL REPORT OF THE REGISTER OF COPYRIGHTS FOR THE FISCAL YEAR ENDING JUNE 30, 1951, at 2 (1952); LIBRARY OF CONG., REPORT OF THE REGISTER OF COPYRIGHTS FOR THE FISCAL YEAR 1924-25, at 196- 97 (1925); LIBRARY OF CONG., REPORT OF THE REGISTER OF COPYRIGHTS FOR THE FISCAL YEAR 1918-19, at 138-39 (1919). 17. See, e.g., Rosen v. Loew's, Inc., 162 F.2d 785 (2d Cir. 1947) (holding the authors of "The Mortal Storm" did not copy "The Mad Dog of Europe"); Marx v. United States, 96 F.2d 204 (9th Cir. 1938) (affirming criminal conviction for infringement of radio script); Davis v. E.I DuPont De Nemours & Co., 240 F. Supp. 612 (S.D.N.Y. 1965) (holding play based on novel "Ethan Frome" to be infringed by TV movie). 18. See Marx v. United States, 96 F.2d 204 (9th Cir. 1938) (holding that twenty-eight year duration for published works should also apply to registered unpublished work, but the date of deposit should be deemed the date of publication). 19. William S. Strauss, Protection of Unpublished Works (1957), in 1 COPYRIGHT SOCIETY OF THE USA, STUDIES ON COPYRIGHT 190, 196 (1963). 20. See, e.g., HOUSE COMM. ON THE JUDICIARY, 88TH CONG., COPYRIGHT LAW REVISION PART 2: DISCUSSION AND COMMENTS ON THE REPORT OF THE REGISTER OF COPYRIGHTS ON THE GENERAL REVISION OF THE U.S. COPYRIGHT LAW 312 (Comm. Print 1963) (letter from Irwin Karp, May 29, 1961); PHILIP WITENBERG, THE PROTECTION AND MARKETING OF LITERARY PROPERTY 11-14 (1937); Benjamin Kaplan, Publication in Copyrght Law: The .uestion of Phonograph Records, 103 U. PA. L. REV. 469, 473-79 (1955). 12. L. Ray Patterson, Free Speech, Copyright and Fair Use, 40 VAND. L. REV. 1, 2-12, 33- 34 (1987); Pamela Samuelson, CONTU Revisited: The Case Against Copynght Protecion for Computer Programs in Machine-Readable Form, 1984 DUKE L.J. 663, 705-06. 12. L. Ray Patterson, Free Speech, Copyright and Fair Use, 40 VAND. L. REV. 1, 2-12, 33- 34 (1987); Pamela Samuelson, CONTU Revisited: The Case Against Copynght Protecion for Computer Programs in Machine-Readable Form, 1984 DUKE L.J. 663, 705-06. 13. See, e.g., BARRETT H. CLARK, 1 AMERICA'S LOST PLAYS v-vi (1940). 14. See Revision of the Copyright Laws: Hearings on Pending Bills to Amend and Consolidate the Acts Respecting Copynght Before the House and Senate Comm. on Patents, 60th Cong. 21-41 (1908) (testimony of witnesses representing the American Dramatists Club and the National Association of Theatrical Managers); 1 LIBRARY OF CONG. COPYRIGHT OFFICE, DRAMATIC COMPOSITIONS COPYRIGHTED IN THE UNITED STATES 1870 TO 1916, at 1 (1918). § p g , 13. See, e.g., BARRETT H. CLARK, 1 AMERICA'S LOST PLAYS v-vi (1940). p p 19. William S. Strauss, Protection of Unpublished Works (1957), in 1 COPYRIGHT SOCIETY OF THE USA, STUDIES ON COPYRIGHT 190, 196 (1963). 13. See, e.g., BARRETT H. CLARK, 1 AMERICA'S LOST PLAYS v-vi (1940). 14. See Revision of the Copyright Laws: Hearings on Pending Bills to Amend and Consolidate the Acts Respecting Copynght Before the House and Senate Comm. on Patents, 60th Cong. 21-41 (1908) (testimony of witnesses representing the American Dramatists Club and the National Association of Theatrical Managers); 1 LIBRARY OF CONG. COPYRIGHT OFFICE, DRAMATIC COMPOSITIONS COPYRIGHTED IN THE UNITED STATES 1870 TO 1916, at 1 (1918). 15. Copyright Act of Mar. 4, 1909 5 11, 35 Stat. 1075, 1078 (later § 12). 16. See LIBRARY OF CONG., FiFrY-FOURTH ANNUAL REPORT OF THE REGISTER OF COPYRIGHTS FOR THE FISCAL YEAR ENDING JUNE 30, 1951, at 2 (1952); LIBRARY OF CONG., REPORT OF THE REGISTER OF COPYRIGHTS FOR THE FISCAL YEAR 1924-25, at 196- 97 (1925); LIBRARY OF CONG., REPORT OF THE REGISTER OF COPYRIGHTS FOR THE FISCAL YEAR 1918-19, at 138-39 (1919). 17. See, e.g., Rosen v. Loew's, Inc., 162 F.2d 785 (2d Cir. 1947) (holding the authors of "The Mortal Storm" did not copy "The Mad Dog of Europe"); Marx v. United States, 96 F.2d 204 (9th Cir. 1938) (affirming criminal conviction for infringement of radio script); Davis v. E.I DuPont De Nemours & Co., 240 F. Supp. 612 (S.D.N.Y. 1965) (holding play based on novel "Ethan Frome" to be infringed by TV movie). 18. See Marx v. United States, 96 F.2d 204 (9th Cir. 1938) (holding that twenty-eight BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 7 Important questions about how the copyright statute applied to works registered as unpublished, such as the duration of rights for unpublished works," remained unresolved for years." Instead of relying on federal statutory copyright, lawyers for dramatists claimed that their clients' work was better protected under state "common law copyright." 20 The 1909 Copyright Act expressly preserved "the right of the author or proprietor of an unpublished work, at common law or in g 18. See Marx v. United States, 96 F.2d 204 (9th Cir. 1938) (holding that twenty-eight year duration for published works should also apply to registered unpublished work, but the date of deposit should be deemed the date of publication). 18. See Marx v. United States, 96 F.2d 204 (9th Cir. 1938) (holding that twenty-eight year duration for published works should also apply to registered unpublished work, but the date of deposit should be deemed the date of publication). p p 19. William S. Strauss, Protection of Unpublished Works (1957), in 1 COPYRIGHT SOCIETY OF THE USA, STUDIES ON COPYRIGHT 190, 196 (1963). p p 19. William S. Strauss, Protection of Unpublished Works (1957), in 1 COPYRIGHT SOCIETY OF THE USA, STUDIES ON COPYRIGHT 190, 196 (1963). 20. See, e.g., HOUSE COMM. ON THE JUDICIARY, 88TH CONG., COPYRIGHT LAW REVISION PART 2: DISCUSSION AND COMMENTS ON THE REPORT OF THE REGISTER OF COPYRIGHTS ON THE GENERAL REVISION OF THE U.S. COPYRIGHT LAW 312 (Comm. Print 1963) (letter from Irwin Karp, May 29, 1961); PHILIP WITENBERG, THE PROTECTION AND MARKETING OF LITERARY PROPERTY 11-14 (1937); Benjamin Kaplan, Publication in Copyrght Law: The .uestion of Phonograph Records, 103 U. PA. L. REV. 469, 473-79 (1955). HeinOnline -- 25 Berkeley Tech. L.J. 1384 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 2010] 1385 equity, to prevent the copying, publication, or use of such unpublished work without his consent. . . .",2' A 1957 Copyright Office study of state common law copyright described it as giving authors "absolute rights in an unpublished work"22 that were "not subject to the limitations imposed by the statute upon copyright." 2 3 The rights "continue perpetually as long as the work remains unpublished, unless ... the owner voluntarily chooses to secure statutory copyright by registration in the Copyright Office."24 The scope of the playwrights' common law right was perceived to be broader than the common law copyright at issue in Wheaton v. 23. Id. at 194. 21. Copyright Act of Mar. 4, 1909 5 2, 35 Stat. 1075, 1075. 21. Copyright Act of Mar. 4, 1909 5 2, 35 Stat. 1075, 1075. 22. Strauss, supra note 19, at 191 (citing DRONE, supra note 1, at 8). 23. Id. at 194. 24. Id. at 191; see George A. Warp, The Posiion of the Dramatist in Copyright Law, 22 B.U. L. REV. 528, 529-30 (1942). 25. Wheaton v. Peters, 33 U.S. (8 Pet.) 591 (1834); see L. RAY PATTERSON, COPYRIGHT IN HISTORICAL PERSPECTIVE (1968). See generaly Craig Joyce, The StoU of Wheaton v. Peters: A Curious Chapter in the HistoU ofJudicature, 42 Hous. L. REV. 325 (2006). 26. DRONE, supra note 1, at 553. Drone used the single word "playright" because of its parallel to "copyright" and the courts following Drone did so as well. In this essay, I use "play right" unless I am quoting to minimize confusion. 27. See, e.g., Fitch v. Young, 230 F. 743, 744 (S.D.N.Y. 1916) (Hand, J.) ("I think that play right and copyright are quite distinct under the statute, in spite of the fact that printed publication will forfeit both, and that one statutory copyright will protect both."); R.R. BOWKER, COPYRIGHT: ITS HISTORY AND ITS LAW 162 (1912) ("[he new American code provides not only for copyright, but for playright or right of performance."); WILLIAM MORRIS COLLES & HAROLD HARDY, PLAYRIGHT AND COPYRIGHT IN ALL COUNTRIES 1 (1906) ("The playright is more valuable, as a rule, than the copyright in a play. . . ."); Charles Collins, Playright and the Common Law, 15 CALIF. L. REV. 381, 381 (1926-1927) ("Playright is a word coined by Mr. Drone .... It has been adopted by the Supreme Court of the United States, and is accordingly preferred . , . ." (citation omitted)); Warp, supra note 24, at 546 ("[The term 'playright' refers to the exclusive right to present dramatic works."). 27. See, e.g., Fitch v. Young, 230 F. 743, 744 (S.D.N.Y. 1916) (Hand, J.) ("I think that play right and copyright are quite distinct under the statute, in spite of the fact that printed publication will forfeit both, and that one statutory copyright will protect both."); R.R. BOWKER, COPYRIGHT: ITS HISTORY AND ITS LAW 162 (1912) ("[he new American code provides not only for copyright, but for playright or right of performance."); WILLIAM MORRIS COLLES & HAROLD HARDY, PLAYRIGHT AND COPYRIGHT IN ALL COUNTRIES 1 (1906) ("The playright is more valuable, as a rule, than the copyright in a play. . . ."); Charles Collins, Playright and the Common Law, 15 CALIF. L. REV. 381, 381 (1926-1927) ("Playright is a word coined by Mr. Drone .... It has been adopted by the Supreme Court of the United States, and is accordingly preferred . , . ." (citation omitted)); Warp, supra note 24, at 546 ("[The term 'playright' refers to the exclusive right to present dramatic works."). Curious Chapter in the HistoU ofJudicature, ous. . V. ( ) 26. DRONE, supra note 1, at 553. Drone used the single word "playright" because of its parallel to "copyright" and the courts following Drone did so as well. In this essay, I use "play right" unless I am quoting to minimize confusion. J ) 24. Id. at 191; see George A. Warp, The Posiion of the Dramatist in Copyright Law, 22 B.U L. REV. 528, 529-30 (1942). py g 22. Strauss, supra note 19, at 191 (citing DRONE, supra note 1, at 8). 23. Id. at 194. py g , , , 22. Strauss, supra note 19, at 191 (citing DRONE, supra note 1, at 8). 25. Wheaton v. Peters, 33 U.S. (8 Pet.) 591 (1834); see L. RAY PATTERSON, COPYRIGH IN HISTORICAL PERSPECTIVE (1968). See generaly Craig Joyce, The StoU of Wheaton v. Peters: A Curious Chapter in the HistoU ofJudicature, 42 Hous. L. REV. 325 (2006). 21. Copyright Act of Mar. 4, 1909 5 2, 35 Stat. 1075, 1075. O 1 8 21. Copyright Act of Mar. 4, 1909 5 2, 35 Stat. 1075, 1075. 22. Strauss, supranote 19, at 191 (citing DRONE, supranote 1, at 8). 21. Copyright Act of Mar. 4, 1909 5 2, 35 Stat. 1075, 1075. 22 Strauss supranote 19 at 191 (citing DRONE supranote 1 at 28. Act of Aug. 18,1856, 11 Stat. 138. 29. E.g., Keene v. Wheatley, 14 F. Cas. 180 (C.C.E.D. Pa. 1861). 30. DRONE, supra note 1, at 553-600; see infra notes 173-89 and accompanying text. 31. See, e.g., Frohman v. Ferris, 87 N.E. 327 (Ill. 1909), afd, 223 U.S. 424 (1912); Tomkins v. Halleck, 133 Mass. 32 (1882). 32. See An Arbiter to Keep Peace in Theatre, N.Y. TIMEs, Mar. 23, 1926, at 24; infra notes 238-45 and accompanying text. 33. Ring v. Spina, 148 F.2d 647 (2d Cir. 1945); see text accompanying infra notes 252- 60. 34. See, e.g., Wendy J. Gordon, Fair Use as Market Failure: A Structural and Economic Analysis of the Betamax Case and its Predecessors, 82 COLUM. L. REv. 1600, 1602 (1982) ("The ultimate goal is not author remuneration, however, but the advancement and dissemination BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 Peters,25 encompassing what we would today identify as adaptation and public performance rights as well as reproduction and distribution rights. In 1879, treatise author Eaton Drone dubbed the common law public performance right in unpublished dramatic works "playright," 26 and, at least for a time, the usage enjoyed favor.27 This Article explores playwrights' common law "play right." Since this conference celebrates the 300th birthday of the Statute of Anne, I begin in England in the 17th Century. I find no trace of a common law playwright's performance right in either the law or the customary practices surrounding 17th and 18th century English theatre. I argue that the nature and degree of royal supervision of theatre companies and performance during the period presented no occasion (and, indeed, left no opportunity) for such a right to arise. I discuss the impetus for Parliament's enactment of a performance HeinOnline -- 25 Berkeley Tech. L.J. 1385 2010 34. See, e.g., Wendy J. Gordon, Fair Use as Market Failure: A Structural and Economic Analysis of the Betamax Case and its Predecessors, 82 COLUM. L. REv. 1600, 1602 (1982) ("The ultimate goal is not author remuneration, however, but the advancement and dissemination 28. Act of Aug. 18,1856, 11 Stat. 138. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1386 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 20101 1387 The evolution of attitudes about the nature of copyright made no appreciable change in playwrights' lives or livelihoods. 37. See JOHN FEATHER, PUBLISHING, PIRACY AND POLITICS: AN HISTORICAL STUDY OF COPYRIGHT IN BRITAIN 15-36 (1994); Edward S. Rogers, Some Historical Matter Concerning Literary Property, 7 MICH. L. REv. 100, 105-09 (2008). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1386 right statute in 1833, and its decision, nine years later, to adopt a law that scuttled any nascent tendencies supporting the development of a common law performance right by equating public performance with publication for the purposes of copyright. right statute in 1833, and its decision, nine years later, to adopt a law that scuttled any nascent tendencies supporting the development of a common law performance right by equating public performance with publication for the purposes of copyright. I then cross the Atlantic to the United States, and note that the colonies and the young nation imported their actors, managers, plays, and customary theatrical practices from England. I again find no trace of a common law play right before 1856. That year was the year that Congress followed the British example and enacted a statutory public performance right for the authors and proprietors of dramatic compositions.28 The first common law performance right cases show up shortly thereafter, as courts sought to respond to formal defects in the copyrights of claimants seeking to enforce their rights under the new statute.29 In 1879, Eaton Drone relied on an expansive natural rights theory about the true nature of copyright to draw from these cases a generous depiction of play right as a perpetual entitlement without exceptions.3 0 Courts adopted Drone's version of common law play right and followed it for the next thirty years.31 The availability of a strong natural right claim, however, made little difference to actual playwrights, who were deemed to have assigned their rights to the producers of their plays. The strong copyright-like rights that playwrights enjoy today are chiefly contractual, secured for them in 1926 by the collective action of members of the Dramatists' Guild, who claimed to be a labor union and thus entitled to an antitrust exemption.32 Courts would later rule that the exemption was not available to the Guild," but its members and the theatre producers it contracts with continued to behave, most of the time, as if they were bound by Guild contracts. Meanwhile, the broad Drone view of natural rights copyright fell out of fashion, to be replaced, first, by a utilitarian public- interest account34 and later by a utilitarian broad property rights account.35 HeinOnline -- 25 Berkeley Tech. L.J. 36. E.g., Howard Abrams, The Historic Foundation of American Copyright Law: Exploding the Myth of Common Law Copyrght, 29 WAYNE L. REV. 1119, 1129-47 (1983). See generally RONAN DEAZLEY, ON THE ORIGINS OF THE RIGHT TO COPY 191-210 (2004). 35. See, e.g., Robert Merges, Contracting Into liability Rules: Intellectual Propery Rights and Collective Rights OrganiZaions, 84 CALIF. L. REV. 1293 (1996). See generally Neil Netanel, Why Has Copynght Expanded? Analysis and Critique, in 6 NEW DIRECTIONS IN COPYRIGHT LAW 3 (Fiona MacMillan ed., 2007) (criticizing trend); Diane Leenheer Zimmerman, Copyright As Incentives: Did We Just Imagine That?, 12 THEORETICAL INQUIRIES IN LAW (forthcoming 2010) (arguing that research in psychology and behavioral economics cast doubt on incentive theory of copyright). of culture and knowledge."). See generally Lloyd L. Weinreb, Copyrght for Functional Expression, 111 HARV. L. REv. 1149, 1215-16 (1998) (remarking trend). 38. See Don-JOHN DUGAS, MARKETING THE BARD: SHAKESPEARE IN PERFORMANCE AND PRINT 1660-1740, 93-94 (2006); MARK ROSE, AUTHORS AND OWNERS: THE INVENTION OF COPYRIGHT 12 (1993). of culture and knowledge."). See generally Lloyd L. Weinreb, Copyrght for Functional Expressio 111 HARV. L. REv. 1149, 1215-16 (1998) (remarking trend). of culture and knowledge."). See generally Lloyd L. Weinreb, Copyrght for Functional Expression, 111 HARV. L. REv. 1149, 1215-16 (1998) (remarking trend). 35. See, e.g., Robert Merges, Contracting Into liability Rules: Intellectual Propery Rights and Collective Rights OrganiZaions, 84 CALIF. L. REV. 1293 (1996). See generally Neil Netanel, Why Has Copynght Expanded? Analysis and Critique, in 6 NEW DIRECTIONS IN COPYRIGHT LAW 3 (Fiona MacMillan ed., 2007) (criticizing trend); Diane Leenheer Zimmerman, Copyright As Incentives: Did We Just Imagine That?, 12 THEORETICAL INQUIRIES IN LAW (forthcoming 2010) (arguing that research in psychology and behavioral economics cast doubt on incentive theory of copyright). 36. E.g., Howard Abrams, The Historic Foundation of American Copyright Law: Exploding the M th fC L C ht 29 WAYNE L REV 1119 1129 47 (1983) SeegenerallyRONAN of culture and knowledge."). See generally Lloyd L. Weinreb, Copyrght for Functional Expression, 111 HARV. L. REv. 1149, 1215-16 (1998) (remarking trend). 35. See, e.g., Robert Merges, Contracting Into liability Rules: Intellectual Propery Rights and Collective Rights OrganiZaions, 84 CALIF. L. REV. 1293 (1996). See generally Neil Netanel, Why Has Copynght Expanded? Analysis and Critique, in 6 NEW DIRECTIONS IN COPYRIGHT LAW 3 (Fiona MacMillan ed., 2007) (criticizing trend); Diane Leenheer Zimmerman, Copyright As Incentives: Did We Just Imagine That?, 12 THEORETICAL INQUIRIES IN LAW (forthcoming 2010) (arguing that research in psychology and behavioral economics cast doubt on incentive theory of copyright). 36. E.g., Howard Abrams, The Historic Foundation of American Copyright Law: Exploding the Myth of Common Law Copyrght, 29 WAYNE L. REV. 1119, 1129-47 (1983). See generally RONAN DEAZLEY, ON THE ORIGINS OF THE RIGHT TO COPY 191-210 (2004). 37. See JOHN FEATHER, PUBLISHING, PIRACY AND POLITICS: AN HISTORICAL STUDY OF COPYRIGHT IN BRITAIN 15-36 (1994); Edward S. Rogers, Some Historical Matter Concerning Literary Property, 7 MICH. L. REv. 100, 105-09 (2008). 38. See Don-JOHN DUGAS, MARKETING THE BARD: SHAKESPEARE IN PERFORMANCE AND PRINT 1660 1740 93 94 (2006) MARK ROSE AUTHORS AND OWNERS: THE 39. See, e.g., PATTERSON, supra note 25, at 47-78; Pamela Samuelson, Copyright and Freedom ofExpression in Historical Perspective, 10 J. INTELL. PROP. L. 319, 323-38 (2003). Paying the owner of a manuscript was not, of course, the same as paying the author. See, e.g., LEO KIRSCHBAUM, SHAKESPEARE AND THE STATIONERS 5-7 (1955). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1388 must by their payment have purchased a perpetual common law literary property right.39 Stationers argued unsuccessfully in court that this perpetual common law right survived the enactment of the Statute of Anne.4 Copyright historians have examined the evidence supporting the existence of a common law literary property right in 17th century England and persuasively debunked it.41 To the extent that common law literary property rights may be said to have existed in some place and at some time, they did not precede the enactment of the Statute of Anne. For most purposes, the inquiry is of only academic interest, because the British courts ultimately held that even if there had been such a common law right, the enactment of the Statute of Anne had abrogated it for all published works.42 Such a right, had it existed, might have offered protection to unpublished scripts, but the British Parliament eventually decided that a play's first public performance should be equated with publication for copyright purposes43 Such a right, had it existed, might have offered protection to unpublished scripts, but the British Parliament eventually decided that a play's first public performance should be equated with publication for copyright purposes.43 , p 5, at 80; ROSE, supra note 38, at 67 91. 41. See, e.g., DEAZLEY, supra note 36, at 73-74, 195-210; RONAN DEAZLEY, RETHINKING COPYRIGHT 63-65 (2006); 1 WILLIAM F. PATRY, PATRY ON COPYRIGHT § 1:11 (2009); ROSE, supra note 38, at 22 ("[here is no evidence that copyright was ever recognized as a common law right of an author in the sixteenth or seventeenth centuries."); Abrams, supra note 36, at 1134 ("As a matter of historical fact, the common law never developed any law of copyright."). But see H. Tomis G6mez-Arostegui, The Untold Stog of the First Copyright Suit Under the Statute of Anne in 1710, 25 BERKELEY TECH. L.J. 1247 (2010) (suggesting that complaints filed in Chancery before 1710 to vindicate stationer's copyright claims may provide some evidence of common law copyright). 42. Donaldson v. Beckett, 1 Eng. Rep. 837, 845 (1774); see DEAZLEY, supra note 36, at 191-220. 43. 1842 Copyright Law Amendment Act, 5 & 6 Vict. c. 45, § XX. The 1842 Copyright Law Amendment includes the first post-mortem copyright term, the first express work- made-for-hire provision, and the first compulsory license for reprinting out-of-print books. II. PLAY RIGHT IN ENGLAND Since our conference is celebrating the 300th birthday of the Statute of Anne, we should begin by examining the common law literary property rights that playwrights enjoyed in the century preceding the enactment of the Statute of Anne. If playwrights' "play right" derives from common law sources, one might expect to be able to find the sources of a common law play right in the case law of the 17th and early 18th century. The examination need not keep us for long, because it seems evident from all sources that playwrights enjoyed no literary property rights in their scripts. Playwrights' common law rights are apparently of more recent origin. First, some common ground: the notion that common law copyright predated statutory copyright in England was first advanced by members of the stationers' guild, after the enactment of the Statute of Anne, to press claims for perpetual printing monopolies. 6 The evidence they offered to support their theory of perpetual common law copyright was longstanding "stationers' copyright," a naked horizontal restraint of trade under which members of the Worshipful Company of Stationers agreed not to reprint any text first claimed by another member." Stationers treated members' assertions of exclusive rights to a text as perpetual, permitting them to be sold to or inherited by other members of the Guild.38 Since stationers commonly (although not invariably) paid the owner of a manuscript something in return for the privilege of printing it, they claimed that they y 38. See Don-JOHN DUGAS, MARKETING THE BARD: SHAKESPEARE IN PERFORMANCE AND PRINT 1660-1740, 93-94 (2006); MARK ROSE, AUTHORS AND OWNERS: THE INVENTION OF COPYRIGHT 12 (1993). HeinOnline -- 25 Berkeley Tech. L.J. 1387 2010 40. See PATTERSON, supra note 25, at 172-80; ROSE, supra note 38, at 67-91. 42. Donaldson v. Beckett, 1 Eng. Rep. 837, 845 (1774); see DEAZLEY, supra note 36, a 191-220. 41. See, e.g., DEAZLEY, supra note 36, at 73-74, 195-210; RONAN DEAZLEY, RETHINKING COPYRIGHT 63-65 (2006); 1 WILLIAM F. PATRY, PATRY ON COPYRIGHT § 1:11 (2009); ROSE, supra note 38, at 22 ("[here is no evidence that copyright was ever recognized as a common law right of an author in the sixteenth or seventeenth centuries."); Abrams, supra note 36, at 1134 ("As a matter of historical fact, the common law never developed any law of copyright."). But see H. Tomis G6mez-Arostegui, The Untold Stog of the First Copyright Suit Under the Statute of Anne in 1710, 25 BERKELEY TECH. L.J. 1247 (2010) (suggesting that complaints filed in Chancery before 1710 to vindicate stationer's copyright claims may provide some evidence of common law copyright). THE INVENTION OF COMMON LAW PLAY RIGHT 1389 2010] More fundamentally, however, proponents of common law copyright conceived of it as a printing right, limited to control over the printing and sale of copies. In the 18th century, not even the most ardent defender of common law copyright suggested it extended to control over public performances of works." Thus, a playwright's common law rights in a script, had they existed, would have empowered the dramatist to sell that script to a printer for publication, or withhold it from publication. It would have given the playwright no recourse, however, if players chose to perform the unpublished script without a license. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 See generally Ronan Deazley, Commentary on Copyright Amendment Act 1842, PRIMARY SOURCES ON COPYRIGHT (1450-1900) (L. Bently & M. Kretschmer eds., 2008), http://www.copyrighthistory.org/cgi-bin/ldeioc/0010/exec/ausgabeCom//22uk_18420/22. Section 20 of the Act extended the dramatic performance right initially enacted in 1833 to musical compositions, and provided that "the first public Representation or Performance of any Dramatic Piece or Musical Composition shall be deemed equivalent, in the Construction of this Act, to the first Publication of any Book." The 1842 Act also permitted the registration of unpublished but publicly performed dramas and musical compositions, without depositing any copies, through recording the title, the name and place of abode of the author or composer, the name and place of abode of the copyright proprietor, and the time and place of the first public performance. 1842 Copyright Law Amendment Act, § XX. The Act also provided that the assignment of the right to print and public a script or musical HeinOnline -- 25 Berkeley Tech. L.J. 1388 2010 THE INVENTION OF COMMON LAW PLAY RIGHT ( ) 47. See, e.g., Judith Milhous, Theatre Companies and Regulation, in 2 THE CAMBRIDGE HISTORY OF BRITISH THEATRE 108 (Joseph Donahue ed., 2004) [hereinafter Milhous, Theatre Companies and Regulation]. composition did not also assign the public performance right unless it said so explicitly. I § XXII. § 44. See DEAZLEY, RETHINKING, supra note 41, at 154-56. composition did not also assign the public performance right unless it said so explicitly. Id. § XXII. 44. See DEAZLEY, RETHINKING, supra note 41, at 154-56. 45. See generally OSCAR G. BROcKETT & FRANKLIN L. HILDY, HISTORY OF THE THEATRE 114-16, 211-14, 220-21 (9th ed. 2003). 46. JOHN RUSSELL STEPHENS, THE CENSORSHIP OF ENGLISH DRAMA 1824-1901, at 7 (2010). 47. See, e.g., Judith Milhous, Theatre Companies and Regulation, in 2 THE CAMBRIDGE HISTORY OF BRITISH THEATRE 108 (Joseph Donahue ed., 2004) [hereinafter Milhous, Theatre Companies and Regulation]. , , p 45. See generally OSCAR G. BROcKETT & FRANKLIN L. HILDY, HISTORY OF TH THEATRE 114-16, 211-14, 220-21 (9th ed. 2003). , , ( ) 46. JOHN RUSSELL STEPHENS, THE CENSORSHIP OF ENGLISH DRAMA 1824-1901, at 7 (2010). § XXII. 44. See DEAZLEY, RETHINKING, supra note 41, at 154-56. 45. See generally OSCAR G. BROcKETT & FRANKLIN L. HILDY, HISTORY OF THE THEATRE 114-16, 211-14, 220-21 (9th ed. 2003). 46. JOHN RUSSELL STEPHENS, THE CENSORSHIP OF ENGLISH DRAMA 1824-1901, at 7 49. See PAULINA KEWES, AUTHORSHIP AND APPROPRIATION: WRITING FOR THE STAGE IN ENGLAND 1660-1710, at 17 (2004); Judith Milhous & Robert D. Hume, Playwrights' Remuneraion in Eighteenth Centug London, 10 HARV. LIBR. BULL. 1 (Summer-Fall 1999) [hereinafter Milhous & Hume, Playwrghts' Remuneraion]; Mary I. Oates & William J. Baumwol, On the Economics of the Theater in Renaissance London, 74 SWED. J. ECoN. 136, 149, 156-58 (1972). 48. See, e.g., JUDITH MILHOUS, THOMAS BETTERTON AND THE MANAGEMENT OF LINCOLN'S INN FIELDS 1695-1708, at 15-19, 66-68, 115-16, 201-02, 209-21 (1979) [hereinafter MILHOUS, THOMAS BETTERTON]; Judith Milhous & Robert D. Hume, The Silencing of Drug Lane in 1709, 32 THEATRE J. 427 (1980) [hereinafter Milhous & Hume, The Silencing of Dmu Lane]. 51. Judith Milhous concludes that theatre companies treated the physical manuscripts of plays as carrying with them the license to produce them. MILHOUS, THOMAS BETTERTON, supra note 48, at 16-17. While one could argue that that practice manifests some precursor common law public performance right in unpublished scripts, the reason is likely more prosaic. Throughout this period, theatrical performances required a license from the Master of the Revels confirming that the play script had met the censor's approval. The Master of the Revels affixed his license stamp to the promptbook, so a company's possession of a promptbook with license stamp was essential evidence of the lawfulness of a production. See JOSEPH LOEWENSTEIN, BEN JOHNSON AND POSSESSIVE AUTHORSHIP 31 (2002). A. We tend to view historical facts through a contemporary lens. When we see a practice that looks familiar, we assume that its context must have matched, or at least resembled, its contemporary analog. Thus, much speculation about common law play right seems to begin by recognizing landmarks that would characterize such a right in recent times, and assuming the context must have followed familiar patterns. A serious exploration of the English theatre before the enactment of the Statute of Anne, though, reveals a context starkly different from one we would recognize. The most fundamental difference is the degree to which, from the 16th through the 18th centuries, all aspects of theatrical performance were regulated by the crown.4 5 A monarchy that reacted to the seditious potential of the printing press by licensing all printing appreciated the subversive possibilities of public theatrical performance.46 The crown responded by criminalizing unlicensed theatrical presentations, limiting the legal theatres to companies owned and run by friends and courtiers, and requiring all scripts to be approved by the royal censor before the initial public performance.47 As the theatre in Tudor and Restoration England existed only by dint of royal privilege, the crown was called upon to decide disputes between rival theatres, and it did so: allocating plays, playwrights, or actors to one company HeinOnline -- 25 Berkeley Tech. L.J. 1389 2010 50. See, e.g., PETER THOMPSON, SHAKESPEARE'S PROFESSIONAL CAREER 27, 82-100 (1992); PETER THOMPSON, SHAKESPEARE'S THEATRE 19-35, 58-87 (2d ed. 1992). 53. See STERN, supra note 52, at 61-79. 52. See TIFFANY STERN, REHEARSAL FROM SHAKESPEARE TO SHERIDAN 10-12, 149- 51 (2000); Judith Milhous, The First Production of Rowe's 'Jane Shore," 38 THEATRE J. 309, 312 (1986). 48. See, e.g., JUDITH MILHOUS, THOMAS BETTERTON AND THE MANAGEMENT OF LINCOLN'S INN FIELDS 1695-1708, at 15-19, 66-68, 115-16, 201-02, 209-21 (1979) [hereinafter MILHOUS, THOMAS BETTERTON]; Judith Milhous & Robert D. Hume, The Silencing of Drug Lane in 1709, 32 THEATRE J. 427 (1980) [hereinafter Milhous & Hume, The Silencing of Dmu Lane]. 49. See PAULINA KEWES, AUTHORSHIP AND APPROPRIATION: WRITING FOR THE STAGE IN ENGLAND 1660-1710, at 17 (2004); Judith Milhous & Robert D. Hume, Playwrights' Remuneraion in Eighteenth Centug London, 10 HARV. LIBR. BULL. 1 (Summer-Fall 1999) [hereinafter Milhous & Hume, Playwrghts' Remuneraion]; Mary I. Oates & William J. Baumwol, On the Economics of the Theater in Renaissance London, 74 SWED. J. ECoN. 136, 149, 156-58 (1972). 50. See, e.g., PETER THOMPSON, SHAKESPEARE'S PROFESSIONAL CAREER 27, 82-100 (1992); PETER THOMPSON, SHAKESPEARE'S THEATRE 19-35, 58-87 (2d ed. 1992). 51. Judith Milhous concludes that theatre companies treated the physical manuscripts of plays as carrying with them the license to produce them. MILHOUS, THOMAS BETTERTON, supra note 48, at 16-17. While one could argue that that practice manifests some precursor common law public performance right in unpublished scripts, the reason is likely more prosaic. Throughout this period, theatrical performances required a license from the Master of the Revels confirming that the play script had met the censor's approval. The Master of the Revels affixed his license stamp to the promptbook, so a company's possession of a promptbook with license stamp was essential evidence of the lawfulness of a production. See JOSEPH LOEWENSTEIN, BEN JOHNSON AND POSSESSIVE AUTHORSHIP 31 (2002). 52. See TIFFANY STERN, REHEARSAL FROM SHAKESPEARE TO SHERIDAN 10-12, 149- 51 (2000); Judith Milhous, The First Production of Rowe's 'Jane Shore," 38 THEATRE J. 309, 312 (1986). 53. See STERN, supra note 52, at 61-79. 54. See LOEWENSTEIN, supra note 51, at 29-31; MILHOUS, THOMAS BETIERTON, supra note 48, at 16-17; Milhous & Hume, Playnaghts' Remuneration, supra note 49, at 35. Scholars have surmised that some of the unauthorized editions were pieced together from actors' parts and actors' memory. See, e.g., GEORGE IAN DUTHIE, THE "BAD" QUARTO OF HAMLET: A CRITICAL STUDY 29-36 (1941); W. Mathews, The Piracies of Love a La Mode, 10 REV. ENG. STUD. 311, 315 (1934); see also 1 JOHN BERNARD, RETROSPECTIONS OF THE STAGE 128 (Boston, Carter & Hendee 1832). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1390 rather than another.48 In that sphere, questions of common law literary property seem to be beside the point. rather than another.48 In that sphere, questions of common law literary property seem to be beside the point. The Elizabethan and Restoration eras of British theatre history have been the subject of extensive study, but extrapolating from surviving documentary evidence requires guesswork and intuition. Current theatre historians have concluded that Elizabethan theatre companies paid playwrights for their services in essentially the same way they paid actors: playwrights received a salary, or participated as partners in the profits of the company, collecting a share of net receipts, or both.49 At least some playwrights (like actors and managers) who were "sharers" or partners in theatre companies were able to earn a living from the theatre, if not from writing plays.o Play scripts were deemed to belong to the company, which decided when to produce them, revive them, or sell them to stationers for printing." After an initial reading of the script to the full cast, companies had scribes copy individual actors' "parts" or "sides"-pages interspersing a single actor's lines with a few words of cues from the end of the immediately preceding lines.5 2 Actors learned their lines from the parts, and did not see the entire play until they came together at the end of the rehearsal period for group rehearsal. Companies were reluctant to authorize printers to publish play scripts while 51. Judith Milhous concludes that theatre companies treated the physical manuscripts of plays as carrying with them the license to produce them. MILHOUS, THOMAS BETTERTON, supra note 48, at 16-17. While one could argue that that practice manifests some precursor common law public performance right in unpublished scripts, the reason is likely more prosaic. Throughout this period, theatrical performances required a license from the Master of the Revels confirming that the play script had met the censor's approval. The Master of the Revels affixed his license stamp to the promptbook, so a company's possession of a promptbook with license stamp was essential evidence of the lawfulness of a production. See JOSEPH LOEWENSTEIN, BEN JOHNSON AND POSSESSIVE AUTHORSHIP 31 (2002). HeinOnline -- 25 Berkeley Tech. L.J. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1390 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 2010] 1391 they were still part of the active repertory, but were more eager to publish in years in which the theatres were closed because of political unrest, royal death, plague, or fire.54 Unauthorized printings were common and typically rife with errors. The Protestant Revolution closed the theatres completely in 1642, and they remained closed until Charles II was restored to the throne in 1660." King Charles granted royal patents56 to two courtiers to run theatre companies in London, and prohibited competing theatrical performances. Charles had enjoyed attending theatrical performances during his exile in France, and ordered his new English theatres to cast women in female roles, after the French fashion.5 ' The crown drew up a list of extant plays and allotted them exclusively to one company or the other, prohibiting each company from performing plays granted to the other." The division was unequal. Thomas Killigrew's King's Company received most of the popular plays. William Davenant's Duke's Company was less fortunate.6 0 Davenant, therefore, needed to commission new scripts. As the Duke's Company staged new plays, the King's Company found that it needed to find new 54. See LOEWENSTEIN, supra note 51, at 29-31; MILHOUS, THOMAS BETIERTON, supra note 48, at 16-17; Milhous & Hume, Playnaghts' Remuneration, supra note 49, at 35. Scholars have surmised that some of the unauthorized editions were pieced together from actors' parts and actors' memory. See, e.g., GEORGE IAN DUTHIE, THE "BAD" QUARTO OF HAMLET: A CRITICAL STUDY 29-36 (1941); W. Mathews, The Piracies of Love a La Mode, 10 REV. ENG. STUD. 311, 315 (1934); see also 1 JOHN BERNARD, RETROSPECTIONS OF THE STAGE 128 (Boston, Carter & Hendee 1832). 55. During this period, theatre companies that had previously refused to publish their scripts made them available to printers. See KEWES, supra note 49, at 183. 56. Royal letters patent were monopoly privileges conferred by the king or queen. They were not limited to technological inventions. See generaly CHRISTINE MACLEOD, INVENTING THE INDUSTRIAL REVOLUTION: THE ENGLISH PATENT SYSTEM 1660-1800, at 10-39 (1988). 57. See Joseph Donahue, The Theatre from 1600 to 1800, in THE CAMBRIDGE HISTORY OF BRITISH THEATRE, supra note 47, at 3, 4-8. Ireland also had a single theatre company by dint of a royal patent granted in 1661. Scotland had none. 57. See Joseph Donahue, The Theatre from 1600 to 1800, in THE CAMBRIDGE HISTORY OF BRITISH THEATRE, supra note 47, at 3, 4-8. Ireland also had a single theatre company by dint of a royal patent granted in 1661. Scotland had none. See Gorel Garlick, Theatre Outside London, 1660-1775, in THE CAMBRIDGE HISTORY OF BRITISH THEATRE, supra note 47, at 165-70. 59. See Robert D. Hume, Theatres and RepertoU, in THE CAMBRIDGE HISTORY OF BRITISH THEATRE, supra note 47, at 53, 55. 58. See generally ELIZABETH HOWE, THE FIRST ENGLISH ACTRESSES: WOMEN AND DRAMA 1660-1700, at 19-24 (1992); MILHOUS, THOMAS BETTERTON, supra note 48, at 4-7. 61. See KEWES, supra note 49, at 13-20. Lacking a ready-made popular repertory, Davenant's Duke's company appears to have competed by offering productions featuring more impressive changeable scenery and special effects. See MILHouS, THOMAS BETTERTON, supra note 48, at 20-25. That strategy proved to be successful; Killigrew shortly had to build a new theatre to accommodate new audience demand for expensive sets. Id at 12. 63. See KEWES, supra note 49, at 13-20; Milhous & Hume, Playwnaghts' Remuneration, supra note 49, at 4. Because the recipient of the benefit paid the costs attributable to that performance, managers could stage benefit performances at little risk to company finances. By the end of the 17th century, theatre companies also offered benefit performances to star actors, as bonuses in addition to or in lieu of salary. See Judith Milhous, United Company Finances, 1682-1692, 7 THEATRE RES. INT'L 37 (1982) (reconstructing the finances of the the theatre company formed in 1682 by combining the Kings Company and the Duke's Company into a single "United Company") [hereinafter Milhous, United Company Finances]. 62. ROBERT D. HUME, THE RAKISH STAGE: STUDIES IN ENGLISH DRAMA 1660-1800, at 277 (1983); KEWES, supra note 49, at 15; MILHOUS, THOMAS BETTERTON, supra note 48, at 15-16. Throughout this period, companies freely adapted plays in their repertory to respond to changing audience tastes, more stringent crown censorship, or changes in acting personnel. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1392 scripts to compete for audiences."1 As before, the new scripts appear to have become the property of the company that first staged them. 6 2 Playwright compensation, however, was no longer by salary or ownership share unless the playwright was also a member of the acting company. Rather, both companies appear to have settled on a compensation system under which the playwright would receive the net proceeds of a single performance, typically the third night, and, if the production proved to run longer than a few days, might be offered the net proceeds of an additional performance or two.63 While the benefit performance might earn a lot of money for an exceptionally successful play, the receipts were usually more modest. The division of the repertory between the companies meant that neither theatre company needed to worry that a competing theatre would perform its scripts. Perhaps as a result, the companies seem to have grown more interested in printing and publishing plays after their initial run. In at least some cases, the companies allowed the playwright to pocket the stationer's payment for the script.6 4 The amount of money printers paid for play scripts, though, was not substantial.s scripts to compete for audiences."1 As before, the new scripts appear to have become the property of the company that first staged them. 6 2 Playwright compensation, however, was no longer by salary or ownership share unless the playwright was also a member of the acting company. Rather, both companies appear to have settled on a compensation system under which the playwright would receive the net proceeds of a single performance, typically the third night, and, if the production proved to run longer than a few days, might be offered the net proceeds of an additional performance or two.63 While the benefit performance might earn a lot of money for an exceptionally successful play, the receipts were usually more modest. The division of the repertory between the companies meant that neither theatre company needed to worry that a competing theatre would perform its scripts. Perhaps as a result, the companies seem to have grown more interested in printing and publishing plays after their initial run. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 See Gorel Garlick, Theatre Outside London, 1660-1775, in THE CAMBRIDGE HISTORY OF BRITISH THEATRE, supra note 47, at 165-70. 58. See generally ELIZABETH HOWE, THE FIRST ENGLISH ACTRESSES: WOMEN AND DRAMA 1660-1700, at 19-24 (1992); MILHOUS, THOMAS BETTERTON, supra note 48, at 4-7. 59. See Robert D. Hume, Theatres and RepertoU, in THE CAMBRIDGE HISTORY OF BRITISH THEATRE, supra note 47, at 53, 55. 60. See KEWES, supra note 49, at 13-15; MILHOUS, THOMAS BETTERTON, supra note 48, at 4-20. Killigrew's company had a much larger proportion of established actors who had been well-known before the revolution. Milhous speculates that Killigrew may have staked his claim to the lion's share of extant English repertory on the basis of his actors' prior connections to those plays. See MILHOUS, THOMAS BETTERTON, supra note 48, at 17. HeinOnline -- 25 Berkeley Tech. L.J. 1391 2010 61. See KEWES, supra note 49, at 13-20. Lacking a ready-made popular repertory, Davenant's Duke's company appears to have competed by offering productions featuring more impressive changeable scenery and special effects. See MILHouS, THOMAS BETTERTON, supra note 48, at 20-25. That strategy proved to be successful; Killigrew shortly had to build a new theatre to accommodate new audience demand for expensive sets. Id at 12. 62. ROBERT D. HUME, THE RAKISH STAGE: STUDIES IN ENGLISH DRAMA 1660-1800, at 277 (1983); KEWES, supra note 49, at 15; MILHOUS, THOMAS BETTERTON, supra note 48, at 15-16. Throughout this period, companies freely adapted plays in their repertory to respond to changing audience tastes, more stringent crown censorship, or changes in acting personnel. 63. See KEWES, supra note 49, at 13-20; Milhous & Hume, Playwnaghts' Remuneration, supra note 49, at 4. Because the recipient of the benefit paid the costs attributable to that performance, managers could stage benefit performances at little risk to company finances. By the end of the 17th century, theatre companies also offered benefit performances to star actors, as bonuses in addition to or in lieu of salary. See Judith Milhous, United Company Finances, 1682-1692, 7 THEATRE RES. INT'L 37 (1982) (reconstructing the finances of the the theatre company formed in 1682 by combining the Kings Company and the Duke's Company into a single "United Company") [hereinafter Milhous, United Company Finances]. 64. See Judith Milhous & Robert D. Hume, Dating Play Premieres from Publication Data 1660-1700, 22 HARv. LIBR. BULL. 374, 395-96 (1974) [hereinafter Milhous & Hume, Dating Play Premieres from Publication Data] (quoting Colley Cibber's 1696 contract). 65. See KEWES, supra note 49, at 24. p y g p y ) 64. See Judith Milhous & Robert D. Hume, Dating Play Premieres from Publication Data 1660-1700, 22 HARv. LIBR. BULL. 374, 395-96 (1974) [hereinafter Milhous & Hume, Dating Play Premieres from Publication Data] (quoting Colley Cibber's 1696 contract). 65. See KEWES, supra note 49, at 24. 66. See MILHoUS, THOMAS BETTERTON, supra note 48, at 31-37. Milhous places most of the blame for the King Company's demise on Killigrew, whom she characterizes as a poor theatrical manager. Id. at 37. 67. See generally Milhous, United Company Finances, supra note 63, at 37 (assessing the "startling amount of money [that] flowed through the United Company"). 69. See MILHOUS, THOMAS BETTERTON, supra note 48, at 59, 62; PAUL SAWYER, CHRISTOPHER RICH OF DRURY LANE: THE BIOGRAPHY OF A THEATRE MANAGER 9-18 (1986). 68. See MILHOUS, THOMAS BETTERTON, supra note 48, at 40-43, 51-62; Milhous, Theatre Companies and Regulation, supra note 47, at 114. 70. See MILHOUS, THOMAS BETTERTON, supra note 48, at 59-66; WATSON NICHOLSON, THE STRUGGLE FOR A FREE STAGE IN LONDON 7-9 (1966). 71. See The Petition of the Players, reprinted in MILHOUS, THOMAS BETTERTON, supra note 48, at 225, 226. 72. See MILHOUS, THOMAS BETIERTON, supra note 48, at 64-67. 73. See id. at 67-72. 66. See MILHoUS, THOMAS BETTERTON, supra note 48, at 31-37. Milhous places most of the blame for the King Company's demise on Killigrew, whom she characterizes as a poor theatrical manager. Id. at 37. 67. See generally Milhous, United Company Finances, supra note 63, at 37 (assessing the "startling amount of money [that] flowed through the United Company"). 68. See MILHOUS, THOMAS BETTERTON, supra note 48, at 40-43, 51-62; Milhous, Theatre Companies and Regulation, supra note 47, at 114. 69. See MILHOUS, THOMAS BETTERTON, supra note 48, at 59, 62; PAUL SAWYER, CHRISTOPHER RICH OF DRURY LANE: THE BIOGRAPHY OF A THEATRE MANAGER 9-18 (1986). 70. See MILHOUS, THOMAS BETTERTON, supra note 48, at 59-66; WATSON NICHOLSON, THE STRUGGLE FOR A FREE STAGE IN LONDON 7-9 (1966). 71. See The Petition of the Players, reprinted in MILHOUS, THOMAS BETTERTON, supra note 48, at 225, 226. 72. See MILHOUS, THOMAS BETIERTON, supra note 48, at 64-67. 73. See id. at 67-72. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 In at least some cases, the companies allowed the playwright to pocket the stationer's payment for the script.6 4 The amount of money printers paid for play scripts, though, was not substantial.s Despite the royal constraints limiting competition, the two companies were unable to sustain their operations. By 1682, the King's Company was insolvent, and the Duke's Company absorbed its remnants, including its p y g p y ) 64. See Judith Milhous & Robert D. Hume, Dating Play Premieres from Publication Data 1660-1700, 22 HARv. LIBR. BULL. 374, 395-96 (1974) [hereinafter Milhous & Hume, Dating Play Premieres from Publication Data] (quoting Colley Cibber's 1696 contract). HeinOnline -- 25 Berkeley Tech. L.J. 1392 2010 66. See MILHoUS, THOMAS BETTERTON, supra note 48, at 31-37. Milhous places most of the blame for the King Company's demise on Killigrew, whom she characterizes as a poor theatrical manager. Id. at 37. 67. See generally Milhous, United Company Finances, supra note 63, at 37 (assessing the "startling amount of money [that] flowed through the United Company"). 68. See MILHOUS, THOMAS BETTERTON, supra note 48, at 40-43, 51-62; Milhous, Theatre Companies and Regulation, supra note 47, at 114. 69. See MILHOUS, THOMAS BETTERTON, supra note 48, at 59, 62; PAUL SAWYER, CHRISTOPHER RICH OF DRURY LANE: THE BIOGRAPHY OF A THEATRE MANAGER 9-18 (1986). 70. See MILHOUS, THOMAS BETTERTON, supra note 48, at 59-66; WATSON NICHOLSON, THE STRUGGLE FOR A FREE STAGE IN LONDON 7-9 (1966). 77. See MILHoUS, THOMAS BETTERTON, supra note 48, at 131-32, 174. 76. Thus, in November of 1704, both companies appear to have staged revivals of Shakepeare's Henry IV. See Season of 1704-1705, in 1 THE LONDON STAGE 1660-1800: PART 2: 1700-1729, at 179, 192, 195-96 (Judith Milhous & Robert D. Hume eds., interim version 2001) (1960), available at http://www.personal.psu.edu/hbl/London/20Stage/ 202001/ londl704.pdf. p g , p p 75. See MILHOUS, THOMAS BETrERTON, supra note 48, at 97-112, 139-50. THE INVENTION OF COMMON LAW PLAY RIGHT THE INVENTION OF COMMON LAW PLAY RIGHT 2010] 1393 theatre building and its repertory of plays." The United Company was now the sole licensed theatre in London, with an ample stock of established drama. The company's theatrical seasons were heavily dominated by revivals from that repertory, and it was well placed to turn a profit." The inheritors of shares of the patent for the original companies, now sharers in the United Company, were able to attract investors to buy their interests." The investors brought in new management with no theatrical experience." The new manager (a lawyer) sought to increase the United Company's profit margins, by reforming practices that seemed to him to be economically unsound. He sought to retire aging actors and cast younger performers in their roles, to present fewer shows, to reduce the compensation and benefits the company traditionally paid its actors, and to rein in spending for popular scenery and special effects.o The actors rebelled. In 1694, fifteen of the United Company's best- known actors petitioned the Lord Chamberlain to hear their complaints, explaining that they found current management of the company was "soe intolerable & heavy that unless relieved wee are not able to act any longer."" The Lord Chamberlain granted the rebels permission to withdraw from the United Company and form a rival company licensed to perform "all manner of Comdyes & Tragedyes, Playes, Interludes & Opera's and to performe all other Theatricall and musicall Entertainments of what kind soever."7 2 The United Company got to keep the theatre buildings, costumes, sets, special effects equipment, and the Killigrew and Davenant patents, but its most famous actors had left, taking with them royal permission to perform any play in the United Company's repertory. 70. See MILHOUS, THOMAS BETTERTON, supra note 48, at 59-66; WATSO NICHOLSON, THE STRUGGLE FOR A FREE STAGE IN LONDON 7-9 (1966). 71. See The Petition of the Players, reprinted in MILHOUS, THOMAS BETTERTON, supra note 48, at 225, 226. HeinOnline -- 25 Berkeley Tech. L.J. 1393 2010 80. See MILHOUS, THOMAS BETTERTON, supra note 48, at 133-44, 188; SAWYER, supra note 69, at 31-35, 43-44. 74. See MILHoUS, THOMAS BEITERTON, supra note 48, at 68-97; Milhous, Theatre Companies and Regulation, supra note 47, at 114-15; SAWYER, supra note 69, at 25-31. 81. See Season of 1704-1705, supra note 76; Milhous, Theatre Companies and Regulation, supra note 47, at 116 79. Id. at 124-25, 154-68; SAWYER, supra note 69, at 40. p g 83. See NICHOLSON, supra note 70, at 8-19; STERN, supra note 52, at 126-28. 78. Id. at 88, 113-24, 161. 78. Id. at 88, 113-24, 161. 74. See MILHoUS, THOMAS BEITERTON, supra note 48, at 68-97; Milhous, Theatre Companies and Regulation, supra note 47, at 114-15; SAWYER, supra note 69, at 25-31. 75. See MILHOUS, THOMAS BETrERTON, supra note 48, at 97-112, 139-50. 76. Thus, in November of 1704, both companies appear to have staged revivals of Shakepeare's Henry IV. See Season of 1704-1705, in 1 THE LONDON STAGE 1660-1800: PART 2: 1700-1729, at 179, 192, 195-96 (Judith Milhous & Robert D. Hume eds., interim version 2001) (1960), available at http://www.personal.psu.edu/hbl/London/20Stage/ 202001/ londl704.pdf. 77. See MILHoUS, THOMAS BETTERTON, supra note 48, at 131-32, 174. 78. Id. at 88, 113-24, 161. 79. Id. at 124-25, 154-68; SAWYER, supra note 69, at 40. 80. See MILHOUS, THOMAS BETTERTON, supra note 48, at 133-44, 188; SAWYER, supra note 69, at 31-35, 43-44. 81. See Season of 1704-1705, supra note 76; Milhous, Theatre Companies and Regulation, supra note 47, at 116 82. See MILHOUS, THOMAS BETTERTON, supra note 48, at 161-62, 167; Milhous, Theatre Companies and Regulation, supra note 47, at 117-18. 83. See NICHOLSON, supra note 70, at 8-19; STERN, supra note 52, at 126-28. 82. See MILHOUS, THOMAS BETTERTON, supra note 48, at 161-62, 167; Milhous, Thea Companies and Regulation, supra note 47, at 117-18. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1394 The rebel actors formed a rival company as an actors' cooperative, while the manager of the patent company continued to treat it as a business investment.74 Both companies relied heavily on productions from their repertory, mounting relatively few new scripts.7" Occasionally they competed by trying to usurp one another's plays," but the strategy proved commercially unrewarding." Neither company prospered. The actors' company lacked the capital to build or renovate a suitable theatre, and its sharers, though famous, were aging. 8 Meanwhile, whether because of mismanagement or fraud, the patent company was unable to pay its investors, its actors or its rent." Audience tastes were changing-increasingly, the crowds preferred spectacle, music, dance, and circus acts to drama or comedy." At the turn of the century, British moral reformers targeted theatres as particular dens of vice and indicted actors for onstage indecency.8 ' Both the actors' company and the patent company let bills go unpaid. Actors and investors complained that the managers of both companies were pocketing money rather than honoring their obligations.8 2 Queen Anne's Lord Chamberlain intervened repeatedly in efforts to reconfigure the two companies into a viable theatre.83 In 1708, he reunited the acting company. He ordered that operas might be performed exclusively in one theatre, that tragedies and comedies might be performed exclusively in the other, and that nobody except for the managers of the two theatres 81. See Season of 1704-1705, supra note 76; Milhous, Theatre Companies and Regulation, supra note 47, at 116 p 82. See MILHOUS, THOMAS BETTERTON, supra note 48, at 161-62, 167; Milhous, Thea Companies and Regulation, supra note 47, at 117-18. p g 83. See NICHOLSON, supra note 70, at 8-19; STERN, supra note 52, at 126-28. HeinOnline -- 25 Berkeley Tech. L.J. 1394 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 2010] 1395 would be permitted to produce theatrical entertainment.84 The opera theatre soon found that its expenses dwarfed its receipts; the manager of the dramatic theatre refused to pay his actors full salaries. In 1709, the actors again petitioned the Lord Chamberlain to intervene. This time, the Lord Chamberlain ordered the patent theatre to close down." He ousted the patent company's manager, and reorganized the company under different managers. To lessen inter-company competition, he restricted each company's performance dates. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 The picture that emerges from a look at 16th and 17th English century theatre history is of an industry closely supervised by the crown, which licensed theatre companies, play performance, and play printing, and intervened in disputes over repertory, personnel, performance schedule, competition, and compensation. Theatre managers valued playwrights as they valued actors, and paid them in the same fashion. Scripts once acquired entered a theatre company's repertory, where they could be revived, adapted, rewritten, performed, and printed without any further license from the writer. 7 Although there is some evidence suggestive of a custom that actors presumptively owned an interest in continuing to play the parts they had performed in the past,88 the limited competition in the English theatre during these years, together with the royal restrictions forbidding actors to defect to rival companies without royal permission, make it difficult to verify whether the custom had much force. It seems clear, though, that at the time Parliament adopted the Statute of Anne, playwrights enjoyed no common 84. See MILHOUS, THOMAS BETTERTON, supra note 48, at 189-221; Robert D. Hume, The Sponsorship of Opera in London 1704-1720, 85 MOD. PHILOLOGY 420, 423-29 (1988); Milhous & Hume, The Silendng of Drug Lane, supra note 48, at 430-32. 85. See Milhous & Hume, The Silencing of Drug Lane, supra note 48, at 435-40. 86 S id t 435 47 87. See, e.g., STERN, supra note 52, at 129-31, 241-45 (George Villier's The Rehearsal written in 1664 and then rewritten repeatedly between 1667 and 1776 to revamp its protagonist to parody different contemporary dramatists); see also COLLEY CIBBER, THE TRAGICAL HISTORY OF KING RIcHARD III (1700) (cut-and-paste job of Shakespeare's Henry VI and Richard III with Cibber's additional dialogue). 88. See STERN, supra note 52, at 149-51 (reporting that actors were deemed to own their parts); see also MILHOUS, THOMAS BETTERTON, supra note 48, at 17 (reporting that in 1660 the King's company laid claim to most of the preexisting repertory on the ground that its veteran actors had performed those plays). The apparent custom may represent no more than the practical difficulty of preventing an actor from continuing to perform a role once she had memorized her part. But see Morris v. Kelly, 37 Eng. Rep. 87. See, e.g., STERN, supra note 52, at 129-31, 241-45 (George Villier's The Rehearsal written in 1664 and then rewritten repeatedly between 1667 and 1776 to revamp its protagonist to parody different contemporary dramatists); see also COLLEY CIBBER, THE TRAGICAL HISTORY OF KING RIcHARD III (1700) (cut-and-paste job of Shakespeare's Henry VI and Richard III with Cibber's additional dialogue). 84. See MILHOUS, THOMAS BETTERTON, supra note 48, at 189-221; Robert D. Hume, The Sponsorship of Opera in London 1704-1720, 85 MOD. PHILOLOGY 420, 423-29 (1988); Milhous & Hume, The Silendng of Drug Lane, supra note 48, at 430-32. g f g p 85. See Milhous & Hume, The Silencing of Drug Lane, supra note 48, at 435-40. 86. See id. at 435-47. g f g p 85. See Milhous & Hume, The Silencing of Drug Lane, supra note 48, at 435-40. 88. See STERN, supra note 52, at 149-51 (reporting that actors were deemed to own their parts); see also MILHOUS, THOMAS BETTERTON, supra note 48, at 17 (reporting that in 1660 the King's company laid claim to most of the preexisting repertory on the ground that its veteran actors had performed those plays). The apparent custom may represent no more than the practical difficulty of preventing an actor from continuing to perform a role once she had memorized her part. But see Morris v. Kelly, 37 Eng. Rep. 451 (1820) (enjoining actress from performing play at a theatre other than the one that alleged ownership of the copyright and could prove ownership of the promptbook). 84. See MILHOUS, THOMAS BETTERTON, supra note 48, at 189-221; Robert D. Hume, The Sponsorship of Opera in London 1704-1720, 85 MOD. PHILOLOGY 420, 423-29 (1988); Milhous & Hume, The Silendng of Drug Lane, supra note 48, at 430-32. 85. See Milhous & Hume, The Silencing of Drug Lane, supra note 48, at 435-40. 86. See id. at 435-47. 87. See, e.g., STERN, supra note 52, at 129-31, 241-45 (George Villier's The Rehearsal written in 1664 and then rewritten repeatedly between 1667 and 1776 to revamp its protagonist to parody different contemporary dramatists); see also COLLEY CIBBER, THE TRAGICAL HISTORY OF KING RIcHARD III (1700) (cut-and-paste job of Shakespeare's Henry VI and Richard III with Cibber's additional dialogue). 88. See STERN, supra note 52, at 149-51 (reporting that actors were deemed to own their parts); see also MILHOUS, THOMAS BETTERTON, supra note 48, at 17 (reporting that in 1660 the King's company laid claim to most of the preexisting repertory on the ground that its veteran actors had performed those plays). The apparent custom may represent no more than the practical difficulty of preventing an actor from continuing to perform a role once she had memorized her part. But see Morris v. Kelly, 37 Eng. Rep. 451 (1820) (enjoining actress from performing play at a theatre other than the one that alleged ownership of the copyright and could prove ownership of the promptbook). 85. See Milhous & Hume, The Silencing of Drug Lane, supra note 48, at 435-40. 86. See id. at 435-47. 89. See, e.g., GEORGE TIcKNOR CURTIS, A TREATISE ON THE LAW OF COPYRIGHT 104-05 (1847); DEAZLEY, RETHINKING, supra note 41, at 155; 2 JAMES APPLETON MORGAN, THE LAW OF LITERATURE 286-87 (1875). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1396 law literary property right in their scripts," and the pervasive regulation of theatre left little room for such a right to arise. law literary property right in their scripts," and the pervasive regulation of theatre left little room for such a right to arise. The Statute of Anne made no immediate observable difference in the lot of dramatists. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 451 (1820) (enjoining actress from performing play at a theatre other than the one that alleged ownership of the copyright and could prove ownership of the promptbook). HeinOnline -- 25 Berkeley Tech. L.J. 1395 2010 , g , , p , 91. See HUME, supra note 62, at 279-307. Debate over the continuing validity of the royal patents, which had not been officially confirmed by Parliament, apparently emboldened several managers to open unlicensed theaters in or near London. The holders of interests in the two theatre patents went to court repeatedly to close these efforts down and to prevent performers from defecting to unlicensed venues. See BROCKETT & HILDY, supra note 45, at 220-21. Further afield, a Dublin theatre successfully fended off a lawsuit seeking to enjoin its production of Sheridan's Duenna brought by the patent theatre that claimed to have purchased all rights froim Sheridan. See The Monthl Chronologer: Ireland - Dublin, April26, THE LONDON MAGAZINE, OR, GENTLEMEN'S MONTHLY INTELLIGENCER, May 1770, at 279 (reporting Chancery decision that plaintiff was not entitled to an injunction). I am indebted to Tomas G6mez-Arostegui for bringing this case to my attention. 92. Theatrical Licensing Act of June 21, 1737, 10 Geo. II c. 28. See DAVID THOMAS, DAVID CARTER, & ANNE ETIENNE, THEATRE CENSORSHIP FROM WALPOLE TO WILSON (2007). According to Brockett, within a few years, managers began to evade the Licensing Act by selling tickets to concerts and offering dramatic entertainment for free. See BROCKETT & HILDY, supra note 45, at 221. By the 1780s, some of the larger towns outside of London had persuaded Parliament to license theatres in their towns, or empower local magistrates to do so. Id. 89. See, e.g., GEORGE TIcKNOR CURTIS, A TREATISE ON THE LAW OF COPYRIGHT 104-05 (1847); DEAZLEY, RETHINKING, supra note 41, at 155; 2 JAMES APPLETON MORGAN, THE LAW OF LITERATURE 286-87 (1875). 90. See, e.g., Hume, supra note 59, at 57. 91. See HUME, supra note 62, at 279-307. Debate over the continuing validity of the royal patents, which had not been officially confirmed by Parliament, apparently emboldened several managers to open unlicensed theaters in or near London. The holders of interests in the two theatre patents went to court repeatedly to close these efforts down and to prevent performers from defecting to unlicensed venues. See BROCKETT & HILDY, supra note 45, at 220-21. Further afield, a Dublin theatre successfully fended off a lawsuit seeking to enjoin its production of Sheridan's Duenna brought by the patent theatre that claimed to have purchased all rights froim Sheridan. See The Monthl Chronologer: Ireland - Dublin, April26, THE LONDON MAGAZINE, OR, GENTLEMEN'S MONTHLY INTELLIGENCER, May 1770, at 279 (reporting Chancery decision that plaintiff was not entitled to an injunction). I am indebted to Tomas G6mez-Arostegui for bringing this case to my attention. 92. Theatrical Licensing Act of June 21, 1737, 10 Geo. II c. 28. See DAVID THOMAS, DAVID CARTER, & ANNE ETIENNE, THEATRE CENSORSHIP FROM WALPOLE TO WILSON (2007). According to Brockett, within a few years, managers began to evade the Licensing Act by selling tickets to concerts and offering dramatic entertainment for free. See BROCKETT & HILDY, supra note 45, at 221. By the 1780s, some of the larger towns outside of London had persuaded Parliament to license theatres in their towns, or empower local magistrates to do so. Id. 90. See, e.g., Hume, supra note 59, at 57. B. ENGLISH THEATRE FROM THE STATUTE OF ANNE TO THE 1833 DRAMATIC LITERARY PROPERTY ACT Queen Anne died in 1714, to be succeeded by George I. The Hanover kings were not particularly interested in theatre, and actors were no longer deemed even honorary members of the royal household." Throughout the 18th century, small theatre companies sprang up to compete with the two patent theatres, but extensive government regulation and uncertain finances kept them from gaining a foothold." In 1737, Parliament passed the Licensing Act, which limited performance of legitimate drama to the two patent theatre companies and required that any script be vetted by the Lord Chamberlain before its first performance. 92 Theatre historians report that the licensing act essentially shut down opportunities for playwrights to place new scripts, and many dramatists shifted their efforts to poetry or novels. Even 91. See HUME, supra note 62, at 279-307. Debate over the continuing validity of the royal patents, which had not been officially confirmed by Parliament, apparently emboldened several managers to open unlicensed theaters in or near London. The holders of interests in the two theatre patents went to court repeatedly to close these efforts down and to prevent performers from defecting to unlicensed venues. See BROCKETT & HILDY, supra note 45, at 220-21. Further afield, a Dublin theatre successfully fended off a lawsuit seeking to enjoin its production of Sheridan's Duenna brought by the patent theatre that claimed to have purchased all rights froim Sheridan. See The Monthl Chronologer: Ireland - Dublin, April26, THE LONDON MAGAZINE, OR, GENTLEMEN'S MONTHLY INTELLIGENCER, May 1770, at 279 (reporting Chancery decision that plaintiff was not entitled to an injunction). I am indebted to Tomas G6mez-Arostegui for bringing this case to my attention. 92. Theatrical Licensing Act of June 21, 1737, 10 Geo. II c. 28. See DAVID THOMAS, DAVID CARTER, & ANNE ETIENNE, THEATRE CENSORSHIP FROM WALPOLE TO WILSON (2007). According to Brockett, within a few years, managers began to evade the Licensing Act by selling tickets to concerts and offering dramatic entertainment for free. See BROCKETT & HILDY, supra note 45, at 221. By the 1780s, some of the larger towns outside of London had persuaded Parliament to license theatres in their towns, or empower local magistrates to do so. Id. 92. Theatrical Licensing Act of June 21, 1737, 10 Geo. II c. 28. See DAVID THOMAS, DAVID CARTER, & ANNE ETIENNE, THEATRE CENSORSHIP FROM WALPOLE TO WILSON (2007). 93. See Hume, supra note 59, at 67-70; HUME, supra note 62, at 307-11; Milhous, Theatre Companies and Regulation, supra note 47, at 108, 122-24. 94. See JOHN RUSSELL STEPHENS, THE CENSORSHIP OF ENGLISH DRAMA 1824-1901, at 7-9 (1980). 95. See Milhous & Hume, Playnwrights' Remuneraion, supra note 49, at 35-43; 3 ALLARDYCE NICOLL, A HISTORY OF ENGLISH DRAMA 1660-1900, at 47 (1952). 96. HOUSE OF COMMONS, REPORT FROM THE SELECT COMM. ON DRAMAnIC LITERATURE WITH MINUTES OF EVIDENCE (1832), http://www.copyrighthistory.org/cgi- bin/kleioc/0010/exec/ausgabe/%22uk_1832%22; Milhouse & Hume, Playwrzghts' Remuneration, supra note 49, at 6-7. 97. Milhouse & Hume, Plawrights' Remuneradon, supra note 49, at 26-27. 98. EDwARD GEORGE BULWER-LYTON, PAUL CLIFFORD 13 (1830). 99. THE BULWER-LYTrON FICTION CONTEST, http://www.bulwer-lytton.com (last visited Feb. 15, 2010). 100. See generally LESLIE MITCHELL, BULWER LYrTON: THE RISE AND FALL OF A VICTORIAN MAN OF LETTERS (2003). THE INVENTION OF COMMON LAW PLAY RIGHT THE INVENTION OF COMMON LAW PLAY RIGHT 2010] 1397 though the vast majority of productions came from the extant repertory, 93 though, the two patent companies produced several new scripts every year. g With their opportunities at home constrained, some British actors formed companies to tour in the New World. Others sought to perform outside of London or in venues on the outskirts of the city, in competition with the patent companies, avoiding the prohibition by styling their productions as "concerts" or "burlettas" rather than dramas, comedies, or plays. 94 p y Toward the end of the 18th century, theatre companies began to contract with some authors to allow them to keep their statutory copyrights, and contract directly with stationers, in return for less money.95 In 1794, the two patent theatres started paying flat fees rather than benefits." Both companies had built larger houses to accommodate larger audiences, and they imposed the change to limit the amounts they needed to pay playwrights for new scripts. Historians Judith Mithous and Robert Hume have examined the account books of one of the patent theatres and concluded that, on average, playwrights collected more money under the new system than they would have under the old. By the end of the 18th century it was becoming possible for at least some playwrights to earn a living writing for the theatre. p 97. Milhouse & Hume, Plawrights' Remuneradon, supra note 49, at 26-27. B. ENGLISH THEATRE FROM THE STATUTE OF ANNE TO THE 1833 DRAMATIC LITERARY PROPERTY ACT According to Brockett, within a few years, managers began to evade the Licensing Act by selling tickets to concerts and offering dramatic entertainment for free. See BROCKETT & HILDY, supra note 45, at 221. By the 1780s, some of the larger towns outside of London had persuaded Parliament to license theatres in their towns, or empower local magistrates to do so. Id. HeinOnline -- 25 Berkeley Tech. L.J. 1396 2010 93. See Hume, supra note 59, at 67-70; HUME, supra note 62, at 307-11; Milhous, Theatre ompanies and Regulation, supra note 47, at 108, 122-24. p g , p , 94. See JOHN RUSSELL STEPHENS, THE CENSORSHIP OF ENGLISH DRAMA 1824-1901, t 7-9 (1980). 99. THE BULWER-LYTrON FICTION CONTEST, http://www.bulwer-lytton.com (last visited Feb. 15, 2010). 97. Milhouse & Hume, Plawrights' Remuneradon, supra note 49, at 26-27. 98 ED ARD GEORGE BULWER LYTON PAUL CLIFFORD 13 (1830) , ) 100. See generally LESLIE MITCHELL, BULWER LYrTON: THE RISE AND FALL OF A VICTORIAN MAN OF LETTERS (2003). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1398 sprung up.'o' Licensed by the crown to perform musical entertainment and burlettas, they insisted that they could produce dramas, comedies, and melodramas so long as they interspersed them in bills with non-dramatic entertainment. The managers of the patent companies claimed that the non- patent theatres bore significant responsibility for patent companies' financial troubles and pursued legal campaigns to shut the non-patent theatres down.10 2 In 1832, residents of London presented a petition protesting the prosecutions of upstart theatre companies. Bulwer-Lytton moved to establish a Select Committee on Dramatic Literature.103 Appointed chairman of the Committee, he held hearings, inviting testimony from playwrights, actors, managers, and government officials.104 The picture that the witnesses presented portrayed a lively environment of non-patent theatres operating under limited licenses but producing dramas and comedies in defiance of the patent companies' assertions of exclusive rights.'s Play scripts were readily available. In order to claim copyright under the extant statute, it was necessary to publish and register a play as a book. Once published, however, a play could be performed by any theatre that could get its hands on a script-the performance rights in published plays were universally understood to be in the public domain. 06 sprung up.'o' Licensed by the crown to perform musical entertainment and burlettas, they insisted that they could produce dramas, comedies, and melodramas so long as they interspersed them in bills with non-dramatic entertainment. The managers of the patent companies claimed that the non- patent theatres bore significant responsibility for patent companies' financial troubles and pursued legal campaigns to shut the non-patent theatres down.10 2 In 1832, residents of London presented a petition protesting the prosecutions of upstart theatre companies. Bulwer-Lytton moved to establish a Select Committee on Dramatic Literature.103 Appointed chairman of the Committee, he held hearings, inviting testimony from playwrights, actors, managers, and government officials.104 The picture that the witnesses presented portrayed a lively environment of non-patent theatres operating under limited licenses but producing dramas and comedies in defiance of the patent companies' assertions of exclusive rights.'s Play scripts were readily available. In order to claim copyright under the extant statute, it was necessary to publish and register a play as a book. Once published, however, a play could be performed by any theatre that could get its hands on a script-the performance rights in published plays were universally understood to be in the public domain. C. STATUTORY PLAY RIGHT IN ENGLAND Edward Bulwer-Lytton is known to modern American audiences as the author of the phrase "[i]t was a dark and Stormy night,"98 and the namesake of an annual contest for overwrought prose." Bulwer-Lytton was a popular novelist, a poet, a baron, and a Member of Parliament.'o By the late 1820s, the patent theatres were deeply in debt and smaller, competing theatres had 99. THE BULWER-LYTrON FICTION CONTEST, http://www.bulwer-lytton.com (last visited Feb. 15, 2010). , ) 100. See generally LESLIE MITCHELL, BULWER LYrTON: THE RISE AND FALL OF A VICTORIAN MAN OF LETTERS (2003). 100. See generally LESLIE MITCHELL, BULWER LYrTON: THE RISE AND FALL OF A VICTORIAN MAN OF LETTERS (2003). HeinOnline -- 25 Berkeley Tech. L.J. 1397 2010 101. See Jane Moody, The Theatrical Revolution, 1776-1843, in THE CAMBRIDGE HISTORY OF BRITISH THEATRE, supra note 47, at 199, 200-10. 102. See REPORT FROM THE SELECT COMM. ON DRAMATIC LITERATURE, supra note 96, at 21-116; see, e.g., Morris v. Kelly, (1820) 37 Eng. Rep. 451. A complicating factor was that managers of the patent theatres mortgaged their shares and defaulted on the loans, giving creditors control of the companies. The creditors frequently failed to pay the actors what they had promised, so the actors had significant incentives to seek additional or alternative work. See id. D l C D i Li P t A t 1833 PRIMARY 101. See Jane Moody, The Theatrical Revolution, 1776-1843, in THE CAMBRIDGE HISTORY OF BRITISH THEATRE, supra note 47, at 199, 200-10. 102. See REPORT FROM THE SELECT COMM. ON DRAMATIC LITERATURE, supra note 96, at 21-116; see, e.g., Morris v. Kelly, (1820) 37 Eng. Rep. 451. A complicating factor was that managers of the patent theatres mortgaged their shares and defaulted on the loans, giving creditors control of the companies. The creditors frequently failed to pay the actors what they had promised, so the actors had significant incentives to seek additional or alternative work. See id. 103. See Ronan Deazley, Commentary on Dramatic Literary Property Act 1833, PRIMARY SOURCES ON COPYRIGHT (1450-1900) (L. Bently & M. Kretschmer, eds., 2008), http://www.copyrighthistory.org/cgi-bin/kleioc/0010/exec/ausgabeCom/%22uk_1833/ 22. 104. See id.; NICHOLSON, supra note 70, at 325-34. 105. See generally REPORT FROM THE SELECT COMM. ON DRAMATIC LITERATURE, supra note 96, at 9 (reproducing hearing transcripts). 106. E.g., DEAZLEY, RETHINKING, supra note 41, at 154-55. 107. See REPORT FROM THE SELECT COMM. ON DRAMATIC LITERATURE, shpra note 96, at 149 (testimony of David Edward Morris, proprietor, Haymarket Theatre). 103. See Ronan Deazley, Commentary on Dramatic Literary Property Act 1833, PRIMARY SOURCES ON COPYRIGHT (1450-1900) (L. Bently & M. Kretschmer, eds., 2008), http://www.copyrighthistory.org/cgi-bin/kleioc/0010/exec/ausgabeCom/%22uk_1833/ 22. 101. See Jane Moody, The Theatrical Revolution, 1776-1843, in THE CAMBRIDGE HISTORY OF BRITISH THEATRE, supra note 47, at 199, 200-10. 104. See id.; NICHOLSON, supra note 70, at 325-34. 105. See generally REPORT FROM THE SELECT COMM. ON DRAMATIC LITERATURE, supra note 96, at 9 (reproducing hearing transcripts). 106. E.g., DEAZLEY, RETHINKING, supra note 41, at 154-55. 107. See REPORT FROM THE SELECT COMM. ON DRAMATIC LITERATURE, shpra note 96, at 149 (testimony of David Edward Morris, proprietor, Haymarket Theatre). 108. See id. at 143 (testimony of Thomas Morton, reader, Drury Lane theatre); id. at 171 (testimony of Charles Mathews, proprietor, Delphi Theatre). 109. Authority on this point is not crystal clear. In Macklin v. Richardson, (1770) 27 Eng. Rep. 451, the author of the unpublished farce Love a la Mode had succeeded in enjoining the publication of the script in a magazine on the strength of his common law copyright in his unpublished script. Macklin was decided after the King's Bench decision in Milar v. Taylor, (1769) 98 Eng. Rep. 201, had recognized a perpetual common law copyright in printed, published texts that survived the Statute of Anne, but before the House of Lords' decision in Donaldson v. Beckett, (1774) 1 Eng. Rep. 837, rejected Milar v. Taylor. After Donaldson v. Beckett repudiated the notion of perpetual common law copyright in published texts, it would in theory have been possible to argue, as playwrights later did in the United States, that so long as the script remained unprinted and unpublished, the common law right had not been abrogated by the statute. The reported cases don't reflect such an argument's being made with success in the period between Donaldson and the 1932 hearings, and Parliament's later decision to equate performance with publication prevented the argument from succeeding later. 110. See, e.g., REPORT FROM THE SELECT COMM. ON DRAMATIC LITERATURE, supra note 96, at 156-59 (testimony of Douglas Jerrold, author); id. at 170 (testimony of Charles Mathews, proprietor, Adelphi Theatre). 112. Id. at 4- 6 . 113. See Deazley, supra note 103; Dewey Ganzel, Patent Wrongs and Patent Theatres: Drama and Lw in the Early Nineteenth Centug, 76 P.M.L.A. 384, 391-92 (1961). 114. 3 & 4 Will.IV, c. 15 (1833). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 06 When there had been only two theatre companies, the risk from competition was small, since the companies appear to have followed an informal practice of declining to poach each other's scripts or actors. In those circumstances, publication of a new play netted the company or playwright some extra money from the publisher without threatening performance revenues. As new theaters sprang up, though, they helped themselves to the plays in the patent companies' repertories.o' Forgoing HeinOnline -- 25 Berkeley Tech. L.J. 1398 2010 , p p p 111. See, e.g., id. at 176 (testimony of W. Thomas Montcrieff, playwright); id. at 182 (testimony of George Bartley, actor). 113. See Deazley, supra note 103; Dewey Ganzel, Patent Wrongs and Patent Theatres: Drama and Lw in the Early Nineteenth Centug, 76 P.M.L.A. 384, 391-92 (1961). 115. Cumberland v. Planche, (1834) 110 Eng. Rep. 1329; Deazley, supra note 103. Eight years later, Parliament enacted a revised copyright law that required assignment of the dramatic public performance right to be separately registered. See supra note 43 and accompanying text. 116. Deazley, supra note 103. Deazley suggests several reasons for the poor fortunes of dramatists following the enactment of the 1833 law, including poor management of theatre companies, high actor salaries and excessive production costs. Id. 117. See sources cited supra note 43. 118. British courts read the law to forfeit common law rights upon public performance even if that performance occurred in the United States. Boucicault v. Chatterton, (1867) 5 Ch.D. 267. 119. Act of May 31, 1790, at ch. 15, 1 Stat. 124. 120. See, e.g., Abrams, supra note 36, at 1178-85. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1400 the authors' assignees in whom the new performance right vested. In Cumberland v. Planchi, the court agreed that the publisher of the play owned the new performance right as part of its prior purchase of the copyright.!15 As importantly, the overall economic climate for theatre in mid-19th century London was not at all playwright-friendly."' In 1842, Parliament enacted a revised copyright statute, which provided that the dramatic performance right could be assigned separately from the copyright,' and equated the first public performance of a dramatic work with publication."' Literary property rights for plays were limited, at least from the time of the first public performance, to the rights available under the copyright statute. Even if one were to credit an argument that at some time in the past, dramatists in England, or the theatres they wrote for, could have claimed some literary property rights at common law to control the performance of their plays (and, as the past discussion indicates, I have concluded they could not), the enactment of the 1842 Act abrogated any such rights going forward. Since performance was the equivalent of publication for performance rights, and publication forfeited any non- statutory copyright, British playwrights lost any common law public performance rights in their scripts upon the initial public performance. THE INVENTION OF COMMON LAW PLAY RIGHT THE INVENTION OF COMMON LAW PLAY RIGHT 1399 2010] publication was no longer a feasible long term option: publishers were notorious for sending stenographers to attend unpublished plays and write down all the dialogue, so that they could publish unlicensed printed versions of the scripts.'" Since unpublished plays were not subject to copyright protection, the publishers risked little.'0 9 Bulwer-Lytton asked all of his witnesses whether the theatre would be improved if the patent theatres lost their monopoly, and whether better plays would result if all theatre companies needed to compensate playwrights for performing plays."0 Witnesses answered variously,"' but Bulwer-Lytton intended to pursue both reforms. In August of 1832, the Committee delivered a report calling for the expansion of copyright in dramatic literature to give writers control of public performances of their plays, and the termination of the patent theatres' monopoly.1 2 Both measures passed the House of Commons, but the House of Lords rejected the limitation on the patent theatres' monopoly, viewing the legislation as an affront to crown prerogative." 3 As of 1833, however, the Dramatic Literary Property Act"4 enabled authors or their assignees to enjoin unlicensed productions. Despite the new law, play writing did not suddenly become more remunerative. First, both theatres and publishers claimed to be 110. See, e.g., REPORT FROM THE SELECT COMM. ON DRAMATIC LITERATURE, supra note 96, at 156-59 (testimony of Douglas Jerrold, author); id. at 170 (testimony of Charles Mathews, proprietor, Adelphi Theatre). 114. 3 & 4 Will.IV, c. 15 (1833). HeinOnline -- 25 Berkeley Tech. L.J. 1399 2010 120. See, e.g., Abrams, supra note 36, at 1178-85. 115. Cumberland v. Planche, (1834) 110 Eng. Rep. 1329; Deazley, supra note 103. Eigh years later, Parliament enacted a revised copyright law that required assignment of the dramatic public performance right to be separately registered. See supra note 43 and accompanying text. 119. Act of May 31, 1790, at ch. 15, 1 Stat. 124. THE INVENTION OF COMMON LAW PLAY RIGHT 1401 2010] Pennsylvania, and, if they had, Congress had abrogated it for printed books when it enacted the 1790 Copyright Act.12 1 (And, had Congress not, the common law right had never been imagined to include public performance rights.) Dramatic works could be protected as books, and some were, but Congress did not add a dramatic performance right to the copyright law until 1856.122 127. Resolution of Oct. 20, 1774, cl. 8. 126. See id. at 6-7. 121. Wheaton v. Peters, 33 U.S. (Pet. 8) 591 (1834). 122. Act of Aug. 18, 1856, 11 Stat. 138. III. PLAY RIGHT IN AMERICA Across the Atlantic Ocean, the story developed differently. The early plot is similar. Congress enacted the first copyright statute in 1790, including only reproduction and distribution rights for books, maps, and charts."' American publishers argued that that law did not abrogate their perpetual common law copyright.'20 In Wheaton v. Peters, the Supreme Court held that it was doubtful that perpetual common law copyright had ever existed in England, but, if it had, the colonists didn't bring it with them when they emigrated to HeinOnline -- 25 Berkeley Tech. L.J. 1400 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 128. See BERNHEIM, supra note 123, at 20. British courts, for their part, ruled that if a play was initially performed in America before being copyrighted in Britain, the initial performance should be deemed a foreign publication, which forever forfeited any British copyright. Boucicault v. Delafield, (1863) 71 Eng. Rep. 261; Boucicault v. Chatterton, (1877) 5 Ch.D. 267. 129. THOMAS JAMES WALSH, PLAYWRIGHTS AND POWER: A HISTORY OF THE DRAMATISTS GUILD 14 (1996); see BEN GRAF HENNEKE, LAURA KEENE: A BIOGRAPHY 51-52 (1990) (explaining that theatre managers and courts interpreted US law before 1856 to permit a manager "to produce any play in print without permission from, or fee to, the author so long as she named him"); THOMAS KrITS, THE THEATRICAL LIFE OF GEORGE HENRY BOKER 116-17 (1994) ("Most touring stars played published texts and were therefore not obligated to seek permission or pay royalties."). 121. Wheaton v. Peters, 33 U.S. (Pet. 8) 591 (1834). 122. Act of Aug. 18, 1856, 11 Stat. 138. 123. ALFRED L. BERNHEIM, THE BUSINESS OF THE THEATRE: AN ECONOMIc HISTORY OF THE THEATRE 1758-1932, at 5-7 (1932). 124 Id 125. See id. at 6. 121. Wheaton v. Peters, 33 U.S. (Pet. 8) 591 (1834). 122. Act of Aug. 18, 1856, 11 Stat. 138. 123. ALFRED L. BERNHEIM, THE BUSINESS OF THE THEATRE: AN ECONOMIc HISTORY OF THE THEATRE 1758-1932, at 5-7 (1932). 124. Id. 125. See id. at 6. 126. See id. at 6-7. 127. Resolution of Oct. 20, 1774, cl. 8. 128. See BERNHEIM, supra note 123, at 20. British courts, for their part, ruled that if a play was initially performed in America before being copyrighted in Britain, the initial performance should be deemed a foreign publication, which forever forfeited any British copyright. Boucicault v. Delafield, (1863) 71 Eng. Rep. 261; Boucicault v. Chatterton, (1877) 5 Ch.D. 267. 129. THOMAS JAMES WALSH, PLAYWRIGHTS AND POWER: A HISTORY OF THE DRAMATISTS GUILD 14 (1996); see BEN GRAF HENNEKE, LAURA KEENE: A BIOGRAPHY BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1402 the script unless a written contract reserved some printing or performance rights to the dramatist.' Playwrights sold all rights in a script for an initial flat fee, and, sometimes, the promise of a benefit performance."' The middle of the nineteenth century saw the rise of American dramatists and the immigration of British ones choosing to become American residents and citizens. Both groups of writers complained about their treatment under U.S. law as compared with the law of Britain or France.' 32 A. EARLY AMERICAN THEATRE Early American theatre was chiefly a British import.'23 British actors formed touring companies to play in the New World, bringing their plays with them.124 Several American colonies, especially in the north, prohibited all theater as immoral,125 but the southern colonies were more receptive to visiting British performers.126 During the Revolution, the Continental Congress passed resolutions banning theatrical performances as displays of "extravagance and dissipation."127 Once independent, the new country had no international copyright relations with any European nation. All British, French, and German scripts, therefore, were free for the taking.128 Most theatre performed in the United States during the first half of the 19th century was old British repertory, new British imports, and American adaptation of scripts that had succeeded in Britain, France, and Germany. Copying was the norm.129 It's not surprising that since America imported its actors and plays from England, it also imported its customary practices. Actors and managers claimed that their payment to a playwright for a script purchased all rights in 125. See id. at 6. 129. THOMAS JAMES WALSH, PLAYWRIGHTS AND POWER: A HISTORY OF THE DRAMATISTS GUILD 14 (1996); see BEN GRAF HENNEKE, LAURA KEENE: A BIOGRAPHY 51-52 (1990) (explaining that theatre managers and courts interpreted US law before 1856 to permit a manager "to produce any play in print without permission from, or fee to, the author so long as she named him"); THOMAS KrITS, THE THEATRICAL LIFE OF GEORGE HENRY BOKER 116-17 (1994) ("Most touring stars played published texts and were therefore not obligated to seek permission or pay royalties."). HeinOnline -- 25 Berkeley Tech. L.J. 1401 2010 B. Nineteenth century American theatre involved artists who traveled frequently between the United States and Europe. English actors starred in American productions. American actors spent a season acting in London. American managers went to London or Paris to acquire scripts to adapt.133 Shortly after the enactment of the English Dramatic Literary Property Act in 1833, American writers began to lobby for a similar law here.134 Dion 130. See CLEMENT E. FOUST, THE LIFE AND DRAMATIC WORKS OF ROBERT MONTGOMERY BIRD 145-47 (1919) (recounting Bird's consulting with lawyer in unsuccessful attempt to reclaim his copyrights from Edwin Forrest). 131. WALSH, supra note 129, at 10; see BERNHEIM, supra note 123, at 20 ("Payments to dramatists did not generally burden our managers of the first half of the nineteenth century."); HENNEKE, supra note 129, at 129 (all rights to OurAmerican Cousin purchased for $1000). Royalties were not unheard of: Henneke reported that Keene offered to pay royalties for American scripts as early as the mid 1850s, 1ENNEKE, supra note 129, at 51-52, and George Henry Boker is believed to have demanded a 5% royalty on all performances of his plays from the 1850s on. WALSH, supra, at 10-11. A royalty arrangement was not necessarily more remunerative. Boker's Leonor de Gugman opened in 1852 in Philadelphia to excellent reviews and then moved to Broadway. Kitts tells us that Boker's 5% royalty on the initial run summed to $159.09. KITTs, supra note 129, at 87. 132. See generally Oren Bracha, Commentary on the U.S. Copyright Act Amendment 1856, PRIMARY SOURCES ON COPYRIGHT (1450-1900) (L. Bently & M. Kretschmer eds., 2008), http://www.copyrighthistory.org/cgi/bin/kleioc/0010/exec/ausgabeCom// 22us_1856%22; sources cited infra note 134. 133. See, e.g., 1 THE CAMBRIDGE HISTORY OF AMERICAN THEATRE 7-16 (Christopher Bigsby & Don B. Wilmeth eds., 1999); Simon E. Williams, European Actors and the Star System in the American Theatre 1752-1870, in THE CAMBRIDGE HISTORY OF AMERICAN THEATRE, supra, at 303; infra notes 134-35, 147-68 and accompanying text. 134. See FOUST, supra note 130, at 147-51; KITTS, supra note 129, at 116-18. Two of the most prolific and litigious playwrights of the 19th century were Dionysus Boucicault, an Irish playwright who became an American citizen, and Augustin Daly, a theatre manager from North Carolina. Boucicault and Daly collided in the most famous of 19th century us copyright cases, Daly v. Palmer, 6 F. Cas. 1132 (C.C.S.D.N.Y. 130. See CLEMENT E. FOUST, THE LIFE AND DRAMATIC WORKS OF ROBERT MONTGOMERY BIRD 145-47 (1919) (recounting Bird's consulting with lawyer in unsuccessful attempt to reclaim his copyrights from Edwin Forrest). 131. WALSH, supra note 129, at 10; see BERNHEIM, supra note 123, at 20 ("Payments to dramatists did not generally burden our managers of the first half of the nineteenth century."); HENNEKE, supra note 129, at 129 (all rights to OurAmerican Cousin purchased for $1000). Royalties were not unheard of: Henneke reported that Keene offered to pay royalties for American scripts as early as the mid 1850s, 1ENNEKE, supra note 129, at 51-52, and George Henry Boker is believed to have demanded a 5% royalty on all performances of his plays from the 1850s on. WALSH, supra, at 10-11. A royalty arrangement was not necessarily more remunerative. Boker's Leonor de Gugman opened in 1852 in Philadelphia to excellent reviews and then moved to Broadway. Kitts tells us that Boker's 5% royalty on the initial run summed to $159.09. KITTs, supra note 129, at 87. 135. See Dion Boucicault, Leaves from a Dramatist's Diag, 149 N. AM. REV. 228, 230 (1889). Oren Bracha doubts that Boucicault had much to do with it. See Bracha, supra note 132. Bracha notes that contemporary sources credit playwrights George Henry Boker, and Robert Montgomery Bird and New York Senator William Henry Seward as the forces behind the amendment. Id. Dr. Bird wrote to Boker in 1853 to encourage his efforts to persuade Congress to enact a dramatic copyright bill, but it is unlikely that he was directly involved in drafting the bill. Bird's health was poor, and he died in 1854. See FOUST, supra note 130, at 147-50 (reprinting letter); id. at 150-52 (describing Bird's final illness and death). Bracha speculates, however, that Bird might have been behind an 1841 effort to secure performance rights for playwrights. Bracha, supra note 132. Boucicault also claimed to have established the entitlement of European playwrights to royalties for performances of their plays. See Boucicault, supra, at 229. 141. CURTIS, supra note 89. Curtis's treatise was later plagiarized by the Bristish treatise author Walter A Copinger, who liberally incorporated Curtis prose into his 1870 book, COPINGER'S LAW OF COPYRIGHT. See id.; Ronan Deazley, Commentay on Copinger's Law of Copyright (1870), PRIMARY SOURCES ON COPYRIGHT (1450-1900) (L. Bently & M. Kretschmer eds., 2008), http://www.copyrighthistory.org/cgi-bin/kleioc/0010/exec/ THE INVENTION OF COMMON LAW PLAY RIGHT 2010] 1403 Boucicault, an Irish playwright who later became a U.S. citizen, claimed at least partial credit for persuading the U.S. Congress to enact the 1856 Dramatic Copyright Act during his initial tour of the States.135 Boucicault may have had some role in acquainting home-grown authors of the differences between U.S. and European laws. The work of persuading Congress, though, appears to have been shouldered by American-born writers George Henry Boker, Robert Taylor Conrad and Robert Montgomer Bird, who convinced Senators Charles Sumner and William Seward to shepherd a bill through the Senate.'36 The 1856 Act gave the author or proprietor of a copyrighted work, or his heirs or assigns, "the sole right . .. to act, perform, or represent" the work.'3 7 As with its British counterpart, the act was initially invoked to vindicate the rights of authors' assignees.'3 8 137. Act of Aug. 18,1856,11 Stat. 138, 139. 139. A.W.B. Simpson, The Rise and Fall of the Legal Treatise: Legal Princples and the Forms Legal Literature, 48 U. CHI. L. REV. 632, 671-74 (1981). g 138. See, e.g., Martinetti v. Maguire, 16 F. Cas. 920 (C.C.D. Cal. 1867); Roberts v. Myer 20 F. Cas. 898 (C.C.D. Mass. 1860). 136. See KITTs, supra note 129, at 118-26; Dramatic Copyright, N.Y. TIMES, Aug. 1, 1856, at 4. B. 1868), in which the court found that Boucicault, the author of After Dark, had plagiarized the railroad rescue scene from Daly's Under the Gasight. Together with Laura Keene, an expatriate Briton who became an American theater manager, and Charles Frohman, an American theatre manager who cornered the market for U.S. rights to produce of British scripts, they were responsible for most of the theatre-related copyright litigation in the 19th and early 20th century. 134. See FOUST, supra note 130, at 147-51; KITTS, supra note 129, at 116-18. Two of the most prolific and litigious playwrights of the 19th century were Dionysus Boucicault, an Irish playwright who became an American citizen, and Augustin Daly, a theatre manager from North Carolina. Boucicault and Daly collided in the most famous of 19th century us copyright cases, Daly v. Palmer, 6 F. Cas. 1132 (C.C.S.D.N.Y. 1868), in which the court found that Boucicault, the author of After Dark, had plagiarized the railroad rescue scene from Daly's Under the Gasight. Together with Laura Keene, an expatriate Briton who became an American theater manager, and Charles Frohman, an American theatre manager who cornered the market for U.S. rights to produce of British scripts, they were responsible for most of the theatre-related copyright litigation in the 19th and early 20th century. HeinOnline -- 25 Berkeley Tech. L.J. 1402 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 137. Act of Aug. 18,1856,11 Stat. 138, 139. C. COMMON LAW RIGHTS In another part of the forest, the 19th century also saw the rise of the American legal treatise." 9 Early treatises were essentially case digests, but their authors began to view their task more ambitiously. 19th century legal treatises sought to organize and explicate entire regions of law, usually by aligning their principles with the principles of natural law.'40 The first U.S. copyright treatise is usually deemed to be Curtis's.'4' Curtis articulated a HeinOnline -- 25 Berkeley Tech. L.J. 1403 2010 ausgabeCom/%22uk1870%22. Because negotiations on international copyright protection had stalled, there was not yet anything illegal about such appropriation. 145. See Roberts v. Myers, 20 F. Cas. 898, 899 (C.C.D. Mass. 1860). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1404 natural rights basis for literary property, deriving from the author's initial exclusive possession of the work that he created. Curtis insisted, however, that the right was inherently limited to "the exclusive multiplication of copies of that particular combination of characters, which exhibits to the eye of another the ideas that he intends shall be received."142 The legislative truncation of its term (from forever to a term of years) was appropriate, he wrote, because after two or three generations, administering it would be too difficult for society to tolerate. 143 Curtis also thought that a dramatist's common law right in a published play would not extend to "the sole right of presentation upon the stage."'" The first reported American cases claiming common law performance rights in plays followed the enactment of a statutory performance right for dramatic works and were initially cast as copyright suits brought under the provisions of the 1856 Act. Plaintiffs claimed copyrights in their plays, but had failed to satisfy the statutory prerequisites. The earliest cases excused the failure to meet statutory formalities,'45 but later courts held that the statute's performance right was available only for properly registered works that met all statutory requirements. 4 6 Dion Boucicault followed the practice of registering and depositing the title page of his plays without publishing his scripts. 47 In England, this would 143. Id. at 24. 144. Id. at 104. 146. The 1831 Copyright Act made copyright available to authors who were U.S. citizens or residents, or the authors' assigns, upon recording of the title of the work, deposit of the title page before publication, publication, payment of a fifty-cent recording fee, notice inserted on the tide page of all published copies, and deposit within three months of publication of a copy of the work with the clerk of the court. Act of Feb. 3, 1831, at ch. 16, 4 Stat. 436 §5 1, 4, 5. Congress replaced the 1831 Act with the general revision of 1870. The major innovations of the 1870 Act were transferring registration and recording functions from the clerks of the courts to the Library of Congress, and expanding copyright rights to allow authors to reserve the exclusive right to dramatize or translate their own works. Copyright remained available only to authors (or the assigns of authors) who were U.S. 144. Id. at 104. 142. CURTIs, supra note 89, at 13. 143. Id. at 24. the Gaslight. See Brady v. Daly, 175 U.S. 148 (1899); Daly v. Webster, 56 F. 483 (2d Cir. 1892); Daly v. Palmer, 6 F. Cas. 1132 (C.C.S.D.N.Y. 1868). Both Boucicault and Daly faced multiple charges of plagiarism in their careers. Many of their plays were adapted or revised versions of material written by others. 149. Koppel v. Downing, 11 App. D.C. 93 (1897); Carillo v. Shook, 5 F. Cas, 68 (C.C.S.D.N.Y. 1876); Boucicault v. Hart, 3 F. Cas. 983 (C.C.S.D.N.Y. 1875); Benn v. LeClercq, 3 F. Cas. 201 (C.C.D. Mass. 1873). But see Roberts, 20 F. Cas. at 899 (enjoining performance of unpublished script). 148. See supra note 43. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 citizens or residents, and continued to require recordation, a fifty-cent fee, publication, deposit, and notice as prerequisites. Act of July 8, 1870, ch. 230, 16 Stat. 198 §§ 86, 90, 92, 93, 97. 147. See, e.g., Boucicault v. Wood, 3 F. Cas. 988 (C.C.N.D. Ill. 1867); Boucicault v. Fox, 3 F. Cas. 977 (C.C.S.D.N.Y. 1862); see also sfpra notes 134-35 and accompanying text. Dionysus Boucicault was an extremely prolific playwright born in Ireland who immigrated to the United States in the latter half of the 19th century. Copyright nerds know him as the author of After Dark, the play found to have plagiarized the railroad rescue scene from Under HeinOnline -- 25 Berkeley Tech. L.J. 1404 2010 153. Boucicault v. Fox, 3 F. Cas. 977 (C.C.S.D.N.Y. 1862). 150. Roberts, 20 F. Cas. at 899. 152. Id. at 899. 151. Id. at 898-99. 154. Id. at 980-81. the Gaslight. See Brady v. Daly, 175 U.S. 148 (1899); Daly v. Webster, 56 F. 483 (2d Cir. 1892); Daly v. Palmer, 6 F. Cas. 1132 (C.C.S.D.N.Y. 1868). Both Boucicault and Daly faced multiple charges of plagiarism in their careers. Many of their plays were adapted or revised versions of material written by others. 148. See supra note 43. 149. Koppel v. Downing, 11 App. D.C. 93 (1897); Carillo v. Shook, 5 F. Cas, 68 (C.C.S.D.N.Y. 1876); Boucicault v. Hart, 3 F. Cas. 983 (C.C.S.D.N.Y. 1875); Benn v. LeClercq, 3 F. Cas. 201 (C.C.D. Mass. 1873). But see Roberts, 20 F. Cas. at 899 (enjoining performance of unpublished script). 150. Roberts, 20 F. Cas. at 899. 151. Id. at 898-99. 152. Id. at 899. 153. Boucicault v. Fox, 3 F. Cas. 977 (C.C.S.D.N.Y. 1862). 154. Id. at 980-81. 155. 3 F. Cas. 983, 986 (C.C.S.D.N.Y. 1875); accord Koppel v. Downing, 11 App. D.C. 93 (1897); Carillo v. Shook, 5 F. Cas. 68 (C.C.S.D.N.Y. 1876); Benn v. LeClercq, 3 F. Cas. 201 (C.C.D. Mass. 1873). 156. Boucicault, 3 F. Cas. at 987-88. The December 17, 1876 issue of the New York Times reports that the case settled out of court. Court Notes, N.Y. TIMEs, Dec. 17, 1876, at 10. 157. The play is best known for being the play being performed at the Ford Theatre the night that John Wilkes Booth assassinated President Lincoln. 158. Keene v. Wheatley, 14 F. Cas. 180, 185, 191 (C.C.E.D. Pa. 1861). Taylor was British. Keene was not yet an American citizen but had been a U.S. resident for many years. She had registered the copyright under the tide Our Country Cousin, in her own name as proprietor on October 2, 1858. See HENNEKE, supra note 129, at 91, 129-43. (Keene had also failed to publish the script as a printed book and deposit copies with the clerk of the court, but the judge excused that on the same reasoning adopted in Roberts v. Myers, see Keene, 14 F. Cas. at 185). Contemporary legal writers interpreted the opinion to hold that a resident alien could not copyright the works of a non-resident alien. See James Appleton Morgan, An International Copynght Law, 3 FORUM 35, 37-38 (1875). Professor Henneke claimed that the case motivated Keene to become a naturalized US citizen to avoid future copyright problems. See HENNEKE, supra note 129, at 111. p 159. HENNEKE, supra note 129, at 188. THE INVENTION OF COMMON LAW PLAY RIGHT THE INVENTION OF COMMON LAW PLAY RIGHT 1405 2010] have sufficed to secure statutory copyright.148 In the United States, however, most courts ultimately insisted that the publication of printed copies was a condition of statutory copyright. 149 Early suits to enforce the new performance right succeeded even with defects in copyright registration. In Roberts v. Myers, the purported assignee of Dion Boucicault's copyright in The Octoroon sought a preliminary injunction under the 1856 Act to prevent an unlicensed production.'" It transpired that Bouccicault had registered the title page of the play with the clerk of the court, but neglected to deposit the required printed copies."' Indeed, he had not yet published the play. Judge Sprague read the statutory language to permit him to grant the preliminary injunction: "The statute requires that such copy shall be deposited within three months after publication. That time has not arrived. There has been no publication."' 52 In a later lawsuit over the same play, the judge expressed more concern that Boucicault had not perfected his statutory copyright.'53 Judge Shipman nonetheless read Roberts v. Myers to stand for the proposition that Boucicault could either rely on a statutory copyright or he could recover on the basis of common law rights in his unpublished manuscript.15 4 In Boucicault v. Hart, however, Judge Hunt ruled the purported copyright registration of another Boucicault play invalid. The judge explained: [1]o secure a copyright of a book or a dramatic composition, the work must be published within a reasonable time after the filing of the title page, and two copies be delivered to the librarian. These two acts are, by the statute, made necessary to be performed, and we can no more take it upon ourselves to say that the latter is not 150. Roberts, 20 F. Cas. at 899. 151. Id. at 898-99. 153. Boucicault v. Fox, 3 F. Cas. 977 (C.C.S.D.N.Y. 1862). 154. Id. at 980-81. HeinOnline -- 25 Berkeley Tech. L.J. 1405 2010 160. Keene, 14 F. Cas. at 201-08. 155. 3 F. Cas. 983, 986 (C.C.S.D.N.Y. 1875); accord Koppel v. Downing, 11 App. D.C. 93 (1897); Carillo v. Shook, 5 F. Cas. 68 (C.C.S.D.N.Y. 1876); Benn v. LeClercq, 3 F. Cas. 201 (C.C.D. Mass. 1873). 156. Boucicault, 3 F. Cas. at 987-88. The December 17, 1876 issue of the New York Times eports that the case settled out of court. Court Notes, N.Y. TIMEs, Dec. 17, 1876, at 10. p 157. The play is best known for being the play being performed at the Ford Theatre the ight that John Wilkes Booth assassinated President Lincoln. 156. Boucicault, 3 F. Cas. at 987-88. The December 17, 1876 issue of the New York Times eports that the case settled out of court. Court Notes, N.Y. TIMEs, Dec. 17, 1876, at 10. 157. The play is best known for being the play being performed at the Ford Theatre the ight that John Wilkes Booth assassinated President Lincoln. 161. Id. at 207. The case recapitulated stands thus: The complainant having the literary proprietorship of this comedy, but no statutory copyright in it, and having publicly performed it at her theatre, with an intention to continue its public performance there, the defendants, against her will, performed it repeatedly at their theatre, without having been, directly or secondarily, enabled so to do through its impression upon the memory of any of her audience. This was an infraction of a proprietary right retained by the complainant. Independently also of such right, she is entitled to redress, because the defendants enabled themselves to represent the play by knowingly taking advantage of a breach of confidence committed by a person in her employment. Id BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1406 an indis ensable requisite to a copyright, than we can say it of the former. Finding that Boucicault had failed to secure a copyright in his play, the court dismissed the case. Judge Hunt expressed the view that defendant had invaded Boucicault's "common law right of ownership in his dramatic composition." Because the parties were citizens of the same state, though, the federal court had no jurisdiction to grant the remedy to which Boucicault was entitled under common law.15 Laura Keene, a New York actor-manager and the assignee of the American rights to Tom Taylor's Our Amencan Cousin,5 1 sued a competitor under the copyright statute to enjoin an unlicensed performance. The court found Keene's copyright invalid because it derived from an author who was neither a citizen nor a resident of the United States.'58 The court allowed Keene to recover damages on a breach of confidence theory: defendants had suborned one of Keene's actors and caused him to reveal all of the alterations Keene had made to the script and all of the stage business performed in her production. 9 As an alternative ground, the court suggested, Keene should be able to recover on a common law literary property theory.6 o The common law cause of action Judge Cadwalader announced was peculiarly cramped. If the proprietor of a play authorized publication, and the printed script enabled a competitor to mount an unlicensed performance, then, absent statutory copyright, the performance was not actionable. Similarly, if the proprietor HeinOnline -- 25 Berkeley Tech. L.J. 1406 2010 162. Keene v. Kimball, 82 Mass. (16 Gray) 545 (1860). 163. Id. at 551. 164. 6 F. Cas. 904 (C.C.N.D. Ill. 1870). 162. Keene v. Kimball, 82 Mass. (16 Gray) 545 (1860). 163. Id. at 551. 164. 6 F. Cas. 904 (C.C.N.D. Ill. 1870). 165. Why the husband? Parliament didn't enact the Married Women's Property Acts until later. See Married Women's Property Act, 1870, 33 & 34 Vict., c. 93 (Eng.); Married Women's Property Act, 1882, 45 & 46 Vict., c. 75 (Eng.). 167. Id. at 907. The rule that it was not copyright infringement to reproduce a script from memory (as distinguished from stenography or purloined parts) persisted, showing up in statutory as well as non statutory play right cases. See, e.g., Shook v. Rankin, 21 F. Cas. 1337 (C.C.D. Minn. 1875). The durability of the rule supports an inference that many 19th century courts continued to view copyright as conferring rights related to making and distributing material objects rather than rights in intangible works of authorship. For a different piece of evidence supporting that inference, see Edward S. Rogers, The Law of Dramac Copynght, 1 MICH. L. REV. pt. 2, 179, at 184 (1902) ("The reason why public performance of an unpublished uncopyrighted play is not a dedication, may be very briefly stated: The author does not part with his manuscript, and no copies are made so that the public can have access to the work itself."). THE INVENTION OF COMMON LAW PLAY RIGHT THE INVENTION OF COMMON LAW PLAY RIGHT 2010] 1407 authorized a public performance, and that performance enabled an unlicensed production, that production was lawful unless the proprietor had a valid statutory copyright in the script. Where, however the public performance was not what enabled the defendant to present an unauthorized production, the proprietor's reserved literary property rights would support a lawsuit. 61 Keene then filed a state court action to enjoin a second unlicensed production of Our American Cousin at the Boston Museum, relying on Massachusetts common law.162 The state court cited Judge Cadwalader's analysis with approval, but dismissed the case nonetheless, holding that [t]hese principles sustain the demurrer to the plaintiffs bill. She has publicly represented the play, Our American Cousin, before audiences consisting of all persons who chose to pay the price charged for admission to her theatre. She has employed actors to commit the various parts to memory; and unless they are restrained by some contract, express or implied, we can perceive no legal reason why they might not repeat what they have learned, before different audiences, and in various places. If persons, by frequent attendance at her theatre, have committed to memory any part or the whole of the play, they have a right to repeat what they heard to others. We know of no right of property in gestures, tones, or scenery, which would forbid such reproduction of them by the spectators as their powers of imitation might enable them to accomplish.163 Crowe v. Aiken'64 involved a second Tom Taylor script, this one titled May Warner. Taylor wrote the play for the actress Kate Bateman, and 164. 6 F. Cas. 904 (C.C.N.D. Ill. 1870). HeinOnline -- 25 Berkeley Tech. L.J. 1407 2010 169. See Robert Brauneis, Copyrght and the World's Most Popular Song, 56 J. COPR. SoC 335, 423-26 (2009). p y 166. Id. at 905. p 168. French v. Maguire, 55 How. Pr. 471 (N.Y. Sup. Ct. 1878). The plaintiff, Samuel French, founded the Samuel French dramatic licensing agency in 1830. See SAMUEL FRENCH, INC., http://www.samuelfrench.com (last visited July 31, 2010). Other cases during this time frame upheld common law first printing rights for unpublished manuscripts. See, e.g., Palmer v. DeWitt, 47 N.Y. 532 (1872) (upholding an injunction against the printing of an unpublished play). 165. Why the husband? Parliament didn't enact the Married Women's Property Acts until later. See Married Women's Property Act, 1870, 33 & 34 Vict., c. 93 (Eng.); Married Women's Property Act, 1882, 45 & 46 Vict., c. 75 (Eng.). 166. Id. at 905. 167. Id. at 907. The rule that it was not copyright infringement to reproduce a script from memory (as distinguished from stenography or purloined parts) persisted, showing up in statutory as well as non statutory play right cases. See, e.g., Shook v. Rankin, 21 F. Cas. 1337 (C.C.D. Minn. 1875). The durability of the rule supports an inference that many 19th century courts continued to view copyright as conferring rights related to making and distributing material objects rather than rights in intangible works of authorship. For a different piece of evidence supporting that inference, see Edward S. Rogers, The Law of Dramac Copynght, 1 MICH. L. REV. pt. 2, 179, at 184 (1902) ("The reason why public performance of an unpublished uncopyrighted play is not a dedication, may be very briefly stated: The author does not part with his manuscript, and no copies are made so that the public can have access to the work itself."). 168. French v. Maguire, 55 How. Pr. 471 (N.Y. Sup. Ct. 1878). The plaintiff, Samuel French, founded the Samuel French dramatic licensing agency in 1830. See SAMUEL FRENCH, INC., http://www.samuelfrench.com (last visited July 31, 2010). Other cases during this time frame upheld common law first printing rights for unpublished manuscripts. See, e.g., Palmer v. DeWitt, 47 N.Y. 532 (1872) (upholding an injunction against the printing of an unpublished play). 169. See Robert Brauneis, Copyrght and the World's Most Popular Song, 56 J. COPR. SoC'Y 335, 423-26 (2009). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1408 assigned the U.S. stage rights to her husband."' After Bateman performed the play in London, she and her husband traveled to New York and performed the play there. Neither Taylor nor plaintiffs husband sought registration under the U.S. copyright statute, nor authorized the play's publication. Defendant purchased a copy of the script and sought to stage the play in Chicago. Bateman's husband sued to enjoin him, "based not upon any copyright statutes, but on the principles of the common law and of equity." Judge Drummond examined the copy of the script, and found it suspicious. "It is not in the usual form of a published play, but consists of printed slips fastened together in pamphlet form, with plats and stage directions as if for dramatic use only."' Concluding that the defendant's script was created in some unauthorized or wrongful way, and not by memory only, the court entered a preliminary injunction.' Finally, in 1878, a New York trial court enjoined a San Francisco production of an unpublished English adaptation of a French play at the behest of the assignee of the assignees of the British adapters, relying on the common law rights attaching to unpublished manuscripts.' 8 Copyright Office records before 1978 are neither as complete nor as reliable as one might wish,'"9 but a Copyright Office publication listing every dramatic composition registered between 1870 and 1916 suggests that many dramatists may have sought to register their copyrights without publishing 167. Id. at 907. The rule that it was not copyright infringement to reproduce a script from memory (as distinguished from stenography or purloined parts) persisted, showing up in statutory as well as non statutory play right cases. See, e.g., Shook v. Rankin, 21 F. Cas. 1337 (C.C.D. Minn. 1875). The durability of the rule supports an inference that many 19th century courts continued to view copyright as conferring rights related to making and distributing material objects rather than rights in intangible works of authorship. For a different piece of evidence supporting that inference, see Edward S. Rogers, The Law of Dramac Copynght, 1 MICH. L. REV. pt. 2, 179, at 184 (1902) ("The reason why public performance of an unpublished uncopyrighted play is not a dedication, may be very briefly stated: The author does not part with his manuscript, and no copies are made so that the public can have access to the work itself."). p 168. 170. 1 LIBRARY OF CONG. COPYRIGHT OFFICE, DRAMATIC COMPOSITIONS COPYRIGHTED IN THE UNITED STATES 1870 TO 1916, at i (1918). 171. See also Frankel v. Irwin, 34 F.2d 142, 143 (S.D.N.Y. 1918) (dismissing suit for infringement of "unprinted play, copyrighted in 1901" on other grounds). 172. During the same fifteen-year time period, only five reported cases with copyrights adjudged to be satisfactory sought to vindicate the 1856 performance right for plays. Three of the five involved unpublished Dion Boucicault plays. See Boucicault v. Wood, 3 F. Cas. 988 (C.C.N.D. Ill. 1867); Boucicault v. Fox, 3 F. Cas. 977 (C.C.S.D.N.Y. 1862); Roberts v. Myers, 20 F. Cas. 898, 899 (C.C.D. Mass. 1860). One, the infamous Daly v. Palmer, 6 F. Cas. 1132 (C.C.S.D.N.Y. 1868), involvcd a copyright in a published play claimed by an American playwright. The fifth involved a translation of a French play that appears not to have been published in print. The assignees of the copyright produced a copyright certificate and an assignment from the copyright proprietor, and neither defendants nor the court appear to have raised the issue of invalidity due to the failure of the registrant to publish and deposit copies. See Shook v. Rankin, 21 F. Cas. 1337 (C.C.D. Minn. 1875). Evidence from published copyright office records suggests that many playwrights and their assignees continued to register the title of their plays without perfecting their copyright with publication and deposit of copies. See supra text accompanying note 170. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1410 presented itself."' Eaton Sylvester Drone, though, was about to publish a magisterial treatise that would change that. presented itself."' Eaton Sylvester Drone, though, was about to publish a magisterial treatise that would change that. D. EATON DRONE'S NATURAL LAW PLAY RIGHT 179. (1774) 1 Eng. Rep. 837; see DRONE, supra note 1, at 42-43. 180. 33 U.S. 591 (1834); see DRONE, supra note 1, at 43-48. 181. Stowe v. Thomas, 23 F. Cas. 201 (C.C.E.D. Pa. 1853); see DRONE, supra note 1, at 450-55. 173. See generaly MORGAN, supra note 89, at 267-370 (summarizing cases); Herman F. Selvin, Should Performance Dedicate?, 42 CALIF. L. REV. 40, 48-50 (1954). 174. DRONE, supra note 1, at vii. 175. Compare, e.g., id. at 2-26, with CURTIS, supra note 89, at 5-25. 176. DRONE, supra note 1, at 8-20, 101-15. 177. Id. at 20-26, 116-27. 178. See, e.g., id. at 22-26, 47-53; see also R.R. BOWKER, COPYRIGHT: ITS LAW AND LITERATURE 2 (1886) ("Property right in unpublished works has never been effectively questioned-a fact which in itself confirms the view that intellectual property is a natural inherent right."). 179. (1774) 1 Eng. Rep. 837; see DRONE, supra note 1, at 42-43. 180. 33 U.S. 591 (1834); see DRONE, supra note 1, at 43-48. 181. Stowe v. Thomas, 23 F. Cas. 201 (C.C.E.D. Pa. 1853); see DRONE, supra note 1, at 450-55. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 French v. Maguire, 55 How. Pr. 471 (N.Y. Sup. Ct. 1878). The plaintiff, Samuel French, founded the Samuel French dramatic licensing agency in 1830. See SAMUEL FRENCH, INC., http://www.samuelfrench.com (last visited July 31, 2010). Other cases during this time frame upheld common law first printing rights for unpublished manuscripts. See, e.g., Palmer v. DeWitt, 47 N.Y. 532 (1872) (upholding an injunction against the printing of an unpublished play). HeinOnline -- 25 Berkeley Tech. L.J. 1408 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 1409 2010] their scripts, despite the fact that the statute did not permit it until 1909. The preface explains: their scripts, despite the fact that the statute did not permit it until 1909. The preface explains: Under the legislation in force from July 8, 1870 to July 1, 1909, it was customary to file the title-page of the drama in advance of the deposit of copies and subsequently deposit the required copies. The result has been that a great many titles were filed for registration which were not followed by the deposit of copies. This was especially so in the case of dramas, and it is estimated that in more than 20,000 cases, while the title has been recorded, no copies have been received.170 We don't know how many of those 20,000 plays were claimed by authors or proprietors who didn't know they needed to print, publish and deposit copies to perfect their rights, how many were deliberately not published to discourage unlicensed performance, and how many simply went through a title change before they were finalized. It seems likely, though, that Boucicault, Keene, and Bateman were not alone in believing themselves to have secured federal copyright protection. 7 1 Thus, by the late-1 870s, the health of a common law public performance right in plays was not yet robust. The claim that common law literary property extended far enough to permit dramatists and their assigns to control public performances had initially been raised as an afterthought to failed statutory copyright claims.7 2 Courts had recognized common law performance rights where statutory copyrights were defective, but cabined them with odd limits. Few cases had arisen, and no general rule had yet HeinOnline -- 25 Berkeley Tech. L.J. 1409 2010 173. See generaly MORGAN, supra note 89, at 267-370 (summarizing cases); Herman F. Selvin, Should Performance Dedicate?, 42 CALIF. L. REV. 40, 48-50 (1954). 174. DRONE, supra note 1, at vii. 175. Compare, e.g., id. at 2-26, with CURTIS, supra note 89, at 5-25. 176. DRONE, supra note 1, at 8-20, 101-15. 177. Id. at 20-26, 116-27. 178. See, e.g., id. at 22-26, 47-53; see also R.R. BOWKER, COPYRIGHT: ITS LAW AND LITERATURE 2 (1886) ("Property right in unpublished works has never been effectively questioned-a fact which in itself confirms the view that intellectual property is a natural inherent right."). , 178. See, e.g., id. at 22-26, 47-53; see also R.R. BOWKER, COPYRIGHT: ITS LAW AND LITERATURE 2 (1886) ("Property right in unpublished works has never been effectively questioned-a fact which in itself confirms the view that intellectual property is a natural inherent right."). 177. Id. at 20-26, 116-27. 175. Compare, e.g., id. at 2-26, with CURTIS, supra note 89, at 5-25. 176. DRONE, supra note 1, at 8-20, 101-15. 173. See generaly MORGAN, supra note 89, at 267-370 (summarizing cases); Herman F. Selvin, Should PerformanceDedicate?, 42 CALIF. L. REV. 40, 48-50 (1954). D. EATON DRONE'S NATURAL LAW PLAY RIGHT Thirty years after Curtis's treatise, Eaton Drone published his thicker treatise. His goal was more ambitious than Curtis's: The task of the juridical writer is to set forth the true principles which govern the law; to point out the proper meaning of the statutes; to show what decisions are rights and what are wrong; to explain what is doubtful or obscure; and, generally, to give the law in a form as true, clear, systematic and harmonious as it is in his power to do. He is without authority to say what construction shall be given to statutes, and he is without power to overrule erroneous decisions. But he may point out the true meaning of the law, and show wherein it has been wrongly interpreted. When this has been done, the judicial affirmance of what is right and the rejection of what is wrong will be in many cases but a question of time. 174 Drone's view of the scope of copyright law was more capacious than the narrow right described by Curtis."' Drone's theory was that literary property rights were natural rights, and, as such, both perpetual and without limitation. Such rights could be expressly abrogated by statute, but not by courts. 77 Beginning with the premise that the author's natural right to his literary property comported with the best theoretical principles and had not been effectively disproved by its doubters, he then considered whether proponents of any restriction of those common law rights had made an 178 effective case for the validity of the limitation. Drone rejected as ill- reasoned cases like Donaldson v. Beckett,"' Wheaton P. Peters,so and Stowe v. Thomas."' Unfortunately, he explained, the British Parliament had abrogated playwrights' common law rights in England by declaring that public HeinOnline -- 25 Berkeley Tech. L.J. 1410 2010 THE INVENTION OF COMMON LAW PIAY RIGHT 1411 20101 performance amounted to publication.18 2 The United States Congress, however, had not yet done any such thing, and indeed, the Octoroon case had held that performance is not publication.' Thus, Drone argued, common law play right survived in all its glory.18 4 Drone's treatise sought to be comprehensive. It was nearly 800 pages long, and cited and discussed most of the copyright cases ever decided. Drone, did not, however, rely on judicial reasoning to tell him what the law was. 182. DRONE, supra note 1, at 575-76. 183. Id. at 555 (citing Boucicault v. Fox, 3 F. Cas. 977 (C.C.S.D.N.Y. 1862)). 184. DRONE, supra note 1, at 573-74. Common law play right cases, Drone cautioned, could be brought only in state court unless diversity existed. Id. at 585. 185. Thus, Drone spent four pages discussing opinions stating that it was lawful to perform a play if one reproduced it entirely from memory, DRONE, supra note 1, at 558-62, and seven pages refuting the reasoning underlying the distinction, id. at 566-72. 186. E.g., id. at 565-74, 582-84. Contemporary reviews criticized Drone for elevating his own theories over judicial decisions. See Culture and Progress: Drone on Copyright, 17 SCRIBNERS MONTHLY 911 (Feb. 1879) ("Mr Drone has no slavish reverence for precedents; he reverses decisions right and left."); Mr. Drone's Treatise on Copyrht, 6 INT'L REv. 699, 702 (1879) ("The chief difficulty .. . is that he does not, for a lawyer, sufficiently distinguish between his own views of what the law ought to be and what the courts have declared it to be."). 182. DRONE, supra note 1, at 575-76. and seven pages refuting the reasoning underlying the distinction, id. at 566 72. 186. E.g., id. at 565-74, 582-84. Contemporary reviews criticized Drone for elevating his own theories over judicial decisions. See Culture and Progress: Drone on Copyright, 17 SCRIBNERS MONTHLY 911 (Feb. 1879) ("Mr Drone has no slavish reverence for precedents; he reverses decisions right and left."); Mr. Drone's Treatise on Copyrht, 6 INT'L REv. 699, 702 (1879) ("The chief difficulty .. . is that he does not, for a lawyer, sufficiently distinguish between his own views of what the law ought to be and what the courts have declared it to be."). 192. The World, a melodrama by Paul Merrit, Henry Pettit, and Augustus Harris, received an elaborate production that ran from March 21, 1882, for eleven weeks. Orlando Tomkins was the senior manager of the Boston Theatre. EUGENE TOMKINS, THE BOSTON THEATRE 1854-1901, at 290 (1908). Thomas Halleck was the proprietor of the Alhambra Theatre, also in Boston. DONALD C. KING, THE THEATRES OF BOSTON: A STAGE AND SCREEN HISTORY 67-68 (2005). D. EATON DRONE'S NATURAL LAW PLAY RIGHT Rather, he derived what the law should be from first principles of natural law, and then criticized the cases that departed from it.' That allowed him to expound on the proper scope of common law play right, and the proper scope according to Drone was exceptionally broad."8 Here, for example, is Drone on the effect of public performance on common law rights: The true principle which governs the question relating to the effect of public representation on the owner's exclusive rights in a manuscript dramatic composition has been wholly overlooked in the recent judicial discussions of the subject. If such rights are lost, restricted, or prejudiced by public representation, it must be either by force of the common law or by operation of some statute. It has been shown elsewhere in this work that by the common law no rights in an intellectual production are forfeited by a publication of any kind. The property in a literary work is not, by the common law, prejudiced even by its publication in print. As far as the common law is concerned, the owner's rights are the same after publication in print as they were before. The now settled doctrine that there can be no copyright after publication except under the statute is based on the ground, not that publication is by the common law an abandonment of the author's rights, but that the common law property in a published work is taken away by operation of the statute. HeinOnline -- 25 Berkeley Tech. L.J. 1411 2010 191. Id. at 46. 190. 133 Mass. 32 (1882). 187. DRONE, supra note 1, at 574. 188. Drone advised that the owner of the common law playright in an unpublished play ould transfer his or her rights "either in whole or in part ... by parol." Id. at 581. 189. Thus Drone concluded, on the basis of reasoning not clear to me, that the authorized print publication of a manuscript play anywhere in the world would abrogate the common law play right of a U.S. rights holder who purchased his exclusive rights prior to the publication. See DRONE, supra note 1, at 577-83. When such a case arose, the New York State court followed Drone, and held that the common law rights in the manuscript play had been destroyed by a subsequent German publication. Daly v. Walrath, 57 N.Y.S. 1125 (App. Div. 1899). 187. DRONE, supra note 1, at 574. 188. Drone advised that the owner of the common law playright in an unpublished play could transfer his or her rights "either in whole or in part ... by parol." Id. at 581. 189. Thus Drone concluded, on the basis of reasoning not clear to me, that the authorized print publication of a manuscript play anywhere in the world would abrogate the common law play right of a U.S. rights holder who purchased his exclusive rights prior to the publication. See DRONE, supra note 1, at 577-83. When such a case arose, the New York State court followed Drone, and held that the common law rights in the manuscript play had been destroyed by a subsequent German publication. Daly v. Walrath, 57 N.Y.S. 1125 (App. Div. 1899). 190. 133 Mass. 32 (1882). 191. Id. at 46. 192. The World, a melodrama by Paul Merrit, Henry Pettit, and Augustus Harris, received an elaborate production that ran from March 21, 1882, for eleven weeks. Orlando Tomkins was the senior manager of the Boston Theatre. EUGENE TOMKINS, THE BOSTON THEATRE 1854-1901, at 290 (1908). Thomas Halleck was the proprietor of the Alhambra Theatre, also in Boston. DONALD C. KING, THE THEATRES OF BOSTON: A STAGE AND SCREEN HISTORY 67-68 (2005). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1412 It is then clear, both on principle and authority, that the property in a manuscript play is not injuriously affected by authorized public representation, unless by the operation of some statute. Now, in the United States, there is no statute which can have this effect, because there has been no legislation relating to manuscript dramatic compositions. Statutory playright is secured in published compositions alone, and representation is not publication within the meaning of the statute. When a dramatic composition is published in print, the owner's common-law rights are destroyed by operation of the statute, to which he must look for protection. But property in a manuscript play is governed exclusively by the common law, and is in no wise affected by any statute. Hence in the United States, the owner's rights in a manuscript pla are not prejudiced by its authorized public representation. Armed with Drone's analysis, the proprietors of scripts filed suit in state court (claiming ownership by assignment of playwright's rights).'" Courts followed Drone. Drone originated the name of "playright." He also invented its scope. In Tomkins v. Halleck,'" the Massachusetts Supreme Court reexamined its ruling in Keene v. Kimball, and overruled it.'9' Orlando Tompkins, one of the managers of the Boston Theatre, purchased the exclusive right to present a British melodrama in New England from its New York producer, who had bought American rights and commissioned a revision to suit American audiences.' 92 During the play's New York run, two individuals ("one Byron and one Mora") attended the play repeatedly. Byron memorized as much as HeinOnline -- 25 Berkeley Tech. L.J. 1412 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 1413 2010] he could, and dictated his recollection to Mora."'3 Byron sold the resulting script to Thomas Halleck, who produced it at his Alhambra Theatre. Tompkins filed suit in state court against Halleck, who insisted that reproducing a script from memory, as Byron and Mora had done, was expressly permitted by Keene v. Kimball. Noting Drone's objection that the privilege to reproduce from memory made no sense, the court repudiated the exception.'94 The most Drone-ish common law play right decision is the United States Supreme Court's opinion in Feris v. Frohman, decided in 1912.195 The case concerned a melodrama, The Fatal Card, written by British authors Charles Haddon Chambers and BC Stephenson, performed in London in September of 1894. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 The authors registered their British copyright in the play in November of 1894 at Stationers Hall. American manager Charles Frohman bought the rights of one of the coauthors, and produced it first in New York and then as a touring production."' George MacFarlane adapted the British play and sold it to theatre manager Richard Ferris, who registered the copyright in the adaptation.' Ferris's production of the MacFarlane 196. See New Bills of the Week, N.Y. TIMEs, Dec. 30, 1894, at 10; The Fatal Card Tonight, DAILY KENNEBECJ., Dec. 1, 1895, at 1, col. 4; Theatrical Gossip, N.Y. TIMES,Jan. 14,1895, at 8. Ultimately, Frohman licensed the script to the Famous Players Film Company, who made it into a motion picture starring John Mason and Hazel Dawn. See The Fatal Card, THE INTERNET MOVIE DATABASE, http://www.imdb.com/title/tt0005305/ (last visited Mar. 4, 2010). 196. See New Bills of the Week, N.Y. TIMEs, Dec. 30, 1894, at 10; The Fatal Card Tonight, DAILY KENNEBECJ., Dec. 1, 1895, at 1, col. 4; Theatrical Gossip, N.Y. TIMES,Jan. 14,1895, at 8. Ultimately, Frohman licensed the script to the Famous Players Film Company, who made it into a motion picture starring John Mason and Hazel Dawn. See The Fatal Card, THE INTERNET MOVIE DATABASE, http://www.imdb.com/title/tt0005305/ (last visited Mar. 4, 2010). ) 197. There is no indication in any of the reported opinions that Ferris actually published the MacFarlane adaptation, so his copyright may have been defective on that ground as well as on the ground that it infringed Frohman's common law rights. See text accompanying supra note 146. According to the special master, the MacFarlane adaptation changed the original script in many ways, and renamed all of the characters, but retained substantial parts of the original plot and dialogue. Ferris v. Frohman, 131 Ill. App. 307, 307 (1907), rev'd, 87 N.E. 327 (Ill. 1909), affd, 223 U.S. 424 (1912). Adapting European scripts for the American stage without permission of the original plays' owners was common throughout the 18th and 19th centuries, see supra notes 128-29 and accompanying text, and nothing in the American copyright statute made it illegal unless the plays had been validly registered in the United States (which became possible in 1891, under the Chace Act, see infra note 213 and accompanying text). Ferris's brief to the U.S. Supreme Court made this argument at some length. See Brief for Petitioner-Appellant in Error at 40-68, Ferris v. 195. Ferris v. Frohman, 223 U.S. 424 (1912) (applying pre-1909 Act law). 194. Id. at 43-44. 193. Tomkins, 133 Mass. at 32. 196. See New Bills of the Week, N.Y. TIMEs, Dec. 30, 1894, at 10; The Fatal Card Tonight, DAILY KENNEBECJ., Dec. 1, 1895, at 1, col. 4; Theatrical Gossip, N.Y. TIMES,Jan. 14,1895, at 8. Ultimately, Frohman licensed the script to the Famous Players Film Company, who made it into a motion picture starring John Mason and Hazel Dawn. See The Fatal Card, THE INTERNET MOVIE DATABASE, http://www.imdb.com/title/tt0005305/ (last visited Mar. 4, 2010). 193. Tomkins, 133 Mass. at 32. 194. Id. at 43-44. 195 F i F h 223 U S 424 (1912) ( l i pre 1909 Act law) 194. Id. at 43-44. 194. Id. at 43-44. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 Frohman, 223 U.S. 424 (1912) (No. 44). Frohman's brief argued that the common law performing right was subject to no such exceptions, and that the only issue for decision was whether the public performance of the play in England forfeited its common law copyright protection worldwide. Brief of Respondant-Appellee in Error at 13-37, Ferris v. Frohman, 223 U.S. 424 (1912) (No. 44). HeinOnline -- 25 Berkeley Tech. L.J. 1413 2010 198. See, e.g., Another Big Hit: The 'Fatal Card" Proves the Strongest Play Yet Seen at Low Prices, CEDAR RAPIDS REPUBLICAN, Aug. 30, 1900, at 1, col. 3. 199. See Ferris, 131 Ill. App. at 308-09. 200. See Ferris, 223 U.S. at 430. 201. Feris, 131 Ill. App. at 314. 198. See, e.g., Another Big Hit: The 'Fatal Card" Proves the Strongest Play Yet Seen at Low Pri CEDAR RAPIDS REPUBLICAN, Aug. 30, 1900, at 1, col. 3. 199. See Ferris, 131 Ill. App. at 308-09. 200. See Ferris, 223 U.S. at 430. 201. Feris, 131 Ill. App. at 314. 202. Id. at 312-14: When the authors published the play at the Adelphi Theatre, London, it was with their consent, as the record shows, That act was done and assented to under the law there in force, which expressly provided that it should be the legal equivalent "to the first publication of any book." It follows, we think, that the English authors of this play had no property rights in the United States, which they could confer upon the appellee, Frohman ... at any time subsequent to the publication in London. We are, therefore, of the opinion that appellees, complainants below, failed to establish an exclusive right to produce the play in the United States .... 332 204. Id. at 328 (citing DRONE, supra note 1, at 101 et. seq.). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1414 adaptation toured the Midwest to enthusiastic reviews.'98 Frohman filed suit in Illinois state court to enjoin Ferris's production on the basis of his common law play right. The trial court referred the case to a special master, who recommended a verdict for the defendant.'99 The trial court instead found for Frohman, entered an injunction and ordered an accounting of profits.20 0 adaptation toured the Midwest to enthusiastic reviews.'98 Frohman filed suit in Illinois state court to enjoin Ferris's production on the basis of his common law play right. The trial court referred the case to a special master, who recommended a verdict for the defendant.'99 The trial court instead found for Frohman, entered an injunction and ordered an accounting of profits.20 0 Ferris appealed, and the court of appeals reversed.2 01 The appellate court reasoned that since British law treated public performance as equivalent to publication, the public performance of the play in London had terminated the authors' common law rights, and left them with only their statutory copyright. The authors, therefore, had no common law rights to sell to Frohman.202 Frohman appealed, and the Illinois Supreme Court reversed.203 The court held, relying on Drone, that at common law, an author had "an absolute property right in his production which he could not be deprived of so long as it remained unpublished, nor could he be compelled to publish it."2 04 Statutory copyright, however, remained unavailable to works that had not been printed and published. Since there was no provision for securing copyright for unpublished plays, common law protection remained available and was not waived by public performance. 205 Ferris appealed to the U.S. 198. See, e.g., Another Big Hit: The 'Fatal Card" Proves the Strongest Play Yet Seen at Low CEDAR RAPIDS REPUBLICAN, Aug. 30, 1900, at 1, col. 3. 199. See Ferris, 131 Ill. App. at 308-09. 200. See Ferris, 223 U.S. at 430. 201. Feris, 131 Ill. App. at 314. 202. Id. 203. Ferris, 87 N.E. at 332. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 at 312-14: When the authors published the play at the Adelphi Theatre, London, it was with their consent, as the record shows, That act was done and assented to under the law there in force, which expressly provided that it should be the legal equivalent "to the first publication of any book." It follows, we think, that the English authors of this play had no property rights in the United States, which they could confer upon the appellee, Frohman ... at any time subsequent to the publication in London. We are, therefore, of the opinion that appellees, complainants below, failed to establish an exclusive right to produce the play in the United States .... It follows, we think, that the English authors of this play had no property rights in the United States, which they could confer upon the appellee, Frohman ... at any time subsequent to the publication in London. We are, therefore, of the opinion that appellees, complainants below, failed to establish an exclusive right to produce the play in the United States .... 203. Ferris, 87 N.E. at 332. 204. Id. at 328 (citing DRONE, supra note 1, at 101 et. seq.). 205. Id. (citing DRONE, supra note 1, at 119). The 1909 Act made it possible for the first time to register copyright in an unpublished play. See infra note 215 and accompanying text. The facts of the Frohman case preceded the enactment of the 1909 Act, even though the Supreme Court's eventual decision was handed down three years later. Nothing in the language of Justice Hughes' opinion, however, suggests that the unavailability of statutory copyright for unpublished works before 1909 figured in the U.S. Supreme Court's reasoning. 205. Id. (citing DRONE, supra note 1, at 119). The 1909 Act made it possible for the first time to register copyright in an unpublished play. See infra note 215 and accompanying text. The facts of the Frohman case preceded the enactment of the 1909 Act, even though the Supreme Court's eventual decision was handed down three years later. Nothing in the language of Justice Hughes' opinion, however, suggests that the unavailability of statutory copyright for unpublished works before 1909 figured in the U.S. Supreme Court's reasoning. HeinOnline -- 25 Berkeley Tech. L.J. 206. Ferris v. Frohman, 223 U.S. 424 (1912). 207. Id. at 430-31. 208. Id. at 431. 209. Id. at 434-35. 210. Id. at 436-37. 211. Id. at 437. None of the published opinions in the case addressed the similarities between the original script and the MacFarland adaptation, so it is difficult to know whether Ferris's "piratical composition" was more similar to Frohman's than the successful Daly and Boucicault adaptations had been similar to their European antecedents. 212. Id. at 434-37. 217. Two of the more litigious claimants, Augustin Daly and Dionysus Bouciault, were in fact playwrights some of the time, but they were also theatre managers. And Daly, although apparently terrifically prolific, appears not to have written many of the plays that he claimed. See generally MARVIN FELHEIM, THE THEATER OF AUGUSTIN DALY: AN ACCOUNT OF THE LATE NINETEENTH CENTURY AMERICAN STAGE 122-40, 144, 220-22 (1956) (identifying some of Daly's unattributed collaborators). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1416 drawn, during the period, to the concept of a theoretically coherent body of natural law that expressed itself in the particulars of common law. Congress, however, had already begun to address the gaps that had made common law play right seem necessary. drawn, during the period, to the concept of a theoretically coherent body of natural law that expressed itself in the particulars of common law. Congress, however, had already begun to address the gaps that had made common law play right seem necessary. In 1891, after heavy and long lobbying by playwrights, Congress passed the Chace international copyright bill, extending copyright to foreign nationals so long as they complied with statutory procedures still including registration, publication with notice, and deposit of copies. 213 In 1897, Congress passed a law imposing criminal penalties for unauthorized public performance of dramatic or musical compositions.214 In 1909, in connection with the general revision of copyright, Congress finally enacted a provision enabling the author or proprietor of a dramatic composition not reproduced for sale to secure federal copyright protection through registration and deposit.215 Thus, the United States had finally added to its law the sort of provisions that playwrights argued were necessary to protect their work. For dramatists who preferred it, moreover, there was the potentially boundless and perpetual protection of common law play right a la Eaton Drone.216 214. Act of Jan. 6, 1897, ch. 4, 29 Stat. 481. 213. Act of Mar. 3, 1891, ch. 565,26 Stat. 1106. 215. Act of Mar. 4, 1909, § 11, 35 Stat. 1075, 1078. The 1909 Act also expressly preserved the right of an "author or proprietor of an unpublished work, at common law or in equity, to prevent the copying, publication, or use of such unpublished work without his consent." Id. § 2. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1414 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 1415 2010] Supreme Court, which, in a unanimous opinion by Justice Hughes, affirmed.206 The basis for the Supreme Court's jurisdiction was not straightforward. The Illinois Supreme Court had decided the case entirely as a matter of Illinois state common law, and Frohman objected that the appeal raised no 207 federal question. justice Hughes held, however, that the fact that Ferris had registered a copyright in MacFarlane's adaptation provided a federal basis for appellate jurisdiction, because the decision denied Ferris a right otherwise secured by federal copyright law.208 Beginning with the observation that "[i]t is not open to dispute that the authors of 'The Fatal Card' had a common law right of property until it was publicly performed," the Court held that neither the English performance nor the English statute had forfeited that right. "The public representation of a dramatic composition, not printed and published, does not deprive the owner of his common law right, save by operation of statute. At common law, the public performance of the play is not an abandonment of it to the public use." 209 The British statutes, the Court continued, did not affect common law rights in the United States, so the authors were free to convey it to Frohman.210 That being the case, MacFarlane's adaptation "was simply a piratical composition," and not entitled to federal copyright protection.2 11 As Drone had, the Court began with the premise that authors had unqualified common law rights and asked whether something had happened to take those rights away.212 Finding that no statute had done so, the Court skipped right past the question of the appropriate scope of the common law right to the determination that defendants had necessarily invaded it. Common law play right was a kluge. Courts had invented it to fill gaps in statutory protection, which applied only to printed, published works by United States citizens or residents. Drone expanded it because it represented a nearly blank canvas on which to paint his theories of the appropriate scope of natural-law based common law literary property. Courts followed Drone both because his treatise seemed comprehensive and because they were HeinOnline -- 25 Berkeley Tech. L.J. 1415 2010 y g 218. See, e.g., WALSH, supra note 129, at 32-34 and sources there cited. 216. See, e.g., O'Neill v. Gen. Film Co., 157 N.Y.S. 1028 (App. Div. 1916). 213. Act of Mar. 3, 1891, ch. 565,26 Stat. 1106. 214. Act of Jan. 6, 1897, ch. 4, 29 Stat. 481. 215. Act of Mar. 4, 1909, § 11, 35 Stat. 1075, 1078. The 1909 Act also expressly preserved the right of an "author or proprietor of an unpublished work, at common law or in equity, to prevent the copying, publication, or use of such unpublished work without his consent." Id. § 2. 216. See, e.g., O'Neill v. Gen. Film Co., 157 N.Y.S. 1028 (App. Div. 1916). 217. Two of the more litigious claimants, Augustin Daly and Dionysus Bouciault, were in fact playwrights some of the time, but they were also theatre managers. And Daly, although apparently terrifically prolific, appears not to have written many of the plays that he claimed. See generally MARVIN FELHEIM, THE THEATER OF AUGUSTIN DALY: AN ACCOUNT OF THE LATE NINETEENTH CENTURY AMERICAN STAGE 122-40, 144, 220-22 (1956) (identifying some of Daly's unattributed collaborators). 218. See, e.g., WALSH, supra note 129, at 32-34 and sources there cited. E. WHOSE PLAY RIGHT? The reader who has been paying attention will have noticed that very few of the complainants in the cases thus discussed were actual playwrights.2 17 Both common law and statutory claims were pressed on behalf of proprietors who bought all rights from authors, typically for a flat fee. Except for playwrights who managed their own theatre companies, neither American nor European dramatists controlled the American rights to their plays. 218 By the early 20th century, the law may have become playwright-friendly, but customary theatre practice had not. HeinOnline -- 25 Berkeley Tech. L.J. 1416 2010 219. See BERNHEIM, supra note 123, at 28-31; OWEN DAvIS, MY FIRST FIFTY YEARS IN THE THEATRE 27-41 (1950); HENNEKE, supra note 129, at 84-96, 175-92; JERRY STAGG, THE BROTHERS SHUBERT 14-16 (1968). 220. See BERNHEIM, supra note 123, at 28-29, 34-37. 221. See id. at 40-50; WALSH, supra note 129, at 19-31. 222. See BERNHEIM, supra note 123, at 44-63; GLENN HUGHES, A HISTORY OF AMERICAN THEATRE 1700-1950, at 317-18 (1951); GEORGE MIDDLETON, THESE THINGS ARE MINE: THE AUTOBIOGRAPHY OF A JOURNEYMAN PLAYWRIGHT 53-55 (1947); STAGG, supra note 219, at 16-17, 29-51; Sean P. Holmes, All The World's a Stage! The Actors' Strike of 1919, 91 J. OF AM. HIST. 1291, 1292-95 (2005); Steve Travis, The Rise and Fall of the Theatrical Syndicate, 10 EDUC. THEATREJ. 35 (1958). 219. See BERNHEIM, supra note 123, at 28-31; OWEN DAvIS, MY FIRST FIFTY YEARS IN HE THEATRE 27-41 (1950); HENNEKE, supra note 129, at 84-96, 175-92; JERRY STAGG, THE BROTHERS SHUBERT 14-16 (1968). 225. See id. at 75-84. y 223. See BERNHEIM, supra note 123, at 64-67. 226. See id. at 85-92; HUGHES, supra note 222, at 319; MIDDLETON, supra note 222, at 374-76. 224. See id. at 67-74. 222. See BERNHEIM, supra note 123, at 44-63; GLENN HUGHES, A HISTORY OF AMERICAN THEATRE 1700-1950, at 317-18 (1951); GEORGE MIDDLETON, THESE THINGS ARE MINE: THE AUTOBIOGRAPHY OF A JOURNEYMAN PLAYWRIGHT 53-55 (1947); STAGG, supra note 219, at 16-17, 29-51; Sean P. Holmes, All The World's a Stage! The Actors' Strike of 1919, 91 J. OF AM. HIST. 1291, 1292-95 (2005); Steve Travis, The Rise and Fall of the Theatrical Syndicate, 10 EDUC. THEATREJ. 35 (1958). 220. See BERNHEIM, supra note 123, at 28-29, 34-37. 219. See BERNHEIM, supra note 123, at 28-31; OWEN DAvIS, MY FIRST FIFTY YEARS IN THE THEATRE 27-41 (1950); HENNEKE, supra note 129, at 84-96, 175-92; JERRY STAGG, THE BROTHERS SHUBERT 14-16 (1968). 220. See BERNHEIM, supra note 123, at 28-29, 34-37. 221. See id. at 40-50; WALSH, supra note 129, at 19-31. 222. See BERNHEIM, supra note 123, at 44-63; GLENN HUGHES, A HISTORY OF AMERICAN THEATRE 1700-1950, at 317-18 (1951); GEORGE MIDDLETON, THESE THINGS ARE MINE: THE AUTOBIOGRAPHY OF A JOURNEYMAN PLAYWRIGHT 53-55 (1947); STAGG, supra note 219, at 16-17, 29-51; Sean P. Holmes, All The World's a Stage! The Actors' Strike of 1919, 91 J. OF AM. HIST. 1291, 1292-95 (2005); Steve Travis, The Rise and Fall of the Theatrical Syndicate, 10 EDUC. THEATREJ. 35 (1958). 223. See BERNHEIM, supra note 123, at 64-67. 224. See id. at 67-74. 225. See id. at 75-84. 226. See id. at 85-92; HUGHES, supra note 222, at 319; MIDDLETON, supra note 222, at 374-76. 219. See BERNHEIM, supra note 123, at 28-31; OWEN DAvIS, MY FIRST FIFTY YEARS IN THE THEATRE 27-41 (1950); HENNEKE, supra note 129, at 84-96, 175-92; JERRY STAGG, THE BROTHERS SHUBERT 14-16 (1968). 220. See BERNHEIM, supra note 123, at 28-29, 34-37. 221. See id. at 40-50; WALSH, supra note 129, at 19-31. 222. See BERNHEIM, supra note 123, at 44-63; GLENN HUGHES, A HISTORY OF AMERICAN THEATRE 1700-1950, at 317-18 (1951); GEORGE MIDDLETON, THESE THINGS p 221. See id. at 40-50; WALSH, supra note 129, at 19-31. 227. BERNHEIM, supra note 123, at 88. 228. See id. at 85-92; BARNARD HEWITT, THEATRE U.S.A. 383 (1959); MIDDLETON, supra note 222, at 376-77. 229. See MIDDLETON, supra note 222, at 45, 53, 60, 68-70, 138-39 (describing problems for playwrights); WALSH, supra note 129, at 32-34 (describing problems for playwrights); Holmes, supra note 222, at 1293-95 (describing poor working conditions for actors). 230. See BERNHEIM, supra note 123, at 67-79. Both the Syndicate and the Shubert allies had invested heavily in vaudeville to fill their theatres, see id. at 67-69, further diminishing the importance of dramatic actors and scripts. 231. See HUGHES, supra note 222, at 377; Holmes, supra note 222, at 1299-1315. 232. See Thomas James Walsh, Playwrights and Power: The Dramatists' Guild Struggle for the 1926 Minimum Basic Agreement, in ART, GLITTER AND GLITZ: MAINSTREAM PLAYWRIGHTS AND POPULAR THEATRE IN 1920s AMERICA 107, 108-113 (Arthur Gewirtz & James J. Kolb eds., 2004). 233. See id.; MIDDLETON, supra note 222, at 311-13. 234. See Walsh, supra note 232 at 108-09. 235. See MIDDLETON, supra note 222, at 304-09. 236. See Walsh, supra note 232, at 107-09. 227. BERNHEIM, supra note 123, at 88. 228. See id. at 85-92; BARNARD HEWITT, THEATRE U.S.A. 383 (1959); MIDDLETON, supra note 222, at 376-77. 229. See MIDDLETON, supra note 222, at 45, 53, 60, 68-70, 138-39 (describing problems for playwrights); WALSH, supra note 129, at 32-34 (describing problems for playwrights); Holmes, supra note 222, at 1293-95 (describing poor working conditions for actors). 230. See BERNHEIM, supra note 123, at 67-79. Both the Syndicate and the Shubert allies had invested heavily in vaudeville to fill their theatres, see id. at 67-69, further diminishing the importance of dramatic actors and scripts. 236. See Walsh, supra note 232, at 107-09. 232. See Thomas James Walsh, Playwrights and Power: The Dramatists' Guild Struggle for the 1926 Minimum Basic Agreement, in ART, GLITTER AND GLITZ: MAINSTREAM PLAYWRIGHTS AND POPULAR THEATRE IN 1920s AMERICA 107, 108-113 (Arthur Gewirtz & James J. Kolb eds., 2004). 233. See id.; MIDDLETON, supra note 222, THE INVENTION OF COMMON LAW PLAY RIGHT 2010] 1417 By the end of the 19th century, theatre managers had discovered that rather than maintaining a stock company to present a season of plays in repertory at a single theatre, it was more profitable to open a show in New York City with a bankable star, run it until audience attention flagged, and then tour it to cities throughout the country.219 The principal actors and the scenery would need to be transported from city to city, but the managers of local houses could provide extra actors and crew.220 Under this model, the economics of an efficient touring route became more important than the script, which was often merely a vehicle for the bankable star of the day. Theatrical producers focused on leasing or purchasing theatres in cities to facilitate lucrative tours. Charles Frohman allied himself with producers Abe Erlanger, Marc Klaw, Al Hayman, Samuel Nixon, and Fred Zimmerman to pool their theatres into a national chain and book tours through a central office.221 Within a short time, the syndicate had an effective monopoly over American play houses, and began to demand concessions from actors, copyright owners, rival managers, and booking agents.222 Unhappy managers formed a splinter association of producers headed by the Shubert brothers to do battle with the trust.223 Both groups competed by acquiring and building new playhouses, investing in vaudeville and novelty attractions, and seeking the allegiance of each others' members.22 4 By the time of the first World War, there were a glut of theatres, and too few attractions to fill them. 225 Empty playhouses and demand for more economical entertainment helped motion pictures eclipse live theatre more quickly than anyone would have believed possible.226 Movies had two huge advantages over stage productions: they appealed to broader audiences and their tickets were much 226. See id. at 85-92; HUGHES, supra note 222, at 319; MIDDLETON, supra note 222, at 374-76. HeinOnline -- 25 Berkeley Tech. L.J. 1417 2010 235. See MIDDLETON, supra note 222, at 304-09. p p 231. See HUGHES, supra note 222, at 377; Holmes, supra note 222, at 1299-1315. 227. BERNHEIM, supra note 123, at 88. 234. See Walsh, supra note 232 at 108 09. 235. See MIDDLETON, supranote 222, at 304-09. eds., ) 233. See id.; MIDDLETON, supra note 222, at 311-13. p g 230. See BERNHEIM, supra note 123, at 67-79. Both the Syndicate and the Shubert allies ad invested heavily in vaudeville to fill their theatres, see id. at 67-69, further diminishing the mportance of dramatic actors and scripts. 234. See Walsh, supra note 232 at 108-09. , p , 228. See id. at 85-92; BARNARD HEWITT, THEATRE U.S.A. 383 (1959); MIDDLETON, pra note 222, at 376-77. See id.; MIDDLETON, supra note 222, at 311-13 237. See id. at 108-10. 238. See MIDDLETON, supra note 222, at 308. Other developments had exacerbated the problem. In 1910, Dam v. Kirk La Shelle Co., 175 F. 902 (2d Cir. 1910), had held that the initial assignee of a copyright was the sole copyright owner, and that an author could not, consistently with the statute, assign some exclusive rights but retain others. Under this ruling, copyrights were not divisible and the first producer of a play claimed not only the public performance rights, but the book publishing and motion picture rights as well. See Copyrgbts: Hearings on H.R 11258 Before the H. Comm. on Patents, 68th Cong. 45, 48 (1925) (statement of Ellis Parker Butler, Authors' League of America). 242. See MIDDLETON, supra note 222, at 322 32; STAGG, supra note 219, at 250. 243. See Hart v. B. F. Keith Vaudeville Exch., 12 F.2d 341, 344 (2d Cir. 1926) (dismissing antitrust suit against agent for vaudeville performers because "the business of acting in a theater is purely a state affair"); People v. Klaw, 106 N.Y.S. 341, 354 (Crim. Ct. 1907) (dismissing antitrust indictment because "I have failed to find any decision, nor has my attention been directed to any decision, classifying theatrical amusements as articles of 'trade' and 'commerce.'") 239. See MIDDLETON, supra note 222, at 298-331. 241. See Shuberts Seek Stay on Authors' Guild: Producers to Carr Piht on 'Tyrannical" Practices Supreme Court Monday, N.Y. TIMES, Apr. 23, 1927, at 15; Shuberts Start War on Dramatists: Ask Injunction Against Guild Charging Secondary Boycott, N.Y. TIMEs, Apr. 22, 1927, at 9. 240. See WALSH, supra note 129, at 114-19. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:138 1418 cheaper. 227 By the 1920s, the motion picture industry had triumphed, and, at least outside of New York City, theatre was already in decline.228 cheaper. 227 By the 1920s, the motion picture industry had triumphed, and, at least outside of New York City, theatre was already in decline.228 The war between the syndicate and the independent producers, meanwhile, had not been kind to working conditions for actors or playwrights.2 29 The commercial producers' focus on theatre buildings and famous stars had led to steady erosion in the perceived importance of ordinary actors and dramatists, and their bargaining power as individuals was modest. 230 In 1919, New York City actors formed a union, and, in August they went on strike, paralyzing Broadway theaters for a month.2 31 When performances resumed, the Actors Equity Association had managed to negotiate a minimum contract for its members. In response, playwrights formed a labor union of their own, and started talking about persuading producers to accept a minimum dramatists' contract.23 2 They compared the different contracts that they had signed, and found wide variation. Some playwrights wrote scripts as works made for hire, or were required to invest their own money in productions of their plays. 233 The Charles Frohman Company insisted on the playwright's ceding at least half of the stock performance and motion picture rights and performance rights in any music. 23 4 Many producers felt no compunction about rewriting lines or miscasting roles. Yet efforts to persuade playwrights to band together to insist on better pay or more artistic control had so far failed.23 1 What brought matters to a head was the question of what we now refer to as subsidiary rights. Producers insisted on controlling licensing of the plays they purchased and sharing any earnings after its initial production. 236 They were particularly eager to sell their scripts to the new movie industry. The HeinOnline -- 25 Berkeley Tech. L.J. 1418 2010 THE INVENTION OF COMMON LAW PLAY RIGHT j g g g y 242. See MIDDLETON, supra note 222, at 322-32; STAGG, supra note 219, at 250. p , 241. See Shuberts Seek Stay on Authors' Guild: Producers to Carr Piht on 'Tyrannical" Practices to Supreme Court Monday, N.Y. TIMES, Apr. 23, 1927, at 15; Shuberts Start War on Dramatists: Ask Injunction Against Guild Charging Secondary Boycott, N.Y. TIMEs, Apr. 22, 1927, at 9. 242. See MIDDLETON, supra note 222, at 322-32; STAGG, supra note 219, at 250. 243. See Hart v. B. F. Keith Vaudeville Exch., 12 F.2d 341, 344 (2d Cir. 1926) (dismissing antitrust suit against agent for vaudeville performers because "the business of acting in a theater is purely a state affair"); People v. Klaw, 106 N.Y.S. 341, 354 (Crim. Ct. 1907) (dismissing antitrust indictment because "I have failed to find any decision, nor has my attention been directed to any decision, classifying theatrical amusements as articles of 'trade' and 'commerce.'") 248. See WALSH, supra note 129, at 134-38 (comparing 1931 and 1936 contracts with 1926 contract); id. at 160-63 (explaining 1955 contracts); Alvin Deutsch, MBPC: Resquesciat in Pace - APC: Quo Vadis, 3 ENT. & SPORTS LAW. 3 (Spring 1985) (comparing 1985 APC with 1961 MBPC); Klein, supra note 244, at 41, 61-62 (1992) (comparing 1954 MBA with earlier contracts); Michael G. Yamin, Legal Problems in Broadway Theatrical Productions, 21 Bus. L. 453 (1966) (explaining 1961 MBPC). The Guild has more recently also promulgated contracts for smaller, nonprofit, and regional theatres. See WALSH, supra note 129, at 190-209; Klein, supra note 244, at 64-65; Mervyn Rothstein, Dramatists Guild Writes a Contract for its Members, N.Y. TIMES, Aug. 30, 1989, at C19. THE INVENTION OF COMMON LAW PLAY RIGHT 1419 2010] Fox Film Corporation approached several major producers, offering to pay the producers a salary in return for giving Fox both a say in the plays they produced and an option to purchase the movie rights.237 Playwrights argued that if they did not gain control of their copyrights, the only plays any theater would be willing to produce would be the plays that seemed suitable for flm. 238 They devised a minimum basic agreement under which the dramatist would keep the copyright to the play and creative control of the production; the producer would agree to pay a royalty in the form of a percentage of the gross box office receipts from the production; and the author would receive not less than half of the proceeds from a motion picture sale. The dramatists agreed with one another that they would refuse to give permission to stage any of their scripts to a producer who did not agree that henceforth, he would use the minimum contract for all American dramatists and all New York productions.239 Producers objected to the playwrights' insistence on copyright ownership and creative control. 24 0 After a significant amount of blustering, and some preliminary skirmishing in court,241 though, the producers agreed.242 Why did the producers go along? Both the idea that theatrical entertainment was exempt from the antitrust laws and the claim that the Dramatists Guild should be treated as a labor union were more colorable in the 192 0s than they are today.243 The majority of established playwrights insisted they would refuse to sign any production contract with a manager who had not agreed to be bound by the Guild's minimum basic agreement HeinOnline -- 25 Berkeley Tech. L.J. 1419 2010 244. William Klein II, Authors and Creators: Up By Their Own Bootstraps, 14 CoMM. & L. 41, 61-62 (1992). BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1420 ("MBA").2" The actors' strike had demonstrated that collective action could close theatres. The new infusion of money from movie studios was significant, playwrights would be necessary parties to the sale of new scripts to the film companies, and producers may have believed them when they threatened to withhold their work. Perhaps, even then, it was clear that the money involved in movies would eventually dwarf the money to be made from live theatre, and devising a way to share the proceeds from the sale of film rights seemed important enough to make compromises on issues of creative control and copyright ownership worthwhile. Or, it may be that playwright/producer relations were always friendlier and less antagonistic than actor/producer relations. A playwright could, after all, take the next play elsewhere.245 In any event, by the fall of 1927, all major commercial producers had agreed with the dramatists' terms. ("MBA").2" The actors' strike had demonstrated that collective action could close theatres. The new infusion of money from movie studios was significant, playwrights would be necessary parties to the sale of new scripts to the film companies, and producers may have believed them when they threatened to withhold their work. Perhaps, even then, it was clear that the money involved in movies would eventually dwarf the money to be made from live theatre, and devising a way to share the proceeds from the sale of film rights seemed important enough to make compromises on issues of creative control and copyright ownership worthwhile. Or, it may be that playwright/producer relations were always friendlier and less antagonistic than actor/producer relations. A playwright could, after all, take the next play elsewhere.245 In any event, by the fall of 1927, all major commercial producers had agreed with the dramatists' terms. Ironically, the 1926-1927 season represented the historical peak for new play production on Broadway, with 188 new plays. 24 6 Broadway productions of new plays have declined steadily since; in 2007-2008, only seven new plays opened on Broadway.247 The MBA was renegotiated in 1931, 1936, 1941, 1946, 1955, 1961, and 1985,248 but the essential terms remained the same. Playwrights keep ownership of their copyrights and creative control of stage productions of their scripts.24 9 Producers pay playwrights a minimum percentage of box office receipts.250 Producers and playwrights share subsequent earnings and split the receipts from sales of film rights. 245. The fledgling Guild was careful to recruit the most lucrative playwrights from each producer's stable. See MIDDLETON, supra note 222, at 316. p 246. TODD LONDON, BEN PESNER & ZANNIE GIRAUD Voss, OUTRAGEOUS FORTUNE: THE LIFE AND TIMES OF THE NEW AMERICAN PLAY 23-34 (2009). 249. See, e.g., Approved Production Contract for Plays §5 1.06, 8.01, reprinted in FARBER, supra note 8, at 177, 179, 201-02. p 250. See id. 4.01-4.04, at 184-91. producer s stable. See MIDDLETON, supra note 222, at 316. 246. TODD LONDON, BEN PESNER & ZANNIE GIRAUD Voss, OUTRAGEOUS FORTUNE: THE LIFE AND TIMES OF THE NEW AMERICAN PLAY 23-34 (2009). 247. Id. at 24. 244. William Klein II, Authors and Creators: Up By Their Own Bootstraps, 14 CoMM. & L. 41, 61-62 (1992). 245. The fledgling Guild was careful to recruit the most lucrative playwrights from each producer's stable. See MIDDLETON, supra note 222, at 316. 246. TODD LONDON, BEN PESNER & ZANNIE GIRAUD Voss, OUTRAGEOUS FORTUNE: THE LIFE AND TIMES OF THE NEW AMERICAN PLAY 23-34 (2009). 247. Id. at 24. 248. See WALSH, supra note 129, at 134-38 (comparing 1931 and 1936 contracts with 1926 contract); id. at 160-63 (explaining 1955 contracts); Alvin Deutsch, MBPC: Resquesciat in Pace - APC: Quo Vadis, 3 ENT. & SPORTS LAW. 3 (Spring 1985) (comparing 1985 APC with 1961 MBPC); Klein, supra note 244, at 41, 61-62 (1992) (comparing 1954 MBA with earlier contracts); Michael G. Yamin, Legal Problems in Broadway Theatrical Productions, 21 Bus. L. 453 (1966) (explaining 1961 MBPC). The Guild has more recently also promulgated contracts for smaller, nonprofit, and regional theatres. See WALSH, supra note 129, at 190-209; Klein, supra note 244, at 64-65; Mervyn Rothstein, Dramatists Guild Writes a Contract for its Members, N.Y. TIMES, Aug. 30, 1989, at C19. 249. See, e.g., Approved Production Contract for Plays §5 1.06, 8.01, reprinted in FARBER, supra note 8, at 177, 179, 201-02. 250. See id. 4.01-4.04, at 184-91. 251. See id. §§ 11.01-11.08, at 216-26. 251. See id. §§ 11.01-11.08, at 216-26. Id. at 643. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 251 There has been controversy, and HeinOnline -- 25 Berkeley Tech. L.J. 1420 2010 THE INVENTION OF COMMON LAW PLAY RIGHT THE INVENTION OF COMMON LAW PLAY RIGHT 2010] 1421 litigation. Producers filed antitrust suits claiming that the Dramatists Guild violates the antitrust laws.252 In 1945, the Court of Appeals for the Second Circuit agreed, 253 but ended up dismissing the case on the ground that the plaintiff had shown no damage.254 In 1982, producers sued again on similar claims.255 The Guild filed a counterclaim asserting that the producers had themselves violated the antitrust laws by conspiring to fix playwright compensation at an artificially low level.256 The litigation settled before anything too final could happen.257 When a new round of contract negotiations raised the specter of a new antitrust suit in 2002, famous g , ( ) ( pp j y ) 254. Author's League of Am., 186 F.2d 637. After the Second Circuit ruled that plaintiff independent producer had made a prima facie showing of an antitrust violation and ordered the trial court to enter a preliminary injunction, Spina, 148 F.2d at 651-54, the trial court submitted the question whether the Guild was a labor union (and thus entitled to the labor antitrust exemption) to the jury, which concluded that it was not, and that the Guild had violated the antitrust laws. Author's League, 186 F.2d at 639. Both Ring and the Authors' League appealed. In an apparent attempt to limit the damage of ruling the Guild to be illegal, Judge Learned Hand declined to revisit the antitrust issue: g , ( ) ( pp j y 254. Author's League of Am., 186 F.2d 637. After the Second Circuit ruled that plaintiff independent producer had made a prima facie showing of an antitrust violation and ordered the trial court to enter a preliminary injunction, Spina, 148 F.2d at 651-54, the trial court submitted the question whether the Guild was a labor union (and thus entitled to the labor antitrust exemption) to the jury, which concluded that it was not, and that the Guild had violated the antitrust laws. Author's League, 186 F.2d at 639. Both Ring and the Authors' League appealed. In an apparent attempt to limit the damage of ruling the Guild to be illegal, Judge Learned Hand declined to revisit the antitrust issue: g [T]he League is naturally concerned that it shall not be held to be a conspiracy in violation of the Anti-Trust Acts. 252. See, e.g., Drama Guild Sued to End Monopoy,'N.Y. TIMEs, Mar. 12, 1941, at 23. 253. Ring v. Spina, 148 F. 2d 647 (2d Cir. 1945) (preliminary injunction); see also Ring v. Author's League of Am., 186 F.2d 637 (2d Cir. 1951) (appeal after jury verdict). 254. Author's League of Am., 186 F.2d 637. After the Second Circuit ruled that plaintiff independent producer had made a prima facie showing of an antitrust violation and ordered the trial court to enter a preliminary injunction, Spina, 148 F.2d at 651-54, the trial court submitted the question whether the Guild was a labor union (and thus entitled to the labor antitrust exemption) to the jury, which concluded that it was not, and that the Guild had violated the antitrust laws. Author's League, 186 F.2d at 639. Both Ring and the Authors' League appealed. In an apparent attempt to limit the damage of ruling the Guild to be illegal, Judge Learned Hand declined to revisit the antitrust issue: 252. See, e.g., Drama Guild Sued to End Monopoy,'N.Y. TIMEs, Mar. 12, 1941, at 23. 253. Ring v. Spina, 148 F. 2d 647 (2d Cir. 1945) (preliminary injunction); see also Ring v. Author's League of Am., 186 F.2d 637 (2d Cir. 1951) (appeal after jury verdict). 254. Author's League of Am., 186 F.2d 637. After the Second Circuit ruled that plaintiff independent producer had made a prima facie showing of an antitrust violation and ordered the trial court to enter a preliminary injunction, Spina, 148 F.2d at 651-54, the trial court submitted the question whether the Guild was a labor union (and thus entitled to the labor antitrust exemption) to the jury, which concluded that it was not, and that the Guild had violated the antitrust laws. Author's League, 186 F.2d at 639. Both Ring and the Authors' League appealed. In an apparent attempt to limit the damage of ruling the Guild to be illegal, Judge Learned Hand declined to revisit the antitrust issue: [T]he League is naturally concerned that it shall not be held to be a conspiracy in violation of the Anti-Trust Acts. It vigorously protests its innocence and its beneficence; it is conscious of no wrongdoing, and asserts that its existence is essential to the protection of authors and composers. Such purposes would of course not protect it, if it is in fact a combination in restraint of trade or an attempted monopoly; but they are relevant in deciding whether we should decide issues in which the plaintiff has only the most shadowy interests. We hold therefore that the judgment should not have decided that, if the authors revive the play they must give the plaintiff an opportunity to "produce" it on an equal footing with anyone else, and that it should have contained no injunction. However, we hasten to add that we leave open all legal questions which such issues involve; we wish to make it entirely clear that we are not be be understood either to throw any doubt upon, or to affirm, what we said when we granted the temporary injunction; we merely decide that the necessity for such affirmance does not arise. 257. Leslie Bennett, Writers and Producers Reach Contract Accord, N.Y. TIMES, Feb. 27, 1985, at C21; see WALSH, supra note 129, at 176-89 (describing negotiations that resulted in the settlement of the litigation). The Guild and League of American Theatres and Producers agreed to characterize the 1985 Approved Production Contracts as recommended rather than mandatory, but the Guild continues its policy of disciplining members who sign a contract that does not substantially comply with the Guild's approved minimum terms. Members who are expelled from the Guild or who resign to avoid being expelled from the Guild may rejoin after a year. 255. Barr v. Dramatists Guild, Inc., 573 F. Supp. 555 (S.D.N.Y. 1983); see Carol Lawson, Dramatists Guild Sued Over Royalty Payments, N.Y. TIMES, July 28, 1982, at C19, col. 1. 256. Barr, 573 F. Supp. at 558-59. THE INVENTION OF COMMON LAW PLAY RIGHT It vigorously protests its innocence and its beneficence; it is conscious of no wrongdoing, and asserts that its existence is essential to the protection of authors and composers. Such purposes would of course not protect it, if it is in fact a combination in restraint of trade or an attempted monopoly; but they are relevant in deciding whether we should decide issues in which the plaintiff has only the most shadowy interests. We hold therefore that the judgment should not have decided that, if the authors revive the play they must give the plaintiff an opportunity to "produce" it on an equal footing with anyone else, and that it should have contained no injunction. However, we hasten to add that we leave open all legal questions which such issues involve; we wish to make it entirely clear that we are not be be understood either to throw any doubt upon, or to affirm, what we said when we granted the temporary injunction; we merely decide that the necessity for such affirmance does not arise. g [T]he League is naturally concerned that it shall not be held to be a conspiracy in violation of the Anti-Trust Acts. It vigorously protests its innocence and its beneficence; it is conscious of no wrongdoing, and asserts that its existence is essential to the protection of authors and composers. Such purposes would of course not protect it, if it is in fact a combination in restraint of trade or an attempted monopoly; but they are relevant in deciding whether we should decide issues in which the plaintiff has only the most shadowy interests. We hold therefore that the judgment should not have decided that, if the authors revive the play they must give the plaintiff an opportunity to "produce" it on an equal footing with anyone else, and that it should have contained no injunction. However, we hasten to add that we leave open all legal questions which such issues involve; we wish to make it entirely clear that we are not be be understood either to throw any doubt upon, or to affirm, what we said when we granted the temporary injunction; we merely decide that the necessity for such affirmance does not arise. 252. See, e.g., Drama Guild Sued to End Monopoy,'N.Y. TIMEs, Mar. 12, 1941, at 23. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 Section 2 of the Bill would have provided: (a) IN GENERAL- Subject to subsection (c), the antitrust laws shall not apply to any joint discussion, consideration, review, action, or agreement for the express purpose of, and limited to, the development of a standard form contract containing minimum terms of artistic protection and levels of compensation for playwrights by means of- (1) meetings, discussions, and negotiations between or among playwrights or their representatives and producers or their representatives; or (2) joint or collective voluntary actions for the limited purposes of developing a standard form contract by playwrights or their representatives. (1) meetings, discussions, and negotiations between or among playwrights or their representatives and producers or their representatives; or (2) joint or collective voluntary actions for the limited purposes of developing a standard form contract by playwrights or their representatives. (b) ADOPTION AND IMPLEMENTATION- Subject to subsection (c), the antitrust laws shall not apply to any joint discussion, consideration, review, or action for the express purpose of, and limited to, reaching a collective agreement among playwrights adopting a standard form contract developed pursuant to subsection (a) as the participating playwrights sole and exclusive means by which participating playwrights shall license their plays to producers. (c) AMENDMENT OF CONTRACT- A standard form of contract developed and implemented under subsections (a) and (b) shall be subject to amendment by individual playwrights and producers consistent with the terms of the standard form contract. (b) ADOPTION AND IMPLEMENTATION- Subject to subsection (c), the antitrust laws shall not apply to any joint discussion, consideration, review, or action for the express purpose of, and limited to, reaching a collective agreement among playwrights adopting a standard form contract developed pursuant to subsection (a) as the participating playwrights sole and exclusive means by which participating playwrights shall license their plays to producers. (c) AMENDMENT OF CONTRACT- A standard form of contract developed and implemented under subsections (a) and (b) shall be subject to amendment by individual playwrights and producers consistent with the terms of the standard form contract. , 260. See generally The Playwvnghts Licensing Antitrust Initiative Act: Safeguarding the Future of American Live Theater Hearing on S. 2349 Before the S. JudidaU Comm., 108th Cong. (2004). Representative Coble reintroduced the Bill in the 109th Congress. See H.R. 532, 109th Cong. (2005). It died there. 261. See 17 U.S.C. §5 102(a), 103 (2006). 258. S. 2082, 107th Cong. (2002); see Jesse McKinley, Legislation to Help Playwnights Negotiate, N.Y. TIMES, Apr. 12, 2002, at E6, col. 1. Id. at 643. 255. Barr v. Dramatists Guild, Inc., 573 F. Supp. 555 (S.D.N.Y. 1983); see Carol Lawson, Dramatists Guild Sued Over Royalty Payments, N.Y. TIMES, July 28, 1982, at C19, col. 1. 256. Barr, 573 F. Supp. at 558-59. 257. Leslie Bennett, Writers and Producers Reach Contract Accord, N.Y. TIMES, Feb. 27, 1985, at C21; see WALSH, supra note 129, at 176-89 (describing negotiations that resulted in the settlement of the litigation). The Guild and League of American Theatres and Producers agreed to characterize the 1985 Approved Production Contracts as recommended rather than mandatory, but the Guild continues its policy of disciplining members who sign a contract that does not substantially comply with the Guild's approved minimum terms. Members who are expelled from the Guild or who resign to avoid being expelled from the Guild may rejoin after a year. HeinOnline -- 25 Berkeley Tech. L.J. 1421 2010 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 S. 2082, § 2. 258. S. 2082, 107th Cong. (2002); see Jesse McKinley, Legislation to Help Playwnights egotiate, N.Y. TIMES, Apr. 12, 2002, at E6, col. 1. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1422 playwrights persuaded Senators Hatch and Schumer to introduce the Playwrights Licensing Relief Act.258 The bill would have exempted playwrights from antitrust liability for participating in discussions or negotiations to facilitate a standard form contract with producers. 25 9 Senators Hatch and Kennedy and Representative Coble reintroduced the bill in the following Congress, where it received a hearing, but no further action.260 The enactment of the 1976 Copyright Act has, in theory, completely obliterated any copyright distinction between published and unpublished plays. To the extent a common law performance right existed at some time under the 1870 or 1909 Acts, the 1976 Act makes the right a thing of the past. The elimination of publication or registration as a condition of statutory copyright also enabled other creative contributors to claim copyright in their authorship as part of any production that is fixed in tangible form.261 258. S. 2082, 107th Cong. (2002); see Jesse McKinley, Legislation to Help Pla Negotiate, N.Y. TIMES, Apr. 12, 2002, at E6, col. 1. 259. Section 2 of the Bill would have provided: (a) IN GENERAL- Subject to subsection (c), the antitrust laws shall not apply to any joint discussion, consideration, review, action, or agreement for the express purpose of, and limited to, the development of a standard form contract containing minimum terms of artistic protection and levels of compensation for playwrights by means of- (1) meetings, discussions, and negotiations between or among playwrights or their representatives and producers or their representatives; or (2) joint or collective voluntary actions for the limited purposes of developing a standard form contract by playwrights or their representatives. (b) ADOPTION AND IMPLEMENTATION- Subject to subsection (c), the antitrust laws shall not apply to any joint discussion, consideration, review, or action for the express purpose of, and limited to, reaching a collective agreement among playwrights adopting a standard form contract developed pursuant to subsection (a) as the participating playwrights sole and exclusive means by which participating playwrights shall license their plays to producers. (c) AMENDMENT OF CONTRACT- A standard form of contract developed and implemented under subsections (a) and (b) shall be subject to amendment by individual playwrights and producers consistent with the terms of the standard form contract. 259. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 Some commentators have argued that a derivative author owns no copyright in a licensed derivative work unless the owner of the copyright in the underlying work has authorized it. See, e.g., William Patry, Copyght in Stage Directions?, PATRY COPYRIGHT BLOG, (Jan. 29, 2006, 8:42 AM), HeinOnline -- 25 Berkeley Tech. L.J. 1422 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 2, 2010): The Council of the Guild has become aware that directors, dramaturgs and other theatrical collaborators have from time to time claimed copyright and other ownership interests in any such changes or contributions for which they claim to be responsible. They have attempted to establish their claims, among other means, by videotaping performances or filing with the Copyright Office the dramatist's script with changes, notations and other additions claimed by these collaborators. Such claims and actions infringe on the rights of dramatists to own and control their plays, and may inhibit the opportunities of other professionals, and audiences, to participate in the re-creation and enjoyment of the play. The Council of the Guild has become aware that directors, dramaturgs and other theatrical collaborators have from time to time claimed copyright and other ownership interests in any such changes or contributions for which they claim to be responsible. They have attempted to establish their claims, among other means, by videotaping performances or filing with the Copyright Office the dramatist's script with changes, notations and other additions claimed by these collaborators. Such claims and actions infringe on the rights of dramatists to own and control their plays, and may inhibit the opportunities of other professionals, and audiences, to participate in the re-creation and enjoyment of the play. Id. THE INVENTION OF COMMON LAW PLAY RIGHT THE INVENTION OF COMMON LAW PLAY RIGHT 2010] 1423 Directors and other collaborators have asserted copyright ownership of their contributions, 262 to be met with protests from the Guild.2 63 The Dramatists Guild insists that recognizing any other contributor as an author would mean the death of drama as we know it.26 4 Although some of these disputes led to litigation, most of the lawsuits have also settled.2 65 http://williampatry.blogspot.com/2006/01/copyright-in-stage-directions.html ("If, as Einhorn apparently represented to the Copyright Office, his work is a derivative work of the play script, under Section 103, he is unlikely to be able to have a valid copyright without the playwright's permission, which presumably here would be denied."). This is careless reading. So long as the derivative work is created lawfully, sections 102 and 103 provide for automatic copyright protection, regardless of the copyright owner's permission. See also H.R. REP. No. 99-1476, at 58 (1976) ("Under this provision, copyright could be obtained as long as the use of the preexisting work was not 'unlawful,' even though the consent of the copyright owner had not been obtained."). Were this not the case, there would be no need to provide expressly in section 115(a)(2) that when a sound recording is made pursuant to a statutory compulsory license (as distinguished from a negotiated license or a Harry Fox license), the new musical arrangement authorized by the statute will not be protected by copyright unless the owner of the copyright in the underlying musical composition consents. py g py g y g p 262. Jesse Green, Exit, Pursued byA Lanyer, N.Y. TIMES, Jan. 29, 2006, at Section 2: Arts & Leisure 1, col. 5; see, e.g., Mullen v. Soc'y of Dirs. & Choreographers, No. 06 C 6818, 2007 U.S. Dist. LEXIS 75235 (N.D. Ill. Sept. 28, 2007) (suit over copyright in direction, choreography and scenic design of Urinetown); Einhorn v. Mergatroyd Prods., 426 F. Supp. 2d 189 (S.D.N.Y. 2006) (suit over copyright in Edward Einhorn's direction of Tam Un); Mantello v. Hall, 947 F. Supp. 92 (S.D.N.Y. 1996) (suit over copyright in Joe Mantello's direction of Love!Valour!Compassion.. 263. See Dramatists Guild of Am., Statement: Dramatist Copynight and Intellectual Properly Rights, http://dramatistsguild.com/about-statementscopyright.aspx (last visited Apr. 2, 2010): 263. See Dramatists Guild of Am., Statement: Dramatist Copynight and Intellectual Properly Rights, http://dramatistsguild.com/about-statementscopyright.aspx (last visited Apr. BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 25:1381 1424 Today, playwrights in America retain both copyright ownership and creative control in their plays. Those strong copyright rights have not, however, made playwriting remunerative. A 2009 study by the Theatre Development Fund concluded that it is no longer possible for even the most successful playwrights to earn a living from productions of their plays.266 Working playwrights need to supplement their incomes with teaching or with writing scripts for film or television under work made for hire contracts. Meanwhile, the Drone-ish form of common law copyright has entirely died out, forgotten by everyone but the record labels who claim it for their pre-1972 recordings.267 g ) 268. Ann Bartow, The Hegemony of the Copyrght Treatise, 71 U. CIN. L. REV. 1 (2004). Id. 264. E.g., Weidman, supra note 10. The Guild argues that its members' contracts require the playwright's approval of any changes to the script, and oblige the producer to assign ownership of any changes to the playwright. Since the production is licensed subject to that clause, the producer has no right to enter into contracts with directors or designers that permit them to retain copyright in their contributions. The Stage Directors and Choreographers Society, which represents directors, has negotiated collective bargaining agreements with Broadway, Off-Broadway, regional, and stock theatres under which both directors and choreographers retain copyright in their contributions. See Stage Dirs. & Choreographers Soc'y, SDC Contracts: Collectively Bargained Agreements, HeinOnline -- 25 Berkeley Tech. L.J. 1423 2010 ttp://sdcweb.org/index.php?option=com-content&task=view&id=40&Itemid=96 (last isited Apr. 2, 2010). visited Apr. 2, 2010). 265. See, e.g., Andrew Gans, Akron Urinetwon Lawsuit Settled, PLAYBILL.COM, July 2, 2008, http://www.playbill.com/news/article/1 19148-Akron-Urinetown-Lawsuit-Settled; Campbell Robertson, A Uinetown Suit is Settled, N.Y. TIMES, Dec. 1, 2007, at B8; Lisa Jo Sagolla, Dance/Movement: Should Stage Direction Be Copynghtable?, BACKSTAGE, May 20, 2008, http://www.backstage.com/bso/esearch/article-display.jsp.vnu-contentid=1003805591. 266. LONDON, PESNER & Voss, supra note 246, at 47-96. 267. See Capitol Records v. Naxos of Am., Inc., 372 F.3d 471 (2d Cir. 2004) (certifying common law copyright question to N.Y. Court of Appeals); Capitol Records, Inc. v. Naxos of Am., Inc., 4 N.Y.2d 540 (2005) (holding that N.Y. common law copyright protects sound recordings made before 1972 even if they are in the public domain in their country of origin). 268. Ann Bartow, The Hegemony of the Copyrght Treatise, 71 U. CIN. L. REV. 1 (2004). 265. See, e.g., Andrew Gans, Akron Urinetwon Lawsuit Settled, PLAYBILL.COM, July 2, 2008, http://www.playbill.com/news/article/1 19148-Akron-Urinetown-Lawsuit-Settled; Campbell Robertson, A Uinetown Suit is Settled, N.Y. TIMES, Dec. 1, 2007, at B8; Lisa Jo Sagolla, Dance/Movement: Should Stage Direction Be Copynghtable?, BACKSTAGE, May 20, 2008, http://www.backstage.com/bso/esearch/article-display.jsp.vnu-contentid=1003805591. , , p , 267. See Capitol Records v. Naxos of Am., Inc., 372 F.3d 471 (2d Cir. 2004) (certifying common law copyright question to N.Y. Court of Appeals); Capitol Records, Inc. v. Naxos of Am., Inc., 4 N.Y.2d 540 (2005) (holding that N.Y. common law copyright protects sound recordings made before 1972 even if they are in the public domain in their country of origin). 269. The most disturbing example, in my view, is probably William S. Strauss's uncritical parroting of Drone in the 1957 copyright office study on protection for unpublished works. See Strauss, supra note 19. Because the series of studies was in general so meticulous, scholars who came after believed Strauss's description was apt. Strauss may have had a reform agenda. In a law review article published at around the same time, he argued that U.S. common law in general, and common law copyright in particular, provided protection of authors' rights that was substantially equivalent with the droit morale required under article 6bis of the Berne Convention. See William S. Strauss, The Moral Rght of the Author, 4 AM. J. COMP. L. 506, 538 (1955). That article, with some additions, became the Copyright Office's 1959 study on moral rights. See WILLIAM S. STRAUSS, STUDY No. 4: THE MORAL RIGHT OF THE AUTHOR, reprinted in S. COMM. ON THE JUDICIARY, 86TH CONG., COPYRIGHT LAW REVISION: STUDIES PREPARED FOR THE SUBCOMM. ON PATENTS, TRADEMARKS AND COPYRIGHTS OF THE S. COMM. ON THE JUDICIARY (Comm. Print 1960); LIBRARY OF CONG., SIXTIETH ANNUAL REPORT OF THE REGISTER OF COPYRIGHTS FOR THE FISCAL YEAR ENDINGJUNE 30, 1957, at 9-10 (1957). IV. CONCLUSION What lessons can we draw from the history of common law play right in America? Several lessons seem evident. Most obviously, for all of the author- centric rhetoric pervading the cases and secondary literature, common law play right was never a playwright's right-it was asserted almost exclusively by and on behalf of the theatre managers and publishers who claimed to be the playwright's assigns. We can miss important information when we take author's rights language at face value, without asking ourselves who is speaking the language and what, exactly, he or she seeks to accomplish. Second, for legal scholars, a particularly interesting subplot in this story represents what Ann Bartow has termed "the Hegemony of the Copyright Treatise."268 Eaton Drone was able to write a thick, comprehensive treatise because, instead of deriving legal principles from statutes and judicial decisions, he created his legal principles out of whole cloth and discussed judicial decisions to demonstrate where they were consistent and inconsistent with the principles he announced. That allowed Drone to have answers for many questions courts had not addressed, and to pick the winners when HeinOnline -- 25 Berkeley Tech. L.J. 1424 2010 THE INVENTION OF COMMON LAW PLAY RIGHT 1425 2010] court decisions demonstrated a conflict of authority. Drone didn't hide his method, or pretend that he was merely describing the law, rather than seeking to reshape it. Courts and other writers nonetheless relied on Drone as a dependable account of what the law was, rather than an argument about what Drone believed the law ought to become.269 Third, the fact that common law performance rights in both England and America did not precede statutory performance rights but followed them, growing up in response to perceived gaps in the rights under the statute, parallels the history of common law printing rights. This history should probably count as another nail in the coffin of the story of copyright's natural right origins in the days of primordial ooze. Further, the histories of the two different flavors of common law copyright suggest that what rights that we perceive as inherent or natural are fundamentally contingent on what rights already have names and a path to enforcement. Finally, the story reminds us that, for most people, the customs that grow up around what people believe the law to be matter more than what the law really is. IV. CONCLUSION Dramatists were able to get exceptional authors' rights through collective action that was not tied to any statute or judicial decision, and they have retained those rights despite (and maybe even in defiance of) later statutes or judicial decisions. Of the 20,000 dramatists who registered the tide of their plays without depositing copies and thought they had thereby perfected a federal copyright, most of them got an apparent copyright that probably worked as well for them as the real thing would have. Today, the strong attribution and integrity rights that playwrights claim, and their insistence on denying that their collaborators author contributions, have everything to do with customs and contracts, and very little to do to with copyright law. HeinOnline -- 25 Berkeley Tech. L.J. 1425 2010
https://openalex.org/W2545188623	https://europepmc.org/articles/pmc5139134?pdf=render	English	null	Cyanobacteria of Greece: an annotated checklist	Biodiversity data jurnal	2,016	cc-by	22,040	Background The checklist of Greek Cyanobacteria was created in the framework of the Greek Taxon Information System (GTIS), an initiative of the LifeWatchGreece Research Infrastructure (ESFRI) that has resumed eﬀorts to compile a complete checklist of species reported from Greece. This list was created from exhaustive search of the scientiﬁc literature of the last 60 years. All records of taxa known to occur in Greece were taxonomically updated. Cyanobacteria of Greece: an annotated checklist ‡ Department of Botany, School of Biology, Aristotle University of Thessaloniki, Thessaloniki, Greece Corresponding author: Spyros Gkelis (sgkelis@bio.auth.gr) Academic editor: Vasilis Gerovasileiou Received: 04 Aug 2016 \| Accepted: 24 Aug 2016 \| Published: 01 Nov 2016 Citation: Gkelis S, Ourailidis I, Panou M, Pappas N (2016) Cyanobacteria of Greece: an annotated checklist. Biodiversity Data Journal 4: e10084. https://doi.org/10.3897/BDJ.4.e10084 © Gkelis S et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Biodiversity Data Journal 4: e10084 doi: 10.3897/BDJ.4.e10084 Taxonomic Paper Biodiversity Data Journal 4: e10084 doi: 10.3897/BDJ.4.e10084 Taxonomic Paper Biodiversity Data Journal 4: e10084 doi: 10.3897/BDJ.4.e10084 Introduction The history of natural science is directly linked to Greek civilization, since the older written references to plants, animals, and algae are found in the Homeric Epics dating back to the 8 century BC; the ﬁrst record of algae in the Western literature appears in Homer’s Iliad: “the dark waves uprear their heads and scatter their sea-wrack (phycos) in all directions” [Iliad IX.1 translated by Batler (1998)]. The Greek philosopher, and ﬁrst marine biologist, Aristotle, who has established scientiﬁc knowledge for many animal and plant taxa (see Voultsiadou 2007) reported “red snow” in 4 century BC: “after lying for some time, snow turns slightly red” (Historia Animalium 552b8 translated by Peck 1970); this is considered as one of the oldest records of snow algae [unicellular Chlorophyta, like Chlamydomonas or Chloromonas according to Duval et al. (1999)]. th th In modern times, research on Greek microﬂora, Cyanobacteria included, started in the 20 century with the studies of freshwater algae by Stanković (1931) and Skuja (1937). In the subsequent two decades, no published work in microﬂora can be traced until the limnological survey by Ananiadis (1956). During the 1960s Konstantinos Anagnostidis performs two extensive studies of Cyanobacteria from thermal springs (Anagnostidis 1961) and sulphur-communities (Sulphuretum) in marine and freshwater habitats of Greece (Anagnostidis 1968). These taxonomic studies revealed the occurrence of a great number of cyanobacteria taxa, led to the description of several new species (e.g. Geitlerinema apolloniae Anagnostidis, 2001; Leptolyngbya phormidioides (Anagnostidis) Anagnostidis & Komárek, 1988; Pseudanabaena lonchoides Anagnostidis, 1961; Pseudophormidium battersii (Gomont) Anagnostidis, 2001; Borzia periklei Anagnostidis in (Anagnostidis and Komárek 1988)), and were later widely used to establish the modern classiﬁcation system of Cyanobacteria (Anagnostidis and Komárek 1985, Anagnostidis and Komárek 1988, Anagnostidis and Komárek 1990, Komárek and Anagnostidis 1986, Komárek and Anagnostidis 1989). Although a lot of the valuable taxonomic, ecological, and biogeographical information recorded in Anagnostidis' monographs is now contained in the “Süsswasserﬂora von Mitteleuropa” series (Komárek 2013, Komárek and Anagnostidis 1999, Komárek and Anagnostidis 2005, Komárek 2013), many details of his work remain unavailable to the broader scientiﬁc community since they were published in Greek. th The systematic research on Greek Cyanobacteria started in the 1980s with a series of publications on cyanobacterial taxonomy and species composition (Anagnostidis and Economou-Amilli 1978, Anagnostidis and Economou–Amilli 1980, Anagnostidis et al. 1981, Anagnostidis et al. 1982, Overbeck et al. 1982, Anagnostidis et al. New information The checklist of Greek Cyanobacteria comprises 543 species, classiﬁed in 130 genera, 41 families, and 8 orders. The orders Synechococcales and Oscillatoriales have the highest number of species (158 and 153 species, respectively), whereas these two orders along with Nostocales and Chroococcales cover 93% of the known Greek cyanobacteria species. It is worth mentioning that 18 species have been initially described from Greek habitats. The marine epilithic Ammatoidea aegea described from Saronikos Gulf is considered endemic to this area. Our bibliographic review shows that Greece hosts a high diversity of cyanobacteria, suggesting that the Mediterranean area is also a hot spot for microbes. © Gkelis S et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Gkelis S et al. Gkelis S et al. 2 blue-green algae, Cyanophyceae, bacteria, microorganisms, Aegean Sea, freshwater, marine, checklist, biodiversity, Mediterranean blue-green algae, Cyanophyceae, bacteria, microorganisms, Aegean Sea, freshwater, marine, checklist, biodiversity, Mediterranean Introduction 1983, Anagnostidis and Pantazidou 1985, Anagnostidis and Pantazidou 1988a, Anagnostidis and Pantazidou 1988b, Anagnostidis and Pantazidou 1988c, Hindak and Moustaka 1988, Anagnostidis et Cyanobacteria of Greece: an annotated checklist 3 al. 1988a, Anagnostidis and Pantazidou 1991a, Anagnostidis and Pantazidou 1991b, Anagnostidis and Roussomoustakaki 1991, Tryfon 1996), as well as on population dynamics (Moustaka-Gouni 1988, Moustaka-Gouni and Nikolaidis 1990, Moustaka-Gouni and Nikolaidis 1992, Tafas and Economou-Amilli 1997, Tryfon et al. 1997, Montesanto et al. 1999, Moustaka-Gouni et al. 2000, Temponeras et al. 2000, Vardaka et al. 2000) regarding both the freshwater and marine waterbodies. The recent research on cyanobacterial diversity in Greece comprises polyphasic or multidisciplinary studies of cyanobacteria strains (Gkelis et al. 2005b, Lamprinou et al. 2011, Lamprinou et al. 2013a, Lamprinou et al. 2013b, Gkelis et al. 2015b, Gkelis and Panou 2016, Bravakos et al. 2016), natural populations (Moustaka-Gouni et al. 2009, Moustaka-Gouni et al. 2010, Papadimitriou et al. 2013, Gkelis and Zaoutsos 2014, Gkelis et al. 2014, Gkelis et al. 2015a), as well as ecological works (e.g. Lamprinou et al. 2009, Katsiapi et al. 2011, Katsiapi et al. 2012). Despite the increased research eﬀorts, knowledge of diversity and distribution of Cyanobacteria in Greece is still inconsistent since no list on local cyanobacterial ﬂora has ever been published. Modern biodiversity research faces an ever increasing rate of data generation and eﬀorts are made to structure, aggregate, link and process these data in a meaningful way (Koureas et al. 2016). Currently, the Greek Taxon Information System (GTIS), an initiative launched by the LifeWatchGreece Research Infrastructure (ESFRI), is resuming eﬀorts to compile a database of all species reported from Greece; a prerequisite for GTIS is to have checklists for each taxonomic group (Bailly et al. 2016). The aim of this study is to create a checklist of cyanobacterial taxa inhabiting the area of Greece on the basis of the recent taxonomic revisions. Gkelis S et al. Gkelis S et al. 4 The taxonomic status of the recorded cyanobacterial taxa was checked and updated, where necessary, to the currently accepted taxonomically species using Komárek and Anagnostidis (1999), Komárek and Anagnostidis (2005), Komárek (2013), and the AlgaeBase (Guiry and Guiry 2016). The species were classiﬁed to suprageneric taxa according to the latest taxonomic revisions on the basis of polyphasic methods (Komárek et al. 2014, Komárek 2016). Not valid species, species of unclear taxonomic status, and species transferred to other phyla were excluded from the list. Also, infraspeciﬁc taxa that are not valid or need revision were not included in the list. The ﬁrst publication mentioning a taxon's occurrence in Greece and the original name used in that record is given under Notes and Nomeclature, respectively. Data resources The data collected were published in Global Biodiversity Information Facility (GBIF) through the Integrated Publishers Toolkit (IPT) according to the Darwin Core Archive biodiversity informatics data standard and are made publicly available through GBIF (UUID: 655027fc- 76ea-447a-a443-f36dd2e853d7, DOI: 10.15468/lkj0mr). Materials and methods The Checklist of Greek Cyanobacteria was created in the framework of the Greek Taxon Information System (GTIS), an initiative of the LifeWatchGreece Research Infrastructure (ESFRI) that has resumed eﬀorts to compile a complete checklist of all species reported from Greece (Bailly et al. 2016). In that publication, a methodology is described to produce Preliminary Checklists only. However, in the present case of Cyanobacteria, the status of the list for Greece was quite advanced, and the recent primary literature was exhaustively searched for this work; the present list is thus considered as an updated, annotated, and archived checklist. The data for this study were collected through extensive search in the scientiﬁc literature of the last 60 years. The search was performed in Scopus and Web of Science using the keywords: Greece AND cyanobacteria, Greece AND phytoplankton, Greece AND cyanoprokaryota, Greece AND blue-green algae. Furthermore, the two exhaustive monographs by Anagnostidis (Anagnostidis 1961, Anagnostidis 1968) were studied and the cyanobacterial taxa were recorded. All geographical areas of Greece and all biomes (freshwater, marine, terrestrial, caves etc.) and biotic forms (planktic, benthic, periphytic, endolithic, epizoic etc.) were considered. 4 Ammatoidea aegaea Anagnostidis & Pantazidou, 1991 Ammatoidea aegaea Anagnostidis & Pantazidou, 1991 Nomenclature: Ammatoidea aegea Notes: Anagnostidis and Pantazidou 1991b Anabaena cf. cylindrica Lemmermann, 1896 Nomenclature: Anabaena cf. cylindrica Notes: Gkelis et al. 2015b Anabaena inaequalis Bornet & Flahault, 1888 Nomenclature: Anabaena inaequalis Notes: Tafas and Economou-Amilli 1997 5 Cyanobacteria of Greece: an annotated checklist 5 Anabaena laxa A. Braun in Bornet & Falhault, 1888 Nomenclature: Anabaena laxa Notes: Anagnostidis 1961 Anabaena oscillarioides Bory ex Bornet & Flahault, 1888 Nomenclature: Anabaena oscillarioides Notes: Anagnostidis 1961 Anabaena perturbata H. Hill, 1976 Nomenclature: Anabaena pertubata Notes: Moustaka 1988 Anabaena planctonica Brunnthaler, 1903 Nomenclature: Anabaena solitaria f. planctonica Notes: Moustaka 1988 Anabaena sphaerica Bornet & Flahault, 1888 Nomenclature: Anabaena sphaerica Notes: Tafas and Economou-Amilli 1997 Anabaena torulosa Lagerheim ex Bornet & Flahault, 1888 Nomenclature: Anabaena torulosa Notes: Anagnostidis 1961 Anabaenopsis arnoldii Aptekar, 1926 Nomenclature: Anabaenopsis arnoldii Cyanobacteria of Greece: an annotated checklist Anabaena laxa A. Braun in Bornet & Falhault, 1888 Nomenclature: Anabaena laxa Notes: Anagnostidis 1961 Anabaena oscillarioides Bory ex Bornet & Flahault, 1888 Nomenclature: Anabaena oscillarioides Notes: Anagnostidis 1961 Anabaena perturbata H. Hill, 1976 Nomenclature: Anabaena pertubata Notes: Moustaka 1988 Anabaena planctonica Brunnthaler, 1903 Nomenclature: Anabaena solitaria f. planctonica Notes: Moustaka 1988 Anabaena sphaerica Bornet & Flahault, 1888 Nomenclature: Anabaena sphaerica Notes: Tafas and Economou-Amilli 1997 Anabaena torulosa Lagerheim ex Bornet & Flahault, 1888 Nomenclature: Anabaena torulosa Notes: Anagnostidis 1961 Anabaenopsis arnoldii Aptekar, 1926 Nomenclature: Anabaenopsis arnoldii Notes: Moustaka-Gouni et al. 2007 Anabaenopsis cf. circularis (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis cf. elenkinii f. circularis Notes: Moustaka 1988 Anabaenopsis circularis (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis cunningtonii Notes: Vardaka et al. 2005 Anabaenopsis elenkinii V. V. Miller, 1923 Nomenclature: Anabaenopsis elenkinii Notes: Vardaka et al. 2005 Anabaenopsis milleri Woronichin, 1929 Nomenclature: Anabaenopsis milleri Notes: Lanaras et al. 1989 Anabaenopsis tanganyikae (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis tanganyikae Notes: Hindak and Moustaka 1988 Anathece clathrata (W. & G. S. West) Komárek, Kastovsky & Jezberová, Nomenclature: Aphanothece clathrata Notes: Anagnostidis 1968 6 Gkelis S et al. Notes: Moustaka-Gouni et al. 2007 Anabaenopsis cf. circularis (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis cf. elenkinii f. circularis Notes: Moustaka 1988 Anabaenopsis circularis (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis cunningtonii Notes: Vardaka et al. 2005 Anabaenopsis elenkinii V. V. Miller, 1923 Nomenclature: Anabaenopsis elenkinii Notes: Vardaka et al. 2005 Anabaenopsis milleri Woronichin, 1929 Nomenclature: Anabaenopsis milleri Notes: Lanaras et al. 1989 Anabaenopsis tanganyikae (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis tanganyikae Notes: Hindak and Moustaka 1988 Anathece clathrata (W. & G. S. West) Komárek, Kastovsky & Jezberová, 20 Nomenclature: Aphanothece clathrata Notes: Anagnostidis 1968 6 Gkelis S et al. Gkelis S et al. 6 Cyanobacteria of Greece: an annotated checklist 7 Anathece endophytica (W. & G. S. West) Komárek, Kastovsky & Jez 2011 Nomenclature: Aphanothece nidulans var. endophytica Notes: Hindak and Moustaka 1988 Anathece minutissima (West) Komárek, Kastovsky & Jezberová, 2011 Nomenclature: Aphanothece minutissima Notes: Tryfon et al. 1997 Aphanizomenon favaloroi S.H.Otaño, 2012 Nomenclature: Aphanizomenon favaloroi Notes: Moustaka-Gouni et al. 2016 Aphanizomenon flos-aquae Ralfs ex Bornet & Flahault, 1886 Nomenclature: Aphanizomenon ﬂos-aquae Notes: Ananiadis 1956 Aphanizomenon gracile Lemmermann, 1907 Nomenclature: Aphanizomenon gracile Notes: Anagnostidis 1968 Aphanocapsa biformis A. Braun in P. Richter, 1879 Nomenclature: Aphanocapsa biformis Notes: Anagnostidis 1961 Aphanocapsa biformis A. Braun in P. Richter, 1879 Nomenclature: Aphanocapsa biformis Notes: Anagnostidis 1961 Gkelis S et al. 8 8 Aphanocapsa cf. holsatica (Lemmermann) G. Cronberg & Komárek, 1994 Nomenclature: Microcystis cf. holsatica Notes: Hindak and Moustaka 1988 Aphanocapsa conferta (W. & G. S. West) Komárková-Legnerová & Cronberg, 1994 Nomenclature: Aphanocapsa elachista var. conferta Notes: Tryfon et al. 1997 Aphanocapsa delicatissima W. & G. S. West, 1912 Nomenclature: Aphanocapsa delicatissima Notes: Moustaka-Gouni 1988 Aphanocapsa elachista W. & G. S. West, 1912 Nomenclature: Aphanocapsa elachista Notes: Anagnostidis 1968 Aphanocapsa fusco-lutea Hansgirg, 1892 Nomenclature: Aphanocapsa fusco-lutea Notes: Lamprinou et al. 2012 Aphanocapsa grevillei (Berkeley) Rabenhorst, 1865 Nomenclature: Aphanocapsa grevillei Notes: Anagnostidis 1961 9 Cyanobacteria of Greece: an annotated checklist Cyanobacteria of Greece: an annotated checklist 9 Aphanocapsa incerta (Lemmermann) G. Cronberg & Komárek, 1994 Nomenclature: Aphanocapsa incerta Notes: Tryfon et al. 1996 Aphanocapsa marina Hansgirg, 1892 Nomenclature: Aphanocapsa marina Notes: Anagnostidis 1968 Aphanocapsa muscicola (Meneghini) Wille, 1919 Nomenclature: Aphanocapsa muscicola Notes: Anagnostidis 1961 Aphanocapsa parietina (Nägeli ex Kützing) Nägeli, 1849 Nomenclature: Aphanocapsa parientina Notes: Lamprinou et al. 2012 Aphanocapsa raspaigellae (Hauck) Frémy in Feldmann, 1933 Nomenclature: Aphanocapsa raspaigellae Notes: Anagnostidis 1968 Aphanocapsa rivularis (Carmichael) Rabenhorst, 1865 Nomenclature: Aphanocapsa anodontae Notes: Anagnostidis 1968 Aphanocapsa salina Woronichin (Voronichin), 1929 Nomenclature: Aphanocapsa salina Gkelis S et al. 10 Notes: Anagnostidis and Golubic 1966 Aphanocapsa sesciacensis Frémy, 1928 Nomenclature: Aphanocapsa sescianensis Notes: Anagnostidis and Golubic 1966 Aphanocapsa thermalis Brügger, 1863 Nomenclature: Aphanocapsa thermalis Notes: Anagnostidis 1961 Aphanothece castagnei (Kützing) Rabenhorst, 1865 Nomenclature: Aphanothece castagnei Notes: Anagnostidis 1961 Aphanothece cf. rubra Liebetanz, 1926 Nomenclature: Aphanothece cf. rubra Notes: Lamprinou et al. 2012 Aphanothece cf. rufescens Hansgrig, 1892 Nomenclature: Aphanothece cf. rufescens Notes: Lamprinou et al. 2012 Aphanothece cf. salina Elenkin & Danilov, 1915 Nomenclature: Aphanothece cf. Notes: Moustaka-Gouni et al. 2007 Anabaenopsis cf. circularis (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis cf. elenkinii f. circularis Notes: Moustaka 1988 Anabaenopsis circularis (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis cunningtonii Notes: Vardaka et al. 2005 Anabaenopsis elenkinii V. V. Miller, 1923 Nomenclature: Anabaenopsis elenkinii Notes: Vardaka et al. 2005 Anabaenopsis milleri Woronichin, 1929 Nomenclature: Anabaenopsis milleri Notes: Lanaras et al. 1989 Anabaenopsis tanganyikae (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis tanganyikae Notes: Hindak and Moustaka 1988 Anathece clathrata (W. & G. S. West) Komárek, Kastovsky & Jezberová, 20 Nomenclature: Aphanothece clathrata Notes: Anagnostidis 1968 6 Gkelis S et al. salina Notes: Anagnostidis and Roussomoustakaki 1991 11 Cyanobacteria of Greece: an annotated checklist 11 Aphanothece marina (Ercegovic) Komárek & Anagnostidis, 1995 Nomenclature: Synechococcus marinus Notes: Anagnostidis 1968 Aphanothece microscopica Nägeli, 1849 Nomenclature: Aphanothece microscopica Notes: Anagnostidis 1968 Aphanothece microspora (Meneghini) Rabenhorst, 1863 Nomenclature: Aphanothece microspora Notes: Anagnostidis 1968 Aphanothece nidulans Richter in Wittrock & Nordstedt, 1884 Nomenclature: Aphanothece nidulans Notes: Anagnostidis 1961 Aphanothece pallida (Kützing) Rabenhorst, 1863 Nomenclature: Aphanothece pallida Notes: Lamprinou et al. 2012 Aphanothece saxicola Nägeli, 1849 Nomenclature: Aphanothece saxicola Notes: Anagnostidis 1968 Aphanothece stagnina (Sprengel) A. Braun in Rabenhorst, 1863 Nomenclature: Aphanothece stagnina Gkelis S et al. 12 Notes: Anagnostidis 1968 Notes: Anagnostidis 1968 Arthrospira fusiformis (Voronikhin) Komárek & J. W. G. Lun Nomenclature: Arthrospira fusiformis Notes: Moustaka-Gouni et al. 2007 Arthrospira platensis Gomont, 1892 Nomenclature: Arthrospira platensis Notes: Vardaka et al. 2005 Asterocapsa aerophytica F. Lederer, 2000 Nomenclature: Asterocapsa aerophytica Notes: Lamprinou et al. 2012 Asterocapsa divina Komárek, 1993 Nomenclature: Asterocapsa divina Notes: Lamprinou et al. 2012 Asterocapsa jilinica H. X. Xiao, 2000 Nomenclature: Asterocapsa jilinica Notes: Lamprinou et al. 2012 Asterocapsa sinica Liang & Chen, 1985 Nomenclature: Asterocapsa sinica Notes: Lamprinou et al. Notes: Moustaka-Gouni et al. 2007 Anabaenopsis cf. circularis (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis cf. elenkinii f. circularis Notes: Moustaka 1988 Anabaenopsis circularis (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis cunningtonii Notes: Vardaka et al. 2005 Anabaenopsis elenkinii V. V. Miller, 1923 Nomenclature: Anabaenopsis elenkinii Notes: Vardaka et al. 2005 Anabaenopsis milleri Woronichin, 1929 Nomenclature: Anabaenopsis milleri Notes: Lanaras et al. 1989 Anabaenopsis tanganyikae (G. S. West) Woloszynska & V.Miller, 1923 Nomenclature: Anabaenopsis tanganyikae Notes: Hindak and Moustaka 1988 Anathece clathrata (W. & G. S. West) Komárek, Kastovsky & Jezberová, 20 Nomenclature: Aphanothece clathrata Notes: Anagnostidis 1968 6 Gkelis S et al. 2012 Cyanobacteria of Greece: an annotated checklist 13 Blennothrix brebissonii (Kützing ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Hydrocoleum brebissonii Notes: Anagnostidis 1968 Blennothrix glutinosa (Gomont ex Gomont) Anagnostidis & Komárek, 2001 Nomenclature: Hydrocoleum glutinosum Notes: Anagnostidis 1968 Blennothrix heterotricha (Gomont ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Hydrocoleum heterotrichum Notes: Anagnostidis 1968 Blennothrix lyngbyacea (Kützing ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Hydrocoleum lyngbyaceum Notes: Anagnostidis and Golubic 1966 Borzia periklei Anagnostidis in Anagnostidis & Komárek, 1988 Nomenclature: Borzia trilocularis Notes: Anagnostidis 1977 Borzia starkii Schiller, 1954 Nomenclature: Borzia starkii Notes: Anagnostidis 1977 Borzia susedana Ercegovic, 1925 Nomenclature: Borzia susedana Blennothrix brebissonii (Kützing ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Hydrocoleum brebissonii Notes: Anagnostidis 1968 Blennothrix glutinosa (Gomont ex Gomont) Anagnostidis & Komárek, 2001 Nomenclature: Hydrocoleum glutinosum Notes: Anagnostidis 1968 Blennothrix heterotricha (Gomont ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Hydrocoleum heterotrichum Notes: Anagnostidis 1968 Blennothrix lyngbyacea (Kützing ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Hydrocoleum lyngbyaceum Notes: Anagnostidis and Golubic 1966 Borzia periklei Anagnostidis in Anagnostidis & Komárek, 1988 Nomenclature: Borzia trilocularis Notes: Anagnostidis 1977 Borzia starkii Schiller, 1954 Nomenclature: Borzia starkii Notes: Anagnostidis 1977 Borzia susedana Ercegovic, 1925 Nomenclature: Borzia susedana Notes: Anagnostidis 1977 Borzia trilocularis Cohn ex Gomont, 1892 Nomenclature: Borzia trilocularis Notes: Anagnostidis 1968 Brachytrichia quoyi Bornet & Flahault, 1886 Nomenclature: Brachytrichia quiyi Notes: Anagnostidis 1968 Calothrix aeruginea Thuret ex Bornet & Flahault, 1886 Nomenclature: Calothrix aeruginea Notes: Anagnostidis 1968 Calothrix braunii Bornet & Flahault, 1886 Nomenclature: Calothrix braunii Notes: Anagnostidis 1968 Calothrix cf. fusca Bornet & Flahault, 1886 Nomenclature: Calothrix cf. fusca Notes: Lamprinou et al. 2012 Calothrix confervicola C. Agardh ex Bornet & Flahault, 188 Nomenclature: Calothrix confervicola Notes: Anagnostidis 1968 14 Gkelis S et al. Notes: Anagnostidis 1977 Borzia trilocularis Cohn ex Gomont, 1892 Nomenclature: Borzia trilocularis Notes: Anagnostidis 1968 Brachytrichia quoyi Bornet & Flahault, 1886 Nomenclature: Brachytrichia quiyi Notes: Anagnostidis 1968 Calothrix aeruginea Thuret ex Bornet & Flahault, 1886 Nomenclature: Calothrix aeruginea Notes: Anagnostidis 1968 Calothrix braunii Bornet & Flahault, 1886 Nomenclature: Calothrix braunii Notes: Anagnostidis 1968 Calothrix cf. fusca Bornet & Flahault, 1886 Nomenclature: Calothrix cf. fusca Notes: Lamprinou et al. 2012 Calothrix confervicola C. Agardh ex Bornet & Flahault, 188 Nomenclature: Calothrix confervicola Notes: Anagnostidis 1968 14 Gkelis S et al. Gkelis S et al. 14 Calothrix confervicola C. Agardh ex Bornet & Flahault, 1886 Nomenclature: Calothrix confervicola Notes: Anagnostidis 1968 15 15 Cyanobacteria of Greece: an annotated checklist Calothrix contarenii Bornet & Flahault, 1886 Nomenclature: Calothrix contarenii Notes: Anagnostidis 1968 Calothrix javanica De Wildeman, 1897 Nomenclature: Calothrix javanica Notes: Anagnostidis et al. 1981 Calothrix marchica Lemmermann, 1914 Nomenclature: Calothrix marchica Notes: Anagnostidis et al. 1981 Calothrix parietina Thuret ex Bornet & Flahault, 1886 Nomenclature: Calothrix parietina Notes: Anagnostidis 1968 Calothrix pulvinata C. Agardh ex Bornet & Flahault, 1886 Nomenclature: Calothrix pulvinata Notes: Anagnostidis and Golubic 1966 Calothrix scopulorum C. Agardh ex Bornet & Flahault, 1886 Nomenclature: Calothrix scopulorum Notes: Anagnostidis and Golubic 1966 Calothrix stagnalis Gomont, 1895 Nomenclature: Calothrix stagnalis Calothrix contarenii Bornet & Flahault, 1886 Nomenclature: Calothrix contarenii Notes: Anagnostidis 1968 Calothrix javanica De Wildeman, 1897 Nomenclature: Calothrix javanica Notes: Anagnostidis et al. 1981 Calothrix marchica Lemmermann, 1914 Nomenclature: Calothrix marchica Notes: Anagnostidis et al. 1981 Calothrix parietina Thuret ex Bornet & Flahault, 1886 Nomenclature: Calothrix parietina Notes: Anagnostidis 1968 Calothrix pulvinata C. Agardh ex Bornet & Flahault, 1886 Nomenclature: Calothrix pulvinata Notes: Anagnostidis and Golubic 1966 Calothrix scopulorum C. Agardh ex Bornet & Flahault, 1886 Nomenclature: Calothrix scopulorum Notes: Anagnostidis and Golubic 1966 Calothrix stagnalis Gomont, 1895 Nomenclature: Calothrix stagnalis Calothrix parietina Thuret ex Bornet & Flahault, 1886 Nomenclature: Calothrix parietina Notes: Anagnostidis 1968 Notes: Anagnostidis et al. 1981 Calothrix thermalis Hasngirg ex Bornet & Flahault, 1886 Nomenclature: Calothrix thermalis Notes: Anagnostidis 1968 Chamaesiphon confervicola A. Braun in Rabenhorst, 1865 Nomenclature: Chamaesiphon curvatus Notes: Anagnostidis 1968 Chamaesiphon geitleri A. Braun in Rabenhorst, 1865 Nomenclature: Chamaesiphon geitleri Notes: Anagnostidis 1968 Chamaesiphon incrustans Grunow in Rabenhorst, 1865 Nomenclature: Chamaesiphon incrustans Notes: Anagnostidis 1968 Chamaesiphon polonicus (Rostafinski) Hansgirg, 1893 Nomenclature: Chamaesiphon polonicus Notes: Anagnostidis 1968 Chamaesiphon polymorphus Geitler, 1925 Nomenclature: Chamaesiphon polymorphus Notes: Anagnostidis 1968 16 Gkelis S et al. Gkelis S et al. 16 17 Cyanobacteria of Greece: an annotated checklist 17 Chlorogloea microcystoides Geitler, 1926 Nomenclature: Chlorogloea microcystoides Notes: Anagnostidis 1961 Chlorogloea novacekii Komárek & Montejano, 1994 Nomenclature: Chlorogloea novacekii Notes: Lamprinou et al. 2012 Chlorogloea rivularis (Hansgirg) Komárek & Anagnostidis, 1995 Nomenclature: Xenococcus rivularis Notes: Anagnostidis 1968 Chroococcidiopsis doonensis R. B. Singh, 1968 Nomenclature: Chroococcidiopsis doonensis Notes: Lamprinou et al. 2009 Chroococcidiopsis kashayi Friedmann, 1961 Nomenclature: Chroococcidiopsis kashaii Notes: Lamprinou et al. 2012 Chroococcidiopsis thermalis Geitler, 1933 Nomenclature: Chroococcidiopsis thermalis Notes: Anagnostidis and Pantazidou 1988a Chroococcidium gelatinosum Geitler, 1933 Nomenclature: Chroococcidium gelatinosum Gkelis S et al. 18 Notes: Metaxatos et al. 2003 Chroococcus cf. aphanocapsoides Skuja ex Joosten, 2006 Nomenclature: Chroococcus cf. aphanocapsoides Notes: Lamprinou et al. 2012 Chroococcus cf. subsphaericus N. L. Gardner, 1927 Nomenclature: Chroococcus cf. subsphaericus Notes: Lamprinou et al. 2012 Chroococcus cohaerens (Brébisson) Nägeli, 1849 Nomenclature: Chroococcus cohaerens Notes: Lamprinou et al. 2009 Chroococcus dispersus (Keissler) Lemmermann, 1904 Nomenclature: Chroococcus dispersus Notes: Anagnostidis 1968 Chroococcus helveticus Nägeli, 1849 Nomenclature: Chroococcus helveticus Notes: Anagnostidis et al. 1981 Chroococcus lithophilus Ercegovic, 1925 Nomenclature: Chroococcus lithophilus Notes: Lamprinou et al. 2012 Notes: Metaxatos et al. 2003 Chroococcus cf. aphanocapsoides Skuja ex Joosten, 2006 Nomenclature: Chroococcus cf. aphanocapsoides Notes: Lamprinou et al. 2012 Chroococcus cf. subsphaericus N. L. Gardner, 1927 Nomenclature: Chroococcus cf. subsphaericus Notes: Lamprinou et al. 2012 Chroococcus cohaerens (Brébisson) Nägeli, 1849 Nomenclature: Chroococcus cohaerens Notes: Lamprinou et al. 2009 Chroococcus dispersus (Keissler) Lemmermann, 1904 Nomenclature: Chroococcus dispersus Notes: Anagnostidis 1968 Chroococcus helveticus Nägeli, 1849 Nomenclature: Chroococcus helveticus Notes: Anagnostidis et al. 1981 Chroococcus lithophilus Ercegovic, 1925 Nomenclature: Chroococcus lithophilus Notes: Lamprinou et al. 2012 19 Cyanobacteria of Greece: an annotated checklist Cyanobacteria of Greece: an annotated checklist Chroococcus microscopicus Komárková-Legnerová & G. Cronberg, 1994 Nomenclature: Chroococcus microscopicus Notes: Tryfon et al. 1997 Chroococcus minor (Kützing) Nägeli, 1849 Nomenclature: Chroococcus minor Notes: Anagnostidis 1961 Chroococcus minutus (Kützing) Nägeli, 1849 Nomenclature: Chroococcus minutus Notes: Anagnostidis 1961 Chroococcus spelaeus Ercegovic, 1925 Nomenclature: Chroococcus spelaeus Notes: Lamprinou et al. 2009 Chroococcus subnudus (Hansgirg) G. Cronberg & J. Komárek, 1994 Nomenclature: Chroococcus subnudus Notes: Lamprinou et al. 2012 Chroococcus tenax (Kirchner) Hieronymus, 1892 Nomenclature: Chroococcus tenax Notes: Lamprinou et al. 2009 Chroococcus thermalis (Meneghini) Nägeli, 1849 Nomenclature: Chroococcus thermalis Cyanobacteria of Greece: an annotated checklist Chroococcus microscopicus Komárková-Legnerová & G. Cronberg, 1 Nomenclature: Chroococcus microscopicus Notes: Tryfon et al. 1997 Chroococcus minor (Kützing) Nägeli, 1849 Nomenclature: Chroococcus minor Notes: Anagnostidis 1961 Chroococcus minutus (Kützing) Nägeli, 1849 Nomenclature: Chroococcus minutus Notes: Anagnostidis 1961 Chroococcus spelaeus Ercegovic, 1925 Nomenclature: Chroococcus spelaeus Notes: Lamprinou et al. 2009 Chroococcus subnudus (Hansgirg) G. Cronberg & J. Komárek, 1994 Nomenclature: Chroococcus subnudus Notes: Lamprinou et al. 2012 Chroococcus tenax (Kirchner) Hieronymus, 1892 Nomenclature: Chroococcus tenax Notes: Lamprinou et al. 2009 Chroococcus thermalis (Meneghini) Nägeli, 1849 Nomenclature: Chroococcus thermalis Notes: Radea et al. 2010 Chroococcus turgidus (Kützing) Nägeli, 1849 Nomenclature: Chroococcus turgidus Notes: Anagnostidis 1961 Chroococcus turicensis (Nägeli) Hansgirg, 1887 Nomenclature: Chroococcus turicensis Notes: Lamprinou et al. 2012 Chroococcus varius A. Braun in Rabenhorst, 1876 Nomenclature: Chroococcus varius Notes: Anagnostidis 1968 Chroococcus westii Boye-Petersen, 1923 Nomenclature: Chroococcus westii Notes: Lamprinou et al. 2012 Chrysosporum bergii (Ostenfeld) E. Zapomelová, O. Skácelo Kopp & E. Janecek, 2012 Nomenclature: Anabaena bergii Notes: Katsiapi et al. 2011 Chrysosporum ovalisporum (Forti) E. Zapomelová, O. Skácelo Kopp & E. Janecek, 2012 Nomenclature: Aphanizomenon ovalisporum Notes: Gkelis et al. 2005a 20 Gkelis S et al. Notes: Radea et al. 2010 Chroococcus turgidus (Kützing) Nägeli, 1849 Nomenclature: Chroococcus turgidus Notes: Anagnostidis 1961 Chroococcus turicensis (Nägeli) Hansgirg, 1887 Nomenclature: Chroococcus turicensis Notes: Lamprinou et al. 2012 Chroococcus varius A. Braun in Rabenhorst, 1876 Nomenclature: Chroococcus varius Notes: Anagnostidis 1968 Chroococcus westii Boye-Petersen, 1923 Nomenclature: Chroococcus westii Notes: Lamprinou et al. 2012 Chrysosporum bergii (Ostenfeld) E. Zapomelová, O. Skácelová, P. Pumann, R Kopp & E. Janecek, 2012 Nomenclature: Anabaena bergii Notes: Katsiapi et al. 2011 Chrysosporum ovalisporum (Forti) E. Zapomelová, O. Skácelová, P. Pumann, R Kopp & E. Janecek, 2012 Nomenclature: Aphanizomenon ovalisporum Notes: Gkelis et al. 2005a 20 Gkelis S et al. Gkelis S et al. 20 Notes: Radea et al. 2010 21 Cyanobacteria of Greece: an annotated checklist Cyanobacteria of Greece: an annotated checklist Clastidium rivulare (Hansgirg) Hansgirg, 1892 Nomenclature: Clastidium rivulare Notes: Anagnostidis 1968 Clastidium setigerum O. Kirchner, 1880 Nomenclature: Clastidium setigerum Notes: Anagnostidis 1968 Coelosphaerium kützingianum Nägeli, 1849 Nomenclature: Coelosphaerium minutissimum Notes: Anagnostidis 1968 Coleofasciculus chthonoplastes (Thuret ex Gomont) M. Siegesmund, J. R. Johansen & T. Friedl in Siegesmund et al. 2008 Nomenclature: Microcoleus chtonoplastes Notes: Anagnostidis 1968 Cuspidothrix issatschenkoi (Usachev) P. Rajaniemi, Komárek, R. Willame, P. Hrouzek, K. Kastovská, L. Hoffmann & K. Sivonen, 2005 Nomenclature: Aphanizomenon issatschenkoi Notes: Moustaka-Gouni 1988 Cyanobacterium cedrorum (Sauvageau) Komárek, Kopecky & Cepák, 1999 Nomenclature: Synechococcus cedrorum Notes: Anagnostidis 1968 Clastidium rivulare (Hansgirg) Hansgirg, 1892 Nomenclature: Clastidium rivulare Notes: Anagnostidis 1968 Clastidium setigerum O. Kirchner, 1880 Nomenclature: Clastidium setigerum Notes: Anagnostidis 1968 Coelosphaerium kützingianum Nägeli, 1849 Nomenclature: Coelosphaerium minutissimum Notes: Anagnostidis 1968 Coleofasciculus chthonoplastes (Thuret ex Gomont) M. Siegesmund, J. R. Johansen & T. Friedl in Siegesmund et al. 2008 Nomenclature: Microcoleus chtonoplastes Notes: Anagnostidis 1968 Cuspidothrix issatschenkoi (Usachev) P. Rajaniemi, Komárek, R. Willame, P. Hrouzek, K. Kastovská, L. Hoffmann & K. Sivonen, 2005 Nomenclature: Aphanizomenon issatschenkoi Notes: Moustaka-Gouni 1988 Cyanobacterium cedrorum (Sauvageau) Komárek, Kopecky & Cepák, 1999 Nomenclature: Synechococcus cedrorum Notes: Anagnostidis 1968 Clastidium rivulare (Hansgirg) Hansgirg, 1892 Nomenclature: Clastidium rivulare Notes: Anagnostidis 1968 Clastidium setigerum O. Kirchner, 1880 Nomenclature: Clastidium setigerum Notes: Anagnostidis 1968 Coelosphaerium kützingianum Nägeli, 1849 Nomenclature: Coelosphaerium minutissimum Notes: Anagnostidis 1968 Coleofasciculus chthonoplastes (Thuret ex Gomont) M. Siegesmund, J. R. Johansen & T. Friedl in Siegesmund et al. 2008 Nomenclature: Microcoleus chtonoplastes Notes: Anagnostidis 1968 Cuspidothrix issatschenkoi (Usachev) P. Rajaniemi, Komárek, R. Willame, P. Hrouzek, K. Kastovská, L. Hoffmann & K. Sivonen, 2005 Nomenclature: Aphanizomenon issatschenkoi Notes: Moustaka-Gouni 1988 Cyanobacterium cedrorum (Sauvageau) Komárek, Kopecky & Cepák, 1999 Nomenclature: Synechococcus cedrorum Notes: Anagnostidis 1968 Gkelis S et al. 22 Cyanobacterium minervae (J. J. Copeland) Komárek, Kopecky & Cepák, 1 Nomenclature: Synechococcus minervae Notes: Anagnostidis 1961 Cyanocatena planctonica Hindák, 1975 Nomenclature: Cyanocatena planctonica Notes: Moustaka-Gouni 1988 Cyanodictyon imperfectum Cronberg & Weibull, 1981 Nomenclature: Cyanodictyon imperfectum Notes: Moustaka-Gouni and Nikolaidis 1990 Cyanodictyon planctonicum B. A. Mayer, 1994 Nomenclature: Cyanodictyon planctonicum Notes: Tryfon et al. 1996 Cyanodictyon reticulatum (Lemmermann) Geitler, 1925 Nomenclature: Cyanodictyon reticulatum Notes: Danielidis et al. 1996 Cyanogranis ferruginea (F. Wawrik) Hindák ex Hindák in Joosen, 2006 Nomenclature: Cyanogranis ferruginea Notes: Moustaka-Gouni 1988 Cyanosaccus aegaeus Anagnostidis & Pantazidou, 1985 Nomenclature: Cyanosaccus aegaeus 22 Gkelis S et al. Notes: Anagnostidis and Pantazidou 1985 Cyanosaccus atticus Anagnostidis & Pantazidou, 1988 Nomenclature: Cyanosaccus atticus Notes: Anagnostidis and Pantazidou 1988b Cyanosarcina cf. burmensis (Skuja) Kovácik, 1988 Nomenclature: Cyanosarcina cf. burmensis Notes: Lamprinou et al. 2012 Cyanosarcina cf. spectabilis (Geitler) Kovácik, 1988 Nomenclature: Cyanosarcina cf. spectabilis Notes: Lamprinou et al. 2012 Cyanosarcina parthenonensis Anagnostidis in Anagnostidis & Pantazidou, 1991 Nomenclature: Cyanosarcina parthenonensis Notes: Lamprinou et al. 2012 Cyanosarcina thalassia Anagnostidis & Pantazidou, 1991 Nomenclature: Cyanosarcina thalassia Notes: Anagnostidis and Pantazidou 1991a Cyanostylon microcystoides Geitler, 1928 Nomenclature: Cyanostylon microcystoides Notes: Anagnostidis 1968 Cyanobacteria of Greece: an annotated checklist 23 23 Cyanobacteria of Greece: an annotated checklist Gkelis S et al. 24 Cyanostylon plancticum Hindák, 1988 Nomenclature: Cyanostylon plancticum Notes: Hindák 1988 Cyanothece aeruginosa (Nägeli) Komárek 1976 Nomenclature: Synechococcus aeruginosus Notes: Anagnostidis 1968 Cyanothece halobia Roussomoustakaki & Anagnostidis 1991 Nomenclature: Cyanothece halobia Notes: Anagnostidis and Roussomoustakaki 1991 Cyanothece major (Schröter) Komárek, 1976 Nomenclature: Synechococcus maior Notes: Anagnostidis 1968 Cylindrospermopsis raciborskii (Woloszynska) Seenayya & Subba Ra Desikachary, 1972 Nomenclature: Cylindrospermopsis raciborskii Notes: Moustaka-Gouni 1988 Cylindrospermum licheniforme Kützing ex Bornet & Flahault, 1886 Nomenclature: Cylindrospermum licheniforme Notes: Economou-Amilli et al. 1984 25 Cyanobacteria of Greece: an annotated checklist Cylindrospermum muscicola Kützing ex Bornet & Flahault, 1886 Nomenclature: Cylindrospermum muscicola Notes: Anagnostidis 1968 Dactylococcopsis echini Rosenvinge in Mortensen & Rosenvinge, 1934 Nomenclature: Dactylococcopsis echini Notes: Anagnostidis 1968 Dactylococcopsis rhaphidioides Hansgirg, 1888 Nomenclature: Dactylococcopsis rhaphidioides Notes: Anagnostidis 1968 Dermocarpella prasina (Reinsch) Komárek & Anagnostidis, 1995 Nomenclature: Dermocarpa prasina Notes: Anagnostidis 1968 Dichothrix compacta Bornet & Flahault, 1886 Nomenclature: Dichothrix compacta Notes: Anagnostidis 1968 Dichothrix gypsophila Bornet & Flahault, 1886 Nomenclature: Dichothrix gypsophila Notes: Anagnostidis 1968 Dolichospermum affine (Lemmermann) Wacklin, L. Hoffmann & Komárek Nomenclature: Anabaena aﬃnis Cyanobacteria of Greece: an annotated checklist Cylindrospermum muscicola Kützing ex Bornet & Flahault, 1886 Nomenclature: Cylindrospermum muscicola Notes: Anagnostidis 1968 Dactylococcopsis echini Rosenvinge in Mortensen & Rosenvinge, 1934 Nomenclature: Dactylococcopsis echini Notes: Anagnostidis 1968 Dactylococcopsis rhaphidioides Hansgirg, 1888 Nomenclature: Dactylococcopsis rhaphidioides Notes: Anagnostidis 1968 Dermocarpella prasina (Reinsch) Komárek & Anagnostidis, 1995 Nomenclature: Dermocarpa prasina Notes: Anagnostidis 1968 Dichothrix compacta Bornet & Flahault, 1886 Nomenclature: Dichothrix compacta Notes: Anagnostidis 1968 Dichothrix gypsophila Bornet & Flahault, 1886 Nomenclature: Dichothrix gypsophila Notes: Anagnostidis 1968 Dolichospermum affine (Lemmermann) Wacklin, L. Hoffmann & Komá Nomenclature: Anabaena aﬃnis Notes: Moustaka-Gouni and Nikolaidis 1990 Dolichospermum crassum (Lemmermann) Wacklin, L. Hoffmann & Komárek, 2009 Nomenclature: Anabaena crassa Notes: Gkelis et al. 2014 Dolichospermum flos-aquae (Brébisson ex Bornet & Flahault) P. Wacklin, L. Hoffmann & J. Komárek, 2009 Nomenclature: Anabaena ﬂos-aquae Notes: Anagnostidis 1968 Dolichospermum lemmermannii (Richter) P. Wacklin, L. Hoffmann & J. Komárek, 2009 Nomenclature: Anabaena lemmermannii var. minor Notes: Tryfon et al. 1997 Dolichospermum mendotae (W. Trelease) Wacklin, L. Hoffmann & Komárek, 2009 Nomenclature: Anabaena mendotae Notes: Moustaka-Gouni et al. 2000 Dolichospermum mucosum (Komárková-Legnerová & Eloranta) Wacklin, L. Hoffmann & Komárek, 2009 Nomenclature: Anabaena mucosa Notes: Tryfon et al. 1997 26 Gkelis S et al. Gkelis S et al. 26 Notes: Moustaka-Gouni and Nikolaidis 1990 Dolichospermum crassum (Lemmermann) Wacklin, L. Hoffmann & Komárek, 2009 Nomenclature: Anabaena crassa Notes: Gkelis et al. 2014 Dolichospermum flos-aquae (Brébisson ex Bornet & Flahault) P. Wacklin, L. Hoffmann & J. Komárek, 2009 Nomenclature: Anabaena ﬂos-aquae Notes: Anagnostidis 1968 Dolichospermum lemmermannii (Richter) P. Wacklin, L. Hoffmann & J. Komárek, 2009 Nomenclature: Anabaena lemmermannii var. minor Notes: Tryfon et al. 1997 Dolichospermum mendotae (W. Trelease) Wacklin, L. Hoffmann & Komárek, 2009 Nomenclature: Anabaena mendotae Notes: Moustaka-Gouni et al. 2000 Dolichospermum mucosum (Komárková-Legnerová & Eloranta) Wacklin, L. Hoffmann & Komárek, 2009 Nomenclature: Anabaena mucosa Notes: Tryfon et al. 1997 Notes: Moustaka-Gouni and Nikolaidis 1990 Cyanobacteria of Greece: an annotated checklist 27 Dolichospermum scheremetievii (Elenkin) Wacklin, L. Hoffmann & Komárek, 2009 Nomenclature: Anabaena scheremetievii Notes: Anagnostidis 1968 Dolichospermum sigmoideum (Nygaard) Wacklin, L. Hoffmann & Komárek, 2009 Nomenclature: Anabaena circinalis Notes: Moustaka-Gouni 1988 Dolichospermum solitarium (Klebahn) Wacklin, L. Hoffmann & Komárek, 2009 Nomenclature: Anabaena solitaria Notes: Hindák 1993 Dolichospermum spiroides (Klebhan) Wacklin, L. Hoffmann & Komárek, 2009 Nomenclature: Anabaena spiroides Notes: Anagnostidis 1968 Dolichospermum viguieri (Denis & Frémy) Wacklin, L. Hoffmann & Komárek, 2009 Nomenclature: Anabaena viguieri Notes: Lanaras et al. 1989 Entophysalis deusta (Meneghini) F. E. Drouet & W. A. Cylindrospermum muscicola Kützing ex Bornet & Flahault, 1886 Nomenclature: Cylindrospermum muscicola Notes: Anagnostidis 1968 Dactylococcopsis echini Rosenvinge in Mortensen & Rosenvinge, 1934 Nomenclature: Dactylococcopsis echini Notes: Anagnostidis 1968 Dactylococcopsis rhaphidioides Hansgirg, 1888 Nomenclature: Dactylococcopsis rhaphidioides Notes: Anagnostidis 1968 Dermocarpella prasina (Reinsch) Komárek & Anagnostidis, 1995 Nomenclature: Dermocarpa prasina Notes: Anagnostidis 1968 Dichothrix compacta Bornet & Flahault, 1886 Nomenclature: Dichothrix compacta Notes: Anagnostidis 1968 Dichothrix gypsophila Bornet & Flahault, 1886 Nomenclature: Dichothrix gypsophila Notes: Anagnostidis 1968 Dolichospermum affine (Lemmermann) Wacklin, L. Hoffmann & Komárek Nomenclature: Anabaena aﬃnis Cyanobacteria of Greece: an annotated checklist Daily 1948: Nomenclature: Entophysalis deusta Notes: Anagnostidis 1968 Gkelis S et al. 28 Entophysalis granulosa Kützing 1843 Nomenclature: Entophysalis granulosa Notes: Anagnostidis and Golubic 1966 Epigloeosphaera cf. glebulenta (Zalessky) J. Komárková-Legnerová, 1991 Nomenclature: Epigloeosphaera cf. grebulenta Notes: Danielidis et al. 1996 Eucapsis cf. terrestris M. Akiyama, 1965 Nomenclature: Eucapsis cf. terrestris Notes: Lamprinou et al. 2012 Eucapsis minor (Skuja) Elenkin, 1933 Nomenclature: Eucapsis minor Notes: Anagnostidis and Economou-Amilli 1978 Geitlerinema acutissimum (Kufferath) Anagnostidis, 1989 Nomenclature: Oscillatoria acutissima Notes: Anagnostidis and Golubic 1966 Geitlerinema amphibium (C. Agardh ex Gomont) Anagnostidis, 1989 Nomenclature: Oscillatoria amphibia f. circinata Notes: Anagnostidis 1961 Geitlerinema apolloniae Anagnostidis, 2001 Nomenclature: Oscillatoria splendida f. major 28 Gkelis S et al. Entophysalis granulosa Kützing 1843 Nomenclature: Entophysalis granulosa Notes: Anagnostidis and Golubic 1966 Epigloeosphaera cf. glebulenta (Zalessky) J. Komárková-Legnerová, 1991 Nomenclature: Epigloeosphaera cf. grebulenta Notes: Danielidis et al. 1996 Eucapsis cf. terrestris M. Akiyama, 1965 Nomenclature: Eucapsis cf. terrestris Notes: Lamprinou et al. 2012 Eucapsis minor (Skuja) Elenkin, 1933 Nomenclature: Eucapsis minor Notes: Anagnostidis and Economou-Amilli 1978 Geitlerinema acutissimum (Kufferath) Anagnostidis, 1989 Nomenclature: Oscillatoria acutissima Notes: Anagnostidis and Golubic 1966 Geitlerinema amphibium (C. Agardh ex Gomont) Anagnostidis, 1989 Nomenclature: Oscillatoria amphibia f. circinata Notes: Anagnostidis 1961 Geitlerinema apolloniae Anagnostidis, 2001 29 Cyanobacteria of Greece: an annotated checklist Notes: Anagnostidis 1961 Geitlerinema cf. lemmermannii (Woloszynska) Anagnostidis, 1989 Nomenclature: Oscillatoria cf. lemmermannii Notes: Hindak and Moustaka 1988 Geitlerinema ionicum (Skuja) Anagnostidis, 1989 Nomenclature: Oscillatoria ionica Notes: Anagnostidis et al. 1981 Geitlerinema lemmermannii (Woloszynska) Anagnostidis, 1989 Nomenclature: Jaaginema cf. lemmermannii Notes: Moustaka-Gouni 1993 Geitlerinema numidicum (Gomont) Anagnostidis, 1989 Nomenclature: Oscillatoria numidica Notes: Anagnostidis 1961 Geitlerinema splendidum (Greville ex Gomont) Anagnostidis, 1989 Nomenclature: Oscillatoria splendida Notes: Anagnostidis and Golubic 1966 Geitlerinema sulphureum (Strzeszewski) Anagnostidis 2001 Nomenclature: Oscillatoria geminata f. sulphurea Notes: Anagnostidis 1961 y Notes: Anagnostidis 1961 Gkelis S et al. 30 Glaucospira laxissima (G.S.West) Simic, Komárek & Dordevic, 2014 Nomenclature: Spirulina laxissima Notes: Anagnostidis 1961 Gloeobacter violaceus Rippka, J. B. Waterbury & Cohen-Bazire 1974 Nomenclature: Aphanothece caldariorum Notes: Anagnostidis 1961 Gloeocapsa aeruginosa Kützing, 1843 Nomenclature: Gloeocapsa aeruginosa Notes: Lamprinou et al. 2014 Gloeocapsa arenaria (Hassall) Rabenhorst, 1865 Nomenclature: Gloeocapsa arenaria Notes: Anagnostidis 1961 Gloeocapsa atrata Kützing, 1843 Nomenclature: Gloeocapsa montana Notes: Economou-Amilli et al. 1984 Gloeocapsa biformis Ercegovic, 1925 Nomenclature: Gloeocapsa biformis Notes: Anagnostidis 1968 Gloeocapsa bituminosa (Bory) Kützing, 1849 Nomenclature: Gloeocapsa bituminosa 30 Gkelis S et al. Glaucospira laxissima (G.S.West) Simic, Komárek & Dordevic, 2014 Nomenclature: Spirulina laxissima Notes: Anagnostidis 1961 Gloeobacter violaceus Rippka, J. B. Waterbury & Cohen-Bazire 1974 Nomenclature: Aphanothece caldariorum Notes: Anagnostidis 1961 Gloeocapsa aeruginosa Kützing, 1843 Nomenclature: Gloeocapsa aeruginosa Notes: Lamprinou et al. 2014 Gloeocapsa arenaria (Hassall) Rabenhorst, 1865 Nomenclature: Gloeocapsa arenaria Notes: Anagnostidis 1961 Gloeocapsa atrata Kützing, 1843 Nomenclature: Gloeocapsa montana Notes: Economou-Amilli et al. 1984 Gloeocapsa biformis Ercegovic, 1925 Nomenclature: Gloeocapsa biformis Notes: Anagnostidis 1968 Gloeocapsa bituminosa (Bory) Kützing, 1849 Nomenclature: Gloeocapsa bit minosa 31 Cyanobacteria of Greece: an annotated checklist 31 Notes: Lamprinou et al. 2012 Gloeocapsa caldariorum Rabenhorst, 1865 Nomenclature: Gloeocapsa caldariorum Notes: Lamprinou et al. 2012 Gloeocapsa compacta Kützing, 1847 Nomenclature: Gloeocapsa compacta Notes: Anagnostidis 1968 Gloeocapsa decorticans (A. Braun) P.Richter in Wille, 1925 Nomenclature: Gloeocapsa decorticans Notes: Anagnostidis et al. 1983 Gloeocapsa gelatinosa Kützing, 1843 Nomenclature: Gloeocapsa gelatinosa Notes: Anagnostidis 1961 Gloeocapsa kuetzingiana Nägeli ex Kützing, 1849 Nomenclature: Gloeocapsa kützingiana Notes: Anagnostidis 1968 Gloeocapsa punctata Nägeli, 1849 Nomenclature: Gloeocapsa punctata Notes: Lamprinou et al. 2012 Notes: Lamprinou et al. 2012 Gloeocapsa caldariorum Rabenhorst, 1865 Nomenclature: Gloeocapsa caldariorum Notes: Lamprinou et al. 2012 Gloeocapsa compacta Kützing, 1847 Nomenclature: Gloeocapsa compacta Notes: Anagnostidis 1968 Gloeocapsa decorticans (A. Braun) P.Richter in Wille, 1925 Nomenclature: Gloeocapsa decorticans Notes: Anagnostidis et al. 1983 Gloeocapsa gelatinosa Kützing, 1843 Nomenclature: Gloeocapsa gelatinosa Notes: Anagnostidis 1961 Gloeocapsa kuetzingiana Nägeli ex Kützing, 1849 Nomenclature: Gloeocapsa kützingiana Notes: Anagnostidis 1968 Gloeocapsa punctata Nägeli, 1849 Nomenclature: Gloeocapsa punctata Notes: Lamprinou et al. 2012 Notes: Lamprinou et al. 2012 Gkelis S et al. 32 Gloeocapsa rupestris Kützing, 1847 Nomenclature: Gloeocapsa rupestris Notes: Anagnostidis 1961 Gloeocapsa sanguinea (C. Agardh) Kützing, 1843 Nomenclature: Gloeocapsa sanguinea Notes: Anagnostidis 1968 Gloeocapsa thermalis Kützing, 1843 Nomenclature: Gloeocapsa thermalis Notes: Anagnostidis 1968 Gloeocapsa thermophila (Wood) Claus, 1959 Nomenclature: Chroococcus thermophilus Notes: Anagnostidis 1961 Gloeocapsopsis crepidinum (Thuret) Geitler ex Komárek, 1993 Nomenclature: Gloeocapsopsis crepidinum Notes: Lamprinou et al. 2012 Gloeocapsopsis cyanea (Krieger) Komárek & Anagnostidis, 1995 Nomenclature: Gloeocapsopsis cyanea Notes: Lamprinou et al. 2012 Gloeocapsa rupestris Kützing, 1847 Nomenclature: Gloeocapsa rupestris Notes: Anagnostidis 1961 Gloeocapsa sanguinea (C. Agardh) Kützing, 1843 Nomenclature: Gloeocapsa sanguinea Notes: Anagnostidis 1968 Gloeocapsa thermalis Kützing, 1843 Nomenclature: Gloeocapsa thermalis Notes: Anagnostidis 1968 Gloeocapsa thermophila (Wood) Claus, 1959 Nomenclature: Chroococcus thermophilus Notes: Anagnostidis 1961 Gloeocapsopsis crepidinum (Thuret) Geitler ex Komárek, 1993 Nomenclature: Gloeocapsopsis crepidinum Notes: Lamprinou et al. 2012 Gloeocapsopsis cyanea (Krieger) Komárek & Anagnostidis, 1995 Nomenclature: Gloeocapsopsis cyanea Notes: Lamprinou et al. 2012 33 Cyanobacteria of Greece: an annotated checklist Gloeocapsopsis pleurocapsoides (Novácek) Komárek & Anagno Komárek, 1993 Nomenclature: Gloeocapsopsis pleurocapsoides Notes: Lamprinou et al. 2012 Gloeothece cf. incerta Skuja, 1964 Nomenclature: Gloeothece cf. incerta Notes: Lamprinou et al. 2014 Gloeothece cf. tepidariorum (A. Braun) Lagerheim 1883 Nomenclature: Gloeothece cf. tepidariorum Notes: Lamprinou et al. 2014 Gloeothece confluens Nägeli, 1849 Nomenclature: Gloeothece conﬂuens Notes: Anagnostidis 1968 Gloeothece fusco-lutea (Nägeli ex Kützing) Nägeli, 1849 Nomenclature: Gloeothece fusco-lutea Notes: Anagnostidis 1968 Gloeothece palea (Kützing) Nägeli, 1849 Nomenclature: Gloeothece palea Notes: Anagnostidis and Golubic 1966 Gloeothece rupestris (Lyngbye) Bornet in Wittrock & Nordsted Nomenclature: Gloeocapsopsis pleurocapsoides (Novácek) Komárek & Anagnosti Komárek, 1993 Nomenclature: Gloeocapsopsis pleurocapsoides Notes: Lamprinou et al. 2012 Gloeothece cf. incerta Skuja, 1964 Nomenclature: Gloeothece cf. incerta Notes: Lamprinou et al. 2014 Gloeothece cf. tepidariorum (A. Braun) Lagerheim 1883 Nomenclature: Gloeothece cf. tepidariorum Notes: Lamprinou et al. 2014 Gloeothece confluens Nägeli, 1849 Nomenclature: Gloeothece conﬂuens Notes: Anagnostidis 1968 Gloeothece fusco-lutea (Nägeli ex Kützing) Nägeli, 1849 Nomenclature: Gloeothece fusco-lutea Notes: Anagnostidis 1968 Gloeothece palea (Kützing) Nägeli, 1849 Nomenclature: Gloeothece palea Notes: Anagnostidis and Golubic 1966 Gloeothece rupestris (Lyngbye) Bornet in Wittrock & Nordstedt Gloeothece cf. tepidariorum (A. Braun) Lagerheim 1883 Nomenclature: Gloeothece cf. tepidariorum Notes: Lamprinou et al. 2014 Gkelis S et al. Gloeothece rupestris Notes: Anagnostidis 1961 Gloeothece tepidariorum (A. Braun) Lagerheim, 1883 Nomenclature: Gloeothece rupestris var. tepidariorum Notes: Anagnostidis 1961 Gloeotrichia echinulata P. G. Richter, 1894 Nomenclature: Gloeotrichia echinulata Notes: Anagnostidis 1968 Gloeotrichia pisum Thuret ex Bornet & Flahault, 1886 Nomenclature: Gloeotrichia pisum Notes: Anagnostidis 1968 Gomphosphaeria aponina Kützing, 1836 Nomenclature: Gomphosphaeria aponina Notes: Anagnostidis 1968 Hapalosiphon intricatus W. & G. S. West, 1894 Nomenclature: Hapalosiphon intricatus Notes: Anagnostidis 1968 Hassallia byssoidea Hassall ex Bornet & Flahault, 1886 Nomenclature: Tolypothrix byssoidea Notes: Anagnostidis 1968 34 Gkelis S et al. 34 Gloeothece rupestris Notes: Anagnostidis 1961 Gloeothece tepidariorum (A. Braun) Lagerheim, 1883 Nomenclature: Gloeothece rupestris var. tepidariorum Notes: Anagnostidis 1961 Gloeotrichia echinulata P. G. Richter, 1894 Nomenclature: Gloeotrichia echinulata Notes: Anagnostidis 1968 Gloeotrichia pisum Thuret ex Bornet & Flahault, 1886 Nomenclature: Gloeotrichia pisum Notes: Anagnostidis 1968 Gomphosphaeria aponina Kützing, 1836 Nomenclature: Gomphosphaeria aponina Notes: Anagnostidis 1968 Hapalosiphon intricatus W. & G. S. West, 1894 Nomenclature: Hapalosiphon intricatus Notes: Anagnostidis 1968 Hassallia byssoidea Hassall ex Bornet & Flahault, 1886 Nomenclature: Tolypothrix byssoidea Notes: Anagnostidis 1968 35 Cyanobacteria of Greece: an annotated checklist Herpyzonema pulverulentum M. Hernández-Mariné & T. Canals, 1994 Nomenclature: Herpyzonema pulverulentum Notes: Lamprinou et al. 2012 Heteroleibleinia infixa (Frémy) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya inﬁxa Notes: Anagnostidis 1968 Heteroleibleinia kuetzingii (Schmidle) Compère, 1985 Nomenclature: Lyngbya kutzingii Notes: Anagnostidis 1968 Homoeothrix caespitosa Kirchner, 1898 Nomenclature: Homoeothrix caespitosa Notes: Anagnostidis 1968 Homoeothrix juliana (Bornet & Flahault ex Gomont) Kirchner,1898 Nomenclature: Homoeothrix juliana Notes: Anagnostidis 1968 Hormoscilla feldmannii Anagnostidis, 2001 Nomenclature: Borzia spongeliae Notes: Anagnostidis 1977 Hydrococcus cesatii Rabenhorst, 1860 Nomenclature: Hydrococcus cesatii Cyanobacteria of Greece: an annotated checklist Herpyzonema pulverulentum M. Hernández-Mariné & T. Canals, 1994 Nomenclature: Herpyzonema pulverulentum Notes: Lamprinou et al. 2012 Heteroleibleinia infixa (Frémy) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya inﬁxa Notes: Anagnostidis 1968 Heteroleibleinia kuetzingii (Schmidle) Compère, 1985 Nomenclature: Lyngbya kutzingii Notes: Anagnostidis 1968 Homoeothrix caespitosa Kirchner, 1898 Nomenclature: Homoeothrix caespitosa Notes: Anagnostidis 1968 Homoeothrix juliana (Bornet & Flahault ex Gomont) Kirchner,1898 Nomenclature: Homoeothrix juliana Notes: Anagnostidis 1968 Hormoscilla feldmannii Anagnostidis, 2001 Nomenclature: Borzia spongeliae Notes: Anagnostidis 1977 Hydrococcus cesatii Rabenhorst, 1860 Nomenclature: Hydrococcus cesatii Gkelis S et al. 36 Notes: Anagnostidis 1968 Hydrococcus rivularis Kützing, 1833 Nomenclature: Hydrococcus rivularis Notes: Anagnostidis 1968 Hydrocoleum homoeotrichum Kützing ex Gomont, 1892 Nomenclature: Hydrocoleum homoeotrichum Notes: Anagnostidis 1968 Hydrocoleum muscicola Hansgirg ex Forti, 1907 Nomenclature: Hydrocoleum muscicolum Notes: Anagnostidis 1968 Hydrocoleum stankovicii Čado, 1958 Nomenclature: Hydrocoleum stankovicii Notes: Lamprinou et al. 2012 Hyella balani Lehmann, 1903 Nomenclature: Hyella balani Notes: Anagnostidis and Pantazidou 1985 Hyella caespitosa Bornet & Flahault, 1888 Nomenclature: Hyella caespitosa Notes: Anagnostidis and Pantazidou 1988a 37 Cyanobacteria of Greece: an annotated checklist 37 Hyella fontana Huber & Jadin, 1892 Nomenclature: Hyella fontana Notes: Anagnostidis and Pantazidou 1988c Hyella inconstans Al-Thukair & Golubic 1991 Nomenclature: Hyella inconstans Notes: Pantazidou et al. 2006 Hyella kalligrammos Anagnostidis & Pantazidou, 1988 Nomenclature: Hyella kalligrammos Notes: Anagnostidis and Pantazidou 1988c Hyella maxima (Geitler) Anagnostidis & Pantazidou, 1988 Nomenclature: Hyella maxima Notes: Anagnostidis and Pantazidou 1988c Hyella reptans Al-Thukair & Golubic, 1991 Nomenclature: Hyella reptans Notes: Pantazidou et al. 2006 Ifinoe spelaea Lamprinou & Pantazidou in Lamprinou et al. 2011 Nomenclature: Iphinoë spelaeobios Notes: Lamprinou et al. 2011 Isactis plana Thuret ex Bornet & Flahault, 1886 Nomenclature: Calothrix plana Gkelis S et al. 38 Notes: Anagnostidis and Golubic 1966 Isocystis halobia K. Anagnostidis & M. Roussomoustakaki, 1991 Nomenclature: Isocystis halobia Notes: Anagnostidis and Roussomoustakaki 1991 Isocystis pallida (Voronichin) Woronichin, 1927 Nomenclature: Isocystis pallida Notes: Anagnostidis 1961 Jaaginema angustissimum (W. & G. S. West) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria angustissima Notes: Anagnostidis 1961 Jaaginema geminatum (Schwabe ex Gomont) Anagnostidis & Komárek 1988 Nomenclature: Oscillatoria geminata Notes: Anagnostidis 1961 Jaaginema mauchanum (Claus) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria cf. mauchiana Notes: Lamprinou et al. 2012 Jaaginema minimum (Gicklhorn) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria minima Notes: Anagnostidis 1968 Notes: Anagnostidis and Golubic 1966 Isocystis halobia K. Anagnostidis & M. Herpyzonema pulverulentum M. Hernández-Mariné & T. Canals, 1994 Nomenclature: Herpyzonema pulverulentum Notes: Lamprinou et al. 2012 Heteroleibleinia infixa (Frémy) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya inﬁxa Notes: Anagnostidis 1968 Heteroleibleinia kuetzingii (Schmidle) Compère, 1985 Nomenclature: Lyngbya kutzingii Notes: Anagnostidis 1968 Homoeothrix caespitosa Kirchner, 1898 Nomenclature: Homoeothrix caespitosa Notes: Anagnostidis 1968 Homoeothrix juliana (Bornet & Flahault ex Gomont) Kirchner,1898 Nomenclature: Homoeothrix juliana Notes: Anagnostidis 1968 Hormoscilla feldmannii Anagnostidis, 2001 Nomenclature: Borzia spongeliae Notes: Anagnostidis 1977 Hydrococcus cesatii Rabenhorst, 1860 Nomenclature: Hydrococcus cesatii Cyanobacteria of Greece: an annotated checklist Roussomoustakaki, 1991 Nomenclature: Isocystis halobia Notes: Anagnostidis and Roussomoustakaki 1991 Isocystis pallida (Voronichin) Woronichin, 1927 Nomenclature: Isocystis pallida Notes: Anagnostidis 1961 Jaaginema angustissimum (W. & G. S. West) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria angustissima Notes: Anagnostidis 1961 Jaaginema geminatum (Schwabe ex Gomont) Anagnostidis & Komárek 1988 Nomenclature: Oscillatoria geminata Notes: Anagnostidis 1961 Jaaginema mauchanum (Claus) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria cf. mauchiana Notes: Lamprinou et al. 2012 Jaaginema minimum (Gicklhorn) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria minima Notes: Anagnostidis 1968 Jaaginema minimum (Gicklhorn) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria minima Notes: Anagnostidis 1968 Cyanobacteria of Greece: an annotated checklist 39 Jaaginema profundum (Schröter & Kirchner) Anagnostidis & Komárek, 1988 Nomenclature: Achroonema projundum Notes: Anagnostidis and Economou–Amilli 1980 Jaaginema pseudogeminatum (G. Schmid) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria pseudogeminata Notes: Anagnostidis and Golubic 1966 Jaaginema subtilissimum (Kützing ex Forti) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria subtilissima Notes: Anagnostidis 1961 Johanseninema constrictum (Szafer) Hasler, Dvorák & Poulícková, 2014 Nomenclature: Pseudanabaena constricta Notes: Anagnostidis and Economou–Amilli 1980 Kamptonema animale (C. Agardh ex Gomont) Strunecký, Komárek & J Smarda, 2014 Nomenclature: Oscillatoria animalis Notes: Anagnostidis 1961 Kamptonema chlorinum (Kützing ex Gomont) Strunecký, Komárek & J Smarda, 2014 Nomenclature: Oscillatoria chlorina Notes: Anagnostidis 1968 Jaaginema profundum (Schröter & Kirchner) Anagnostidis & Komárek, 1988 Nomenclature: Achroonema projundum Notes: Anagnostidis and Economou–Amilli 1980 Jaaginema pseudogeminatum (G. Schmid) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria pseudogeminata Notes: Anagnostidis and Golubic 1966 Jaaginema subtilissimum (Kützing ex Forti) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria subtilissima Notes: Anagnostidis 1961 Johanseninema constrictum (Szafer) Hasler, Dvorák & Poulícková, 2014 Nomenclature: Pseudanabaena constricta Notes: Anagnostidis and Economou–Amilli 1980 Kamptonema animale (C. Agardh ex Gomont) Strunecký, Komárek & J Smarda, 2014 Nomenclature: Oscillatoria animalis Notes: Anagnostidis 1961 Kamptonema chlorinum (Kützing ex Gomont) Strunecký, Komárek & J Smarda, 2014 Nomenclature: Oscillatoria chlorina Notes: Anagnostidis 1968 Gkelis S et al. 40 Kamptonema cortianum (Meneghini ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria cortiana Notes: Anagnostidis 1961 Kamptonema formosum (Bory ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria formosa Notes: Anagnostidis 1961 Kamptonema jasorvense (Vouk) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria jasorvensis Notes: Anagnostidis and Golubic 1966 Kamptonema laetevirens (H.M.Crouan & P.L.Crouan ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Phormidium laetevirens Notes: Anagnostidis and Roussomoustakaki 1988 Kamptonema okenii (C.Agardh ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria okenii Notes: Anagnostidis 1961 Komvophoron cf. Herpyzonema pulverulentum M. Hernández-Mariné & T. Canals, 1994 Nomenclature: Herpyzonema pulverulentum Notes: Lamprinou et al. 2012 Heteroleibleinia infixa (Frémy) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya inﬁxa Notes: Anagnostidis 1968 Heteroleibleinia kuetzingii (Schmidle) Compère, 1985 Nomenclature: Lyngbya kutzingii Notes: Anagnostidis 1968 Homoeothrix caespitosa Kirchner, 1898 Nomenclature: Homoeothrix caespitosa Notes: Anagnostidis 1968 Homoeothrix juliana (Bornet & Flahault ex Gomont) Kirchner,1898 Nomenclature: Homoeothrix juliana Notes: Anagnostidis 1968 Hormoscilla feldmannii Anagnostidis, 2001 Nomenclature: Borzia spongeliae Notes: Anagnostidis 1977 Hydrococcus cesatii Rabenhorst, 1860 Nomenclature: Hydrococcus cesatii Cyanobacteria of Greece: an annotated checklist crassum (Vozzennikova) Anagnostidis & Komárek, 1988 Nomenclature: Komvophoron cf. crassum Notes: Anagnostidis and Roussomoustakaki 1991 40 Gkelis S et al. Kamptonema cortianum (Meneghini ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria cortiana Notes: Anagnostidis 1961 Kamptonema formosum (Bory ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria formosa Notes: Anagnostidis 1961 Kamptonema jasorvense (Vouk) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria jasorvensis Notes: Anagnostidis and Golubic 1966 Kamptonema laetevirens (H.M.Crouan & P.L.Crouan ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Phormidium laetevirens Notes: Anagnostidis and Roussomoustakaki 1988 Kamptonema okenii (C.Agardh ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria okenii Notes: Anagnostidis 1961 Komvophoron cf. crassum (Vozzennikova) Anagnostidis & Komárek, 1988 Nomenclature: Komvophoron cf. crassum Notes: Anagnostidis and Roussomoustakaki 1991 Kamptonema cortianum (Meneghini ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria cortiana Notes: Anagnostidis 1961 Kamptonema formosum (Bory ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria formosa Notes: Anagnostidis 1961 Kamptonema jasorvense (Vouk) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria jasorvensis Notes: Anagnostidis and Golubic 1966 Kamptonema laetevirens (H.M.Crouan & P.L.Crouan ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Phormidium laetevirens Notes: Anagnostidis and Roussomoustakaki 1988 Kamptonema okenii (C.Agardh ex Gomont) Strunecký, Komárek & J. Smarda, 2014 Nomenclature: Oscillatoria okenii Notes: Anagnostidis 1961 Komvophoron cf. crassum (Vozzennikova) Anagnostidis & Komárek, 1988 Nomenclature: Komvophoron cf. crassum Cyanobacteria of Greece: an annotated checklist 41 41 Komvophoron pallidum (Skuja) Anagnostidis & Komárek, 1988 Nomenclature: Pseudanabaena pallida Notes: Anagnostidis and Economou–Amilli 1980 Komvophoron schmidlei (Jaag) Anagnostidis & Komárek, 1988 Nomenclature: Pseudanabaena schmidlei Notes: Anagnostidis 1968 Kyrtuthrix dalmatica Ercegović, 1929 Nomenclature: Kyrtuthrix dalmatica Notes: Anagnostidis 1968 Leibleinia gracilis (Rabenhorst ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya gracilis Notes: Anagnostidis and Golubic 1966 Leibleinia nordgaardii Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya nordagardhii Notes: Anagnostidis 1968 Lemmermanniella pallida (Lemmermann) Geitler, 1942 Nomenclature: Lemmermanniella pallida Notes: Tryfon et al. 1997 Cyanobacteria of Greece: an annotated checklist Komvophoron pallidum (Skuja) Anagnostidis & Komárek, 1988 Nomenclature: Pseudanabaena pallida Notes: Anagnostidis and Economou–Amilli 1980 Komvophoron schmidlei (Jaag) Anagnostidis & Komárek, 1988 Nomenclature: Pseudanabaena schmidlei Notes: Anagnostidis 1968 Kyrtuthrix dalmatica Ercegović, 1929 Nomenclature: Kyrtuthrix dalmatica Notes: Anagnostidis 1968 Leibleinia gracilis (Rabenhorst ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya gracilis Notes: Anagnostidis and Golubic 1966 Leibleinia nordgaardii Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya nordagardhii Notes: Anagnostidis 1968 Lemmermanniella pallida (Lemmermann) Geitler, 1942 Nomenclature: Lemmermanniella pallida Notes: Tryfon et al. 1997 Komvophoron pallidum (Skuja) Anagnostidis & Komárek, 1988 Nomenclature: Pseudanabaena pallida Notes: Anagnostidis and Economou–Amilli 1980 Komvophoron schmidlei (Jaag) Anagnostidis & Komárek, 1988 Nomenclature: Pseudanabaena schmidlei Notes: Anagnostidis 1968 Kyrtuthrix dalmatica Ercegović, 1929 Nomenclature: Kyrtuthrix dalmatica Notes: Anagnostidis 1968 Leibleinia gracilis (Rabenhorst ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya gracilis Notes: Anagnostidis and Golubic 1966 Leibleinia nordgaardii Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya nordagardhii Notes: Anagnostidis 1968 Lemmermanniella pallida (Lemmermann) Geitler, 1942 Nomenclature: Lemmermanniella pallida Notes: Tryfon et al. 1997 Gkelis S et al. 42 Leptolyngbya amplivaginata (van Goor) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya amplivaginata Notes: Anagnostidis 1968 Leptolyngbya angustissima (W. & G. S. West) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium angustissimum Notes: Anagnostidis 1961 Leptolyngbya antarctica (W. & G. S. West) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium antarcticum Notes: Anagnostidis et al. 1982 Leptolyngbya boryana (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Plectonema boryanum Notes: Anagnostidis et al. 1981 Leptolyngbya calotrichoides (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Plectonema calotrichoides Notes: Anagnostidis 1968 Leptolyngbya cebennensis (Gomont) I. Umezaki & M. Watanabe, 1994 Nomenclature: Leptolyngbya cebennensis Notes: Lamprinou et al. 2012 Cyanobacteria of Greece: an annotated checklist 43 Leptolyngbya cf. carnea (Kützing ex Lemmermann) Anagnostidis & Komárek, 1988 Nomenclature: Leptolyngbya cf. carnea Notes: Lamprinou et al. 2012 Leptolyngbya cf. laminosa (Gomont ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Leptolyngbya cf. laminosa Notes: Lamprinou et al. 2014 Leptolyngbya compacta (Kützing ex Hansgirg) Komárek in Anagnostidis, 2001 Nomenclature: Leptolyngbya compacta Notes: Lamprinou et al. 2012 Leptolyngbya ectocarpi (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium ectocarpi Notes: Anagnostidis 1968 Leptolyngbya ercegovicii (Čado) Anagnostidis & Komárek, 1988 Nomenclature: Leptolyngbya ercegovicii Notes: Lamprinou et al. 2012 Leptolyngbya foveolara (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium foveolarum Notes: Anagnostidis 1961 Gkelis S et al. 44 Leptolyngbya fragilis (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium fragile Notes: Anagnostidis 1961 Leptolyngbya gracillima (Hansgirg) Anagnostidis & Komárek, 1988 Nomenclature: Plectonema gracillimum Notes: Anagnostidis et al. Komvophoron pallidum (Skuja) Anagnostidis & Komárek, 1988 Nomenclature: Pseudanabaena pallida Notes: Anagnostidis and Economou–Amilli 1980 Komvophoron schmidlei (Jaag) Anagnostidis & Komárek, 1988 Nomenclature: Pseudanabaena schmidlei Notes: Anagnostidis 1968 Kyrtuthrix dalmatica Ercegović, 1929 Nomenclature: Kyrtuthrix dalmatica Notes: Anagnostidis 1968 Leibleinia gracilis (Rabenhorst ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya gracilis Notes: Anagnostidis and Golubic 1966 Leibleinia nordgaardii Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya nordagardhii Notes: Anagnostidis 1968 Lemmermanniella pallida (Lemmermann) Geitler, 1942 Nomenclature: Lemmermanniella pallida Notes: Tryfon et al. 1997 Cyanobacteria of Greece: an annotated checklist 1982 Leptolyngbya halophila (Hansgirg ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya halophila Notes: Anagnostidis 1968 Leptolyngbya lagerheimii (Gomont ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya lagerheimii Notes: Anagnostidis 1961 Leptolyngbya laminosa (Gomont ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium laminosum Notes: Anagnostidis 1961 Leptolyngbya lurida (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium luridum Notes: Anagnostidis 1961 Leptolyngbya nana (Tilden) Anagnostidis & Komárek, 1988 Nomenclature: Leptolyngbya lagerheimii (Gomont ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya lagerheimii Notes: Anagnostidis 1961 Notes: Lamprinou et al. 2012 Leptolyngbya norvegica (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Plectonema norvegicum Notes: Anagnostidis 1968 Leptolyngbya nostocorum (Bornet ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Plectonema nostocorum Notes: Anagnostidis and Golubic 1966 Leptolyngbya notata (Schmidle) Anagnostidis & Komárek, 1988 Nomenclature: Plectonema notatum Notes: Anagnostidis 1968 Leptolyngbya ochracea (Thuret ex Gomont) Anagnostidis & Komárek 1988 Nomenclature: Lyngbya ochracea Notes: Anagnostidis 1968 Leptolyngbya perelegans (Lemmermann) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya perelegans Notes: Anagnostidis and Golubic 1966 Leptolyngbya perforans (Geitler) Anagnostidis & Komárek, 1988 Nomenclature: Schizothrix perforans st. typicus Notes: Anagnostidis 1968 Cyanobacteria of Greece: an annotated checklist 4 45 Cyanobacteria of Greece: an annotated checklist Gkelis S et al. 46 Leptolyngbya phormidioides (Anagnostidis) Anagnostidis & Komárek, 198 Nomenclature: Oscillatoria acuminata f. phormidioides Notes: Anagnostidis 1961 Leptolyngbya purpurascens (Gomont ex Gomont) Anagnostidis & Kom 1988 Nomenclature: Leptolyngbya purpurascens Notes: Lamprinou et al. 2012 Leptolyngbya rivulariarum (Gomont) Anagnostidis & Komárek 1988 Nomenclature: Lyngbya rivulariarum Notes: Anagnostidis 1968 Leptolyngbya subtilissima (Hansgirg) Komárek in Anagnostidis, 2001 Nomenclature: Leptolyngbya subtilissima Notes: Lamprinou et al. 2014 Leptolyngbya subuliformis (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium subuliforme Notes: Anagnostidis 1961 Leptolyngbya tenuis (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium tenue f. circinata Notes: Anagnostidis 1961 Cyanobacteria of Greece: an annotated checklist Cyanobacteria of Greece: an annotated checklist 47 Leptolyngbya terebrans (Bornet & Flahault ex Gomont) Anagnost Komárek, 1988 Nomenclature: Plectonema terebrans Notes: Anagnostidis and Golubic 1966 Leptolyngbya truncata (Lemmermann) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium truncatum Notes: Anagnostidis 1961 Leptolyngbya undosa (Čado) Anagnostidis & Komárek, 1988 Nomenclature: Leptolyngbya undosa Notes: Lamprinou et al. 2012 Leptolyngbya valderiana (Gomont) Anagnostidis & Komárek 1988 Nomenclature: Phormidium valderianum Notes: Anagnostidis 1961 Leptolyngbya weedii (Drouet) Anagnostidis 2001 Nomenclature: Phormidium weedii Notes: Anagnostidis and Economou-Amilli 1978 Limnococcus limneticus (Lemmermann) Komárková, Jezberová, O. Kom Zapomelová, 2010 Nomenclature: Chroococcus limneticus Notes: Anagnostidis 1968 Gkelis S et al. 48 Limnothrix brachynema (Skuja) Hindák & Trifonova, 1989 Nomenclature: Lyngbya brachynema Notes: Anagnostidis 1968 Limnothrix obliqueacuminata (Skuja) Meffert 1988 Nomenclature: Oscillatoria obliqueacuminata Notes: Anagnostidis 1968 Limnothrix redekei (Goor) Meffert, 1988 Nomenclature: Limnothrix redekei Notes: Vardaka et al. 2000 Lyngbya f. minor Elenkin, 1949 Nomenclature: Lyngbya aerugineo-coerulea f. minor Notes: Anagnostidis 1961 Lyngbya aestuarii Liebman ex Gomont, 1892 Nomenclature: Lyngbya aestuarii Notes: Anagnostidis 1968 Lyngbya agardhii Gomont, 1892 Nomenclature: Lyngbya agardhii Notes: Anagnostidis and Golubic 1966 Lyngbya boryana Kirchner ex Hansgirg, 1892 Nomenclature: Phormidium boryanum Limnothrix brachynema (Skuja) Hindák & Trifonova, 1989 Nomenclature: Lyngbya brachynema Notes: Anagnostidis 1968 Limnothrix obliqueacuminata (Skuja) Meffert 1988 Nomenclature: Oscillatoria obliqueacuminata Notes: Anagnostidis 1968 Limnothrix redekei (Goor) Meffert, 1988 Nomenclature: Limnothrix redekei Notes: Vardaka et al. 2000 Lyngbya f. minor Elenkin, 1949 Nomenclature: Lyngbya aerugineo-coerulea f. minor Notes: Anagnostidis 1961 Lyngbya aestuarii Liebman ex Gomont, 1892 Nomenclature: Lyngbya aestuarii Notes: Anagnostidis 1968 Lyngbya agardhii Gomont, 1892 Nomenclature: Lyngbya agardhii Notes: Anagnostidis and Golubic 1966 Lyngbya boryana Kirchner ex Hansgirg, 1892 Nomenclature: Phormidi m bor an m Lyngbya aestuarii Liebman ex Gomont, 1892 Nomenclature: Lyngbya aestuarii Notes: Anagnostidis 1968 49 Cyanobacteria of Greece: an annotated checklist 49 Notes: Anagnostidis 1961 Lyngbya cf. spiralis Geitler, 1932 Nomenclature: Lyngbya cf. spiralis Notes: Overbeck et al. 1982 Lyngbya confervoides C. Agardh ex Gomont, 1892 Nomenclature: Lyngbya confervoides Notes: Anagnostidis and Golubic 1966 Lyngbya lutea Gomont ex Gomont, 1892 Nomenclature: Lyngbya lutea Notes: Anagnostidis and Golubic 1966 Lyngbya majuscula Harvey ex Gomont, 1892 Nomenclature: Lyngbya majuscula Notes: Anagnostidis 1968 Lyngbya martensiana Meneghini ex Gomont, 1892 Nomenclature: Lyngbya martensiana Notes: Anagnostidis 1961 Lyngbya palikiana Claus, 1955 Nomenclature: Leptolyngbya palikiana Notes: Lamprinou et al. 2012 Notes: Anagnostidis 1961 Lyngbya cf. spiralis Geitler, 1932 Nomenclature: Lyngbya cf. spiralis Notes: Overbeck et al. 1982 Lyngbya confervoides C. Agardh ex Gomont, 1892 Nomenclature: Lyngbya confervoides Notes: Anagnostidis and Golubic 1966 Lyngbya lutea Gomont ex Gomont, 1892 Nomenclature: Lyngbya lutea Notes: Anagnostidis and Golubic 1966 Lyngbya majuscula Harvey ex Gomont, 1892 Nomenclature: Lyngbya majuscula Notes: Anagnostidis 1968 Lyngbya martensiana Meneghini ex Gomont, 1892 Nomenclature: Lyngbya martensiana Notes: Anagnostidis 1961 Lyngbya palikiana Claus, 1955 Nomenclature: Leptolyngbya palikiana Notes: Lamprinou et al. 2012 Gkelis S et al. 50 Lyngbya semiplena J. Agardh ex Gomont, 1892 Nomenclature: Lyngbya semiplena Notes: Anagnostidis and Golubic 1966 Lyngbya sordida Gomont, 1892 Nomenclature: Lyngbya sordida Notes: Anagnostidis and Golubic 1966 Mastigocladus laminosus Cohn ex Kirchner, 1898 Nomenclature: Mastigocladus laminosus Notes: Anagnostidis 1961 Mastigocoleus testarum Lagerheim ex Bornet & Flahault, 1887 Nomenclature: Mastigocoleus testarum Notes: Anagnostidis 1968 Merismopedia elegans A. Braun ex Kützing, 1849 Nomenclature: Merismopedia elegans Notes: Anagnostidis 1968 Merismopedia glauca (Ehrenberg) Kützing, 1845 Nomenclature: Merismopedia glauca Notes: Anagnostidis 1968 Merismopedia mediterranea Nägeli, 1849 Nomenclature: Merismopedia glauca f. mediterranea 51 Cyanobacteria of Greece: an annotated checklist 51 Notes: Anagnostidis and Golubic 1966 Merismopedia minima G. Beck 1897 Nomenclature: Merismopedia minima Notes: Anagnostidis 1968 Merismopedia punctata Meyen, 1839 Nomenclature: Merismopedia punctata Notes: Anagnostidis 1968 Merismopedia smithii De Toni, 1939 Nomenclature: Merismopedia maior Notes: Anagnostidis 1968 Merismopedia tenuissima Lemmermann 1898 Nomenclature: Merismopedia tenuissima Notes: Anagnostidis 1968 Merismopedia trolleri Bachmann, 1920 Nomenclature: Merismopedia trolleri Notes: Anagnostidis 1968 Merismopedia warmingiana (Lagerheim) Forti, 1883 Nomenclature: Merismopedia warmingiana Notes: Anagnostidis 1968 Notes: Anagnostidis and Golubic 1966 Merismopedia minima G. Beck 1897 Nomenclature: Merismopedia minima Notes: Anagnostidis 1968 Merismopedia punctata Meyen, 1839 Nomenclature: Merismopedia punctata Notes: Anagnostidis 1968 Merismopedia smithii De Toni, 1939 Nomenclature: Merismopedia maior Notes: Anagnostidis 1968 Merismopedia tenuissima Lemmermann 1898 Nomenclature: Merismopedia tenuissima Notes: Anagnostidis 1968 Merismopedia trolleri Bachmann, 1920 Nomenclature: Merismopedia trolleri Notes: Anagnostidis 1968 Merismopedia warmingiana (Lagerheim) Forti, 1883 Nomenclature: Merismopedia warmingiana Notes: Anagnostidis 1968 Gkelis S et al. 52 Microchaete grisea Thuret ex Bornet & Flahault, 1886 Nomenclature: Microchaete grisea Notes: Anagnostidis 1968 Microchaete tenera Thuret ex Bornet & Flahault, 1886 Nomenclature: Microchaete tenera Notes: Anagnostidis 1968 Microcoleus amoenus (Gomont) Strunecky, Komárek & J. R. Johansen, 2013 Nomenclature: Oscillatoria amoena Notes: Anagnostidis 1968 Microcoleus amoenus (Gomont) Strunecky, Komárek & J. R. Johansen, 2013 Nomenclature: Phormidium amoenum Notes: Anagnostidis and Roussomoustakaki 1991 Microcoleus autumnalis (Gomont) Strunecky, Komárek & J. R. Johansen, 2013 Nomenclature: Phormidium autumnale Notes: Anagnostidis 1961 Microcoleus ferrugineus Frémy, 1936 Nomenclature: Microcoleus ferrugineus Notes: Anagnostidis 1968 Microcoleus lacustris Farlow ex Gomont, 1892 Nomenclature: Microcoleus lacustris Microchaete grisea Thuret ex Bornet & Flahault, 1886 Nomenclature: Microchaete grisea Notes: Anagnostidis 1968 Microchaete tenera Thuret ex Bornet & Flahault, 1886 Nomenclature: Microchaete tenera Notes: Anagnostidis 1968 Microcoleus amoenus (Gomont) Strunecky, Komárek & J. R. Johansen, 2013 Nomenclature: Oscillatoria amoena Notes: Anagnostidis 1968 Microcoleus amoenus (Gomont) Strunecky, Komárek & J. R. Johansen, 2013 Nomenclature: Phormidium amoenum Notes: Anagnostidis and Roussomoustakaki 1991 Microcoleus autumnalis (Gomont) Strunecky, Komárek & J. R. Johansen, 2013 Nomenclature: Phormidium autumnale Notes: Anagnostidis 1961 Microcoleus ferrugineus Frémy, 1936 Nomenclature: Microcoleus ferrugineus Notes: Anagnostidis 1968 Microcoleus lacustris Farlow ex Gomont, 1892 Nomenclature: Microcoleus lacustris 53 Cyanobacteria of Greece: an annotated checklist Notes: Anagnostidis 1968 Microcoleus paludosus Gomont, 1892 Nomenclature: Microcoleus paludosus Notes: Anagnostidis 1968 Microcoleus setchellianus (Gomont) Strunecky, Komárek & J. R. Jo 2013 Nomenclature: Phormidium setchellianum Notes: Lamprinou et al. 2012 Microcoleus steenstrupii J. B. Petersen, 1923 Nomenclature: Microcoleus steenstrupii Notes: Lamprinou et al. 2012 Microcoleus vaginatus Gomont ex Gomont, 1892 Nomenclature: Microcoleus vaginatus Notes: Anagnostidis 1968 Microcoleus vulgaris Strunecky, Komárek & J. R. Johansen 2013 Nomenclature: Phormidium vulgare Notes: Lamprinou et al. 2012 Microcystis flos-aquae (Wittrock) Kirchner, 1898 Nomenclature: Microcystis ﬂos-aquae Notes: Anagnostidis 1961 Notes: Anagnostidis 1968 Microcoleus paludosus Gomont, 1892 Nomenclature: Microcoleus paludosus Notes: Anagnostidis 1968 Microcoleus setchellianus (Gomont) Strunecky, Komárek & J. R. Johansen 2013 Nomenclature: Phormidium setchellianum Notes: Lamprinou et al. 2012 Microcoleus steenstrupii J. B. Petersen, 1923 Nomenclature: Microcoleus steenstrupii Notes: Lamprinou et al. 2012 Microcoleus vaginatus Gomont ex Gomont, 1892 Nomenclature: Microcoleus vaginatus Notes: Anagnostidis 1968 Microcoleus vulgaris Strunecky, Komárek & J. R. Johansen 2013 Nomenclature: Phormidium vulgare Notes: Lamprinou et al. 2012 Microcystis flos-aquae (Wittrock) Kirchner, 1898 Nomenclature: Microcystis ﬂos-aquae Notes: Anagnostidis 1968 Gkelis S et al. Notes: Anagnostidis and Golubic 1966 Merismopedia minima G. Beck 1897 Nomenclature: Merismopedia minima Notes: Anagnostidis 1968 Merismopedia punctata Meyen, 1839 Nomenclature: Merismopedia punctata Notes: Anagnostidis 1968 Merismopedia smithii De Toni, 1939 Nomenclature: Merismopedia maior Notes: Anagnostidis 1968 Merismopedia tenuissima Lemmermann 1898 Nomenclature: Merismopedia tenuissima Notes: Anagnostidis 1968 Merismopedia trolleri Bachmann, 1920 Nomenclature: Merismopedia trolleri Notes: Anagnostidis 1968 Merismopedia warmingiana (Lagerheim) Forti, 1883 Nomenclature: Merismopedia warmingiana Notes: Anagnostidis 1968 54 Microcystis aeruginosa (Kützing) Kützing, 1846 Nomenclature: Microcystis aeruginosa Notes: Anagnostidis 1968 Microcystis ichthyoblabe (G.Kunze) Kützing, 1846 Nomenclature: Microcystis ichthyoblabe Notes: Tryfon et al. 1997 Microcystis incerta (Lemmermann) Lemmermann, 1903 Nomenclature: Microcystis incerta Notes: Overbeck et al. 1982 Microcystis marginata (Meneghini) Kützing, 1846 Nomenclature: Microcystis marginata Notes: Anagnostidis 1968 Microcystis novacekii (Komárek) Compère, 1974 Nomenclature: Microcystis novacekii Notes: Vardaka et al. 2000 Microcystis panniformis Komárek, Komárková-Legnerová, S Azevedo, & P. A. C. Senna, 2002 Nomenclature: Microcystis panniformis Notes: Papadimitriou et al. 2013 55 Cyanobacteria of Greece: an annotated checklist Microcystis pulverea (H. C. Wood) Forti, 1907 Nomenclature: Microcystis pulverea Notes: Anagnostidis 1968 Microcystis smithii Komárek & Anagnostidis, 1995 Nomenclature: Aphanocapsa pulchra Notes: Anagnostidis 1968 Microcystis viridis (A. Braun) Lemmermann, 1903 Nomenclature: Microcystis viridis Notes: Anagnostidis 1968 Microcystis wesenbergii (Komárek) Komárek ex Komárek, 1968 Nomenclature: Microcystis wesenbergii Notes: Tryfon et al. 1996 Myxosarcina concinna Printz, 1921 Nomenclature: Myxosarcina concinna Notes: Anagnostidis et al. 1983 Nodularia harveyana Thuret ex Bornet & Flahault, 1886 Nomenclature: Nodularia harveyana Notes: Anagnostidis 1968 Nodularia spumigena Mertens ex Bornet & Flahault 1888 Nomenclature: Nodularia spumigena Cyanobacteria of Greece: an annotated checklist Microcystis pulverea (H. C. Wood) Forti, 1907 Nomenclature: Microcystis pulverea Notes: Anagnostidis 1968 Microcystis smithii Komárek & Anagnostidis, 1995 Nomenclature: Aphanocapsa pulchra Notes: Anagnostidis 1968 Microcystis viridis (A. Braun) Lemmermann, 1903 Nomenclature: Microcystis viridis Notes: Anagnostidis 1968 Microcystis wesenbergii (Komárek) Komárek ex Komárek, 1968 Nomenclature: Microcystis wesenbergii Notes: Tryfon et al. 1996 Myxosarcina concinna Printz, 1921 Nomenclature: Myxosarcina concinna Notes: Anagnostidis et al. 1983 Nodularia harveyana Thuret ex Bornet & Flahault, 1886 Nomenclature: Nodularia harveyana Notes: Anagnostidis 1968 Nodularia spumigena Mertens ex Bornet & Flahault 1888 Nomenclature: N d l i i Notes: Anagnostidis 1968 Nostoc calcicola Brébisson ex Bornet & Flahault, 1886 Nomenclature: Nostoc calcicola Notes: Anagnostidis 1961 Nostoc commune Vaucher ex Bornet & Flahault, 1888 Nomenclature: Nostoc commune Notes: Economou-Amilli et al. 1984 Nostoc humifusum Carmichael ex Bornet & Flahault, 1888 Nomenclature: Nostoc humifusum Notes: Anagnostidis 1961 Nostoc letestui Frémy, 1930 Nomenclature: Nostoc letestui Notes: Lamprinou et al. 2012 Nostoc linckia Bornet ex Bornet & Flahault, 1888 Nomenclature: Nostoc linckia Notes: Anagnostidis 1968 Nostoc microscopicum Carmichael ex Bornet & Flahault, 1888 Nomenclature: Nostoc microscopicum Notes: Anagnostidis 1968 56 Gkelis S et al. Gkelis S et al. 56 Notes: Anagnostidis 1968 Nostoc calcicola Brébisson ex Bornet & Flahault, 1886 Nomenclature: Nostoc calcicola Notes: Anagnostidis 1961 Nostoc commune Vaucher ex Bornet & Flahault, 1888 Nomenclature: Nostoc commune Notes: Economou-Amilli et al. 1984 Nostoc humifusum Carmichael ex Bornet & Flahault, 1888 Nomenclature: Nostoc humifusum Notes: Anagnostidis 1961 Nostoc letestui Frémy, 1930 Nomenclature: Nostoc letestui Notes: Lamprinou et al. 2012 Nostoc linckia Bornet ex Bornet & Flahault, 1888 Nomenclature: Nostoc linckia Notes: Anagnostidis 1968 Nostoc microscopicum Carmichael ex Bornet & Flahault, 1888 Nomenclature: Nostoc microscopicum Notes: Anagnostidis 1968 56 Gkelis S et al. 57 Cyanobacteria of Greece: an annotated checklist Nostoc minutissimum Kützing ex Bornet & Flahault, 1888 Nomenclature: Nostoc minutissimum Notes: Anagnostidis 1968 Nostoc paludosum Kützing ex Bornet & Flahault, 1888 Nomenclature: Nostoc paludosum Notes: Anagnostidis 1961 Nostoc punctiforme Hariot 1891 Nomenclature: Nostoc punctiforme Notes: Anagnostidis et al. 1981 Nostoc sphaericum Vaucher ex Bornet & Flahault, 1888 Nomenclature: Nostoc sphaericum Notes: Anagnostidis 1968 Oculatella subterranea Zammit, Billi & Albertano, 2012 Nomenclature: Oculatella subterranea Notes: Christodoulou et al. 2015 Oscillatoria anguina Bory ex Gomont 1892 Nomenclature: Oscillatoria anguina Notes: Anagnostidis 1961 Oscillatoria bonnemaisonii P. Crouan & H. Crouan ex Gomon Nomenclature: Oscillatoria bonnemaisonii Nostoc minutissimum Kützing ex Bornet & Flahault, 1888 Nomenclature: Nostoc minutissimum Notes: Anagnostidis 1968 Nostoc paludosum Kützing ex Bornet & Flahault, 1888 Nomenclature: Nostoc paludosum Notes: Anagnostidis 1961 Nostoc punctiforme Hariot 1891 Nomenclature: Nostoc punctiforme Notes: Anagnostidis et al. 1981 Nostoc sphaericum Vaucher ex Bornet & Flahault, 1888 Nomenclature: Nostoc sphaericum Notes: Anagnostidis 1968 Oculatella subterranea Zammit, Billi & Albertano, 2012 Nomenclature: Oculatella subterranea Notes: Christodoulou et al. 2015 Oscillatoria anguina Bory ex Gomont 1892 Nomenclature: Oscillatoria anguina Notes: Anagnostidis 1961 Oscillatoria bonnemaisonii P. Crouan & H. Crouan ex Gomont Nomenclature: O ill t i b i ii Nostoc punctiforme Hariot 1891 Notes: Anagnostidis 1968 Oscillatoria corallinae Gomont ex Gomont, 1892 Nomenclature: Oscillatoria corallinae Notes: Anagnostidis and Golubic 1966 Oscillatoria crassa (C. B. Rao) Anagnostidis, 2001 Nomenclature: Oscillatoria ornata var. crassa Notes: Anagnostidis 1961 Oscillatoria curviceps C. Agardh ex Gomont, 1892 Nomenclature: Oscillatoria curviceps Notes: Anagnostidis 1968 Oscillatoria laetevirens Hofman-Bang ex Forti 1892 Nomenclature: Oscillatoria laetevirens Notes: Anagnostidis 1961 Oscillatoria limosa C. Agardh ex Gomont 1892 Nomenclature: Oscillatoria limosa Notes: Anagnostidis and Golubic 1966 Oscillatoria margaritifera Kützing ex Gomont, 1892 Nomenclature: Oscillatoria margaritifera Notes: Anagnostidis 1968 58 Gkelis S et al. Gkelis S et al. 58 Oscillatoria margaritifera Kützing ex Gomont, 1892 Nomenclature: Oscillatoria margaritifera Notes: Anagnostidis 1968 59 Cyanobacteria of Greece: an annotated checklist 59 Oscillatoria okenii var. gracilis (Kützing) Kützing ex Forti Nomenclature: Oscillatoria okenii f. gracilis Notes: Anagnostidis 1961 Oscillatoria ornata Kützing ex Gomont, 1892 Nomenclature: Oscillatoria ornata Notes: Anagnostidis 1961 Oscillatoria princeps Vaucher ex Gomont, 1892 Nomenclature: Oscillatoria princeps Notes: Anagnostidis 1961 Oscillatoria pseudoangusta Claus, 1955 Nomenclature: Oscillatoria pseudoangusta Notes: Anagnostidis 1961 Oscillatoria putrida Schmidle, 1901 Nomenclature: Oscillatoria putrida Notes: Anagnostidis 1968 Oscillatoria rhamphoidea Anagnostidis, 2001 Nomenclature: Oscillatoria salina f. major Notes: Anagnostidis et al. 1981 Oscillatoria rupicola (Hansgirg) Hansgirg ex Forti, 1907 Nomenclature: Oscillatoria r picola Gkelis S et al. 60 Notes: Lamprinou et al. 2012 Oscillatoria sancta Kützing ex Gomont 1892 Nomenclature: Oscillatoria sancta Notes: Anagnostidis 1961 Oscillatoria simplicissima Gomont, 1892 Nomenclature: Oscillatoria simplicissima Notes: Anagnostidis 1961 Oscillatoria subbrevis Schmidle, 1901 Nomenclature: Oscillatoria subbrevis Notes: Anagnostidis et al. 1981 Oscillatoria tenuis C. Agardh ex Gomont, 1892 Nomenclature: Oscillatoria tenuis Notes: Anagnostidis 1961 Oscillatoria var. phormidioides Hansgirg ex Drouet, 1957 Nomenclature: Oscillatoria okenii f. phormidioides Notes: Anagnostidis 1961 Oscillatoria trichoides Szafer, 1910 Nomenclature: Oscillatoria trichoides Notes: Anagnostidis and Golubic 1966 60 Gkelis S et al. Notes: Lamprinou et al. 2012 Oscillatoria sancta Kützing ex Gomont 1892 Nomenclature: Oscillatoria sancta Notes: Anagnostidis 1961 Oscillatoria simplicissima Gomont, 1892 Nomenclature: Oscillatoria simplicissima Notes: Anagnostidis 1961 Oscillatoria subbrevis Schmidle, 1901 Nomenclature: Oscillatoria subbrevis Notes: Anagnostidis et al. 1981 Oscillatoria tenuis C. Agardh ex Gomont, 1892 Nomenclature: Oscillatoria tenuis Notes: Anagnostidis 1961 Oscillatoria var. phormidioides Hansgirg ex Drouet, 1957 Nomenclature: Oscillatoria okenii f. phormidioides Notes: Anagnostidis 1961 Oscillatoria trichoides Szafer, 1910 Nomenclature: Oscillatoria trichoides Notes: Anagnostidis and Golubic 1966 60 Gkelis S et al. Cyanobacteria of Greece: an annotated checklist Cyanobacteria of Greece: an annotated checklist 61 Oxynema acuminatum (Gomont) Chatchawan, Komárek, Strunecky, Smarda & Peerapornpisal, 2012 Nomenclature: Oscillatoria acuminata Notes: Anagnostidis 1961 Pannus spumosus B.Hickel, 1991 Nomenclature: Pannus spumosus Notes: Hindak and Moustaka 1988 Petalonema densum (Bornet ex Bornet & Flahault) Migula, 1907 Nomenclature: Petalonema densum Notes: Anagnostidis 1968 Phormidesmis molle (Gomont) Turicchia, Ventura, Komárková & Komárek, 2009 Nomenclature: Phormidium molle Notes: Anagnostidis 1961 Phormidiochaete nordstedtii (Bornet et Flahault) Komárek in Anagnostidis, 2001 Nomenclature: Phormidiochaete nordstedtii Notes: Lamprinou et al. 2012 Phormidium acuminatum var. longe-attenuatum (Geitler & Ruttner) I. Umezaki & M. Watanabe, 1994 Nomenclature: Oscillatoria acuminata f. longe-attenuata Notes: Anagnostidis 1961 Gkelis S et al. 62 Phormidium aerugineo-caeruleum (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya aerugineo-coerulea Notes: Anagnostidis 1961 Phormidium ambiguum Gomont, 1892 Nomenclature: Phormidium ambiguum Notes: Anagnostidis 1961 Phormidium articulatum (N. L. Gardner) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium articulatum Notes: Lamprinou et al. 2012 Phormidium boryanum Kützing, 1843 Nomenclature: Oscillatoria boryana Notes: Anagnostidis and Golubic 1966 Phormidium breve (Kützing ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria brevis Notes: Anagnostidis 1961 Phormidium cf. griseoviolaceum (Skuja) Anagnostidis, 2001 Nomenclature: Phormidium cf. griseoviolaceum Notes: Lamprinou et al. 2012 Phormidium chalybeum (Mertens ex Gomont) Anagnostidis & Komárek, 198 Nomenclature: Oscillatoria chalybea 63 Cyanobacteria of Greece: an annotated checklist Notes: Anagnostidis 1961 Phormidium corium Gomont ex Gomont, 1892 Nomenclature: Phormidium corium Notes: Anagnostidis 1961 Phormidium diguetii (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya diguetii Notes: Anagnostidis 1961 Phormidium edessae Skuja, 1937 Nomenclature: Phormidium edessae Notes: Anagnostidis 1968 Phormidium endophyticum (W. & G. S. West) Skuja, 1956 Nomenclature: Phormidium endophyticum Notes: Anagnostidis 1968 Phormidium favosum Gomont, 1892 Nomenclature: Phormidium favosum Notes: Anagnostidis 1968 Phormidium feldmanni Frémy, 1937 Nomenclature: Phormidium feldmanni Notes: Anagnostidis 1968 Cyanobacteria of Greece: an annotated checklist Notes: Anagnostidis 1961 Phormidium corium Gomont ex Gomont, 1892 Nomenclature: Phormidium corium Notes: Anagnostidis 1961 Phormidium diguetii (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya diguetii Notes: Anagnostidis 1961 Phormidium edessae Skuja, 1937 Nomenclature: Phormidium edessae Notes: Anagnostidis 1968 Phormidium endophyticum (W. & G. S. West) Skuja, 1956 Nomenclature: Phormidium endophyticum Notes: Anagnostidis 1968 Phormidium favosum Gomont, 1892 Nomenclature: Phormidium favosum Notes: Anagnostidis 1968 Phormidium feldmanni Frémy, 1937 Nomenclature: Phormidium feldmanni Notes: Anagnostidis 1968 Notes: Anagnostidis 1961 Gkelis S et al. 64 Phormidium grunowianum (Gomont) Anagnostidis & Komárek, 19 Nomenclature: Oscillatoria grunowiana Notes: Anagnostidis 1961 Phormidium henningsii (Lemmermann) Anagnostidis, 2001 Nomenclature: Leptolyngbya henningsii Notes: Lamprinou et al. 2012 Phormidium incrustatum Gomont ex Gomont, 1892 Nomenclature: Phormidium incrustans Notes: Anagnostidis 1968 Phormidium interruptum Kützing ex Gomont, 1892 Nomenclature: Phormidium interruptum Notes: Lamprinou et al. 2012 Phormidium inundatum Kützing ex Gomont, 1892 Nomenclature: Phormidium inundatum Notes: Anagnostidis 1961 Phormidium janthiphorum Elenkin, 1949 Nomenclature: Oscillatoria janthiphora Notes: Anagnostidis and Golubic 1966 Phormidium jenkelianum Kützing ex Gomont, 1892 Nomenclature: Phormidium jenkelianum Cyanobacteria of Greece: an annotated checklist 65 Notes: Anagnostidis 1961 Phormidium kuetzingianum (Kirchner ex Gomont) Anagnostid 1988 Nomenclature: Phormidium kuetzingianum Notes: Lamprinou et al. 2012 Phormidium lacustre (Čado) Anagnostidis, 2001 Nomenclature: Phormidium lacustre Notes: Lamprinou et al. 2012 Phormidium litorale Golubic, 1960 Nomenclature: Phormidium litorale Notes: Anagnostidis and Golubic 1966 Phormidium macedonicum (Čado), 1959 Nomenclature: Phormidium macedonicum Notes: Lamprinou et al. 2012 Phormidium melanochroun Lamprinou & Pantazidou, 2013 Nomenclature: Phormidium melanochroun Notes: Lamprinou et al. 2013b Phormidium minutum Lindstedt, 1943 Nomenclature: Phormidium minutum Notes: Anagnostidis 1968 y Notes: Anagnostidis 1961 Phormidium kuetzingianum (Kirchner ex Gomont) Anagnostid 1988 Nomenclature: Phormidium kuetzingianum Notes: Lamprinou et al. 2012 Phormidium lacustre (Čado) Anagnostidis, 2001 Nomenclature: Phormidium lacustre Notes: Lamprinou et al. 2012 Phormidium litorale Golubic, 1960 Nomenclature: Phormidium litorale Notes: Anagnostidis and Golubic 1966 Phormidium macedonicum (Čado), 1959 Nomenclature: Phormidium macedonicum Notes: Lamprinou et al. 2012 Phormidium melanochroun Lamprinou & Pantazidou, 2013 Nomenclature: Phormidium melanochroun Notes: Lamprinou et al. 2013b Phormidium minutum Lindstedt, 1943 Nomenclature: Phormidium minutum Notes: Anagnostidis 1968 Notes: Anagnostidis 1961 Gkelis S et al. 66 Phormidium molischii (Vouk) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya molischi Notes: Anagnostidis 1968 Phormidium molle var. tenuior W. & G. S. West ex Geitler, 1925 Nomenclature: Phormidium molle f. tenuior Notes: Anagnostidis 1961 Phormidium nigroviride (Thwaites ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria nigroviridis Notes: Anagnostidis and Golubic 1966 Phormidium nigrum (Vaucher ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria nigra Notes: Anagnostidis 1968 Phormidium pachydermaticum Frémy, 1930 Nomenclature: Phormidium pachydermaticum Notes: Anagnostidis 1961 Phormidium papyraceum Gomont ex Gomont, 1892 Nomenclature: Phormidium papyraceum Notes: Anagnostidis 1961 Phormidium priestleyi F. E. Fritsch, 1917 Nomenclature: Phormidium priestleyi Phormidium molischii (Vouk) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya molischi Notes: Anagnostidis 1968 Phormidium molle var. tenuior W. & G. S. West ex Geitler, 1925 Nomenclature: Phormidium molle f. tenuior Notes: Anagnostidis 1961 Phormidium nigroviride (Thwaites ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria nigroviridis Notes: Anagnostidis and Golubic 1966 Phormidium nigrum (Vaucher ex Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria nigra Notes: Anagnostidis 1968 Phormidium pachydermaticum Frémy, 1930 Nomenclature: Phormidium pachydermaticum Notes: Anagnostidis 1961 Phormidium papyraceum Gomont ex Gomont, 1892 Nomenclature: Phormidium papyraceum Notes: Anagnostidis 1961 Phormidium priestleyi F. E. Fritsch, 1917 67 Cyanobacteria of Greece: an annotated checklist 67 Notes: Lamprinou et al. 2012 Phormidium retzii Kützing ex Gomont, 1892 Nomenclature: Phormidium retzii Notes: Anagnostidis 1961 Phormidium subfuscum Kützing ex Gomont. 1892 Nomenclature: Phormidium subfuscum Notes: Anagnostidis et al. 1981 Phormidium submembranaceum Gomont, 1892 Nomenclature: Phormidium submembranaceum Notes: Anagnostidis 1968 Phormidium tanganyikae (G. S. West) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria tanganyikae Notes: Anagnostidis et al. 1981 Phormidium terebriforme (C. Agardh ex Gomont) Anagnostidis & Kom 1988 Nomenclature: Oscillatoria terebriformis Notes: Anagnostidis 1961 Phormidium tergestinum Kützing ex Anagnostidis & Komárek, 1988 Nomenclature: Phormidium tergestinum Notes: Lamprinou et al. 2012 Cyanobacteria of Greece: an annotated checklist Notes: Lamprinou et al. 2012 Phormidium retzii Kützing ex Gomont, 1892 Nomenclature: Phormidium retzii Notes: Anagnostidis 1961 Phormidium subfuscum Kützing ex Gomont. 1892 Nomenclature: Phormidium subfuscum Notes: Anagnostidis et al. 1981 Phormidium submembranaceum Gomont, 1892 Nomenclature: Phormidium submembranaceum Notes: Anagnostidis 1968 Phormidium tanganyikae (G. S. West) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria tanganyikae Notes: Anagnostidis et al. 1981 Phormidium terebriforme (C. Agardh ex Gomont) Anagnostidis & Komárek 1988 Nomenclature: Oscillatoria terebriformis Notes: Anagnostidis 1961 Phormidium tergestinum Kützing ex Anagnostidis & Komárek, 1988 Nomenclature: Phormidium tergestinum Notes: Lamprinou et al. 2012 Gkelis S et al. 68 Phormidium thwaitesii I. Umezaki & M. Watanabe, 1994 Nomenclature: Oscillatoria subuliformis Notes: Anagnostidis 1968 Phormidium uncinatum Gomont ex Gomont, 1892 Nomenclature: Phormidium uncinatum Notes: Anagnostidis 1968 Phormidium willei (N. L. Gardner) Anagnostidis & Komárek, 1988 Nomenclature: Oscillatoria willei Notes: Anagnostidis and Golubic 1966 Planktolyngbya bipunctata (Lemmermann) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya bipunctata Notes: Anagnostidis and Golubic 1966 Planktolyngbya circumcreta (G. S. West) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya circumcreta Notes: Moustaka-Gouni 1988 Planktolyngbya contorta (Lemmermann) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya contorta Notes: Anagnostidis 1968 Planktolyngbya contorta (Lemmermann) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya contorta Notes: Anagnostidis 1968 Cyanobacteria of Greece: an annotated checklist 69 Planktolyngbya limnetica (Lemmermann) Komárková- Legnerová & Cr 1992 Nomenclature: Lyngbya limnetica Notes: Anagnostidis 1961 Planktothrix agardhii (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Planktothrix agardhii Notes: Papadimitriou et al. 2013 Planktothrix isothrix (Skuja) Komárek & Komárková, 2004 Nomenclature: Oscillatoria cf. agardhii var. isothrix Notes: Overbeck et al. 1982 Planktothrix mougeotii Anagnostidis & Komárek, 1988 Nomenclature: Planktothrix mougeotii Notes: Anagnostidis and Pantazidou 1988c Planktothrix rubescens (De Candolle ex Gomont) Anagnostidis & Ko 1988 Nomenclature: Planktothrix rubescens Notes: Vareli et al. 2009 Plectonema araucanum Schwabe, 1960 Nomenclature: Plectonema araucanum Notes: Lamprinou et al. 2012 Gkelis S et al. 70 Plectonema tomasinianum Bornet ex Gomont, 1893 Nomenclature: Plectonema tomassinianum Notes: Anagnostidis and Economou-Amilli 1978 Pleurocapsa fuliginosa Hauck, 1885 Nomenclature: Pleurocapsa fuliginosa Notes: Lamprinou et al. 2012 Pleurocapsa minor Hansgirg, 1891 Nomenclature: Scopulonema minus Notes: Anagnostidis 1968 Porphyrosiphon versicolor (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya versicolor Notes: Anagnostidis 1968 Pseudanabaena amphigranulata (Van Goor) Anagnostidis, 2001 Nomenclature: Pseudanabaena galeata f. tenuis Notes: Anagnostidis 1961 Pseudanabaena arcuata (Skuja) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium arcuatum Notes: Anagnostidis 1968 Pseudanabaena biceps Böcher, 1946 Nomenclature: Pseudanabaena biceps Plectonema tomasinianum Bornet ex Gomont, 1893 Nomenclature: Plectonema tomassinianum Notes: Anagnostidis and Economou-Amilli 1978 Pleurocapsa fuliginosa Hauck, 1885 Nomenclature: Pleurocapsa fuliginosa Notes: Lamprinou et al. 2012 Pleurocapsa minor Hansgirg, 1891 Nomenclature: Scopulonema minus Notes: Anagnostidis 1968 Porphyrosiphon versicolor (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Lyngbya versicolor Notes: Anagnostidis 1968 Pseudanabaena amphigranulata (Van Goor) Anagnostidis, 2001 Nomenclature: Pseudanabaena galeata f. tenuis Notes: Anagnostidis 1961 Pseudanabaena arcuata (Skuja) Anagnostidis & Komárek, 1988 Nomenclature: Phormidium arcuatum Notes: Anagnostidis 1968 Pseudanabaena biceps Böcher, 1946 Nomenclature: Pseudanabaena biceps 71 Cyanobacteria of Greece: an annotated checklist 71 Notes: Anagnostidis 1968 Pseudanabaena catenata Lauterborn, 1915 Nomenclature: Pseudanabaena catenata Notes: Anagnostidis 1961 Pseudanabaena galeata Böcher, 1949 Nomenclature: Pseudanabaena galeata Notes: Anagnostidis 1961 Pseudanabaena limnetica (Lemmermann) Komárek, 1974 Nomenclature: Oscillatoria limnetica Notes: Anagnostidis 1968 Pseudanabaena lonchoides Anagnostidis, 1961 Nomenclature: Pseudanabaena lonchoides Notes: Anagnostidis 1961 Pseudanabaena mucicola (Naumann & Huber-Pestalozzi) Schwabe, 1964 Nomenclature: Phormidium mucicola Notes: Anagnostidis 1968 Pseudanabaena raphidioides (Geitler) Anagnostidis & Komárek, 1988 Nomenclature: Pseudanabaena raphidioides Notes: Montesanto et al. 1999 Notes: Anagnostidis 1968 Pseudanabaena catenata Lauterborn, 1915 Nomenclature: Pseudanabaena catenata Notes: Anagnostidis 1961 Pseudanabaena galeata Böcher, 1949 Nomenclature: Pseudanabaena galeata Notes: Anagnostidis 1961 Pseudanabaena limnetica (Lemmermann) Komárek, 1974 Nomenclature: Oscillatoria limnetica Notes: Anagnostidis 1968 Pseudanabaena lonchoides Anagnostidis, 1961 Nomenclature: Pseudanabaena lonchoides Notes: Anagnostidis 1961 Pseudanabaena mucicola (Naumann & Huber-Pestalozzi) Schwabe, 196 Nomenclature: Phormidium mucicola Notes: Anagnostidis 1968 Pseudanabaena raphidioides (Geitler) Anagnostidis & Komárek, 1988 Nomenclature: Pseudanabaena raphidioides Notes: Montesanto et al. 1999 Notes: Anagnostidis 1968 Gkelis S et al. 72 Pseudocapsa dubia Ercegovic, 1925 Nomenclature: Pseudocapsa dubia Notes: Anagnostidis and Pantazidou 1991a Pseudophormidium battersii (Gomont) Anagnostidis, 2001 Nomenclature: Plectonema battersii Notes: Anagnostidis 1968 Pseudophormidium golenkinianum (Gomont) Anagnostidis, 2001 Nomenclature: Plectonema golenkinianum Notes: Anagnostidis and Golubic 1966 Pseudophormidium hollerbachianum (Elenkin) Anagnostidis, 2001 Nomenclature: Plectonema boryanum f. hollerbachianum Notes: Anagnostidis et al. 1983 Pseudophormidium radiosum (Gomont) Anagnostidis & Komárek, 1988 Nomenclature: Plectonema radiosum Notes: Anagnostidis 1968 Pseudophormidium spelaeoides (Čado) Anagnostidis, 2001 Nomenclature: Pseudophormidium spelaeoides Notes: Lamprinou et al. 2012 Pseudophormidium tenue (Thuret ex Gomont) Anagnostidis & Komárek, Nomenclature: Plectonema tenue Cyanobacteria of Greece: an annotated checklist 73 73 Notes: Anagnostidis and Economou-Amilli 1978 Pseudoscytonema endolithicum (Ercegovic) Anagnostidis in Kom Anagnostidis, 2005 Nomenclature: Plectonema endolithicum Notes: Anagnostidis 1968 Radiocystis geminata Skuja, 1948 Nomenclature: Radiocystis geminata Notes: Anagnostidis 1968 Raphidiopsis mediterranea Skuja, 1937 Nomenclature: Raphidiopsis mediterranea Notes: Anagnostidis 1968 Rhabdoderma curtum (Setchell) Komárek & Anagnostidis, 1995 Nomenclature: Synechococcus curtus Notes: Anagnostidis and Golubic 1966 Rhabdoderma lineare Schmidle & Lauterborn in Schmidle, 1900 Nomenclature: Synechococcus linearis Notes: Moustaka 1988 Rhabdogloea cf. elenkinii (Roll) Komárek & Anagnostidis, 1995 Nomenclature: Dactylococcopsis cf. elenkinii Notes: Hindak and Moustaka 1988 Notes: Anagnostidis and Economou-Amilli 1978 Pseudoscytonema endolithicum (Ercegovic) Anagnostidis in Ko Anagnostidis, 2005 Nomenclature: Plectonema endolithicum Notes: Anagnostidis 1968 Radiocystis geminata Skuja, 1948 Nomenclature: Radiocystis geminata Notes: Anagnostidis 1968 Raphidiopsis mediterranea Skuja, 1937 Nomenclature: Raphidiopsis mediterranea Notes: Anagnostidis 1968 Rhabdoderma curtum (Setchell) Komárek & Anagnostidis, 1995 Nomenclature: Synechococcus curtus Notes: Anagnostidis and Golubic 1966 Rhabdoderma lineare Schmidle & Lauterborn in Schmidle, 1900 Nomenclature: Synechococcus linearis Notes: Moustaka 1988 Rhabdogloea cf. elenkinii (Roll) Komárek & Anagnostidis, 1995 Nomenclature: Dactylococcopsis cf. elenkinii Notes: Hindak and Moustaka 1988 Gkelis S et al. 74 Rivularia atra Roth ex Bornet & Flahault, 1886 Nomenclature: Rivularia atra Notes: Anagnostidis 1968 Rivularia biasolettiana Meneghini ex Bornet & Flahault, 1886 Nomenclature: Rivularia biasoletiana Notes: Anagnostidis 1968 Rivularia bullata Berkeley ex Bornet & Flahault, 1886 Nomenclature: Rivularia bullata Notes: Anagnostidis 1968 Rivularia calcarata (Woronichin) Poljankij in Hollerbach et al. 1953 Nomenclature: Dichothrix compacta var. calcarata Notes: Anagnostidis 1968 Rivularia haematites C. Agardh ex Bornet & Flahault, 1886 Nomenclature: Rivularia haematites Notes: Anagnostidis 1968 Rivularia nitida C. Agardh ex Bornet & Flahault, 1886 Nomenclature: Rivularia nitida Notes: Anagnostidis 1968 Rivularia polyotis Roth ex Bornet & Flahault, 1886 Nomenclature: Rivularia polyotis 75 Notes: Anagnostidis and Golubic 1966 Romeria gracilis (Koczwara) Koczwara in Geitler, 1932 Nomenclature: Romeria gracilis Notes: Tryfon et al. 1996 Romeria simplex (Hindák) Hindák, 1988 Nomenclature: Romeria simplex Notes: Katsiapi et al. 2012 Schizothrix affinis Lemmermann, 1905 Nomenclature: Schizothrix aﬃnis Notes: Anagnostidis 1968 Schizothrix arenaria Gomont, 1892 Nomenclature: Schizothrix arenaria Notes: Anagnostidis 1968 Schizothrix bosniaca (Forti) Geitler, 1932 Nomenclature: Schizothrix bosniaca Notes: Anagnostidis et al. 1983 Schizothrix calcicola Gomont, 1892 Nomenclature: Schizothrix calcicola Notes: Anagnostidis 1968 Cyanobacteria of Greece: an annotated checkl Cyanobacteria of Greece: an annotated checklist Notes: Anagnostidis and Golubic 1966 Romeria gracilis (Koczwara) Koczwara in Geitler, 1932 Nomenclature: Romeria gracilis Notes: Tryfon et al. 1996 Romeria simplex (Hindák) Hindák, 1988 Nomenclature: Romeria simplex Notes: Katsiapi et al. 2012 Schizothrix affinis Lemmermann, 1905 Nomenclature: Schizothrix aﬃnis Notes: Anagnostidis 1968 Schizothrix arenaria Gomont, 1892 Nomenclature: Schizothrix arenaria Notes: Anagnostidis 1968 Schizothrix bosniaca (Forti) Geitler, 1932 Nomenclature: Schizothrix bosniaca Notes: Anagnostidis et al. 1983 Schizothrix calcicola Gomont, 1892 Nomenclature: Schizothrix calcicola Notes: Anagnostidis 1968 Gkelis S et al. 76 Schizothrix coriacea Gomont, 1892 Nomenclature: Schizothrix coriacea Notes: Anagnostidis 1968 Schizothrix delicatissima W. & G. S. West, 1897 Nomenclature: Schizothrix delicatissima Notes: Anagnostidis 1968 Schizothrix fasciculata Gomont ex Gomont, 1892 Nomenclature: Schizothrix fasciculata Notes: Anagnostidis 1968 Schizothrix heufleri Grunow ex Gomont, 1892 Nomenclature: Schizothrix heuﬂeri Notes: Anagnostidis 1968 Schizothrix lacustris A. Braun ex Gomont, 1892 Nomenclature: Schizothrix lacustris Notes: Anagnostidis 1968 Schizothrix lardacea Gomont, 1892 Nomenclature: Schizothrix lardacea Notes: Anagnostidis 1968 Schizothrix lateritia Gomont, 1892 Nomenclature: Schizothrix lateritia 77 77 Cyanobacteria of Greece: an annotated checklist Notes: Anagnostidis 1968 Schizothrix lenormandiana Gomont, 1892 Nomenclature: Schizothrix lenormandiana Notes: Katsiapi et al. 2012 Schizothrix simplicior Skuja, 1964 Nomenclature: Schizothrix simplicior Notes: Anagnostidis 1968 Schizothrix tenuis Woronichin, 1923 Nomenclature: Schizothrix tenuis Notes: Anagnostidis 1968 Scytonema crispum Bornet ex De Toni, 1907 Nomenclature: Scytonema crispum st. chiastus Notes: Anagnostidis 1968 Scytonema hofmanii var. calcicola Agardh ex Bornet & Flahault, 1887 Nomenclature: Scytonema hofmanii var. calcicola Notes: Lamprinou et al. 2012 Scytonema julianum Meneghini ex B. A. Whitton, 2011 Nomenclature: Scytonema julianum Notes: Anagnostidis et al. 1982 Notes: Anagnostidis 1968 Schizothrix lenormandiana Gomont, 1892 Nomenclature: Schizothrix lenormandiana Notes: Katsiapi et al. 2012 Schizothrix simplicior Skuja, 1964 Nomenclature: Schizothrix simplicior Notes: Anagnostidis 1968 Schizothrix tenuis Woronichin, 1923 Nomenclature: Schizothrix tenuis Notes: Anagnostidis 1968 Scytonema crispum Bornet ex De Toni, 1907 Nomenclature: Scytonema crispum st. chiastus Notes: Anagnostidis 1968 Scytonema hofmanii var. calcicola Agardh ex Bornet & Flahault, 1887 Nomenclature: Scytonema hofmanii var. calcicola Notes: Lamprinou et al. 2012 Scytonema julianum Meneghini ex B. A. Whitton, 2011 Nomenclature: Scytonema julianum Notes: Anagnostidis et al. 1982 Scytonema julianum Meneghini ex B. A. Whitton, 2011 Nomenclature: Scytonema julianum Notes: Anagnostidis et al. 1982 Gkelis S et al. 78 Scytonema myochrous C. Agardh ex Bornet & Flahault, 1886 Nomenclature: Scytonema myochrous Notes: Anagnostidis 1968 Scytonema polycystum Bornet & Flahault, 1886 Nomenclature: Scytonema polycystum Notes: Anagnostidis 1968 Scytonema wolleanum Forti, 1907 Nomenclature: Scytonema mirabile Notes: Anagnostidis 1968 Scytonematopsis crustacea (Thuret ex Bornet & Flahault) Koválik & Komárek, 1988. Nomenclature: Calothrix crustacea Notes: Anagnostidis and Golubic 1966 Siphonema polonicum (Raciborski) Geitler, 1925 Nomenclature: Siphononema polonicum Notes: Anagnostidis 1968 Snowella atomus Komárek & Hindák, 1988 Nomenclature: Snowella atomus Notes: Danielidis et al. 1996 79 Cyanobacteria of Greece: an annotated checklist Snowella lacustris (Chodat) Komárek & Hindák, 1988 Nomenclature: Gomphosphaeria lacustris Notes: Anagnostidis 1968 Snowella litoralis (Häyrén) Komárek & Hindák, 1988 Nomenclature: Snowella litoralis Notes: Hindak and Moustaka 1988 Snowella septentrionalis Komárek & Hindák, 1988 Nomenclature: Snowella septentrionalis Notes: Montesanto et al. 1999 Solentia paulocellularis (Ercegovic) LeCampion-Alsumard & Golubic ex Belyakova, 1988 Nomenclature: Hormathonema paulocellulare Notes: Anagnostidis and Pantazidou 1988a Sphaerocavum microcystiforme (Hindák) Azevedo & Sant' Anna, 2003 Nomenclature: Pannus microcystiformis Notes: Hindák 1993 Sphaerospermopsis aphanizomenoides (Forti) Zapomelová, Jezberová, Hrouzek, Hisem, Reháková & Komárková, 2010 Nomenclature: Anabaena aphanizomenoides Notes: Moustaka-Gouni 1988 y Snowella lacustris (Chodat) Komárek & Hindák, 1988 Nomenclature: Gomphosphaeria lacustris Notes: Anagnostidis 1968 Snowella litoralis (Häyrén) Komárek & Hindák, 1988 Nomenclature: Snowella litoralis Notes: Hindak and Moustaka 1988 Snowella septentrionalis Komárek & Hindák, 1988 Nomenclature: Snowella septentrionalis Notes: Montesanto et al. 1999 Solentia paulocellularis (Ercegovic) LeCampion-Alsumard & Golubic ex Belyakova, 1988 Nomenclature: Hormathonema paulocellulare Notes: Anagnostidis and Pantazidou 1988a Sphaerocavum microcystiforme (Hindák) Azevedo & Sant' Anna, 2003 Nomenclature: Pannus microcystiformis Notes: Hindák 1993 Sphaerospermopsis aphanizomenoides (Forti) Zapomelová, Jezberová, Hrouzek, Hisem, Reháková & Komárková, 2010 Nomenclature: Anabaena aphanizomenoides Notes: Moustaka-Gouni 1988 Snowella lacustris (Chodat) Komárek & Hindák, 1988 Nomenclature: Gomphosphaeria lacustris Notes: Anagnostidis 1968 Gkelis S et al. 80 Spirulina caldaria Tilden, 1898 Nomenclature: Spirulina caldaria Notes: Anagnostidis 1961 Spirulina corakiana Playfair, 1914 Nomenclature: Spirulina corakiana Notes: Anagnostidis 1961 Spirulina labyrinthiformis Gomont, 1892 Nomenclature: Spirulina labyrinthiformis Notes: Anagnostidis and Golubic 1966 Spirulina major Kützing ex Gomont, 1892 Nomenclature: Spirulina major Notes: Anagnostidis 1961 Spirulina meneghiniana Zanardini ex Gomont, 1892 Nomenclature: Spirulina meneghiniana Notes: Anagnostidis and Golubic 1966 Spirulina princeps W. & G. S. West, 1902 Nomenclature: Spirulina princeps Notes: Anagnostidis 1961 Spirulina robusta H. Welsh, 1965 Nomenclature: Spirulina subsalsa var. crassior Spirulina caldaria Tilden, 1898 Nomenclature: Spirulina caldaria Notes: Anagnostidis 1961 Spirulina corakiana Playfair, 1914 Nomenclature: Spirulina corakiana Notes: Anagnostidis 1961 Spirulina labyrinthiformis Gomont, 1892 Nomenclature: Spirulina labyrinthiformis Notes: Anagnostidis and Golubic 1966 Spirulina major Kützing ex Gomont, 1892 Nomenclature: Spirulina major Notes: Anagnostidis 1961 Spirulina meneghiniana Zanardini ex Gomont, 1892 Nomenclature: Spirulina meneghiniana Notes: Anagnostidis and Golubic 1966 Spirulina princeps W. & G. S. West, 1902 Nomenclature: Spirulina princeps Notes: Anagnostidis 1961 Spirulina robusta H. Welsh, 1965 Nomenclature: S i li b l i 81 Cyanobacteria of Greece: an annotated checklist 81 Notes: Anagnostidis 1961 Notes: Anagnostidis 1961 Spirulina subsalsa Oersted ex Gomont, 1892 Nomenclature: Spirulina subsalsa Notes: Anagnostidis 1961 Spirulina subtilissima Kützing ex Gomont, 1892 Nomenclature: Spirulina subtilissima Notes: Anagnostidis 1961 Spirulina tenerrima Kützing ex Gomont, 1892 Nomenclature: Spirulina tenerrima Notes: Anagnostidis 1961 Spirulina versicolor Cohn ex Gomont, 1892 Nomenclature: Spirulina subsalsa st. versicolor Notes: Anagnostidis 1968 Stanieria sublitoralis (Lindstedt) Anagnostidis & Pantazidou, 1991 Nomenclature: Staniera sublitoralis Notes: Anagnostidis and Pantazidou 1991a Stanieria sphaerica (Setchel & Gardner) Anagnostidis & Pantazido Nomenclature: Dermocarpa sphaerica Notes: Anagnostidis and Golubic 1966 Gkelis S et al. 82 Stichosiphon pseudopolymorphus (F. E. Fritsch) Komárek, 1989 Nomenclature: Chamaesiphon pseudopolymorphus Notes: Anagnostidis 1968 Stigonema hormoides Bornet & Flahault, 1886 Nomenclature: Stigonema hormoides Notes: Anagnostidis and Economou-Amilli 1978 Stigonema mamillosum C. Agardh ex Bornet & Flahault, 1887 Nomenclature: Stigonema mamillosum Notes: Anagnostidis 1968 Stigonema minutum Hassall ex Bornet & Flahault, 1887 Nomenclature: Stigonema minutum Notes: Anagnostidis 1968 Symploca dubia Gomont, 1892 Nomenclature: Symploca dubia Notes: Anagnostidis 1968 Symploca elegans Kützing ex Gomont, 1892 Nomenclature: Symploca elegans Notes: Anagnostidis 1968 Symploca hydnoides Kützing ex Gomont, 1892 Nomenclature: Symploca hydnoides 82 Gkelis S et al. Stichosiphon pseudopolymorphus (F. E. Fritsch) Komárek, 1989 Nomenclature: Chamaesiphon pseudopolymorphus Notes: Anagnostidis 1968 Stigonema hormoides Bornet & Flahault, 1886 Nomenclature: Stigonema hormoides Notes: Anagnostidis and Economou-Amilli 1978 Stigonema mamillosum C. Agardh ex Bornet & Flahault, 1887 Nomenclature: Stigonema mamillosum Notes: Anagnostidis 1968 Stigonema minutum Hassall ex Bornet & Flahault, 1887 Nomenclature: Stigonema minutum Notes: Anagnostidis 1968 Symploca dubia Gomont, 1892 Nomenclature: Symploca dubia Notes: Anagnostidis 1968 Symploca elegans Kützing ex Gomont, 1892 Nomenclature: Symploca elegans Notes: Anagnostidis 1968 Symploca hydnoides Kützing ex Gomont, 1892 Symploca dubia Gomont, 1892 83 Cyanobacteria of Greece: an annotated checklist Notes: Anagnostidis 1968 Symploca muscorum Gomont ex Gomont, 1892 Nomenclature: Symploca muscorum Notes: Anagnostidis 1968 Symploca radians (Kützing) Rabenhorst ex Gomont, 1892 Nomenclature: Symploca radians Notes: Lamprinou et al. 2012 Symplocastrum penicillatum (Gomont) Anagnostidis, 2001 Nomenclature: Schizothrix penicillata Notes: Anagnostidis 1968 Gkelis S et al. 84 Synechococcus salinarum Komárek, 1956 Nomenclature: Synechoccus salinarum Notes: Anagnostidis and Roussomoustakaki 1991 Synechococcus bigranulatus Skuja, 1933 Nomenclature: Synechococcus elongatus f. thermalis Notes: Anagnostidis 1961 Synechococcus elongatus Nägeli, 1849 Nomenclature: Synechococcus elongatus Notes: Anagnostidis and Golubic 1966 Synechocystis aquatilis Sauvageau, 1892 Nomenclature: Synechocystis aquatilis Notes: Anagnostidis 1961 Synechocystis crassa Woronichin, 1929 Nomenclature: Synechocystis crassa Notes: Anagnostidis 1968 Synechocystis diplococca (Pringsheim) Bourrelly, 1970 Nomenclature: Synechocystis diplococcus Notes: Anagnostidis and Economou-Amilli 1978 Synechocystis endobiotica (Elenkin & Hollerbach) Elenkin, Nomenclature: Aphanocapsa endophytica 84 Gkelis S et al. Synechococcus salinarum Komárek, 1956 Nomenclature: Synechoccus salinarum Notes: Anagnostidis and Roussomoustakaki 1991 Synechococcus bigranulatus Skuja, 1933 Nomenclature: Synechococcus elongatus f. thermalis Notes: Anagnostidis 1961 Synechococcus elongatus Nägeli, 1849 Nomenclature: Synechococcus elongatus Notes: Anagnostidis and Golubic 1966 Synechocystis aquatilis Sauvageau, 1892 Nomenclature: Synechocystis aquatilis Notes: Anagnostidis 1961 Synechocystis crassa Woronichin, 1929 Nomenclature: Synechocystis crassa Notes: Anagnostidis 1968 Synechocystis diplococca (Pringsheim) Bourrelly, 197 Nomenclature: Synechocystis diplococcus Notes: Anagnostidis and Economou-Amilli 1978 Synechocystis endobiotica (Elenkin & Hollerbach) Ele Nomenclature: Synechocystis aquatilis Sauvageau, 1892 Nomenclature: Synechocystis aquatilis Notes: Anagnostidis 1961 85 Cyanobacteria of Greece: an annotated checklist Notes: Anagnostidis 1968 Synechocystis minuscula Woronichin, 1926 Nomenclature: Synechocystis minuscula Notes: Anagnostidis 1961 Synechocystis pevalekii Ercegovic 1925 Nomenclature: Synechocystis pevalekii Notes: Anagnostidis 1968 Synechocystis salina Wislouch, 1924 Nomenclature: Synechocystis salina Notes: Anagnostidis 1961 Synechocystis sallensis Skuja, 1930 Nomenclature: Synechocystis sallensis Notes: Metaxatos et al. 2003 Synechocystis thermalis J. J. Copeland, 1936 Nomenclature: Synechocystis thermalis Notes: Anagnostidis 1961 Tapinothrix crustacea (Woronichin) Bohunická & Johansen Johansen & Fuciková, 2011 Nomenclature: Homoeothrix crustacea Notes: Anagnostidis 1968 Notes: Anagnostidis 1968 Notes: Anagnostidis 1968 Synechocystis minuscula Woronichin, 1926 Nomenclature: Synechocystis minuscula Notes: Anagnostidis 1961 Synechocystis pevalekii Ercegovic 1925 Nomenclature: Synechocystis pevalekii Notes: Anagnostidis 1968 Synechocystis salina Wislouch, 1924 Nomenclature: Synechocystis salina Notes: Anagnostidis 1961 Synechocystis sallensis Skuja, 1930 Nomenclature: Synechocystis sallensis Notes: Metaxatos et al. 2003 Synechocystis thermalis J. J. Copeland, 1936 Nomenclature: Synechocystis thermalis Notes: Anagnostidis 1961 Tapinothrix crustacea (Woronichin) Bohunická Johansen & Fuciková, 2011 Nomenclature: Homoeothrix crustacea Notes: Anagnostidis 1968 Gkelis S et al. 86 Tapinothrix varians (Geitler) Bohunická & J. R. Johanse Johansen & Fuciková, 2011 Nomenclature: Homoeothrix varians Notes: Anagnostidis 1968 Tolypothrix cavernicola Weber-van Bosse, 1913 Nomenclature: Tolypothrix cavernicola Notes: Lamprinou et al. 2012 Tolypothrix distorta Kützing ex Bornet & Flahault, 1886 Nomenclature: Tolypothrix distorta Notes: Anagnostidis 1968 Tolypothrix elenkinii Hollerbach, 1923 Nomenclature: Tolypothrix elenkinii Notes: Anagnostidis 1968 Tolypothrix fragilissima Ercegovic, 1925 Nomenclature: Tolypothrix fragilissima Notes: Lamprinou et al. 2012 Tolypothrix penicillata Thuret ex Bornet & Flahault, 1886 Nomenclature: Tolypothrix penicillata Notes: Anagnostidis 1968 87 Cyanobacteria of Greece: an annotated checklist 87 Tolypothrix rivularis Hansgirg, 1892 Nomenclature: Tolypothrix rivularis Notes: Lamprinou et al. 2012 Toxopsis calypsus Lamprinou, Skaraki, Kotoulas, Economou Pantazidou, 2012 Nomenclature: Toxopsis calypsus Notes: Lamprinou et al. 2012 Trichocoleus delicatulus (W. & G. S. West) Anagnostidis, 2001 Nomenclature: Microcoleus delicatulus Notes: Anagnostidis 1968 Trichocoleus minimus (Frémy) Anagnostidis, 2001 Nomenclature: Microcoleus minimus Notes: Anagnostidis et al. 1982 Trichocoleus sociatus (W. & G. S. West) Anagnostidis, 2001 Nomenclature: Microcoleus sociatus Notes: Anagnostidis 1961 Trichocoleus tenerrimus (Gomont) Anagnostidis, 2001 Nomenclature: Microcoleus tenerrimus Notes: Anagnostidis and Economou-Amilli 1978 Tolypothrix rivularis Hansgirg, 1892 & Gkelis S et al. 88 Trichocoleus voukii (Frémy) Anagnostidis, 2001 Nomenclature: Microcoleus voukii Notes: Anagnostidis 1968 Trichodesmium cf. iwanoffianum Nygaard, 1926 Nomenclature: Oscillatoria cf. iwanoﬃana Notes: Moustaka 1988 Trichodesmium erythraeum Ehrenberg ex Gomont, 1892 Nomenclature: Trichodesmium erythraeum Notes: Metaxatos et al. 2003 Trichormus azollae (Strasburger) Komárek & Anagnostidis, 1989 Nomenclature: Anabaena azollae Notes: Anagnostidis et al. 1988b Trichormus thermalis (V.Vouk) Komárek & Anagnostidis, 1989 Nomenclature: Anabaena thermalis Notes: Anagnostidis 1961 Trichormus variabilis (Kützing ex Bornet & Flahault) Komárek & An 1989 Nomenclature: Anabaena variabilis Notes: Anagnostidis 1968 88 Gkelis S et al. 89 Cyanobacteria of Greece: an annotated checklist 89 Woronichinia compacta (Lemmermann) Komárek & Hindák, 1988 Nomenclature: Woronichinia compacta Notes: Tryfon et al. 1996 Woronichinia naegeliana (Unger) Elenkin, 1933 Nomenclature: Coelosphaerium nägelianum Notes: Anagnostidis 1968 Xenococcus cf. pyriformis Setchell & N. L. Gardner in N. L. Gardner, 1918 Nomenclature: Xenococcus cf. pyriformis Notes: Radea et al. 2010 Xenococcus schousboei Thuret in Bornet & Thuret, 1880 Nomenclature: Xenococcus schousboei Notes: Anagnostidis and Golubic 1966 Xenotholos kerneri (Hansgirg) M. Gold-Morgan, G. Montejano & J. Komáre 1994 Nomenclature: Xenococcus kerneri Notes: Anagnostidis and Golubic 1966 Yonedaella lithophila (Ercegovic) Umezaki, 1962 Nomenclature: Sphaeronema lithophilum Notes: Anagnostidis 1968 Cyanobacteria of Greece: an annotated checklist Woronichinia compacta (Lemmermann) Komárek & Hindák, 1988 Nomenclature: Woronichinia compacta Notes: Tryfon et al. 1996 Woronichinia naegeliana (Unger) Elenkin, 1933 Nomenclature: Coelosphaerium nägelianum Notes: Anagnostidis 1968 Xenococcus cf. pyriformis Setchell & N. L. Gardner in N. L. Gardner, 1918 Nomenclature: Xenococcus cf. pyriformis Notes: Radea et al. 2010 Xenococcus schousboei Thuret in Bornet & Thuret, 1880 Nomenclature: Xenococcus schousboei Notes: Anagnostidis and Golubic 1966 Xenotholos kerneri (Hansgirg) M. Gold-Morgan, G. Montejano & J. Komárek 1994 Nomenclature: Xenococcus kerneri Notes: Anagnostidis and Golubic 1966 Yonedaella lithophila (Ercegovic) Umezaki, 1962 Nomenclature: Sphaeronema lithophilum Notes: Anagnostidis 1968 Gkelis S et al. Gkelis S et al. Gkelis S et al. 90 Analysis A total of 543 species and 85 infraspeciﬁc taxa have been found in Greece up to date. The species are classiﬁed in 130 genera, 41 families, and 8 orders. The orders Synechococcales and Oscillatoriales have the highest number of species (158 and 153 species, respectively), whereas these two orders along with Nostocales and Chroococcales cover 93% of the known Greek cyanobacteria species (Fig. 1). The families Oscillatoriaceae, Leptolyngbyaceae, Microcoleaceae, and Chroococcaceae have the highest number of species (>30 each) (Fig. 2a); the genera Phormidium, Leptolyngbya, Oscillatoria, Chroococcus, Aphanocapsa, and Gloeocapsa have the highest number of species (>15 each) (Fig. 2b). Fourteen species were found to include 85 infraspeciﬁc taxa ( Table 1); those infraspeciﬁc taxa were not included in the checklist. Furthermore, 21 species recorded in Greece are not valid or of uncertain taxonomic status (Table 2) and were not included in the checklist. List of infraspeciﬁc cyanobacterial taxa known to occur in Greece. For each taxon the ﬁ publication mentioning its occurrence in Greece is given. Species Infraspeciﬁc epithet Reference Anabaena torulosa var. tenuis Anagnostidis 1961 Anabaena oscillarioides fa. circinalis Anagnostidis 1968 Aphanothece microscopica fa. endophytica Anagnostidis 1968 Entophysalis deusta st. solentioides Anagnostidis 1968 st. dermatocarpoides Anagnostidis 1968 st. dermocarpoides Anagnostidis 1968 st. gloeocapsoides Anagnostidis 1968 st. hormathonematoides Anagnostidis 1968 st. hyelloides Anagnostidis 1968 st. pleurocapsoides Anagnostidis 1968 st. tryponematoides Anagnostidis 1968 st. typicus Anagnostidis 1968 Gloeocapsa biformis st. derm. col. ralfsianus Anagnostidis 1968 st. derm. magma Anagnostidis 1968 st. dermochrous Anagnostidis 1968 st. nannocytosus Anagnostidis 1968 st. perdurans Anagnostidis 1968 st. punctatus Anagnostidis 1968 Gloeocapsa compacta st. perdurans Anagnostidis 1968 st. lam. col. magma Anagnostidis 1968 Cyanobacteria of Greece: an annotated checklist The current taxonomic status is Nostoc punctiforme fa. populorum 1 Cyanobacteria of Greece: an annotated checklist 91 st. lam. col. typicus Anagnostidis 1968 st. lam. coloratus Anagnostidis 1968 st. nannocytosus Anagnostidis 1968 st. perdurans Anagnostidis 1968 st. simplex Anagnostidis 1968 Gloeocapsa kützingiana st. lam. col. magma Anagnostidis 1968 st. lam. coloratus Anagnostidis 1968 st. perdurans Anagnostidis 1968 st. rupestris Anagnostidis 1968 Gloeocapsa sanguinea st. col. alp. magma Anagnostidis 1968 st. col. alp. ralfsianus Anagnostidis 1968 st. col. ralfsianus Anagnostidis 1968 st. lam. col. alp. magma Anagnostidis 1968 st. lam. col. alpinus Anagnostidis 1968 st. lam. col. magma Anagnostidis 1968 st. lam. coloratus Anagnostidis 1968 st. perdurans Anagnostidis 1968 st. simplex Anagnostidis 1968 Mastigocladus laminosus fa. oscillariodes subfa. subrecta Anagnostidis 1961 fa. oscillariodes subfa. circinata Anagnostidis 1961 fa. oscillariodes subfa. spiroides Anagnostidis 1961 fa. pseudanabaenoides Anagnostidis 1961 fa. pseudanabaenoides subfa. subrecta Anagnostidis 1961 fa. pseudanabaenoides subfa. circinata Anagnostidis 1961 fa. pseudanabaenoides subfa. spiroides Anagnostidis 1961 fa. lyngbyoides Anagnostidis 1961 fa. lyngbyoides subfa. subrecta Anagnostidis 1961 fa. lyngbyoides subfa. circinata Anagnostidis 1961 fa. phormidioides Anagnostidis 1961 fa. phormidioides subfa. subrecta Anagnostidis 1961 fa. phormidioides subfa. circinata Anagnostidis 1961 fa. plectonematoides Anagnostidis 1961 fa. symplocoides Anagnostidis 1961 fa. anabaenoides Anagnostidis 1961 fa. anabaenoides subfa. subrecta Anagnostidis 1961 fa. anabaenoides subfa. circinata Anagnostidis 1961 fa. anabaenoides subfa. spiroides Anagnostidis 1961 92 Gkelis S et al. Gkelis S et al. Gkelis S et al. fa. nostocoides Anagnostidis 1961 fa. aulosiroides Anagnostidis 1961 fa. stigonematoides Anagnostidis 1961 fa. microchaetoides Anagnostidis 1961 fa. tolypotrichoides Anagnostidis 1961 fa. scytonematoides Anagnostidis 1961 fa. typica Anagnostidis 1961 fa. typica subfa. normalis Anagnostidis 1961 fa. typica subfa. brachytrichoides Anagnostidis 1961 fa. pallidus Anagnostidis 1961 fa. oscillariodes Anagnostidis 1961 fa. chroococcoides Anagnostidis 1961 fa. pleurocapsoides Anagnostidis 1961 Nostoc punctiforme var. populorum Anagnostidis and Economou-Amilli 1978 Pseudanabaena articulata fa. Anagnostidis 1968 Pseudanabaena lonchoides fa. circinata Anagnostidis 1961 fa. crassior Anagnostidis 1968 fa. tenuis Anagnostidis 1968 Scytonema myochrous st. crustaceus Anagnostidis 1968 st. petalonema Anagnostidis 1968 st. petalonema Anagnostidis 1968 st. petalonema Anagnostidis 1968 st. typicus Anagnostidis 1968 Siphononema polonicum st. chamaesiphonoides Anagnostidis 1968 st. juvenilis Anagnostidis 1968 st. scopulonematoides Anagnostidis 1968 st. stigonematoides Anagnostidis 1968 Synechocystis aquatilis var. minor Anagnostidis 1961 The current taxonomic status is Nostoc punctiforme fa. populorum 1 2 1 Cyanobacteria of Greece: an annotated checklist Cyanobacteria of Greece: an annotated checklist 93 Table 2. List of cyanobacterial taxa known to occur in Greece with uncertain or not valid taxonomic status. Cyanobacteria of Greece: an annotated checklist For each taxon the publication mentioning its occurrence in Greece and the current taxonomic status is given. List of cyanobacterial taxa known to occur in Greece with uncertain or not valid taxonomic status. For each taxon the publication mentioning its occurrence in Greece and the current taxonomic status is given. Taxon Reference Current taxonomic status Achroonema lentum Anagnostidis and Economou–Amilli 1980 unclassiﬁed below the Kingdom level (Guiry and Guiry 2016) Achroonema proteiforme Anagnostidis 1968 Achroonema angustum Anagnostidis 1968 Achroonema atriculatum Anagnostidis 1968 Achroonema splendens Anagnostidis 1968 Achroonema subsalsum Anagnostidis 1968 Desmanthos thiocrenophilum (?) Anagnostidis 1968 Pelonema aphane Anagnostidis 1968 Pelonema cf. subtilissimum Anagnostidis and Roussomoustakaki 1991 Pelonema pseudovacuolatum Anagnostidis 1968 Pelonema subtilissimum Anagnostidis 1968 Pelonema tenue Anagnostidis 1968 Peloploca ferruginea (?) Anagnostidis 1968 Peloploca taeniata Anagnostidis 1968 Peloploca undulata Anagnostidis 1968 Pleurocapsa crepidinum Anagnostidis and Golubic 1966 needs revision; possibly a Chroococcidiopsis species (Komárek and Anagnostidis 2005 ) Aphanothece nidulans var. nidulans Hindak and Moustaka 1988 uncertain; Algabase (Guiry and Guiry 2016) contains only two valid A. nidulans varieties: var. longissima Nash and var. thermalis Hansgirg Hydrocoleus homoeotrichus Anagnostidis et al. 1983 needs revision; sensu Anagnostidis et al 1983, probably belongs to the Oscillatoriaceae according to trichome structure, and into the genus Blennothrix (Komárek and Anagnostidis 2005) Achroonema pseudangustum Anagnostidis and Economou–Amilli 1980 not valid (Guiry and Guiry 2016) Scolecia ﬁlosa Färber et al. 2015 Ichnotaxa produced by unknown heterotrophic organisms (Heindel et al. 2009) Gloeocapsa polydermatica Anagnostidis 1961 Sporotetras polydermatica [Chlorophyta] (Guiry and Guiry 2016 ) a b Figure 1. Classiﬁcation of cyanobacteria of Greece in orders. Figure 2. Classiﬁcation of cyanobacteria of Greece in families (a) and genera (b). Only families and genera including more than ﬁve species are shown. 94 Gkelis S et al. 94 Gkelis S et al. Classiﬁcation of cyanobacteria of Greece in orders. a b b a Classiﬁcation of cyanobacteria of Greece in families (a) and genera (b). Only families and genera including more than ﬁve species are shown. Discussion A thorough review of the relevant literature revealed that 543 species of cyanobacteria have been reported up to date from Greece. The species recorded from this area are classiﬁed in 130 genera and 41 families, all of the 8 orders of the class Cyanophyceae (Komárek et al. 2014) being represented. In this checklist we have not included 85 infraspeciﬁc taxa mostly from the genera Entophysalis, Gloeocapsa, Mastigocladus, Scytonema, and Siphononema recorded in several habitats such as gulfs, waterfalls, caves, and lakes by Anagnostidis (Anagnostidis 1961, Anagnostidis 1968). In the genus Mastigocladus numerous taxa were described, which do not belong to this genus (Komárek 2013). Among the formae described by Anagnostidis, some are subject to revision [e.g. M. Cyanobacteria of Greece: an annotated checklist 95 fa. microchaetoides is "surely not Mastigocladus, as it is morphologically similar to Aulosira" according to Komárek (2013)]. The great morphological variability of this chroococcalean and nostocalean species should be further investigated in the future. A number of species reported in Greece were found to be of not valid or of uncertain taxonomic status: The species Pleurocapsa crepidinum is possibly a Chroococcidiopsis species and needs revision (Komárek and Anagnostidis 2005). The species Hydrocoleus homoeotrichus sensu Anagnostidis et al. (1983), probably belongs to the Oscillatoriaceae and into the genus Blennothrix (Komárek and Anagnostidis 2005) according to its trichome structure. The taxon Aphanothece nidulans var. nidulans found in Lake Volvi (Hindak and Moustaka 1988) is not included either in the Chroococcales volume of Süsswasserﬂora von Mitteleuropa (Komárek and Anagnostidis 1999) nor in the AlgaeBase (Guiry and Guiry 2016); Algaebase contains only two valid A. nidulans varieties: var. longissima Nash and var. thermalis Hansgirg. Gloeocapsa polydermatica Kützing found in Nigrita hot springs (Anagnostidis 1961) is currently regarded as a taxonomic synonym of the chlorophyte Sporotetras polydermatica (Kützing) I.Kostikov, T.Darienko, A.Lukesová, & L.Hoﬀmann (Guiry and Guiry 2016). The taxon Scolecia ﬁlosa found in microbial bioerosion at an underwater marine site in Peloponnese (Färber et al. 2015) belongs to ichnotaxa produced by unknown heterotrophic organisms (Heindel et al. 2009), not cyanobacteria. The remaining 17 species belong to the family Pelonemataceae; this family, consisting of the genera Achroonema, Pelonema, Peloploca, and Desmanthos, was created to accommodate some colorless bacteria with multicellular ﬁlaments (trichomes), which were observed in hypolimnia of lakes and ponds or in sulfur springs (Hirsch 1981). Gkelis S et al. 96 Gkelis S et al. To the best of our knowledge, the list of Greek cyanobacteria compiled for the needs of the present study is the ﬁrst list of both freshwater and marine Cyanophyceae in the Mediterranean. There is another list presenting 327 species from Israel (Tsarenko et al. 2000), but this refers only to the continental part of that country. The Egyptian list (Nassar 2014) contains 19 cyanobacteria species, whereas a phytoplankton checklist from the Eastern Adriatic (Vilicic et al. 2002) contains no cyanobacteria. Other local cyanobacterial lists were published in the past 15 years from continental parts of Europe, e.g. Slovakia, Sweden, Poland, Romania, Ukraine, Slovenia, the Netherlands, and the Czech Republic (Kaštovský et al. 2010 and references therein), as well as a checklist from the Baltic Sea reporting 172 cyanobacterial species (Hällfors 2004). The Czech Republic list comprises 505 species, 392 of which are considered recently present according to Kaštovský et al. 2010. The cyanobacteria checklist of Nepal (Rai et al. 2011) presents 274 species. Therefore, it appears that Greece hosts a comparatively high diversity of cyanobacteria, suggesting that the Mediterranean area is also a hot spot for the “substantially underestimated or unexplored” (Coll et al. 2010) microbes. To the best of our knowledge, the list of Greek cyanobacteria compiled for the needs of the present study is the ﬁrst list of both freshwater and marine Cyanophyceae in the Mediterranean. There is another list presenting 327 species from Israel (Tsarenko et al. 2000), but this refers only to the continental part of that country. The Egyptian list (Nassar 2014) contains 19 cyanobacteria species, whereas a phytoplankton checklist from the Eastern Adriatic (Vilicic et al. 2002) contains no cyanobacteria. Other local cyanobacterial lists were published in the past 15 years from continental parts of Europe, e.g. Slovakia, Sweden, Poland, Romania, Ukraine, Slovenia, the Netherlands, and the Czech Republic (Kaštovský et al. 2010 and references therein), as well as a checklist from the Baltic Sea reporting 172 cyanobacterial species (Hällfors 2004). The Czech Republic list comprises 505 species, 392 of which are considered recently present according to Kaštovský et al. 2010. The cyanobacteria checklist of Nepal (Rai et al. 2011) presents 274 species. Therefore, it appears that Greece hosts a comparatively high diversity of cyanobacteria, suggesting that the Mediterranean area is also a hot spot for the “substantially underestimated or unexplored” (Coll et al. 2010) microbes. Acknowledgements This paper is dedicated to the memory of Professor Konstantinos Anagnostidis (12 December 1924-9 December 1994), whose pioneer work on cyanobacterial diversity has paved the way for modern cyanobacterial taxonomy and the study of Phycology in Greece. This study was supported by the LifeWatchGreece Infrastructure, (MIS 384676), funded by the Greek Government under the General Secretariat of Research and Technology (GSRT), ESFRI Projects, National Strategic Reference Framework (NSRF). SG gratefully acknowledges Prof. Eleni Voultsiadou for critically reading the manuscript, useful suggestions, and fruitful discussions. We thank Dr. Vasilis Gerovasileiou for his support. Author contributions SG conceived, designed and supervised this project, and prepared the manuscript. IO and NP collected the data. IO, MP, and SG and built the database. Discussion At present, members of this family are considered unclassiﬁed below the Kingdom level (Guiry and Guiry 2016). It is worth mentioning that 18 of the species included in the list have been initially described from Greek habitats. The marine epilithic Ammatoidea aegea described from Saronikos Gulf (Aegean Sea) (Anagnostidis and Pantazidou 1991a) is considered endemic to this area (Komarek and Anagnostidis 2005). Interestingly, most of these new species were found in rarely sampled habitats, such as caves [e.g. Toxopsis calypsus, Iﬁnoe spelaea (Lamprinou et al. 2011)], marine coastal rocks [e.g. epilithic Cyanosarcina thalassia (Anagnostidis and Pantazidou 1991b) or endolithic Cyanosaccus aegaeus and Cyanosaccus atticus (Anagnostidis and Pantazidou 1985, Anagnostidis and Pantazidou 1988c)], archaeological sites [e.g. aerophytic, chasmoendolithic Cyanosarcina parthenonensis (Anagnostidis and Pantazidou 1991b)], heliothermal saltwork mats [Cyanothece halobia (Anagnostidis and Roussomoustakaki 1991)], and waterfalls [Phormidium edessae (Skuja 1937)], highlighting the need for further research to reveal the unknown diversity in these environments. A recent study by Bravakos et al. (2016) on cyanobacteria strains isolated from mats of thermal springs of Greece indicates at least six new taxa at the genus level. Furthermore, Konstantinou et al. (2016) report possible new species on cyanobacteria strains isolated from marine sponges from the Aegean Sea, thus corroborating the abundant diversity of lesser-known environments. References • Anagnostidis K (1961) Investigations of the Cyanophyceae in some thermal springs in Greece. Institute of Systematic Botany and Phytogeography of the Aristotle University of Thessaloniki, Thessaloniki, 7: 1-322 pp. [In Greek]. • Anagnostidis K (1968) Investigations of Sulphur-communities (Sulphuretum) in marine and freshwater habitats. A taxonomic, ﬂoristic, ecological, phytosociological, Cyanobacteria of Greece: an annotated checklist 97 phytogeographic study. 10. Scientiﬁc annals of the Faculty of Sciences, Aristotle University of Thessaloniki, Thessaloniki, 406-866 pp. [In Greek]. • Anagnostidis K (1977) Some remarks on the taxonomy of the genus Borzia Cohn ex Gom. from Greece. Schweizerische Zeitschrift für Hydrologie 39: 121‑123. • Anagnostidis K, Economou-Amilli A (1978) Microorganisms from the volcano of Nea Kammeni Island (Santorini). Thera and the Aegean World 1: 707‑723. • Anagnostidis K, Economou–Amilli A (1980) Limnological studies on Lake Pamvotis (Ioannina), Greece I. Hydroclimatology, phytoplankton - periphyton with special reference to the valency of some microorganisms from sulphureta as bioindicators. Archiv für Hydrobiologie 89: 313‑342. Anagnostidis K, Economou–Amilli A (1980) Limnological studies on Lake Pamvotis (Ioannina), Greece I. Hydroclimatology, phytoplankton - periphyton with special reference to the valency of some microorganisms from sulphureta as bioindicators. Archiv für Hydrobiologie 89: 313‑342. • Anagnostidis K, Economou–Amilli A (1980) Limnological studies on Lake Pamvotis (Ioannina), Greece I. Hydroclimatology, phytoplankton - periphyton with special reference to the valency of some microorganisms from sulphureta as bioindicators. Archiv für Hydrobiologie 89: 313‑342. • Anagnostidis K, Golubic S (1966) Über die Okologie einiger Spirulina Arten. Nova Hedwigia 11: 309‑335. [In German]. • Anagnostidis K, Komárek J (1985) Modern approach to the classiﬁcation system of cyanophytes. 1- Introduction. Arch. Hydrobiol. Suppl. 71: 291‑302. Anagnostidis K, Komárek J (1988) Modern approach to the classiﬁcation system of cyanophytes. 3- Oscillatoriales. Archiv für Hydrobiologie, Supplement 80: 327‑472. • Anagnostidis K, Komárek J (1988) Modern approach to the classiﬁcation system of cyanophytes. 3- Oscillatoriales. Archiv für Hydrobiologie, Supplement 80: 327‑472. Anagnostidis K, Komárek J (1990) Modern approach to the classiﬁcation system of cyanophytes. 5- Stigonematales. Archiv für Hydrobiologie, Supplement 86: 1‑73. • Anagnostidis K, Komárek J (1990) Modern approach to the classiﬁcation system of cyanophytes. 5- Stigonematales. Archiv für Hydrobiologie, Supplement 86: 1‑73. • Anagnostidis K, Pantazidou A (1985) Cyanosaccus aegaeus n. sp., a new marine endolithic cyanophyte from the Aegean Sea, Hellas (Greece). Archiv für Hydrobiologie (Algological Studies) 38/39 (suppl. 71): 105‑114. Anagnostidis K, Pantazidou A (1985) Cyanosaccus aegaeus n. References sp., a new marine endolithic cyanophyte from the Aegean Sea, Hellas (Greece). Archiv für Hydrobiologie (Algological Studies) 38/39 (suppl. 71): 105‑114. • Anagnostidis K, Pantazidou A (1988a) Endolithic cyanophytes from the saline thermal springs of Aedipsos, Hellas (Greece). Algological Studies/Archiv für Hydrobiologie, Supplement 50-53 (suppl. 80): 555‑559. • Anagnostidis K, Pantazidou A (1988a) Endolithic cyanophytes from the saline thermal springs of Aedipsos, Hellas (Greece). Algological Studies/Archiv für Hydrobiologie, Supplement 50-53 (suppl. 80): 555‑559. • Anagnostidis K, Pantazidou A (1988b) Hyella kalligrammos sp. nov., Hyella maxima (Geitl.) comb. nov., and other freshwater morphotypes of the genus Hyella Born. et Flah. (Chroococcales, Cyanophyceae). Algological Studies/Archiv für Hydrobiologie, Supplement 50-53 (suppl. 80): 227‑247. Anagnostidis K, Pantazidou A (1988c) Cyanosaccus atticus, a new marine euendolithic chroococcoid cyanophyte in relation to the epilithic Podocapsa Erceg. Algological Studies/Archiv für Hydrobiologie, Supplement 48 (suppl. 78): 279‑302. • Anagnostidis K, Pantazidou A (1988c) Cyanosaccus atticus, a new marine euendolithic chroococcoid cyanophyte in relation to the epilithic Podocapsa Erceg. Algological Studies/Archiv für Hydrobiologie, Supplement 48 (suppl. 78): 279‑302. Anagnostidis K, Pantazidou A (1991a) Ammatoidea aegaea (Oscillatoriales), a new marine epilithic species from the Aegean Sea, Hellas, with a reference to the validity of the genus Ammatoidea . Algological Studies/Archiv für Hydrobiologie, Supplement 64: 281‑297. • Anagnostidis K, Pantazidou A (1991a) Ammatoidea aegaea (Oscillatoriales), a new marine epilithic species from the Aegean Sea, Hellas, with a reference to the validity of the genus Ammatoidea . Algological Studies/Archiv für Hydrobiologie, Supplement 64: 281‑297. • Anagnostidis K, Pantazidou A (1991b) Marine and aerophytic Cyanosarcina, Stanieria and Pseudocapsa (Chroococcales) species from Hellas (Greece). Algological Studies/ Archiv für Hydrobiologie, Supplement 64: 141‑157. • Anagnostidis K, Pantazidou A (1991b) Marine and aerophytic Cyanosarcina, Stanieria and Pseudocapsa (Chroococcales) species from Hellas (Greece). Algological Studies/ Archiv für Hydrobiologie, Supplement 64: 141‑157. • Anagnostidis K, Roussomoustakaki M (1988) Cyanophytes from metal burdened substrates. Algological Studies/Archiv für Hydrobiologie, Supplement 50-53 (suppl. 80): 561‑564. • Anagnostidis K, Roussomoustakaki M (1988) Cyanophytes from metal burdened substrates. Algological Studies/Archiv für Hydrobiologie, Supplement 50-53 (suppl. 80): 561‑564. • Anagnostidis K, Roussomoustakaki M (1991) Isocystis halobia spec. nova, a benthic nostocalean cyanophyte from the heliothermal saltwork mats of Messolongion, Hellas (Greece). Algological Studies 64: 299‑332. • Anagnostidis K, Roussomoustakaki M (1991) Isocystis halobia spec. nova, a benthic nostocalean cyanophyte from the heliothermal saltwork mats of Messolongion, Hellas (Greece). Algological Studies 64: 299‑332. References URL: http:// sparks eserver org/books/iliad pdf • Batler S (1998) Homer's Iliad. Orange Street Press Classics, 495 pp. URL: http:// sparks.eserver.org/books/iliad.pdf • Bravakos P, Kotoulas G, Skaraki K, Pantazidou A, Economou-Amilli A (2016) A polyphasic taxonomic approach in isolated strains of Cyanobacteria from thermal springs of Greece. Molecular Phylogenetics and Evolution 98: 147‑160. https:// • Christodoulou M, Meletiou-Christou M, Parmakelis A, Economou-Amilli A, Pantazidou A (2015) Further ﬁndings from Daveli Cave (Attica, Greece) enhancing the establishment of the genusOculatella (Pseudanabaenaceae, Cyanobacteria). Phytotaxa 202 (3): 169. https://doi.org/10.11646/phytotaxa.202.3.1 • Christodoulou M, Meletiou-Christou M, Parmakelis A, Economou-Amilli A, Pantazidou A (2015) Further ﬁndings from Daveli Cave (Attica, Greece) enhancing the establishment of the genusOculatella (Pseudanabaenaceae, Cyanobacteria). Phytotaxa 202 (3): 169. htt //d i /10 11646/ h t t 202 3 1 • Coll M, Piroddi C, Steenbeek J, Kaschner K, Rais Lasram FB, Aguzzi J, Ballesteros E, Bianchi CN, Corbera J, Dailianis T, Danovaro R, Estrada M, Froglia C, Galil B, Gasol J, Gertwagen R, Gil J, Guilhaumon F, Kesner-Reyes K, Kitsos M, Koukouras A, Lampadariou N, Laxamana E, la Cuadra CLd, Lotze H, Martin D, Mouillot D, Oro D, Raicevich S, Rius-Barile J, Saiz-Salinas JI, Vicente CS, Somot S, Templado J, Turon X, Vaﬁdis D, Villanueva R, Voultsiadou E (2010) The Biodiversity of the Mediterranean Sea: Estimates, Patterns, and Threats. PLoS ONE 5 (8): e11842. https:// • Coll M, Piroddi C, Steenbeek J, Kaschner K, Rais Lasram FB, Aguzzi J, Ballesteros E, Bianchi CN, Corbera J, Dailianis T, Danovaro R, Estrada M, Froglia C, Galil B, Gasol J, Gertwagen R, Gil J, Guilhaumon F, Kesner-Reyes K, Kitsos M, Koukouras A, Lampadariou N, Laxamana E, la Cuadra CLd, Lotze H, Martin D, Mouillot D, Oro D, Raicevich S, Rius-Barile J, Saiz-Salinas JI, Vicente CS, Somot S, Templado J, Turon X, Vaﬁdis D, Villanueva R, Voultsiadou E (2010) The Biodiversity of the Mediterranean Sea: Estimates, Patterns, and Threats. PLoS ONE 5 (8): e11842. https:// • Danielidis DB, Spartinou M, Economou-Amilli A (1996) Limnological survey of Lake Amvrakia, western Greece. Hydrobiologia 318 (3): 207‑218. https://doi.org/10.1007/ bf00016682 • Danielidis DB, Spartinou M, Economou-Amilli A (1996) Limnological survey of Lake Amvrakia, western Greece. Hydrobiologia 318 (3): 207‑218. https://doi.org/10.1007/ bf00016682 • Duval B, Duval E, Hoham RW (1999) Snow algae of the Sierra Nevada, Spain, and High Atlas mountains of Morocco. International microbiology : the oﬃcial journal of the Spanish Society for Microbiology 2 (1): 39‑42. References • Anagnostidis K, Economou-Amilli A, Makris K (1988a) On the morphotypes of Phormidium boryanum (Bory ex Gom.) Anagn. et Kom. and Phormidium janthiphorum • Anagnostidis K, Economou-Amilli A, Makris K (1988a) On the morphotypes of Phormidium boryanum (Bory ex Gom.) Anagn. et Kom. and Phormidium janthiphorum 98 Gkelis S et al. (Fiori-Mazz. ex Gom.) Elenk. A taxonomic consideration. Algological Studies/Archiv für Hydrobiologie, Supplement 50-53 (suppl. 80): 71‑91. (Fiori-Mazz. ex Gom.) Elenk. A taxonomic consideration. Algological Studies/Archiv für Hydrobiologie, Supplement 50-53 (suppl. 80): 71‑91. • Anagnostidis K, Economou-Amilli A, Overbeck J (1988b) Anabaena azollae Strasb. and the periphyton of Azolla ﬁliculoides Lam. in lake Trichonis and the lagoon of Aetolikon, Hellas (Greece). Archiv für Hydrobiologie. 133 (3): 321‑343. • Anagnostidis K, Economou-Amilli A, Pantazidou A (1982) Studies on microﬂora of Perama Cave, Ioannina, Greece. Bulletin de la Société Spéléologique de Grèce 18 (1-2): 458‑530. • Anagnostidis K, Economou-Amilli A, Roussomoustakaki M (1983) Epilithic and chasmolithic microﬂora (Cyanophyta, Bacillariophyta) from marbles of the Parthenon- Acropolis, Athens, Greece. Nova Hedwigia 38: 227‑288. p g • Anagnostidis K, Economou-Amilli A, Tsangridis A (1981) Taxonomic and ﬂoristic studies of algae from rice-ﬁelds of Kalochorion-Thessaloniki, Greece. Nova Hedwigia 34: 1‑66. • Ananiadis CI (1956) Limnological study of Lake Karla Bulletin de L'Institut Anagnostidis K, Economou-Amilli A, Tsangridis A (1981) Taxonomic and ﬂoristic studies of algae from rice-ﬁelds of Kalochorion-Thessaloniki, Greece. Nova Hedwigia 34: 1‑66. • Anagnostidis K, Economou-Amilli A, Tsangridis A (1981) Taxonomic and ﬂoristic studies of algae from rice-ﬁelds of Kalochorion-Thessaloniki, Greece. Nova Hedwigia 34: 1‑66. • Ananiadis CI (1956) Limnological study of Lake Karla. Bulletin de L'Institut Océanographique 1083: 1‑19. of algae from rice-ﬁelds of Kalochorion-Thessaloniki, Greece. Nova Hedwigia 34: 1‑66. Ananiadis CI (1956) Limnological study of Lake Karla. Bulletin de L'Institut Océanographique 1083: 1‑19. • Ananiadis CI (1956) Limnological study of Lake Karla. Bulletin de L'Institut Océanographique 1083: 1‑19. • Bailly N, Gerovasileiou V, Arvanitidis C, Legakis A (2016) Introduction to the Greek Taxon Information System (GTIS) in LifeWatchGreece: the construction of the Preliminary Checklists of Species of Greece. Biodiversity Data Journal in press: e. • Bailly N, Gerovasileiou V, Arvanitidis C, Legakis A (2016) Introduction to the Greek Taxon Information System (GTIS) in LifeWatchGreece: the construction of the Preliminary Checklists of Species of Greece. Biodiversity Data Journal in press: e. • Batler S (1998) Homer's Iliad. Orange Street Press Classics, 495 pp. References • Economou-Amilli A, Anagnostidis K, Roussomoustakaki M (1984) Structural Aspects of the Adaptation of Some Blue-Green Algae and Diatoms to Desiccation. Tasks for vegetation science. https://doi.org/10.1007/978-94-009-6578-2_13 Färber C, Wisshak M, Pyko I, Bellou N, Freiwald A (2015) Eﬀects of Water Depth, Seasonal Exposure, and Substrate Orientation on Microbial Bioerosion in the Ionian • Färber C, Wisshak M, Pyko I, Bellou N, Freiwald A (2015) Eﬀects of Water Depth, Seasonal Exposure, and Substrate Orientation on Microbial Bioerosion in the Ionian Cyanobacteria of Greece: an annotated checklist 99 • Gkelis S, Zaoutsos N (2014) Cyanotoxin occurrence and potentially toxin producing cyanobacteria in freshwaters of Greece: A multi-disciplinary approach. Toxicon 78: 1‑9. https://doi.org/10.1016/j.toxicon.2013.11.010 • Gkelis S, Lanaras T, Sivonen K (2015a) Cyanobacterial Toxic and Bioactive Peptides in Freshwater Bodies of Greece: Concentrations, Occurrence Patterns, and Implications for Human Health. Marine Drugs 13 (10): 6319‑6335. https://doi.org/10.3390/ md13106319 • Gkelis S, Lanaras T, Sivonen K (2015a) Cyanobacterial Toxic and Bioactive Peptides in Freshwater Bodies of Greece: Concentrations, Occurrence Patterns, and Implications for Human Health. Marine Drugs 13 (10): 6319‑6335. https://doi.org/10.3390/ md13106319 • Gkelis S, Tussy PF, Zaoutsos N (2015b) Isolation and preliminary characterization of cyanobacteria strains from freshwaters of Greece. Open Life Sciences 10 (1): 52‑60. htt ps://doi.org/10.1515/biol-2015-0006 • Gkelis S, Tussy PF, Zaoutsos N (2015b) Isolation and preliminary characterization of cyanobacteria strains from freshwaters of Greece. Open Life Sciences 10 (1): 52‑60. htt ps://doi.org/10.1515/biol-2015-0006 • Gkelis S, Moustaka-Gouni M, Sivonen K, Lanaras T (2005) First report of the cyanobacterium Aphanizomenon ovalisporum Forti in two Greek lakes and cyanotoxin occurrence. Journal of Plankton Research 27 (12): 1295‑1300. https://doi.org/10.1093/ plankt/fbi085 • Gkelis S, Moustaka-Gouni M, Sivonen K, Lanaras T (2005) First report of the cyanobacterium Aphanizomenon ovalisporum Forti in two Greek lakes and cyanotoxin occurrence. Journal of Plankton Research 27 (12): 1295‑1300. https://doi.org/10.1093/ plankt/fbi085 • Gkelis S, Papadimitriou T, Zaoutsos N, Leonardos I (2014) Anthropogenic and climate- induced change favors toxic cyanobacteria blooms: Evidence from monitoring a highly eutrophic, urban Mediterranean lake. Harmful Algae 39: 322‑333. https:// doi org/10 1016/j hal 2014 09 002 • Gkelis S, Papadimitriou T, Zaoutsos N, Leonardos I (2014) Anthropogenic and climate- induced change favors toxic cyanobacteria blooms: Evidence from monitoring a highly eutrophic, urban Mediterranean lake. Harmful Algae 39: 322‑333. References https:// d i /10 1016/j h l 2014 09 002 • Gkelis S, Rajaniemi P, Vardaka E, Moustaka-Gouni M, Lanaras T, Sivonen K (2005) Limnothrix redekei (Van Goor) Meﬀert (Cyanobacteria) Strains from Lake Kastoria, Greece Form a Separate Phylogenetic Group. Microbial Ecology 49 (1): 176‑182. http s://doi.org/10.1007/s00248-003-2030-7 • Gkelis S, Rajaniemi P, Vardaka E, Moustaka-Gouni M, Lanaras T, Sivonen K (2005) Limnothrix redekei (Van Goor) Meﬀert (Cyanobacteria) Strains from Lake Kastoria, Greece Form a Separate Phylogenetic Group. Microbial Ecology 49 (1): 176‑182. http s://doi.org/10.1007/s00248-003-2030-7 • Guiry MD, Guiry MD (2016) AlgaeBase. World-wide electronic publication, National University of Ireland, Galway. http://www.algaebase.org. Accession date: 2016 7 16. • Hällfors G (2004) Checklist of Baltic Sea Phytoplankton Species. 95. Baltic Sea • Guiry MD, Guiry MD (2016) AlgaeBase. World-wide electronic publication, National University of Ireland, Galway. http://www.algaebase.org. Accession date: 2016 7 16. • Hällfors G (2004) Checklist of Baltic Sea Phytoplankton Species. 95. Baltic Sea Environment Proceedings, 210 pp. • Guiry MD, Guiry MD (2016) AlgaeBase. World-wide electronic publication, National University of Ireland, Galway. http://www.algaebase.org. Accession date: 2016 7 16. • Hällfors G (2004) Checklist of Baltic Sea Phytoplankton Species. 95. Baltic Sea Environment Proceedings, 210 pp. • Hällfors G (2004) Checklist of Baltic Sea Phytoplankton Species. 95. Baltic Sea Environment Proceedings, 210 pp. • Heindel K, Westphal H, Wisshak M (2009) Data report: bioerosion in the reef framework, IODP Expedition 310 oﬀ Tahiti (Tiarei, Maraa, and Faaa sites). Proceedings of the IODP, epub pp. https://doi.org/10.2204/iodp.proc.310.201.2009 • Hindak F, Moustaka M (1988) Planktic cyanophytes of Lake Volvi, Greece. Algological Studies/Archiv für Hydrobiologie, Supplement Volumes 50-53: 497‑528. • Hindák F (1988) Taxonomic position of two cyanophyte genera, Cyanostylon Geitler and Siphonosphaera gen. nov. Archiv für Hydrobiologie/Algological Studies 78: 411‑423. • Hindák F (1993) To the taxonomy of the chroococcal genus Pannus Hickel 1991 (Cyanophyta/Cyanobacteria). Archiv für Hydrobiologie/Algological Studies 69: 1‑10. • Hirsch P (1981) The Family Pelonemataceae. The Prokaryotes. https:// doi.org/10.1007/978-3-662-13187-9_26 • Kaštovský J, Hauer T, Komarek J, Skacelova O (2010) The list of cyanobacterial species of the Czech Republic to the end of 2009. Fottea 10: 245‑249. https:// doi.org/10.5507/fot.2010.015 g pp • Heindel K, Westphal H, Wisshak M (2009) Data report: bioerosion in the reef framework, IODP Expedition 310 oﬀ Tahiti (Tiarei, Maraa, and Faaa sites). Proceedings of the IODP, epub pp. https://doi.org/10.2204/iodp.proc.310.201.2009 • Hindak F, Moustaka M (1988) Planktic cyanophytes of Lake Volvi, Greece. Algological Studies/Archiv für Hydrobiologie, Supplement Volumes 50-53: 497‑528. References Teil: Chroococcales. Süsswasserﬂora von Mitteleuropa., 19/1. Spektrum Akademischer Verlag, 548 pp. • Komárek J, Anagnostidis K (2005) Cyanoprokaryota 2. Teil/ 2nd Part: Oscillatoriales. In: Büdel B, Krienitz L, Gärtner G, Schagerl M (Eds) Süsswasserﬂora von Mitteleuropa, 19/2. Elsevier/Spektrum, Heidelberg, Germany, 759 pp. [ISBN 978-3827409195] • Komárek J, Kaštovský J, Mareš J, Johansen JR (2014) Taxonomic classiﬁcation of • Komárek J, Anagnostidis K (1999) Cyanoprokaryota 1. Teil: Chroococcales. Süsswasserﬂora von Mitteleuropa., 19/1. Spektrum Akademischer Verlag, 548 pp. • Komárek J, Anagnostidis K (1999) Cyanoprokaryota 1. Teil: Chroococcales. Süsswasserﬂora von Mitteleuropa., 19/1. Spektrum Akademischer Verlag, 548 pp. • Komárek J, Anagnostidis K (2005) Cyanoprokaryota 2. Teil/ 2nd Part: Oscillatoriales. In: Büdel B Krienitz L Gärtner G Schagerl M (Eds) Süsswasserﬂora von Mitteleuropa • Komárek J, Anagnostidis K (2005) Cyanoprokaryota 2. Teil/ 2nd Part: Oscillatoriales. In: Büdel B, Krienitz L, Gärtner G, Schagerl M (Eds) Süsswasserﬂora von Mitteleuropa, 19/2. Elsevier/Spektrum, Heidelberg, Germany, 759 pp. [ISBN 978-3827409195] • Komárek J, Anagnostidis K (2005) Cyanoprokaryota 2. Teil/ 2nd Part: Oscillatoriales. In: Büdel B, Krienitz L, Gärtner G, Schagerl M (Eds) Süsswasserﬂora von Mitteleuropa, 19/2. Elsevier/Spektrum, Heidelberg, Germany, 759 pp. [ISBN 978-3827409195] • Komárek J, Kaštovský J, Mareš J, Johansen JR (2014) Taxonomic classiﬁcation of cyanoprokaryotes (cyanobacterial genera) 2014, using a polyphasic approach. Preslia 86: 295‑335. • Komárek J, Kaštovský J, Mareš J, Johansen JR (2014) Taxonomic classiﬁcation of cyanoprokaryotes (cyanobacterial genera) 2014, using a polyphasic approach. Preslia 86: 295‑335. • Konstantinou D, Gerovasileiou V, Voultsiadou E, Gkelis S (2016) 41st CIESM Congress Proceedings. 41st CIESM Congress. 1 pp. Konstantinou D, Gerovasileiou V, Voultsiadou E, Gkelis S (2016) 41st CIESM Congress Proceedings. 41st CIESM Congress. 1 pp. • Koureas D, Hardisty A, Vos R, Agosti D, Arvanitidis C, Bogatencov P, Buttigieg PL, Jong Yd, Horvath F, Gkoutos G, Groom Q, Kliment T, Kõljalg U, Manakos I, Marcer A, Marhold K, Morse D, Mergen P, Penev L, Pettersson L, Svenning J, de Putte Av, Smith V (2016) Unifying European Biodiversity Informatics (BioUnify). Research Ideas and Outcomes 2: e7787. https://doi.org/10.3897/rio.2.e7787 • Lamprinou V, Economou-Amilli A, Danielidis D, Pantazidou A (2012) Distribution survey of Cyanobacteria in three Greek caves of Peloponnese. International Journal of Lamprinou V, Economou-Amilli A, Danielidis D, Pantazidou A (2012) Distribution survey of Cyanobacteria in three Greek caves of Peloponnese. International Journal of • Lamprinou V, Economou-Amilli A, Danielidis D, Pantazidou A (2012) Distribution survey of Cyanobacteria in three Greek caves of Peloponnese. References • Hindák F (1988) Taxonomic position of two cyanophyte genera, Cyanostylon Geitler and Siphonosphaera gen. nov. Archiv für Hydrobiologie/Algological Studies 78: 411‑423. • Hindák F (1993) To the taxonomy of the chroococcal genus Pannus Hickel 1991 (Cyanophyta/Cyanobacteria). Archiv für Hydrobiologie/Algological Studies 69: 1‑10. • Hirsch P (1981) The Family Pelonemataceae. The Prokaryotes. https:// doi.org/10.1007/978-3-662-13187-9_26 • Kaštovský J, Hauer T, Komarek J, Skacelova O (2010) The list of cyanobacterial species of the Czech Republic to the end of 2009. Fottea 10: 245‑249. https:// d / /f 100 Gkelis S et al. • Katsiapi M, Mazaris A, Charalampous E, Moustaka-Gouni M (2012) Watershed land use types as drivers of freshwater phytoplankton structure. Hydrobiologia 698 (1): 121‑131. https://doi.org/10.1007/s10750-012-1095-z • Katsiapi M, Moustaka-Gouni M, Michaloudi E, Kormas K (2011) Phytoplankton and water quality in a Mediterranean drinking-water reservoir (Marathonas Reservoir, Greece). Environmental Monitoring and Assessment 181: 563‑575. https:// d i /10 1007/ 10661 010 1851 3 • Komárek J (2013) Phenotypic characters of heterocytous cyanobacteria. Süßwasserﬂora von Mitteleuropa, Bd. 19/3: Cyanoprokaryota. https:// doi.org/10.1007/978-3-8274-2737-3 2 • Komárek J (2013) Phenotypic characters of heterocytous cyanobacteria. Süßwasserﬂora von Mitteleuropa, Bd. 19/3: Cyanoprokaryota. https:// doi.org/10.1007/978-3-8274-2737-3_2 g _ • Komárek J (2016) Review of the cyanobacterial genera implying planktic species after recent taxonomic revisions according to polyphasic methods: state as of 2014. Hydrobiologia 764 (1): 259‑270. https://doi.org/10.1007/s10750-015-2242-0 • Komárek J, Anagnostidis K (1986) Modern approach to the classiﬁcation system of cyanophytes. 2- Chroococcales. Archiv für Hydrobiologie, Supplement Volumes 73: 157‑226. • Komárek J (2016) Review of the cyanobacterial genera implying planktic species after recent taxonomic revisions according to polyphasic methods: state as of 2014. Hydrobiologia 764 (1): 259‑270 https://doi org/10 1007/s10750-015-2242-0 Hydrobiologia 764 (1): 259 270. https://doi.org/10.1007/s10750 015 2242 0 • Komárek J, Anagnostidis K (1986) Modern approach to the classiﬁcation system of cyanophytes. 2- Chroococcales. Archiv für Hydrobiologie, Supplement Volumes 73: 157‑226. y g ( ) p g • Komárek J, Anagnostidis K (1986) Modern approach to the classiﬁcation system of cyanophytes. 2- Chroococcales. Archiv für Hydrobiologie, Supplement Volumes 73: 157‑226. • Komárek J, Anagnostidis K (1989) Modern approach to the classiﬁcation system of cyanophytes. 4- Nostocales. Archiv für Hydrobiologie, Supplement 82: 247‑345. • Komárek J, Anagnostidis K (1989) Modern approach to the classiﬁcation system of cyanophytes. 4- Nostocales. Archiv für Hydrobiologie, Supplement 82: 247‑345. g ( ) pp y cyanophytes. 4- Nostocales. Archiv für Hydrobiologie, Supplement 82: 247‑345. • Komárek J, Anagnostidis K (1999) Cyanoprokaryota 1. SYSTEMATIC AND EVOLUTIONARY MICROBIOLOGY 61 (12): 2907‑2915. https:// doi.org/10.1099/ijs.0.029223-0 • Lamprinou V, Hernandez-Marine M, Pachiadaki M, Kormas K, Economou-Amilli A, Pantazidou A (2013a) New ﬁndings on the true-branched monotypic genus Iphinoe (Cyanobacteria) from geographically isolated caves (Greece). Fottea 13: 15‑23. https:// doi.org/10.5507/fot.2013.002 • Lamprinou V, Hernandez-Marine M, Pachiadaki M, Kormas K, Economou-Amilli A, Pantazidou A (2013a) New ﬁndings on the true-branched monotypic genus Iphinoe (Cyanobacteria) from geographically isolated caves (Greece). Fottea 13: 15‑23. https:// doi.org/10.5507/fot.2013.002 • Lamprinou V, Skaraki K, Kotoulas G, Anagnostidis K, Economou-Amilli A, Pantazidou A (2013b) A new species of Phormidium (Cyanobacteria, Oscillatoriales) from three Greek Caves: morphological and molecular analysis. Fundamental and Applied Limnology / Archiv für Hydrobiologie 182 (2): 109‑116. https://doi.org/10.1127/1863-9135/2013/0323 • Lanaras T, Tsitsamis S, Chlichlia C, Cook CM (1989) Toxic cyanobacteria in Greek freshwaters. Journal of Applied Phycology 1 (1): 67‑73. https://doi.org/10.1007/ bf00003537 • Metaxatos A, Panagiotopoulos C, Ignatiades L (2003) Monosaccharide and aminoacid composition of mucilage material produced from a mixture of four phytoplanktonic taxa. Journal of Experimental Marine Biology and Ecology 294 (2): 203‑217. https:// doi.org/10.1016/s0022-0981(03)00269-7 • Montesanto B, Ziller S, Danielidis D, Economou-Amilli A (1999) Phytoplankton community structure in the lower reaches of a Mediterranean river (Aliakmon, Greece). Fundamental and Applied Limnology 147 (2): 171‑191. https://doi.org/10.1127/archiv- hydrobiol/147/1999/171 • Montesanto B, Ziller S, Danielidis D, Economou-Amilli A (1999) Phytoplankton community structure in the lower reaches of a Mediterranean river (Aliakmon, Greece). Fundamental and Applied Limnology 147 (2): 171‑191. https://doi.org/10.1127/archiv- hydrobiol/147/1999/171 • Moustaka-Gouni M (1988) The structure and dynamics of the phytoplankton assemblages in Lake Volvi, Greece I. Phytoplankton composition and abundance during the period March 1984-March 1985. Archiv für Hydrobiologie 112: 251‑264. • Moustaka-Gouni M (1988) The structure and dynamics of the phytoplankton assemblages in Lake Volvi, Greece I. Phytoplankton composition and abundance during the period March 1984-March 1985. Archiv für Hydrobiologie 112: 251‑264. • Moustaka-Gouni M (1993) Phytoplankton succession and diversity in a warm monomictic, relatively shallow lake: Lake Volvi, Macedonia, Greece. Intermediate Disturbance Hypothesis in Phytoplankton Ecology. https:// doi.org/10.1007/978-94-017-1919-3 5 • Moustaka-Gouni M, Nikolaidis G (1990) Phytoplankton of a warm monomictic lake - Lake Vegoritis, Greece. Archiv für Hydrobiologie 119: 299‑313. • Moustaka-Gouni M, Nikolaidis G (1990) Phytoplankton of a warm monomictic lake - Lake Vegoritis, Greece. Archiv für Hydrobiologie 119: 299‑313. • Moustaka-Gouni M, Nikolaidis G (1992) Phytoplankton and physical-chemical features of Tavropos Reservoir, Greece. Hydrobiologia 228 (2): 141‑149. References International Journal of Speleology 41 (2): 267‑272. https://doi.org/10.5038/1827-806x.41.2.12 p gy ( ) p g • Lamprinou V, Danielidis D, Pantazidou A, Oikonomou A, Economou-Amilli A (2014) The show cave of Diros vs. wild caves of Peloponnese, Greece - distribution patterns of Cyanobacteria. International Journal of Speleology 43 (3): 335‑342. https:// • Lamprinou V, Danielidis D, Pantazidou A, Oikonomou A, Economou-Amilli A (2014) The show cave of Diros vs. wild caves of Peloponnese, Greece - distribution patterns of Cyanobacteria. International Journal of Speleology 43 (3): 335‑342. https:// • Lamprinou V, Pantazidou A, Papadogiannaki G, Radea C, Economou-Amilli A (2009) Cyanobacteria and associated invertebrates in Leontari Cave. Fottea 9: 155‑164. https:// doi.org/10.5507/fot.2009.014 • Lamprinou V, Pantazidou A, Papadogiannaki G, Radea C, Economou-Amilli A (2009) Cyanobacteria and associated invertebrates in Leontari Cave. Fottea 9: 155‑164. https:// doi.org/10.5507/fot.2009.014 • Lamprinou V, Hernandez-Marine M, Canals T, Kormas K, Economou-Amilli A, Pantazidou A (2011) Morphology and molecular evaluation of Iphinoe spelaeobios gen. nov., sp. nov. and Loriellopsis cavernicola gen. nov., sp. nov., two stigonematalean cyanobacteria from Greek and Spanish caves. INTERNATIONAL JOURNAL OF Cyanobacteria of Greece: an annotated checklist 101 SYSTEMATIC AND EVOLUTIONARY MICROBIOLOGY 61 (12): 2907‑2915. https:// doi.org/10.1099/ijs.0.029223-0 https://doi.org/10.1007/ bf00006202 • Moustaka-Gouni M, Albanakis K, Mitrakas M, Psilovikos A (2000) Planktic autotrophs and environmental conditions in the newly-formed hydroelectric Thesaurus reservoir, Greece. Fundamental and Applied Limnology 149 (3): 507‑526. https://doi.org/10.1127/ archiv-hydrobiol/149/2000/507 • Moustaka-Gouni M, Michaloudi E, Katsiapi M, Genitsaris S (2007) Greece : The coincidence of an Arthrospira -Anabaenopsis bloom and the mass mortality of birds in Lake Koronia. Harmful algae news 35: 6‑7. • Moustaka-Gouni M, Michaloudi E, Katsiapi M, Genitsaris S (2007) Greece : The coincidence of an Arthrospira -Anabaenopsis bloom and the mass mortality of birds in Lake Koronia. Harmful algae news 35: 6‑7. Moustaka-Gouni M, Michaloudi E, Katsiapi M, Genitsaris S (2007) Greece : The coincidence of an Arthrospira -Anabaenopsis bloom and the mass mortality of birds in • Moustaka-Gouni M, Kormas K, Vardaka E, Katsiapi M, Gkelis S (2009) Raphidiopsis mediterranea Skuja represents non-heterocytous life-cycle stages of Cylindrospermopsis raciborskii (Woloszynska) Seenayya et Subba Raju in Lake Kastoria (Greece), its type locality: Evidence by morphological and phylogenetic analysis. Harmful Algae 8 (6): 864‑872. https://doi.org/10.1016/j.hal.2009.04.003 • Moustaka-Gouni M, Kormas K, Vardaka E, Katsiapi M, Gkelis S (2009) Raphidiopsis mediterranea Skuja represents non-heterocytous life-cycle stages of Cylindrospermopsis raciborskii (Woloszynska) Seenayya et Subba Raju in Lake Kastoria (Greece), its type locality: Evidence by morphological and phylogenetic analysis. Harmful Algae 8 (6): 864‑872. https://doi.org/10.1016/j.hal.2009.04.003 • Moustaka-Gouni M, Kormas KA, Polykarpou P, Gkelis S, Bobori DC, Vardaka E (2010) Polyphasic evaluation of Aphanizomenon issatschenkoi and Raphidiopsis mediterranea • Moustaka-Gouni M, Kormas KA, Polykarpou P, Gkelis S, Bobori DC, Vardaka E (2010) Polyphasic evaluation of Aphanizomenon issatschenkoi and Raphidiopsis mediterranea 102 Gkelis S et al. g p q • Moustaka-Gouni M, Hiskia A, Genitsaris S, Katsiapi M, Manolidi K, Zervou S, Christophoridis C, Triantis T, Kaloudis T, Orfanidis S (2016) First report of Aphanizomenon favaloroi occurrence in Europe associated with saxitoxins and a massive ﬁsh kill in Lake Vistonis, Greece. Marine and Freshwater Research 67: 1‑8. htt ps://doi.org/10.1071/mf16029 • Moustaka M (1988) SEASONAL VARIATIONS, ANNUAL PERIODICITY AND SPATIAL DISTRIBUTION OF PHYTOPLANKTON IN LAKE VOLVI. Aristotle University of Thessaloniki, 363 pp. [In Greek]. • Nassar M (2014) Checklist of Phytoplankton Species in the Egyptian Waters of the Red Sea and Some Surrounding Habitats (1990-2010). Annual Research & Review in Biology 4 (23): 3566‑3585 https://doi org/10 9734/arrb/2014/10866 • Overbeck J, Anagnostidis K, Economou-Amilli A (1982) A limnological survey of three Greek lakes, Trichonis, Lyssimachia and Amvrakia. Archiv fur Hydrobiologie 95: 365‑394. SYSTEMATIC AND EVOLUTIONARY MICROBIOLOGY 61 (12): 2907‑2915. https:// doi.org/10.1099/ijs.0.029223-0 • Overbeck J, Anagnostidis K, Economou-Amilli A (1982) A limnological survey of three Greek lakes, Trichonis, Lyssimachia and Amvrakia. Archiv fur Hydrobiologie 95: 365‑394. • Pantazidou A, Louvrou I, Economou-Amilli A (2006) Euendolithic shell-boring cyanobacteria and chlorophytes from the saline lagoon Ahivadolimni on Milos Island, Greece. European Journal of Phycology 41 (2): 189‑200. https:// doi.org/10.1080/09670260600649420 • Pantazidou A, Louvrou I, Economou-Amilli A (2006) Euendolithic shell-boring cyanobacteria and chlorophytes from the saline lagoon Ahivadolimni on Milos Island, Greece. European Journal of Phycology 41 (2): 189‑200. https:// doi.org/10.1080/09670260600649420 • Papadimitriou T, Katsiapi M, Kormas KA, Moustaka-Gouni M, Kagalou I (2013) Artiﬁcially-born “killer” lake: Phytoplankton based water quality and microcystin aﬀected ﬁsh in a reconstructed lake. Science of The Total Environment 452-453: 116‑124. http s://doi.org/10.1016/j.scitotenv.2013.02.035 • Papadimitriou T, Katsiapi M, Kormas KA, Moustaka-Gouni M, Kagalou I (2013) Artiﬁcially-born “killer” lake: Phytoplankton based water quality and microcystin aﬀected ﬁsh in a reconstructed lake. Science of The Total Environment 452-453: 116‑124. http s://doi org/10 1016/j scitotenv 2013 02 035 • Peck AL (1970) Aristotle. History of Animals. Books IV-VI. English translation, Introduction and Comments. Cambridge, Massachusetts: Harvard University Press., 409 pp. • Peck AL (1970) Aristotle. History of Animals. Books IV-VI. English translation, Introduction and Comments. Cambridge, Massachusetts: Harvard University Press., 409 pp. • Radea C, Louvrou I, Pantazidou A, Amilli AE (2010) Photosynthetic microorganisms as epibionts and euendoliths on biotic substrates in a thermal spring with ferric-iron deposits. Fottea 10: 129‑140. https://doi.org/10.5507/fot.2010.007 • Radea C, Louvrou I, Pantazidou A, Amilli AE (2010) Photosynthetic microorganisms as epibionts and euendoliths on biotic substrates in a thermal spring with ferric-iron deposits. Fottea 10: 129‑140. https://doi.org/10.5507/fot.2010.007 • Rai SK, Rai RK, Jha S (2011) Cyanobacteria of Nepal: A Checklist with Distribution. Our Nature 8 (1): 336‑354. https://doi.org/10.3126/on.v8i1.4342 • Rai SK, Rai RK, Jha S (2011) Cyanobacteria of Nepal: A Checklist with Distribution. Our Nature 8 (1): 336‑354. https://doi.org/10.3126/on.v8i1.4342 • Skuja H (1937) Süsswasseralgen aus Griechenland und Kleinasien. Hedwigia 77: 15‑73. [In German]. Stanković S (1931) Sur les particularités limnologiques de lacs égéens. Verhandlungen des Internationalen Verein Limnologie 5: 158‑196. • Tafas T, Economou-Amilli A (1997) Limnological survey of the warm monomictic lake Trichonis (central western Greece)II. Seasonal phytoplankton periodicity – a community approach . Hydrobiologia 344: 141‑153. https://doi.org/10.1023/a:1002962513146 , ( ) g y Trichonis (central western Greece)II. Seasonal phytoplankton periodicity – a community approach . Hydrobiologia 344: 141‑153. SYSTEMATIC AND EVOLUTIONARY MICROBIOLOGY 61 (12): 2907‑2915. https:// doi.org/10.1099/ijs.0.029223-0 https://doi.org/10.1023/a:1002962513146 • Temponeras M, Kristiansen J, Moustaka-Gouni M (2000) Seasonal variation in phytoplankton composition and physical-chemical features of the shallow Lake Doïrani, Macedonia, Greece. The Trophic Spectrum Revisited. https:// doi.org/10.1007/978-94-017-3488-2_10 • Tryfon E (1996) Pannus spumosus (Chroococcales, Cyanoprocaryota) from Lake Mikri Prespa, Greece. Phycologia 35 (3): 222‑224. https://doi.org/10.2216/ i0031-8884-35-3-222.1 Cyanobacteria of Greece: an annotated checklist 103 • Tryfon E, Moustaka-Gouni M, Nikolaidis G (1996) Phytoplankton and Nutrients in the River Strymon, Greece. Internationale Revue der gesamten Hydrobiologie und Hydrographie 81 (2): 281‑292. https://doi.org/10.1002/iroh.19960810210 • Tryfon E, Moustaka-Gouni M, Nikolaidis G (1996) Phytoplankton and Nutrients in the River Strymon, Greece. Internationale Revue der gesamten Hydrobiologie und Hydrographie 81 (2): 281‑292. https://doi.org/10.1002/iroh.19960810210 • Tryfon E, Moustaka-Gouni M, Nikolaidis G (1996) Phytoplankton and Nutrients in the River Strymon, Greece. Internationale Revue der gesamten Hydrobiologie und Hydrographie 81 (2): 281‑292. https://doi.org/10.1002/iroh.19960810210 • Tryfon E, Moustaka-Gouni M, Nikolaidis G (1997) Planktic cyanophytes and their ecology in the shallow Lake Mikri Prespa, Greece. Nordic Journal of Botany 17 (4): 439‑448. https://doi.org/10.1111/j.1756-1051.1997.tb00340.x • Tsarenko PM, Vinogradova ON, Stupina VV, Wasser S, Nevo E (2000) Diversity of algae in the continental part of Israel. International Journal on Algae 2 (4): 20‑39. http s://doi.org/10.1615/interjalgae.v2.i4.20 • Vardaka E, Moustaka-Gouni M, Lanaras T (2000) Temporal and spatial distribution of planktic cyanobacteria in Lake Kastoria, Greece, a shallow, urban lake. Nordic Journal of Botany 20 (4): 501‑511. https://doi.org/10.1111/j.1756-1051.2000.tb01594.x • Vardaka E, Moustaka-Gouni M, Cook C, Lanaras T (2005) Cyanobacterial blooms and water quality in Greek waterbodies. Journal of Applied Phycology 17 (5): 391‑401. http s://doi.org/10.1007/s10811-005-8700-8 • Vareli K, Briasoulis E, Pilidis G, Sainis I (2009) Molecular conﬁrmation of Planktothrix rubescens as the cause of intense, microcystin—Synthesizing cyanobacterial bloom in Lake Ziros, Greece. Harmful Algae 8 (3): 447‑453. https://doi.org/10.1016/ j.hal.2008.09.005 • Vilicic D, Marasovic I, Miokovic D (2002) Checklist of phytoplankton in the eastern Adriatic Sea. Acta Bot. Croat. 61: 57‑91. • Voultsiadou E (2007) Sponges: an historical survey of their knowledge in Greek antiquity. Journal of the Marine Biological Association of the UK 87 (6): 1757‑1763. http s://doi.org/10.1017/s0025315407057773 • Tryfon E, Moustaka-Gouni M, Nikolaidis G (1997) Planktic cyanophytes and their ecology in the shallow Lake Mikri Prespa, Greece. Nordic Journal of Botany 17 (4): 439‑448. https://doi.org/10.1111/j.1756-1051.1997.tb00340.x • Tsarenko PM, Vinogradova ON, Stupina VV, Wasser S, Nevo E (2000) Diversity of algae in the continental part of Israel. International Journal on Algae 2 (4): 20‑39. SYSTEMATIC AND EVOLUTIONARY MICROBIOLOGY 61 (12): 2907‑2915. https:// doi.org/10.1099/ijs.0.029223-0 http s://doi.org/10.1615/interjalgae.v2.i4.20 • Vardaka E, Moustaka-Gouni M, Lanaras T (2000) Temporal and spatial distribution of planktic cyanobacteria in Lake Kastoria, Greece, a shallow, urban lake. Nordic Journal of Botany 20 (4): 501‑511. https://doi.org/10.1111/j.1756-1051.2000.tb01594.x • Vardaka E, Moustaka-Gouni M, Lanaras T (2000) Temporal and spatial distribution of planktic cyanobacteria in Lake Kastoria, Greece, a shallow, urban lake. Nordic Journal of Botany 20 (4): 501‑511. https://doi.org/10.1111/j.1756-1051.2000.tb01594.x planktic cyanobacteria in Lake Kastoria, Greece, a shallow, urban lake. Nordic Journal of Botany 20 (4): 501‑511. https://doi.org/10.1111/j.1756-1051.2000.tb01594.x • Vardaka E, Moustaka-Gouni M, Cook C, Lanaras T (2005) Cyanobacterial blooms and water quality in Greek waterbodies. Journal of Applied Phycology 17 (5): 391‑401. http s://doi.org/10.1007/s10811-005-8700-8 • Vareli K, Briasoulis E, Pilidis G, Sainis I (2009) Molecular conﬁrmation of Planktothrix rubescens as the cause of intense, microcystin—Synthesizing cyanobacterial bloom in Lake Ziros, Greece. Harmful Algae 8 (3): 447‑453. https://doi.org/10.1016/ j h l 2008 09 00 • Vardaka E, Moustaka-Gouni M, Cook C, Lanaras T (2005) Cyanobacterial blooms and water quality in Greek waterbodies. Journal of Applied Phycology 17 (5): 391‑401. http s://doi.org/10.1007/s10811-005-8700-8 • Vardaka E, Moustaka-Gouni M, Cook C, Lanaras T (2005) Cyanobacterial blooms and water quality in Greek waterbodies. Journal of Applied Phycology 17 (5): 391‑401. http s://doi.org/10.1007/s10811-005-8700-8 • Vareli K, Briasoulis E, Pilidis G, Sainis I (2009) Molecular conﬁrmation of Planktothrix rubescens as the cause of intense, microcystin—Synthesizing cyanobacterial bloom in Lake Ziros, Greece. Harmful Algae 8 (3): 447‑453. https://doi.org/10.1016/ j.hal.2008.09.005 • Vareli K, Briasoulis E, Pilidis G, Sainis I (2009) Molecular conﬁrmation of Planktothrix rubescens as the cause of intense, microcystin—Synthesizing cyanobacterial bloom in Lake Ziros, Greece. Harmful Algae 8 (3): 447‑453. https://doi.org/10.1016/ j hal 2008 09 005 • Vilicic D, Marasovic I, Miokovic D (2002) Checklist of phytoplankton in the eastern Adriatic Sea. Acta Bot. Croat. 61: 57‑91. • Voultsiadou E (2007) Sponges: an historical survey of their knowledge in Greek antiquity. Journal of the Marine Biological Association of the UK 87 (6): 1757‑1763. http s://doi.org/10.1017/s0025315407057773
https://openalex.org/W4248763807	https://www.researchsquare.com/article/rs-27666/latest.pdf	English	null	Elective removal vs. retaining of hardware after osteosynthesis in asymptomatic patients – A scoping review.	Research Square (Research Square)	2,020	cc-by	6,365	Elective removal vs. retaining of hardware after osteosynthesis in asymptomatic patients – A scoping review. Barbara Prediger (  Barbara.Prediger@uni-wh.de ) Universitat Witten/Herdecke Tim Mathes Martin-Luther-Universitat Halle-Wittenberg Philosophische Fakultat I Christian Probst Klinikum Oberberg GmbH Dawid Pieper Universitat Witten/Herdecke Results We identified 13476 titles/abstracts. Of these we obtained 115 full-text publications which were assessed in detail against the inclusion criteria. We included 13 studies (1 RCT, 4 cohort studies, 8 before-after studies) and identified two ongoing RCTs. Nine assessed the removal of the internal fixation in the lower extremities (six of these syndesmotic screws in ankle fractures only) and two in the upper extremities. One study analysed the effectiveness of hardware removal in children in all types of extremity fractures. Outcomes reported included various scales measuring functionality, pain and clinical assessments (e.g. range of motion) and health-related quality of life. Research License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Version of Record: A version of this preprint was published on October 2nd, 2020. See the published version at https://doi.org/10.1186/s13643-020-01488-2. Page 1/16 Page 1/16 Page 1/16 Background Osteosynthesis is the internal fixation of fractures or osteotomy by mechanical devices (also called hardware). After bone healing, there are two options: One is to remove the hardware, the other is to leave it in place. The removal of the hardware in patients without medical indication (elective) is controversially discussed. We performed a scoping review to identify evidence on the elective removal of hardware in asymptomatic patients compared to retaining of the hardware to check feasibility of performing a health technology assessment. In addition, we wanted to find out which type of evidence is available. Conclusions We identified 13 studies that evaluated the effectiveness/safety of hardware removal in the extremities. The follow up times were short, the patient groups small and the ways of measurement differed. In general clinical heterogeneity was high. Evidence on selected topics e.g. syndesmotic screw removal is available nevertheless not sufficient to allow a meaningful assessment of effectiveness. Methods A systematic literature search was performed in Pubmed, Embase, EconLit and Cinahl (11/2019). We included studies comparing asymptomatic patients with an internal fixation in the lower or upper extremities whose internal fixation was electively (without medical indication) removed or retained. We did not restrict inclusion to any effectiveness/safety outcome and considered any comparative study design as eligible. Study selection and data extraction was performed by two reviewers. Background Osteosynthesis is the internal fixation of fractures or osteotomy by mechanical devices (also called hardware). After bone healing there is either the option to remove the osteosynthetic material or to leave it in place. In case the internal fixation causes symptoms (e.g. strong pain, reduced physical functioning) or complications (wound infection, device failure) the indication for removal is apparent. In contrast, the Page 2/16 Page 2/16 decision for or against elective removal of the osteosynthetic material in asymptomatic patients is difficult. (1) The decision for removal should be primarily based on individual patient factors (e.g. age, physical activity), considering the possible future outcomes associated with removal or non-removal.(2) The outcomes, that are important for decision making, include chronic pain, physical functioning, The decision for removal should be primarily based on individual patient factors (e.g. age, physical activity), considering the possible future outcomes associated with removal or non-removal.(2) The outcomes, that are important for decision making, include chronic pain, physical functioning, complications, reoperations, negative body sensation and spatial limitation. However, the benefits of hardware removal in asymptomatic patients are not sufficiently analysed.(3, 4) Thus, there is an ongoing debate on the justification of osteosynthetic material removal in general and on the patients groups that might benefit most by the removal.(5, 6) While some surgeons never extract the hardware in symptom free patients, other surgeons removes the hardware to prevent future complications. A national survey performed in the United Kingdom revealed that only 7 % of polled surgeons had departmental or unit policies.(7) Nevertheless, removal of implants is one of the most common surgical procedures.(8) According to an analysis from Bostman et al. approximately 180 procedures per 100,000 person years could be estimated in western countries.(9) In 2018 176,257 surgeries on hardware removal were performed in Germany, which means in about 80 % of fractures treated with osteosynthesis, material was removed.(10) Similar numbers are known from the USA.(11) In Germany,it is estimated that expenses exceed about 430 Mio Euro per year.(12) Most (>90%) of the internal fixations are removed within 24 months after the initial surgery.(3) However, there is no clear timing for removing the osteosynthetic material, the time-point of removal depends mainly on the time-point of bone healing.(13) Furthermore, the time-point of bone healing again depends on many factors including localization, type of fracture, severity, type of fixation device/s used and patient characteristic. Methods We developed a protocol for the review (in German), which is available from the corresponding author. It was written following the structure of PROSPERO and finalized on the 30th of April 2018. We did not register the protocol anywhere. Reporting on the findings of this scoping review we followed the PRISMA- ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses for scoping reviews).(14) Previously we reported this scoping review to the Swiss Federal Office of Public Health. Considering this to be the feasibility test of performing HTA we also sighted the other domains of an HTA like costs, legal, social and organisational issues.(17) We did not find any evidence on these issues. Which is why we only concentrate on effectiveness in this manuscript. Please refer to the report provided to the Swiss Federal Office of Public Health for any detail regarding the results for other domains than effectiveness. That publication contains results from the systematic search conducted in October 2018 (containing 10 studies) only. (17) Objectives The Swiss Federal Office of Public Health delegated us to perform a scoping review to produce a basis for the decision about the feasibility of a full Health Technology Assessment (HTA). This approach is based on the idea to perform a scoping review to assess the feasibility of performing a systematic review. (14) The aim of our contracting authority (Swiss Federal Office of Public Health) was the generation of an evidence base allowing to decide about the feasibility of a full-HTA questioning if hardware removal in patients without medical indication (elective) is effective and safe compared to retaining of hardware covering all parts of the upper and lower extremities. So far no systematic review or meta-analysis comparing non-indicated removal of hardware to retaining of hardware in various parts of the extremities exists. However two reviews are known that specialize on syndesmotic screw removal.(15, 16) To generate an overview if there are other body parts of which a full HTA of hardware removal might be useful and to evaluate where and in what extent primary studies are needed we performed a scoping review keeping the surgical site broad. Based on the quality and quantity Page 3/16 of the available evidence to answer the central research questions, additional or modified questions can be determined for performing a full HTA. We performed a scoping review to identify evidence on the elective removal of hardware in asymptomatic patients compared to retaining of the hardware. In addition, we wanted to find out which type of evidence is available. Data charting process and data items Data was collected in an a priori-piloted abstraction table by one reviewer, the other reviewer monitored all entries for completeness and accuracy. We extracted following study characteristics: author, publication year, region, setting, study design, recruitment period, inclusion/exclusion criteria, patient characteristics (age, body-mass index, comorbidities, fracture characteristics, surgery characteristics), time points measured, outcomes. Selection of sources of evidence Records identified through the searches were added to an Endnote X7 database and duplicates were removed. All titles/abstracts identified in the electronic databases were screened by one reviewer and a second reviewer screened all excluded titles/abstracts (liberal acceleration). All articles judged to be potentially relevant were obtained. The full-texts off all potentially relevant articles were screened by two reviewers independently. Any disagreement in the study selection process was resolved in a discussion until consensus. Information sources We performed a systematic literature search in Pubmed, Embase, EconLit and CINAHL (all from inception) in October 2018 and updated in November 2019. The search strategy for Pubmed is displayed in box 1. The search strategies for the other databases can be found in appendix I. The strategy was developed by our information specialist and checked by another reviewer by consulting the Peer Review of Electronic Search Strategies (PRESS) criteria.(21) We deviated from the protocol and searched the database CINAHL instead of CENTRAL due to the overlap of results within PubMed, Embase and CENTRAL. We performed the literature search without limiting the publication date. We applied search limitations to English and German articles and excluded comments, editorials, letters and research on animals in our electronic database search. We cross-checked all references of included articles and systematic reviews on similar topics known to us. Critical appraisal of individual sources of evidence As this is a scoping review, there was no risk of bias assessment. This is consistent with guidance on the conduct of scoping reviews.(20) Eligibility criteria We included studies analysing asymptomatic patients with an internal fixation in the lower or upper extremities. All age groups were eligible. We considered the elective removal of the internal fixation device/s as experimental intervention and non-removal of the internal fixation as control intervention. All types of health outcomes were considered. This includes mortality, morbidity (pain, satisfaction, physical functioning and clinical events), clinical measures (e.g. range of motion), health-related quality of life and adverse events/complications. The inclusion was limited to studies in WHO Stratum A. This covers states with very low child and very low adult mortality including western Europe, North-America and various Western-Pacific states.(16) We chose this criterion for two reasons. First, access to and health services are comparable in these countries as is morbidity and mortality. Transferability of technological appraisals might be restricted from countries which are non WHO Stratum A.(18) Second is a pragmatic reason, elective removal of asymptomatic implants seems to be a novel trend in recent years in less developed countries and we did not expect many research yet.(19) We did not define any other exclusion criteria regarding the population. We considered the following comparative study designs: Randomized controlled trials, non-randomized controlled trials, cohort studies, case-control studies, before-after studies, interrupted-time-series, controlled before-after studies. We did not make any restrictions regarding publication date. We only included studies written in German or English as the reviewers could only ensure to review these languages in duplicate. We followed the framework of Arksey and O’Malley for scoping reviews.(20) Page 4/16 Characteristics of sources of evidence The studies included adults (12 studies)(22-24, 26-31, 33-35) and children compared elective removal versus non-removal of osteosynthetic materia screws, nails and staples. Sometimes a combination of implants was use between 6 weeks and 27 months after surgery. Ankle screws were typical bearing (6-12 weeks) whereas plates and nails were removed after longer studies included a total of 588 patients. Primary efficacy and safety outc mobility scores (e.g., Olerud-Molander ankle (OMAS) or American Orthopa ankle-hind foot scores (AOFAS)) and pain scores (e.g., Visual Analogue S outcomes included surgery related complications, quality of life and retur from 9 weeks to 31 months. Table 1 in the appendix shows the detailed c studies. The studies included adults (12 studies)(22-24, 26-31, 33-35) and children (one study).(32) All studies compared elective removal versus non-removal of osteosynthetic material. The implants involved plates, screws, nails and staples. Sometimes a combination of implants was used. Implants were removed between 6 weeks and 27 months after surgery. Ankle screws were typically removed before weight bearing (6-12 weeks) whereas plates and nails were removed after longer periods of bone healing. The studies included a total of 588 patients. Primary efficacy and safety outcomes included functional mobility scores (e.g., Olerud-Molander ankle (OMAS) or American Orthopaedic Foot and Ankle Society ankle-hind foot scores (AOFAS)) and pain scores (e.g., Visual Analogue Scale (VAS)). Secondary outcomes included surgery related complications, quality of life and return to work. Follow-up ranged from 9 weeks to 31 months. Table 1 in the appendix shows the detailed characteristics of the included studies. Selection of sources of evidence Figure 1 shows the study selection process. The literature search identified 13 eligible studies.)(22-35) The studies were conducted in Germany,(26-28) the USA,(29, 30, 32, 33) Switzerland,(22, 34) Singapore, (23) New Zealand,(24) the UK,(31) and Japan.(35) Moreover we identified a protocol on an ongoing RCT regarding syndesmotic screw removal conducted in the Netherlands(36) and an entry on clinicaltrials.gov about an RCT regarding removal/retaining material in Lisfranc fractures conducted in Canada. There was one RCT comparing removal after three months postoperatively to non-removal.(24) We identified four cohort studies, one comparing patients with removal before weight bearing versus non removal,(23) and three other comparing patients with removal with non removal after three to seven months.(26, 29, 31) Furthermore we identified eight before-after studies, comparing the same patients before implant removal and after implant removal.(22, 27, 28, 30, 32-35) Implant removal proceeded after 3 until 27 months postoperatively. Ten studies assessed the effectiveness and safety of removal/non-removal of the osteosynthetic material at the lower extremities with six on ankle,(23, 24, 29-31, 33) three on tibia(26, 27, 35) and one on femoral fractures.(28) Two studies investigated the effectiveness and safety of removal/non-removal in the proximal humerus.(22, 34) One study analysed the effectiveness of osteosynthetic material removal in children in all body parts.(32) Synthesis of results We used Arksey and O’Malley’s methods and provide a descriptive analysis of the extent, nature, and distribution of the studies included in the review as well as a narrative, thematic summary of the data collected.(20) For this we summarized the literature considering study types, location/type of the material and outcomes. On this basis we analyzed similarities and differences within and in between studies to Page 5/16 identify patterns and themes and postulate explanations for findings. Furthermore we developed an evidence map illustrating the type of evidence, indications and outcomes. Synthesis of results Figure 2 shows an evidence map of the included literature. Six studies analysed removal/retain of syndesmotic screws. Three of them used the OMAS(24, 30, 31) and two used the AOFAS for functional assessment.(24). Three studies analysed functional outcomes of hardware removal in the tibia.(26, 27, 35) The results in general are in some parts homogenous in between different studies for the retention or removal group (e.g. functional outcomes). The other locations and types of materials differed too much to synthesise them in any way. They were all pictured as single studies. Results of individual sources of evidence We identified five studies analysing pain, three of them used VAS(24, 26, 29) and the other reported pain as a count.(23, 32) Pain was measured similar in the retention and the removal groups of the cohort studies as in the before removal/after removal study from Chu et al..(32) Except for Briceno et al.(33) all Page 6/16 Page 6/16 studies measured general functioning, some in various ways, Functioning was assessed with a (specific) score, for example the OMAS for studies analysing syndesmotic screws, except in both studies from Gosling et al..(27, 28) The cohort studies reported similar results in both groups for each score. According to the scores used, the before after studies showed better functional outcomes, after removal of the implant. Clinical outcomes, like range of motion, were reported in six studies.(23, 24, 30, 33-35) For the cohort studies and one before-after study, there was no difference in the outcomes,(23, 24, 35) while the other three before-after studies showed advanced range of motion.(30, 33, 34) Garner et al. measured two components of the Short Form 36 scale for quality of life, with better outcomes for physical components in the removal group.(26) Please see table 1 for individual study results. Summary of evidence We identified 13 articles that evaluated the effectiveness/safety of hardware removal in the extremities. One of the included studies is an RCT which is about syndesmotic screw fixation in ankle fractures. There were four cohort studies examining hardware removal in the lower extremities and further eight studies with a before after design, two examining the upper extremities and one including diverse indications, the remaining examining the lower extremities. We also identified two ongoing RCT on syndesmotic screw removal and removal/retaining of material in Lisfranc fractures.(36) Three studies were conducted in Germany,(26-28) four in the USA.(29, 30, 32, 35) The other studies were conducted in Switzerland,(22, 34) Singapore,(23) New Zealand,(24)the UK(31) and Japan.(35) The ongoing RCTs will be performed in the Netherlands and Canada.(25) There are six studies examining the removal of syndesmotic screws(23-25, 29-31, 33), two studies on plate removal from the proximal humerus,(22, 34) two studies on plate removal in the tibia,(26, 35) and one study on nail removal in the tibia(27) and nail removal in the femur.(28) Chu et al. included diverse indications treated with different materials. This is the only study treating children. (32) As we did not specify the outcome in this scoping review, we identified multiple outcomes concerning function, pain, clinical measures and undesired events. Three studies collected their outcomes with the OMAS as a score evaluating functional impairment, three studies used the VAS for the assessment of pain and two used the AOFAS for the functional evaluation. Moreover there were other scores used for functional and pain evaluation. FIve studies assessed flexion as a clinical outcome.(23, Page 7/16 Page 7/16 24, 30, 33) There were small positive effects in the removal group for various outcomes. However the effects differ and are not statistical significant. Four before after studies show a statistical significant effect in favour of removal regarding functional and/or clinical measures after removal.(22, 30, 32, 34) 24, 30, 33) There were small positive effects in the removal group for various outcomes. However the effects differ and are not statistical significant. Four before after studies show a statistical significant effect in favour of removal regarding functional and/or clinical measures after removal.(22, 30, 32, 34) So far the approach of asymptomatic removal of implants is based on either the wish of the patient or on habits of the surgeon/hospital policies. Summary of evidence To help making more standardized decisions in the future, more research is needed especially considering the variety of patients and indications.(7) Along with diagnostic information, prognosis should guide clinical decision making. Right now decision for elective removal is partly based on prognostic factors, as for example patients with a younger age are recommended to have the hardware removed because the material will longer stay in the body.(37, 38) Research to identify prognostic factors (e.g. patient characteristics) for determining standardized decisions seems feasible and necessary. The identified studies showed high heterogeneity regarding surgical site and patients but also regarding study design and methods. It seems feasible to perform a systematic review on either a more specific question (e.g. Is the standardized removal of syndesmotic screws effective considering pain?) or analysing other aspects like prognostic factors. Due to the high numbers of implant removal and associated costs there is a need for more valuable evidence. The underlying evidence uses too short follow-up times to guide decision making reasonably. Furthermore the heterogeneity of outcomes is a problem in the interpretation of results. The use of uniform or generic measurement instruments (e.g. measuring quality of life with a generic instrument) is necessary alongside with longer follow-up times. Analysing this in RCTs or cohort studies fulfilling methodological standards (e.g. considering confounding) is required and potentially feasible. Finally the identified knowledge gaps and challenges show the need but also the possibility of performing more primary research on elective removal of hardware. Authors' contributions BP, TM and DP designed the study. BP wrote the first draft of the manuscript. DP developed the search strategy draft. BP and TM screened the titles and abstracts and full text independently. BP and TM extracted the data. All authors have been involved in revising the manuscript and approved the final manuscript. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. Competing interests The authors declare that they have no competing interests Limitations One of the main issues of this scoping review is that we included both, before-after and cohort studies. Interpretation of these types of studies is different and if a systematic review would be performed there should be various analyses for each study type. Methodological quality seems very heterogeneous even in the cohort studies. Above all, Dingemans et al. showed the feasibility of performing a RCT which should be the aim for future trials.(36) We limited publications to only high developed countries (WHO Stratum A) and also made restrictions to language. These restrictions could have excluded significant literature. But on the one hand we assumed less literature on surgery of asymptomatic patients in countries other than WHO Stratum A dealing with more serious health issues and doubt in their general transferability. And on the other hand we were not able to provide translations of other languages than English and German in the extent of a scoping review. Both should be fitted in a systematic review. Page 8/16 Conclusions We identified 13 studies that evaluated the effectiveness/safety of hardware removal in the extremities. The follow up times were short, the patient groups small and the ways of measurement differed. In general clinical heterogeneity was high. Evidence on selected topics e.g. syndesmotic screw removal is available nevertheless not sufficient to allow a meaningful assessment of effectiveness. Acknowledgements Mark Finlayson and Klazien Matter-Walstra (Federal Office of Public Health, Switzerland) contributed to this manuscript with their expertise in HTA and knowledge of the specific health care conditions in Switzerland. They both read and accepted the manuscript. Availability of data and materials Not applicable Ethics approval and consent to participate Ethics approval and consent to participate Not applicable Declarations Funding This work was funded by the Swiss Federal Office of Public Health. Abbreviations Page 9/16 Page 9/16 AOFAS American Orthopaedic Foot and Ankle Society ankle-hind foot score HTA Health Technology Assessment OMAS Olerud-Molander ankle score PRESS Peer Review of Electronic Search Strategies PRISMA-ScR Preferred Reporting Items for Systematic Reviews and Meta-Analyses for scoping reviews RCT Randomized Controlled Trial VAS Visual Analogue Scale WHO World Health Organization AOFAS American Orthopaedic Foot and Ankle Society ankle-hind foot score HTA Health Technology Assessment OMAS Olerud-Molander ankle score PRESS Peer Review of Electronic Search Strategies PRISMA-ScR Preferred Reporting Items for Systematic Reviews and Meta-Analyses for scoping reviews RCT Randomized Controlled Trial VAS Visual Analogue Scale WHO World Health Organization AOFAS American Orthopaedic Foot and Ankle Society ankle-hind foot score HTA Health Technology Assessment OMAS Olerud-Molander ankle score PRESS Peer Review of Electronic Search Strategies PRISMA-ScR Preferred Reporting Items for Systematic Reviews and Meta-Analyses for scoping reviews RCT Randomized Controlled Trial VAS Visual Analogue Scale WHO World Health Organization References 1. Busam ML, Esther RJ, Obremskey WT. Hardware removal: indications and expectations. The Journal of the American Academy of Orthopaedic Surgeons. 2006;14(2):113-20. 1. Busam ML, Esther RJ, Obremskey WT. Hardware removal: indications and expectations. The Journal of the American Academy of Orthopaedic Surgeons. 2006;14(2):113-20. 2. Suda AJ, Heilgeist E, Tinelli M, Bischel OE. High early post-operative complication rate after elective aseptic orthopedic implant removal of upper and lower limb. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 2018;36(3):1035-9. 2. Suda AJ, Heilgeist E, Tinelli M, Bischel OE. High early post-operative complication rate after elective aseptic orthopedic implant removal of upper and lower limb. Journal of orthopaedic research : official publication of the Orthopaedic Research Society. 2018;36(3):1035-9. 3. Reith G, Schmitz-Greven V, Hensel KO, Schneider MM, Tinschmann T, Bouillon B, et al. Metal implant removal: benefits and drawbacks – a patient survey. BMC Surgery. 2015;15(1):96. 3. Reith G, Schmitz-Greven V, Hensel KO, Schneider MM, Tinschmann T, Bouillon B, et al. Metal implant removal: benefits and drawbacks – a patient survey. BMC Surgery. 2015;15(1):96. 4. Sanderson PL, Ryan W, Turner PG. Complications of metalwork removal. Injury. 1992;23(1):29-30. 4. Sanderson PL, Ryan W, Turner PG. Complications of metalwork removal. Injury. 1992;23(1):29-30. 5. Unno Veith F, Ladermann A, Hoffmeyer P. [Is hardware removal a necessity?]. Revue medicale suisse. 2009;5(201):977-80. 5. Unno Veith F, Ladermann A, Hoffmeyer P. [Is hardware removal a necessity?]. Revue medicale suisse. 2009;5(201):977-80. 6. Richards RH, Palmer JD, Clarke NMP. Observations on removal of metal implants. Injury. 1992;23(1):25-8 6. Richards RH, Palmer JD, Clarke NMP. Observations on removal of metal implants. Injury. 1992;23(1):25-8. 7. Jamil W, Allami M, Choudhury MZ, Mann C, Bagga T, Roberts A. Do orthopaedic surgeons need a policy on the removal of metalwork? A descriptive national survey of practicing surgeons in the United Kingdom. Injury. 2008;39(3):362-7. 8. Brown RM, Wheelwright EF, Chalmers J. Removal of metal implants after fracture surgery-- indications and complications. Journal of the Royal College of Surgeons of Edinburgh. 1993;38(2):96-100. 9. Bostman O, Pihlajamaki H. Routine implant removal after fracture surgery: a potentially reducible consumer of hospital resources in trauma units. The Journal of trauma. 1996;41(5):846-9. 10. Destatis. Vollstationär behandelte Patientinnen und Patienten in Krankenhäuser 2018. Destatis Statistisches Bundesamt; 2019. Contract No.: 09.10.2019. 11. Minkowitz RB, Bhadsavle S, Walsh M, Egol KA. Removal of Painful Orthopaedic Implants After Fracture Union. JBJS. 2007;89(9):1906-12. 11. References Minkowitz RB, Bhadsavle S, Walsh M, Egol KA. Removal of Painful Orthopaedic Implants After Fracture Union. JBJS. 2007;89(9):1906-12. 12. Müller M, Mückley T, Hofmann GO. Kosten und Komplikationen der Materialentfernung. Trauma und Berufskrankheit. 2007;9(3):S297-S301. Page 10/16 Page 10/16 13. Hidaka S, Gustilo RB. Refracture of bones of the forearm after plate removal. The Journal of bone and joint surgery American volume. 1984;66(8):1241-3. 14. Tricco AC, Lillie E, Zarin W, O’Brien K, Colquhoun H, Kastner M, et al. A scoping review on the conduct and reporting of scoping reviews. BMC Medical Research Methodology. 2016;16(1):15. 15. Schepers T, Van Lieshout EM, Van der Linden HJ, De Jong VM, Goslings JC. Aftercare following syndesmotic screw placement: a systematic review. The Journal of foot and ankle surgery : official publication of the American College of Foot and Ankle Surgeons. 2013;52(4):491-4. 16. Vos DI, Verhofstad MH. Indications for implant removal after fracture healing: a review of the literature. European journal of trauma and emergency surgery : official publication of the European Trauma Society. 2013;39(4):327-37. 17. Prediger B, Mathes T, Pieper D. Scoping Report for the Evaluation of the Removal of Osteosynthesis Ma-terials in Switzerland. Bern: Bundesamt für Gesundheit Sektion Health Technology Assessment; 2019. 18. Pichon-Riviere A, Augustovski F, Garcia Marti S, Sullivan SD, Drummond M. Transferability of health technology assessment reports in Latin America: an exploratory survey of researchers and decision makers. International journal of technology assessment in health care. 2012;28(2):180-6. 19. Mue DD, Yongu WT, Elachi IC, Salihu MN. Surgeons’ Perceptions about Removal of Orthopaedic Implants in Nigeria. IOSR Journal of Dental and Medical Sciences. 2014;13(9):8-13. 20. Arksey H, O'Malley L. Scoping studies: towards a methodological framework. International Journal of Social Research Methodology. 2005;8(1):19-32. 21. McGowan J, Sampson M, Lefebvre C. An Evidence Based Checklist for the Peer Review of Electronic Search Strategies (PRESS EBC). 2010. 2010;5(1):6. 21. McGowan J, Sampson M, Lefebvre C. An Evidence Based Checklist for the Peer Review of Electronic Search Strategies (PRESS EBC). 2010. 2010;5(1):6. 22. Acklin YP, Michelitsch C, Sommer C. Elective implant removal in symptomatic patients after internal fixation of proximal humerus fractures improves clinical outcome. BMC musculoskeletal disorders. 2016;17:119. 22. Acklin YP, Michelitsch C, Sommer C. Elective implant removal in symptomatic patients after internal fixation of proximal humerus fractures improves clinical outcome. BMC musculoskeletal disorders. 2016;17:119. 23. Bell DP, Wong MK. Syndesmotic screw fixation in Weber C ankle injuries--should the screw be removed before weight bearing? References Injury. 2006;37(9):891-8. 23. Bell DP, Wong MK. Syndesmotic screw fixation in Weber C ankle injuries--should the screw be removed before weight bearing? Injury. 2006;37(9):891-8. 24. Boyle MJ, Gao R, Frampton CM, Coleman B. Removal of the syndesmotic screw after the surgical treatment of a fracture of the ankle in adult patients does not affect one-year outcomes: a randomised controlled trial. Bone Joint J. 2014;96-b(12):1699-705. 24. Boyle MJ, Gao R, Frampton CM, Coleman B. Removal of the syndesmotic screw after the surgical treatment of a fracture of the ankle in adult patients does not affect one-year outcomes: a randomised controlled trial. Bone Joint J. 2014;96-b(12):1699-705. 25. Dingemans SA, Birnie MFN, Sanders FRK, Van Den Bekerom MPJ, Backes M, Van Beeck E, et al. Routine versus on demand removal of the syndesmotic screw; A protocol for an international randomised controlled trial (RODEO-trial). BMC Musculoskeletal Disorders. 2018;19(1). 25. Dingemans SA, Birnie MFN, Sanders FRK, Van Den Bekerom MPJ, Backes M, Van Beeck E, et al. Routine versus on demand removal of the syndesmotic screw; A protocol for an international randomised controlled trial (RODEO-trial). BMC Musculoskeletal Disorders. 2018;19(1). 26. Garner MR, Thacher RR, Ni A, Berkes MB, Lorich DG. Elective removal of implants after open reduction and internal fixation of Tibial Plateau fractures improves clinical outcomes. Archives of orthopaedic and trauma surgery. 2015;135(11):1491-6. 26. Garner MR, Thacher RR, Ni A, Berkes MB, Lorich DG. Elective removal of implants after open reduction and internal fixation of Tibial Plateau fractures improves clinical outcomes. Archives of orthopaedic and trauma surgery. 2015;135(11):1491-6. 27. Gösling T, Hüfner T, Hankemeier S, Müller U, Richter M, Krettek C. Indikation zur Entfernung von Tibiamarknägeln. Der Chirurg. 2005;76(8):789-94. 27. Gösling T, Hüfner T, Hankemeier S, Müller U, Richter M, Krettek C. Indikation zur Entfernung von Tibiamarknägeln. Der Chirurg. 2005;76(8):789-94. Page 11/16 28. Gosling T, Hufner T, Hankemeier S, Zelle BA, Muller-Heine A, Krettek C. Femoral nail removal should be restricted in asymptomatic patients. Clinical orthopaedics and related research. 2004(423):222-6. 29. Hamid N, Loeffler BJ, Braddy W, Kellam JF, Cohen BE, Bosse MJ. Outcome after fixation of ankle fractures with an injury to the syndesmosis: the effect of the syndesmosis screw. The Journal of bone and joint surgery British volume. 2009;91(8):1069-73. 30. Miller AN, Paul O, Boraiah S, Parker RJ, Helfet DL, Lorich DG. Functional outcomes after syndesmotic screw fixation and removal. Journal of orthopaedic trauma. 2010;24(1):12-6. 31. References Tucker A, Street J, Kealey D, McDonald S, Stevenson M. Functional outcomes following syndesmotic fixation: A comparison of screws retained in situ versus routine removal - Is it really necessary? Injury. 2013;44(12):1880-4. 32. Chu A, Madou MR, Sala DA, Chorney GS, Feldman DS. Outcomes analysis after routine removal of implants in healthy pediatric patients. Journal of pediatric orthopedics Part B. 2009;18(6):381-7. 33. Briceno J, Wusu T, Kaiser P, Cronin P, Leblanc A, Miller C, et al. Effect of Syndesmotic Implant Removal on Dorsiflexion. Foot & ankle international. 2019;40(5):499-505. 33. Briceno J, Wusu T, Kaiser P, Cronin P, Leblanc A, Miller C, et al. Effect of Syndesmotic Implant Removal on Dorsiflexion. Foot & ankle international. 2019;40(5):499-505. 34. Dimitriou D, Waldmann S, Antoniadis A, Liebhauser M, Helmy N, Riede U. Early locking plate removal following open reduction and internal fixation of proximal humeral fractures could prevent secondary implant-related complications. Journal of Orthopaedics. 2020;17:106-9. 35. Goshima KA-Ohoo, Sawaguchi T, Shigemoto K, Iwai S, Fujita K, Kataoka T. Plate removal without loss of correction after open-wedge high tibial osteotomy is possible when posterior cortex bone union reaches osteotomy gap center even in incompletely filled gaps. LID - 10.1007/s00167-019-05615-y [doi]. (1433-7347 (Electronic)). 36. Dingemans SA, Birnie MFN, Sanders FRK, van den Bekerom MPJ, Backes M, van Beeck E, et al. Routine versus on demand removal of the syndesmotic screw; a protocol for an international randomised controlled trial (RODEO-trial). BMC musculoskeletal disorders. 2018;19(1):35. 36. Dingemans SA, Birnie MFN, Sanders FRK, van den Bekerom MPJ, Backes M, van Beeck E, et al. Routine versus on demand removal of the syndesmotic screw; a protocol for an international randomised controlled trial (RODEO-trial). BMC musculoskeletal disorders. 2018;19(1):35. 37. Muller-Farber J. [Metal removal after osteosyntheses. Indications and risks]. Der Orthopade. 2003;32(11):1039-57; quiz 58. 38. Manjoo A, Sanders DW, Tieszer C, MacLeod MD. Functional and radiographic results of patients with syndesmotic screw fixation: implications for screw removal. Journal of orthopaedic trauma. 2010;24(1):2-6. 38. Manjoo A, Sanders DW, Tieszer C, MacLeod MD. Functional and radiographic results of patients with syndesmotic screw fixation: implications for screw removal. Journal of orthopaedic trauma. 2010;24(1):2-6. Table Table 1 Individual Study outcomes Table 1 Individual Study outcomes Page 12/16 Pain Follow-up (IG/CG) Bell 2006 Pain free walking (n(%); p): 11(48)/2(29); p>0.05 15/16 months postoperativeb Boyle 2014 VAS (mean; MD; [95% CI]): 0.66/1.03; -0.38; [-1.01–0.26] 12 months postoperativeb Chu 2009 Pain (n(%); p): 7(28)/3(12); p=0.92 16.5 months postoperativeb (mean) Garner 2015 VAS (median; p): 0.6/0.5; p=0.64 15.4 / 40.6 months (median) Hamid 2009 VAS (mean; p): 0.074/2.02; p=0.268 30 months postoperativeb (mean) Functional Acklin 2016 Constant-Murley score (mean; MD; [95% CI]): 70.8/75.6; 4.8; [1.8– 7.8]. 9 ± 4 weeks Bell 2006 Baird and Jackson ankle score (mean, p): 88/86, p=0.79. Return to work (n(%), p): 13(57)/4(57), p>0.05 15/16 months postoperativeb Boyle 2014 Olerud–Molander ankle score (mean; MD; [95% CI]): 86.7/82.4; 4.3; [-5.2–13.9] American Orthopaedic Foot and Ankle Society ankle-hind foot score (mean; MD; [95% CI]): 90.1/88.6; 1.5; [-6.0–9.1] American Academy of Orthopaedic Surgeons foot and ankle score (mean; MD; [95% CI]): 91.8/87.0; 4.8; [ -3.5–13.2] 12 months postoperativeb Chu 2009 Pediatric Outcomes Data Collection Instrument: global functioning improved p=0.012 16.5 months postoperativeb (mean) Dimitrou 2020 Subjective increase of function n(%): 54 (96) 12 months postoperativeb Garner 2015 Knee Outcome Survey (median; p): 85/78.8; p=0.12 Lower Extremity Functional Scale (median; p): 80/66.3; p<0.05 15.4/40.6 months (median) Goshima 2019 Japanese Orthopedic Association score (mean, SD, p): 93.9(7.2)/94.7(6.2); p>0.05 Oxford Knee Score (mean, SD, p): 41.0(5.1)/43.1(4.7); p=0.03 12 months postoperativeb Gosling 2004 Complaints after nail removal (n(%)): 3(17) 6.3 years (mean) Gosling 2005 Complaints after nail removal (n(%)): 10(20) 7.4 years (mean) Hamid 2009 American Orthopaedic Foot and Ankle Society ankle-hind foot score (mean; p): 85.8/85.59; p=0.714 30 months postoperativeb (mean) Miller 2010 Olerud-Molander Ankle Score (mean; p): 42/75; p=0.002 Foot and Ankle Outcome Score (mean): Symptoms: 58/75 Pain: 65/79 Activities of daily living: 74/87 7 months postoperativeb Page 13/16 postoperative (mean) Functional Acklin 2016 Constant-Murley score (mean; MD; [95% CI]): 70.8/75.6; 4.8; [1.8– 7.8]. 9 ± 4 weeks Bell 2006 Baird and Jackson ankle score (mean, p): 88/86, p=0.79. Table Return to work (n(%), p): 13(57)/4(57), p>0.05 15/16 months postoperativeb Boyle 2014 Olerud–Molander ankle score (mean; MD; [95% CI]): 86.7/82.4; 4.3; [-5.2–13.9] American Orthopaedic Foot and Ankle Society ankle-hind foot score (mean; MD; [95% CI]): 90.1/88.6; 1.5; [-6.0–9.1] American Academy of Orthopaedic Surgeons foot and ankle score (mean; MD; [95% CI]): 91.8/87.0; 4.8; [ -3.5–13.2] 12 months postoperativeb Chu 2009 Pediatric Outcomes Data Collection Instrument: global functioning improved p=0.012 16.5 months postoperativeb (mean) Dimitrou 2020 Subjective increase of function n(%): 54 (96) 12 months postoperativeb Garner 2015 Knee Outcome Survey (median; p): 85/78.8; p=0.12 Lower Extremity Functional Scale (median; p): 80/66.3; p<0.05 15.4/40.6 months (median) Goshima 2019 Japanese Orthopedic Association score (mean, SD, p): 93.9(7.2)/94.7(6.2); p>0.05 Oxford Knee Score (mean, SD, p): 41.0(5.1)/43.1(4.7); p=0.03 12 months postoperativeb Gosling 2004 Complaints after nail removal (n(%)): 3(17) 6.3 years (mean) Gosling 2005 Complaints after nail removal (n(%)): 10(20) 7.4 years (mean) Hamid 2009 American Orthopaedic Foot and Ankle Society ankle-hind foot score (mean; p): 85.8/85.59; p=0.714 30 months postoperativeb (mean) Miller 2010 Olerud-Molander Ankle Score (mean; p): 42/75; p=0.002 Foot and Ankle Outcome Score (mean): Symptoms: 58/75 Pain: 65/79 Activities of daily living: 74/87 Sports and recreation: 49/62 Quality of life: 40/53 7 months postoperativeb Tucker 2013 Olerud-Molander Ankle Score (mean; p): 75.0/81.5; p=0.107 Excellent overall functional outcome (%, mean adjusted difference, [95% CI]): 23.26/25; –9.3; [-18.5 – -0.2] 31 months (mean) Page 13/16 Page 13/16 measures Bell 2006 Ankle range of motion (mean motion deficit in degrees compared to normal ankle, p) Flexion: 11.5/12.1, p>0.05 Inversion: 10.4/10.0, p>0.05 15/16 months postoperativeb Boyle 2014 Ankle dorsiflexion [degree] (mean; MD; [95% CI]): 13.0/10.2; 2.7; [-1.4–6.9] Ankle plantar flexion [degree] (mean; MD; [95% CI]): 31.2/33.6; -2.3; [-9.3–4.6] Calf girth loss [cm] (mean; MD; [95% CI]): 0.04/0.07M -0.21; [-0.69– 0.26] Tibiofibular clear space [mm] (mean; MD; [95% CI]):5.3/5.0; 0.34; [-0.28–0.95] 12 months postoperativeb Briceno 2019 Ankle dorsiflexion [degree] (mean; SE; p): 13.8(1.5)/10.1(2.4); p=0.129. Table Subjective improvement of dorsiflexion (n(%)): 10 (48) 3 months postoperativeb Dimitrou 2020 Range of motion [degree] (mean; SD; p): External rotation: 38(NR)/41(8.3); p=0.07 Abduction: 125(29)/140(25); p=0.001 Flexion: 130(27)/150(20); p=0 .001 12 months postoperativeb Goshima 2019 Radiological evaluation [degree] (mean; SD; p) Hip–knee–ankle angle: 4.1(2.5)/3.9(2.7); p=ns Medial proximal tibial angle: 94.0(3.0)/93.7(3.0);p=ns Posterior tibial slope: 9.2(3.2)/9.4(3.3); p=ns Weightbearing line ratio [%] (mean; SD; p): 67.8(10.0)/65.7(10.6); p=ns 12 months postoperativeb Miller 2010 Average range of motion [degree] (mean; p): Dorsiflexion: 10/20; p<0.05 Plantarflexion: 35/45; p<0.05 7 months postoperativeb Quality of life Garner 2015 Short Form-36 Survey (median; p): Mental Component Summary: 57.6/55.6; p=0.78 Physical Component: 50.9/44.9; p<0.05 15.4 / 40.6 months median) Adverse events Bell 2006 Syndesmotic screw malposition (n): 1/0 Syndesmotic screw breakage (n; p): 0a/2; p < 0.025 15/16 months postoperativeb Dimitrou 2020 Avascular necrosis n(%): 7(13) Other complications n(%): 0(0) 29 months postoperative b Hamid 2009 Syndesmotic screw breakage (n): 0a/10 30 months postoperativeb (mean) a=Screws have been removed before assessment; b= After index procedure CG: control group; IG: intervention group; MD: mean difference; SE: standard error; VAS: visual analogue scale; Figures Figures Page 14/16 Page 14/16 Page 14/16 Figure 2 Evidence map of identified literature The different colors depict the study types. The size of the bubbles are equivalent to patient numbers in total. The fragments show the different study types and the size of the fragments are equivalent to the patient numbers according to study type. The y-axis shows the indication and the x-axis shows the different outcomes evidence was identified for. Figure 1 Study selection flow diagram for the literature search on all indications Page 15/16 Page 15/16 Figure 2 Evidence map of identified literature The different colors depict the study types. The size of the bubbles are equivalent to patient numbers in total. The fragments show the different study types and the size of the fragments are equivalent to the patient numbers according to study type. The y-axis shows the indication and the x-axis shows the different outcomes evidence was identified for. Supplementary Files his is a list of supplementary files associated with this preprint. Click to dow PRISMAScRScopingRemoval0520.docx Appendices.docx Page 16/16 Page 16/16
https://openalex.org/W3199402362	https://www.jois.eu/files/14_988_Koziuk.pdf	English	null	Does political regime matter for abnormal hoarding of international exchange reserves?	Journal of International Studies	2,021	cc-by	12,616	Journal of International Studies Scientific Papers © Foundation of International Studies, 2021 © CSR, 2021 Journal of International Studies Scientific Papers © Foundation of International Studies, 2021 © CSR, 2021 Koziuk, V. (2021). Does political regime matter for abnormal hoarding of international exchange reserves?. Journal of International Studies, 14(2), 208-227. doi:10.14254/2071-8330.2021/14-2/14 DOI: 10.14254/2071- 8330.2021/14-2/14 Keywords: international exchange reserves hoarding, central banks, political regimes, financial development, quality of institutions. Received: May, 2020 1st Revision: March, 2021 Accepted: May, 2021 JEL Classification: E58, E59, O23, Q33 1. INTRODUCTION The tendency to maintain external assets of the central banks on a scale far exceeding traditional metrics is still evident. For example, median ratio of reserve holdings exceeds IMF ARA metrics. At the same time, reserves hoarding, particularly its large amounts, is heterogeneous. There are significant differences in the ability of some countries to benefit from competitiveness and resistance to financial shocks. Heterogeneity in the ability of reserves accumulation is determined directly by political and economic factors. Such factors are based on the quality of institutions that contribute to financial development, which reduces the need for reserves to absorb shocks. Although macroeconomic drivers of reserves accumulation are well known (openness, exchange rate regime, capital flow regime, commodity exports, etc.), political and economic factors remain a matter of debate. Political economy factors, in conjunction with the structural characteristics of the economy (e.g. resource wealth), determine the choice of macroeconomic policy regime (democracies tend to float, and autocracies vice versa) regarding reserves accumulation. On the other hand, the institutional characteristics of a country present a profile of macroeconomic policy, within which either the ability to reserves accumulation is made possible, or vice versa, the country is prone to lose them. The role of the central bank in the political economy of the reserve accumulation should reflect the broader political and economic choice of the design of the macro-financial stability mechanism. Given this, the question of the role of political economy drivers of abnormal exchange reserves hoarding is still relevant. At the beginning of the millennium, countries with different political regimes (as a group) were closer to maintaining reserves at the level of 4 months of import. However, later autocracies reached the level of nine months of import while democracies were closer to six respectively. This mean that during the recent period some structural changes in global economy affect the countries with different political regimes in the way that required a different vision of the efficient channels of macroeconomic adjustment in a new global environment. In the paper, the idea that different political regimes form different institutional preconditions for particular macroeconomic adjustment mechanism is explored. The main hypothesis of the paper is that the central bank independence as well as the institutions that evolve from the political regime may be crucial to our understanding why countries differ in their demand for abnormal reserves. Does political regime matter for abnormal hoarding of international exchange reserves? Viktor Koziuk Economics Department, West Ukrainian National University, Ukraine viktorkoziuk@wunu.edu.ua ORCID 0000-0002-5715-2983 Abstract. It is shown that political business-cycle approach to a political and institutional analysis of large exchange reserves hoarding is not full story. It is stressed, the choice of the central bank to accumulate reserves or not is not defined per se. Such choice is a consequence of political and economic consensus on the design of the mechanism of macrofinancial stability and the channels of macroeconomic adjustment. It is shown that political regimes are important proxies to understand why countries differ in incentives to hoard as well as accumulate in the long-run. At the same time, while autocracies outperforme democracies in reserves hoarding, the situation is not so simple. Institutional factors matter. This means that heterogeneity across political regimes may also play a role. The analysis of reserves distribution shows that autocracies are less homogenous. These results are substantiated by general multi-factor regression. While large hoarding is more typical of the countries with large GDP per capita, less democratic political regime and less independent central banks do not choose flexible exchange rates; i.e. only autocracy with some threshold level of development ( including political and macroeconomic stability) can hoard reserves sustainably. Autocracies have better preconditions for reserves hoarding only when they are able to push a prudent macroeconomic policy. At the same time, democracies have better preconditions for alternative institutional paths, under which some alternative channels of macroeconomic adjustment could potentially arise. This finding has strong practical implication. The independence of the central banks as well as the structural reforms addressed to financial development support macroeconomic adjustment with lower level of exchange reserves. JEL Classification: E58, E59, O23, Q33 208 Does political regime matter for abnormal hoarding of international exchange reserves? Viktor Koziuk 2. LITERATURE REVIEW International exchange reserves hoarding has been in the focus of global monetary order analysis for a long time and the global financial crisis only enhanced the interest to the issues. The growing role of global factors in the macro-financial vulnerability profile of countries has highlighted the importance of reserves as shock absorbers of capital outflows (Obstfeld et al., 2008). There are several key explanations for hypertrophy of external assets. The new (monetary) mercantilism means that international exchange reserves are being hoarded in order to avoid the movements in an exchange rate towards the PPP, thereby supporting the export-oriented growth (Durdu et al., 2007). Some sorts of these views are financial mercantilism (Aizenman and Lee (2008)), in which international exchange reserves are seen as a factor protecting the country’s financial sector from external stress, and the Bretton Woods II approach (Dooley et al., 2003, 2004). Capital inflows smoothing is further consideration to hoard the reserves. The advantages of price competition is thus the by-product attempt to reduce the vulnerability of the economy to capital flow reversals shocks (ECB, 2006; Wijnholds and Sondergaard, 2007; Choi et al., 2007). Precautionary demand or self-insurance became a very popular and influential way to understand the nature of reserve accumulation. The more vulnerable a country is to capital flow shocks, the more important is the role of reserves in ensuring macro-financial stability (Aizenman and Lee, 2008). Offsetting productivity shocks and smoothing exchange rate fluctuations in terms of financial sector weakness is also a very popular view. In the case where the latter cannot actively play the role of adjustment to shocks, and fluctuations in the exchange rate may deepen them, international exchange reserves are the best alternative (Aizenman and Riera-Crichton, 2006). In the line of continuous discussions about exchange rates and macroeconomic stability, Sanusa, Meyer and Hassan (2019) explain that the “fear of floating” drivers of reserve accumulation are still important even in the time of expended capital mobility and declined external fragility in many emerging countries. While they provide Southern African experience, such results are important form the view that the stage of development may affect the motivation for reserves accumulation. Therefore, mercantilist vs precautionary motives may be more important for countries that are at higher levels of economic development. Such conclusion helps to validate the GDP per capita variable as a proxy for reserve accumulation. 1. INTRODUCTION The paper demonstrates that central bank independence, flexible rates and financial development are kind of an institutional background for macroeconomic adjustment mechanism that requires less reserves hoarding in democracies. This does not mean that political business cycle approach is not relevant. This means that democracies can “invest” in institutions favourable for central bank independence and financial development, while autocracies should “invest” in the instruments supporting macroeconomic stability as a guaranty for political stability and saving the rent. At the same time, it is shown that autocracies are more heterogeneous in terms of reserve hoarding. Sustainable hoarding is thus possible in the environment when political and macroeconomic stability reinforce each other. This idea is proved with regression analysis revealing an important role of the country’s GDP level as a proxy for hoarding. While the relation between hoarding and development level is viewed as normal, such relation should be viewed through the lenses of specific institutional setup in countries with different political regimes that are relevant for the central bank independence, exchange rate flexibility and financial development. The paper provides literature review and further discussion related to the problem of political regime relevance to the institutions of macroeconomic adjustment. Empirical part is based on min-max analysis for countries with a homogeneity test and multiple regression methods for deeper understanding the role of factors that affect large reserve hoarding. The final part provides the conclusion. 209 209 Journal of International Studies Vol.14, No.2, 2021 2. LITERATURE REVIEW Competitive hoarding (Aizenman, 2007) is an evolutionary vision of how incentives for hoarding evolveover time. If one (more) country (ies) gains an obvious advantage over others through the reserves hoarding, others begin to follow it (them). This leads to a collective upward shift in demand for reserves, which nourishes individual demand. The country, which has lower losses from holding large reserves will win the competition in reserve accumulation, or this will be the country for which such losses are not significant from political and economic reasons. Another view on reserves accumulation is related to the debate on the “trilemma” of goals in an open economy. The incompatibility of pursuing monetary policy in accordance with the priority of internal equilibrium, gaining financial integration benefits, and seeking to limit exchange rate fluctuations can be substantially mitigated by the possession of significant external assets. The latter makes it possible to reduce the dependence of interest rates on external shocks with a lower risk of vulnerability to reversals in capital flows, and with the simultaneous ability to smooth exchange rate fluctuations (Aizenamn et al., 2015; Aizenman et al., 2013). However this view is “macroeconomically versatile” and empirically oriented, one can see hidden political motives and more complex institutional nuances that determine monetary policy priorities. The political and economic analysis of reserves accumulation presented in the literature is not homogeneous. 210 Does political regime matter for abnormal hoarding of international exchange reserves? Viktor Koziuk Aizenman and Marion (2004) have shown that the following features characterise most emerging market countries: reserves accumulation, external borrowing, inelastic spending behaviour, government heterogeneity in terms of the ability to follow the principles of maximising social well-being vs rent seeking groups. A number of important political and economic issues derive from this sort of analysis. First, autocracies had better accumulated reserves. In addition to the democratic regime factor, the tendency to reserves accumulation is determined by the size of the country’s economy, external vulnerability, and the IMF quota as a proxy for rapid access to international liquidity. Second, the level of corruption is inversely related to reserves. In other words, corrupted governments lose reserves and accumulate external debt, redistributing welfare to the benefit of individual rent-seeking groups. Third, governments are conditionally divided into tough – focused on maximising public welfare – and soft – prone to opportunistic behaviour. 2. LITERATURE REVIEW If the likelihood of re-election of the former is reduced, they tend to behave like the latter, while the latter, regardless of the electoral cycle, tend to lose reserves and increase debt. In other words, under the rule of soft governments, the social maximum is reached at lower reserves, since this means reducing the amount of theft in favour of an individual group and vice versa. However, there are other empirical data. For example, Han (2006) shows that reserves accumulation may be associated with macro-financial stability and the success of the government, which will make democracies inclined to hoard them in response to a public preference. The electoral dimension of the ability to accumulate external assets is shown in other works. Thus, the behavior of reserves is sensitive to the deployment of the electoral cycle. In emerging economies, as highlighted in Dreher and Vaubel (2009), the political business cycle is often represented by changes in reserves that offset spending increases and accelerate money supply. In particular, as highlighted in McGrath (2016), when governments change after crises, successors will be prone to increase demand for reserves as an additional toolkit to keep the macroeconomic situation favorable for re-election in the future. Jager (2016) implements a somewhat more complex approach to electoral analysis of reserve behavior. He focuses on the political business cycle issues across the political regimes. Empirically, autocracies are more successful in reserves accumulation than democracies. There are several reasons for this. The political business cycle is stronger feature of democracies, with the corresponding consequences for the external assets accumulation. Central banks in democratic countries are more independent and more focused on price stability, with which reserves accumulation may conflict (in the absence of sterilization). Similarly, in democratic countries, the number of veto-players is higher, which does not always guarantee the elimination of preconditions for the political business cycle, but raises the likelihood that the central bank will not only be formally but actually independent. Other types of publications focusing on motives that are ahead of demand on larger reserves in countries with different political regimes. For example, Leblang and Pepinsky (2008) showed that nature of political regime forming different motives for reserves accumulation. Democracies hoarding due to mercantilist pressure while autocracies mobilize external assets because of precautionary pressure. Democracies accumulate more reserves in the environment of strong grouping exporter’s interests that may capture the monetary policy. 2. LITERATURE REVIEW It must be acknowledged that a number of important issues remain unresolved. If democracies / autocracies are not homogeneous, what is the role of other institutional factors? For example, the lack of homogeneity in society can generate political conflicts that are incompatible with the reserves accumulation. The rule of law can have a positive effect on forces that allow adaptation to shocks that do not require significant reserves. Similarly, central bank independence can be an important factor in prioritizing price stability over reserves accumulation, but equally important for protecting reserves against “capture” by rent-seeking elites. Important is the fact about reserves hoarding. As the latter generate significant fiscal losses (and additional macro-financial risks), the question of society's tolerance to them arises. Can the central bank accept such losses taking into account the political regime in which it operates? Fractionalization may also be significant. If it is not compensated by rule of law or rigid dictatorship, it could affect political instability as well as political differences. In view of the above, it is hypothesized that: Level of development in emerging markets still important proxy for reserve hoarding but it should be viewed through the lense of how political and macroeconomic stability reinforce each over; Political regimes form basic preconditions for particular mechanism of macroeconomic adjustment from institutional viewpoint. Central banks independence with flexible exchange rates all together with financial development have better opportunities to evolve in the democracies adherent to rule of law that doesn’t require large hoarding. Autocracies may prefer alternative model of macroeconomic adjustment and preconditions for central bank independence. The same is for financial development. In this alternate setup, larger hoarding is needed to reinforce macroeconomic and financial loop of stability; Political stability, driven by variables of social homogeneity, plays a role in minimising political conflicts that do not correspond to larger hoardings in democracies due to macroeconomic mismanagement and in autocracies because of political coups. Both political regimes are vulnerable to rent seeking by elite groups but solve the problem differently which may affect hoarding. Therefore, it is assumed that social homogeneity as a prerequisite for minimising political conflicts between competing groups for capturing political and economic rent is significant for the external assets accumulation. The purpose of this paper is to systematize theoretical discussions around the political and economic drivers of reserves accumulation and to test empirically the above hypothesis. 2. LITERATURE REVIEW For autocracies, countries with macroeconomic stability are more important, so they are more prone to hoard in terms of prestige and political stability. They confirm this hypothesis on the evidence of 82 developing countries. Son (2020) represents the similar conclusions on the base of analysis of 127 countries. According to Son (2020) more outward oriented elites in autocracies prefer large reserves and this effect is stronger compare with democracies with large exporting sector. At the same time, size of exporting sector is important for democracies. In contrast to these approaches, Koziuk (2016a, 2016b, 2018) showed that both democracies and autocracies are not homogeneous due to the ability to reserves accumulation. There is also a more 211 Journal of International Studies Vol.14, No.2, 2021 sophisticated modality for choosing the design of a macro-financial stability mechanism. Autocracies, focused on converting resource richness into payments for loyalty, are prone to fixed exchange rates, dependent central banks, significant reserves and fiscal buffers. Democracies with quality institutions are prone to independent central banks, floating rates, and financial markets developing. Both expropriation autocracies and weak democracies are equally incapable of macro-financial stability, in particular because of a macroeconomic policy conflict with the ability to international exchange reserves hoarding. These conclusions drew from the example of resource-rich countries, but can be extrapolated on a wider sample. sophisticated modality for choosing the design of a macro-financial stability mechanism. Autocracies, focused on converting resource richness into payments for loyalty, are prone to fixed exchange rates, dependent central banks, significant reserves and fiscal buffers. Democracies with quality institutions are prone to independent central banks, floating rates, and financial markets developing. Both expropriation autocracies and weak democracies are equally incapable of macro-financial stability, in particular because of a macroeconomic policy conflict with the ability to international exchange reserves hoarding. These conclusions drew from the example of resource-rich countries, but can be extrapolated on a wider sample. This view is not in contrast to business cycle approach by Jager (2016) or motives ahead of hoarding explained in Leblang and Pepinsky (2008) and Son (2020). However, we differ in understanding the role of institutions that affects nature of macroeconomic adjustment in real political environment. The initial political choice of preferable channels of macroeconomic adjustment, quality of institutions as well as social fractionalization may determine efficiency of macroeconomic policy independly from motives that are ahead of reserves demand. 3. POLITICAL AND ECONOMIC ANALYSIS OF RESERVE HOARDING The fact of abnormal reserves hoarding began to be actively debated after 2002. Graph 1 shows that tendency to accumulate large reserves was prominent during 1990-s and early 2000s. Moreover, in medium term perspective such phenomenon is structural, not cyclical. The speed of hoarding decelerated since 2015 events of some distress in Chinese economy and previous commodity prices correction. However, it is obvious that many countries try to maintain exchange reserves that far exceed so called ordinary needs. Deviation from the ARA metrics, introduced by IMF, is a good sign that tendency to maintaining large reserves still actual and require further attention. Over time, abnormal hoarding is not strictly viewed as a witnessing of global imbalances, but rather as a precautionary normal. Crisis episodes on that time-path contributes to incentive for further accumulation contradicting in some way with efforts of transition toward more flexible exchange rates. Since many countries gained from increased exchange rates flexibility, the role of exchange reserves demonstrates no sign of diminishing. G raph 1. Median ratio between E xchange reserves d l t AR A it Graph 1. Median ration between Exchange reserves and relevance to ARA criteria Source: Constructed on IMF ARA Statistics. and relevance to AR A criterea 0 0,2 0,4 0,6 0,8 1 1,2 1,4 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Graph 1. Median ration between Exchange reserves and relevance to ARA criteria Source: Constructed on IMF ARA Statistics. Most Asian exporters exceeded the conventional reserve adequacy metrics at the time, covering three months of imports or covering 100% of short-term external debt (Disyatat 2003; Gosselin and Parent, 2005). Changes in the global economy are also important, but they do not explain by themselves the significant asymmetries in the ability of different countries to accumulate external assets in large scale. Commodity exporters have joined the tendency to accumulate large-scale reserves a little later in response to the dramatic increase in prices for natural resources. This clarification is not fundamental to the overall logic of reserves hoarding process. However, it is crucial to understand that reserves hoarding requires a favorable political and economic basis, not solely the fact of maintaining a balance of payments surplus. There are two lines of arguments here. 2. LITERATURE REVIEW It is found that level of development still important for large hoarding. This means that reinforced political and macroeconomic stability that positively affected sustainable growth is a strong precondition for reserves accumulation in large amounts. At the same time, central bank independence enforced by rule of law is strong driver of motivation to maintain smaller external assets. This corresponds to the notion that better institutional quality positively affects external sustainability by exchange rate flexibility and financial development. 212 Does political regime matter for abnormal Does political regime matter for abnormal hoarding of international exchange reserves? Viktor Koziuk g f hoarding of international exchange reserves? However, variables of society structure are not significant, meaning that better rule of law and political preconditions for economic development serve better needs to hoarding reserves. However, variables of society structure are not significant, meaning that better rule of law and political preconditions for economic development serve better needs to hoarding reserves. 3. POLITICAL AND ECONOMIC ANALYSIS OF RESERVE HOARDING The first line of arguments relate to the political and economic factors that come to the fore that the large scale reserves accumulation takes place in the medium and long term. In its purest form, this is contrary to standard open-economy macroeconomic models. In other words, hoarding that enabling reserves build-up should be based on appropriate macroeconomic policy mechanism for a relatively long time. In order for such a mechanism to function relatively sustainably, several conditions are necessary. 213 Journal of International Studies Vol.14, No.2, 2021 For both monetary and fiscal policies to comply with the principles of prudence (discipline), otherwise changes in aggregate demand induced by policy changes will inevitably affect the balance of payments and capital flows. In order for the central bank and the government to pursue prudent policies, they must be subject to some political immunity from possible changes in public perceptions of short-term priorities. Such conditions seem rather technical. However, taking into account that macroeconomic policy is implemented in the political environment, relevant to such environment institutions determine the extent to which the reserves hoarding will be demanded and viable in the political and economic sense. The viability of external assets accumulation, in turn, is also determined by a number of institutional factors that should facilitate the inflow of foreign currency as a main direction into the country rather than the outflow. This is possible due to incentives for export-oriented growth, sectoral competitive advantages (such as Singapore or Hong Kong’s financial system), or mobilized commodity rents. Regardless of the initial motive, reserves hoarding will entail the emergence of redistributive conflict between individual groups. The presence of political formal or informal institutions in which such a conflict is neutralized is an important precondition for maintaining a focus on reserve growth over time. Significant reserves should be associated with certain public benefit to form the basis for resolving conflict over the implementation of the hoarding policy. For democracy, this can be the macroeconomic stability in the form of better adaptability to shocks. The same applies to the autocracies, but with the notion that such reserves should serve as an additional toolkit for offsetting the exchange rate risks of the “payment for loyalty” policy. The reserves accumulation solely for the purpose of promoting geopolitical or leader ambitions, that is, without a certain relation to social welfare enhancement, is politically vulnerable. 3. POLITICAL AND ECONOMIC ANALYSIS OF RESERVE HOARDING Otherwise, political competition will inevitably lead to such a contest for a voter, which will result in a policy that makes it impossible to maintain large amounts of reserves. Autocracies in fragmented societies can maintain authority through a distribution of resource stock as “payment for loyalty”. Otherwise, they face the problem of violence. In the latter case, the country’s political landscape is unlikely to help maintain over- optimal reserves. the preconditions for a different perception of the role of social welfare reserves. Otherwise, political competition will inevitably lead to such a contest for a voter, which will result in a policy that makes it impossible to maintain large amounts of reserves. Autocracies in fragmented societies can maintain authority through a distribution of resource stock as “payment for loyalty”. Otherwise, they face the problem of violence. In the latter case, the country’s political landscape is unlikely to help maintain over- optimal reserves. Large-scale reserves generate significant losses on sterilization and maintaining. The risk of their increase is compounded by the degree of real flexibility of floating exchange rate. According to Rodrick (2006) and Turner (2006), social costs of abnormal reserves are very high. It follows that the fiscal price of excess central bank external assets should be taken at the political level. Alternatively, it should not be considered as a “price” in the eyes of elite groups that tolerant to holding reserves. Or, such a price should be a better policy alternative than when it is unattainable or undesirable to initiate changes in the quality of institutions that would lead to alternative ways of adjustment to shocks, such as the development of the financial sector or a floating exchange rate. The second line of argument relates to the problem of choosing in favor of reserves hoarding rather than the development of institutions that would be compatible with alternative mechanisms of adjustment to shocks. Alternative channels of shocks adjustment – the development of the financial market and the flexible exchange rate – do not work by themselves separately but rather complement each other. However, for them to work as channels of macroeconomic adjustment, for which there will be some political consensus, institutional preconditions are needed. In order to get shock-absorbing function of the financial sector, it is necessary to enter a stage of financial development at which the financial system will effectively shares risks and provide access to financing. 3. POLITICAL AND ECONOMIC ANALYSIS OF RESERVE HOARDING In democracies, this vulnerability stems from the competition for voters, who may be offered an alternative view of economic policy, which does not necessarily sacrifice individual consumption for the sake of large reserves as such. Autocracies are also not immune to political coups, especially when it comes to the struggle for commodity rents. Significant reserves may look like “prey”. The greater the amount of reserves diverges from conventional metrics, the greater the question of economic policy optimality is. In democracies, there must be a strong image of large-scale reserves as an important component of macro-financial stability. Accordingly, hoarding on a potentially suboptimal scale should be based on some political legitimation shared by competing social actors. In a competitive political environment, there must be rational arguments for why large-scale reserves are needed rather than just sufficient reserves. Moreover, such arguments should be relatively equally accepted by representatives of competing social groups. In autocracies, hoarding in large amount relies on a slightly different algorithm of legitimation. Significant reserves can be a “by-product” of a policy that, on the one hand, contributes to economic growth, on the other, is unlikely to be realized under another political regime. The austerity of export-oriented growth or the monopolisation of access to commodity rents as a way of preventing its wasting is typical examples. Moreover, even as they become a sign of the “success” of an autocratic regime in economic policy, the legitimation of significant reserves is transformed into a criterion for assessing whether autocracy is somehow “paying for loyalty” or turning into a tyranny of the elite. The latter is exacerbated by the negative value of net external assets (reserves minus external liabilities) due to the voluntary redistribution of social welfare in favor of the current unproductive consumption. Similar effects are well known in references on predatory rent capture and the corresponding degradation of macroeconomic discipline (Auty, 2001; Aizenamn & Marion, 2004; Arezki & Bruckner, 2010). The homogeneity of society is an important prerequisite for non-competing ethnic or social groups. Political tensions very often achieve a level that is incompatible with the implementation of economic policies conducive to reserves hoarding. In the case of democracy, the homogeneity of society eliminates 214 Does political regime matter for abnormal hoarding of international exchange reserves? Viktor Koziuk the preconditions for a different perception of the role of social welfare reserves. 3. POLITICAL AND ECONOMIC ANALYSIS OF RESERVE HOARDING It is no coincidence that in autocracies 215 Journal of International Studies Vol.14, No.2, 2021 the level of central bank independence is lower than that in democracies, but the latter are not by themselves a success guarantee of formally independent monetary authorities (Koziuk, 2016; Koziuk, 2018). The combination of the rule of law and the democratic regime brings the formal and de facto independence of the central bank closer. On the other hand, where price stability is the subject of political consensus, it is possible to achieve it even if the central banks independence is lower if there is a high level of collective confidence in society. If the central bank is in fact independent and there is a public consensus on price stability, the international exchange reserves hoarding on a large scale becomes unlikely in the long run. This implies that central banks which are more independent are not inclined to reserves hoarding, as noted by Jager (2016). They rely on achieving price stability not through hypertrophy of external assets but sound monetary policy, which in turn is protected from political interference. Moreover, the reason why stable and long-lasting political consensuses around the priority of price stability over reserve hypertrophy is possible largely due to the political interpretation of the role of the exchange rate in ensuring macro- financial stability. Since reserves accumulation and exchange rate flexibility are inversely dependent, but the exchange rate is a channel of macroeconomic adjustment, political consensus on a more flexible exchange rate will take place when the cumulative benefits of doing so will outweigh the losses from the fixation. The greater the effectiveness of a flexible exchange rate is supported by a more developed financial system, the more devastating the effects of the exchange rate fixation, the faster the political consensus will be reached toward exchange rate flexibility, a policy of price stability and supporting rule of law. And within the framework of the latter, financial system participants’ rights guarantee as well as guarantee of politicians’ non-interference in the central bank’s price stability mandate will occur. It is natural to assume that such a combination is more likely for democracies. Conversely, the more destructive for economy the exchange rate fluctuations are, the less legal institutions contribute to deepening the financial system, the more likely that political consensus will take place for greater exchange rate fixation. 3. POLITICAL AND ECONOMIC ANALYSIS OF RESERVE HOARDING On this basis, consumption is to be smoothed in time. Institutions to protect the investors and creditors rights are required to move toward this stage of financial development. Moreover, the political system itself must be non-expropriatory in its essence. Macroeconomic stability is also required for financial development (Aizenman et al., 2015; Cihak et al., 2012; Claessens & Laeven, 2003). Low inflation and the absence of drastic and significant changes in the macroeconomic policy are preconditions for generating returns on financial instruments consistent with the performance of the economy. Otherwise, there will be serious problems with the demand for funding. Given that, apart from legal institutions and the basic modality of the political regime, financial stability requires price stability, the question of the central bank independence as an institutional decision comes to the fore. Equally important is how such a solution affects the ability to choose the optimal exchange rate regime. The institutional choice in favor of an independent central bank would be unpredictable if reserves hoarding was a priority. It can be assumed that, in more ordinary situations, central bank independence is more likely to help maintain a certain optimal level of reserves, if this is not contrary to price stability, but not in the case of hoarding. There are several points here. Achieving price stability is largely based on the relevant mandate of the monetary authorities, their decision-making autonomy and political consensus that key social actors respect the central bank independence. All of this requires the rule of law. Some papers emphasize on this (Nurbayev, 2017; Hielscher & Markwardt, 2012). Rule of law by definition limits the actions of politicians and ensures that the delegated independence of the central banks is not abolished / replaced / distorted in accordance with electoral processes or other public challenges. Moreover, the rule of law is largely achieved under a democratic political regime. However, this does not mean that, there is no specific rule of law in autocracies. For the most part, it focuses on private law. Issues governed by public law fall under a somewhat different modality of regulation, consistent with the principle of balancing the preservation of autocratic power and the provision of rent as a payment for loyalty. 3. POLITICAL AND ECONOMIC ANALYSIS OF RESERVE HOARDING However, maintaining such a mix also requires appropriate macroeconomic policies. If the general weakness of institutions extends to the central bank, its ability to maintain a fixed exchange rate is limited in time. Otherwise, the influence of weak institutions on monetary policy should be eliminated. In democracies with weak institutions, this is difficult, if not impossible. Weak institutions open a wide field for competitive populism and subordination of monetary policy to electoral goals, or in some cases redistributive goals by “captured” regulator. In autocracies, this problem is solved easier. Under this political regime, there is an opportunity radically reduce political costs of budget negotiations. Moreover, this opens up another possibility in which a more dependent central bank focuses on supporting fixed exchange rate, and price stability is achieved through counter-cyclical fiscal policy. The reserves accumulation is a continuation of the link between autocratic mode, fixed rates and a weak central bank. The case of reserves hoarding in commodity autocracies indicates why in a certain political regime more often choose a fixed rate (Aliyev, 2014; Wills & van der Ploeg, 2014). On the one hand, the answer to this question is that this is the easiest way to ensure macroeconomic preconditions for regime loyalty by a controlled central bank. On the other hand, there is a pattern that autocracy is a way to counteract the waste of commodity rent by monopolising access to it. Furthermore, significant foreign exchange inflows to commodity exporters will be absorbed by foreign exchange reserves and sovereign wealth funds, which allows for the support of fixed exchange rates, the implementation of counter-cyclical fiscal policies, which together solve the problem of ensuring macroeconomic stability. The tendency of commodity economies to fixed rate to some extent derived from the amplitude of fluctuations in world prices. Their negative effects are better offset by changes in reserves rather than changes in the nominal exchange rate, taking into account its impact on other sectors of the economy with 216 Does political regime matter for abnormal hoarding of international exchange reserves? Viktor Koziuk weak financial system (Aliyev, 2012; Aliyev, 2013; Aliyev, 2014). Such weakness in commodity economies stems from the fact that extractive industries are prone to significant external borrowing. Moreover, the national financial sector cannot meet the large-scale resource requirements because of the problem of institutional quality (Kuronen, 2012). 4. DEMOCRACIES VS AUTOCRACIES: PRELIMINARY CONSIDERATIONS AND STYLIZED FACTS The above analysis clearly shows that the political regime can be a good example of understanding why countries can differ not only through success in accumulating foreign exchange reserves, but also through incentives to do so on a large scale. Jager (2016) mainly emphasizes on the autocracies leadership in reserves hoarding concerning political business cycle. Leblang and Pepinsky (2008) and Son (2020) consider motives to accumulate reserves as for group interest meaning that such interests may vary across the countries with different political regimes. Such views are important, but not fit all possible cases. It is true that strict control over veto-players or budget redistribution issues support autocracies to accumulate reserves. The same is for political internalization of maintaining costs problem. However, it is very important to note that democracies as well as autocracies are not homogenous. This means that different political regimes are not incentives for hoarding and opportunities in the long run. Democracies and autocracies should generate institutional principle due to which large hoarding would be possible. This principle can go beyond the egoistic principles of interest groups. Democracy with strong rule of law may restrain egoistic interests of elite groups. While autocratic regime may introduce control elite groups and compensate them for loyalty. This also means that without strong institutional support countries with different political regimes are similarly vulnerable to the problem of exchange reserves lost. Table 1 summarizes the political and economic analysis of reserves hoarding problem in the context of alternative political regimes. Table 1 presents that political regime is not only precondition for reserves accumulation. However, at the same time, each institutional factor support or not support large hoarding in different ways. Jager (2016) put it well, describing that electoral challenge for reserves accumulation is different across countries with different political regimes. The approach of Leblang and Pepinsky (2008) and Son (2020) is also important because it postpones the problem of group interests in reserve accumulation. Some empirical facts confirm the idea that reserves accumulation is more successful in the general autocracies. Graph 2 demonstrates that autocracies advantage over democracies was insignificant before the financial imbalances and the era of the commodity boom. At the same time, countries with democratic regimes demonstrate more stable ratio of exchange reserves in month of import that slightly rose during two decades. Simultaneously, autocracies demonstrate dramatic difference in reserves hoarding, but mostly in times of global expansion. 3. POLITICAL AND ECONOMIC ANALYSIS OF RESERVE HOARDING It is no coincidence that the commodity boom of the early 2000s was a turning point when autocracies were ahead of democracies hoarding large-scale reserves ( Jager, 2016). Naturally, it is the specifics of the institutions that make it possible to come to the fore, bearing in mind that the fight for rent is in the opposite direction. On the other hand, the example of Asian exporters emphasizes that the extraction factor may not be decisive. 4. DEMOCRACIES VS AUTOCRACIES: PRELIMINARY CONSIDERATIONS AND STYLIZED FACTS It is very likely that autocratic countries run with advantage in reserves accumulation mostly in the case of commodity windfalls. This mean that commodity prices, interacting with specific political and institutional dimension of resource abundance, support the idea that autocratic regimes are better tailored for reserves accumulation. However, this statement also means that not the political regime itself play key role. Politically vulnerable autocracy may lose reserves during times of political instability and resource richness may support that due to bigger temptation of competing groups to monopolize the rent. 217 Journal of International Studies Vol.14, No.2, 2021 Table 1 Institutional dimension of reserves hoarding across the political regimes Autocracies Democracies The motive of hoarding An “by-product” of economic policy that makes it possible to “pay for loyalty”, which subsequently becomes a symbol of regime power Creating an “airbag” turns into an image of a key element of macro-financial stability, for which there is relative public agreement Electoral cycle Sensitive dynamics of reserves volume to the electoral cycle occurs only in some cases The sensitivity of reserves to the electoral cycle is inversely proportional to the power of public consent to their key importance as a macro-financial buffer The role of the exchange rate As autocracies gravitate to fixed rates, the reserves accumulation replaces the process of macroeconomic adjustment As democracies are more typical of a floating exchange rate, reserves play the role of an optimal mix of different channels of macroeconomic adjustment The role of financial sector The financial sector can act as a shock absorber in some cases where autocracy creates specific institutions for the development of the financial system. In general, the relative weakness of the financial sector leads to an alternative macroeconomic adjustment in the form of fluctuations in reserves or changes in fiscal buffer stocks A strong financial sector is able to act as a shock absorber, but this requires high quality institutions, which together allow for alternative macroeconomic adjustment. In other cases, reserves offset institutional failures of financial development. In the case of weak institutions, the country is unable to either develop the financial sector or hoard reserves. The role of central bank independence The central bank institutional weakness in maintaining price stability is offset by an active fiscal policy (fiscal buffers). 4. DEMOCRACIES VS AUTOCRACIES: PRELIMINARY CONSIDERATIONS AND STYLIZED FACTS The central bank is subordinated to the needs of maintaining a stable exchange rate and appropriate management of significant reserves In its purest form, an institutionally weak central bank is not able to “protect” reserves in case of irresponsible government policy. On the other hand, a strong central bank can guarantee price stability with a lower volume of reserves. To hoard reserves in large amount, a specific combination of the “metered” independence of the central bank with the political prerequisites for a prudent macroeconomic policy is required Fiscal losses They may be ignored by the absence of an alternative political view on economic policy May be accepted if there is a general tolerance for reserve hypertrophy Social fragmentation It is neutralized by the specificity of the political regime and its ability to “pay for loyalty”. In other cases, significant reserves can turn into a “prey” of competing social fractions It complicates the reserves hoarding, since the redistributive effects and fiscal losses of such policies are easily transformed into a toolkit for political struggle between conflicting social groups Resource wealth It is a basic precondition for reserves hypertrophy. Sovereign wealth funds are a functional continuation of an active fiscal policy It is a basic precondition for reserves hypertrophy only when reaching a certain level of institutional maturity. Creating a sovereign wealth fund may be a better alternative Export-oriented growth It is an important precondition for reserves hypertrophy, as it is easily made possible by the political regime, which limits political requests for greater current consumption It can only lead to hoarding if there is strong public agreement to abstain from current consumption. However, this situation is always vulnerable to alternative options for stimulating growth in a competitive political environment, making reserves hypertrophy unlikely Institutional dimension of reserves hoarding across the political regimes 218 Does political regime matter for abnormal hoarding of international exchange reserves? Viktor Koziuk While Jager (2016) used reserves to GDP metrics, graph 2 demonstrates another way to express the imilar tendency on the data of 164 countries. Graph 2. Exchange reserves in month of import across different political regimes Source: Authors calculations based on IMF Statistics. Countries are divided on different groups according to Democracy Index. Higher score than 4 means the democracy. G raph 2. 4. DEMOCRACIES VS AUTOCRACIES: PRELIMINARY CONSIDERATIONS AND STYLIZED FACTS E xchange res erves in month of import acros s different political regimes 0 1 2 3 4 5 6 7 8 9 10 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Autocracies D emocracies While Jager (2016) used reserves to GDP metrics, graph 2 demonstrates another way to express the similar tendency on the data of 164 countries. G raph 2. E xchange res erves in month of import acros s different political regimes 10 Graph 2. Exchange reserves in month of import across different political regimes Source: Authors calculations based on IMF Statistics. Countries are divided on different groups according t Democracy Index. Higher score than 4 means the democracy. While it is important to recognize that political regimes are not preconditions for reserves hoarding, Graphs 3 and 4 show that heterogeneity of the democracies and autocracies is important. Min-max analysis demonstrates some differences in reserves distribution across the countries with different political regimes. Graphs 3-4 demonstrate some novelty on the area of reserves accumulation across different political regimes. At first, democracies were relatively more successful before the commodity boom. On initial stages of hoarding process, they were much closer to autocracies as it is shown in graph 2. However, evolving, process of reserve accumulation went different ways in two groups of countries. Democracies are less volatile in that regard. This also supports the idea of commodity prices role. Secondly, democracies are more homogenous. Distance between minimum, maximum and average levels of reserves accumulation is much smaller than in autocracies. This means that democratic political regime is more open for searching for alternative adjustment channels and more sensitive to costs of maintaining. Minimum levels of reserves are higher in democracies. While average levels are similar in both cases, such levels only in particular years after global financial crisis are higher in autocracies substantially. However, maximum levels are dramatically higher in autocracies. Figure 3. Democracies. Exchange reserves in month of import Source: Authors calculations based on IMF Statistics. Countries are divided on different groups according to Democracy Index. Higher score than 4 means the democracy. Grapth 3. Democracies. E xchange reserves in month of import 0 5 10 15 20 25 30 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Max Min Median Figure 3. Democracies. 4. DEMOCRACIES VS AUTOCRACIES: PRELIMINARY CONSIDERATIONS AND STYLIZED FACTS Exchange reserves in month of import Figure 3. Democracies. Exchange reserves in month of import Source: Authors calculations based on IMF Statistics. Countries are divided on different groups according to Democracy Index. Higher score than 4 means the democracy. g g p Source: Authors calculations based on IMF Statistics. Countries are divided on different group Democracy Index. Higher score than 4 means the democracy. calculations based on IMF Statistics. Countries are divided on different groups according to Democracy Index. Higher score than 4 means the democracy. 219 Journal of International Studies Vol.14, No.2, 2021 Figure 4. Autocracies. Exchange reserves in month of import Source: Authors calculations based on IMF Statistics. Countries are divided on different groups according to Democracy Index. Higher score than 4 means the democracy. Grapth 4. Autocracies. E xchange reserves in month of import 0 10 20 30 40 50 60 70 80 90 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Max Min Median Figure 4. Autocracies. Exchange reserves in month of import Source: Authors calculations based on IMF Statistics. Countries are divided on different groups according to Democracy Index. Higher score than 4 means the democracy. Such evidence is a strong argument that institutions are important. Weak and expropriation-oriented autocracy cannot accumulate reserves as well as weak democracy vulnerable to populism. Simultaneously, large hoarding is also possible in autocracies due to the peculiarities of the external adjustment mechanism, the institutional set-up of central banks or the inconsistency with the content of the problem of expenditures. These considerations are supported by more formal proofs in next part of the paper. Such evidence is a strong argument that institutions are important. Weak and expropriation-oriented autocracy cannot accumulate reserves as well as weak democracy vulnerable to populism. Simultaneously, large hoarding is also possible in autocracies due to the peculiarities of the external adjustment mechanism, the institutional set-up of central banks or the inconsistency with the content of the problem of expenditures. These considerations are supported by more formal proofs in next part of the paper. THODOLOGY: AN EVALUATION OF RESERVES ACCUMULATION Some heterogeneity across countries described in previous chapter requires deeper further analysis. Patterns of uneven distribution of reserves among the countries are better understood using the institutional powers of attorney as independed variables. Simultaneously, there are no consensus in references which variable is a good proxy for abnormal accumulation. Empirical research does not show agreement on which variable is chosen to be dependent. In some cases, these are reserves per unit of population, reserves to M2, or reserves to GDP (Jager, 2016; Han, 2006; Aizenman & Marion, 2004). The same for independent variables, the choice of which is dictated by specific research needs. In order to emphasize the fact of hoarding, constructing a multifactor regression model (tested in nine specifications, table 2), and the independent variable is specified as “reserves in months of import – 3” and “ratio between reserves and short-term external debt – 100%”. In both cases, the source of data is World Development Indicators provided by World Bank. The latter variable is hypothetically better suited to the idea of precautionary demand, but is available in far fewer countries. It was used to test the basic equation, which is generally similar, so these results are not shown. For reasons of statistical validity, the indicator “reserves in months of imports – 3” was chosen because it covers a much larger number of countries. The sample covers 164 countries for the period 2000-2018. Availability of data is important factor of time length selection. The choice of independent variables in this paper directly follows from the theoretical considerations. Such considerations are derived from assumptions about the potential influence of political-economic / institutional factors on the ability of central banks to accumulate external assets. Because of this, the 220 Does political regime matter for abnormal hoarding of international exchange reserves? Viktor Koziuk selected dependent variable is tested for relation with composition of independent variables that are not presented in previous studies. Controlled independent variables are: Ln GDP per capita. According to most theoretical explorations, it is assumed that reserves hoarding requires a certain scale effect. Moreover, many poor countries do not have the ability to reserves accumulation due to the weakness of institutions and the atrophy of foreign currency inflows. Economic development is also an indirect indicator of the political and macroeconomic loop of stability. Accordingly, a positive correlation is expected between the GDP per capita and the selected dependent variable. THODOLOGY: AN EVALUATION OF RESERVES ACCUMULATION Logarithmisation is performed to neutralize the effect of significant variation in the sample. Data source – World Bank Statistics (World Development Indicators); FO – financial openness. Theoretically, the more financially open a country is, the more flexible is an exchange rate, the more capital flows are adjusted to domestic macroeconomic processes, the less likely it is to be vulnerable to significant capital flow shocks, which reduces demand for reserves. The relationship between financial openness and the dependent variable is expected to be negative. The normalized Chinn- Ito Index value was selected from the Aizenman-Chinn-Ito Trilemma Statistics as statistical quantitative indicator of financial openness; FO – financial openness. Theoretically, the more financially open a country is, the more flexible is an exchange rate, the more capital flows are adjusted to domestic macroeconomic processes, the less likely it is to be vulnerable to significant capital flow shocks, which reduces demand for reserves. The relationship between financial openness and the dependent variable is expected to be negative. The normalized Chinn- Ito Index value was selected from the Aizenman-Chinn-Ito Trilemma Statistics as statistical quantitative indicator of financial openness; Ln FD – financial depth. It is assumed that the deeper the financial system is, the better it absorbs shocks, reducing demand for reserves. This variable is represented by the logarithmic indicator of the ratio of domestic credit to private sector to GDP according to World Bank Statistics (World Development Indicators). Although this variable does not fully reflect financial developments and may not take into account the financial system model (market-based/banking-based), this indicator is more representative for most low- and middle-income countries. The theoretically predictable is inverse correlation; Ln Com – share of non-industrial exports. This variable is selected as an indicator for resource wealth, with the adjustment that the nature of the vulnerability is similar for countries with developed agrarian sectors. Data source – World Bank Statistics (World Development Indicators). Theoretically, resource wealth should be directly related to the demand for reserves, but this assumption is not rigid, since it requires additional institutional considerations. Furthermore, not all countries that have hoarded reserves are commodity exporters. The Economic Complexity Index (ECI) variable is used for additional control. As commodity specialization is inversely related to economic complexity, it is likely that the latter may better explain the ability to accumulate reserves in countries not burdened with resource wealth. THODOLOGY: AN EVALUATION OF RESERVES ACCUMULATION The interaction of this indicator with the rule of law and the exchange rate flexibility index (ExRF), which is a normalized value of exchange rate volatility from statistical base the Aizenman- Chinn-Ito Trilemma Statistics, is additionally tested to assess the actual independence of the central bank. The exchange rate variable is not selected directly for testing because of the potential for nonlinear links to the reserves accumulation process. However, in conjunction with central bank independence, it better reflects the profile of the monetary regime, which should see a lower demand for foreign exchange reserves. THODOLOGY: AN EVALUATION OF RESERVES ACCUMULATION In this case, the connection is assumed; DI – political mode variable represented by the Democracy Index, calculated by the Heritage Foundation. An index value of < 4 means an autocratic regime, > 4 means a democratic regime (with different variants of the quality of democracy, which, however, is not the focus of the research). The link may theoretically be direct and inverse depending on additional institutional considerations, although the coefficient is expected to be negative in other empirical evidence. The variables of the institutional analysis are presented as follows: RL - rule of law index. An indicator of the rule of law on World Bank Statistics (World Development Indicators) aims to represent the quality of institutions, as the country is most general but capacious institutional characteristic. The link is tested directly (although its direction can theoretically be both direct and inverse) and in conjunction with the central bank independence index; FE – indicator for social heterogeneity. This variable is presented as the Fractionalized Elites sub- index of the Fragile State Index, which attests to the elite fragmentation or the distance between the interests of different social groups, pointing to potential tensions in society caused by its socio-economic heterogeneity. The Peace Fund calculates this indicator. It is suggested that a lack of homogeneity in society should weaken the ability to reserves accumulation; FE – indicator for social heterogeneity. This variable is presented as the Fractionalized Elites sub- index of the Fragile State Index, which attests to the elite fragmentation or the distance between the interests of different social groups, pointing to potential tensions in society caused by its socio-economic heterogeneity. The Peace Fund calculates this indicator. It is suggested that a lack of homogeneity in society should weaken the ability to reserves accumulation; 221 Journal of International Studies Vol.14, No.2, 2021 EF – alternative indicator for social heterogeneity. In this case, it is an indicator of ethnic fragmentation calculated to analyze the impact of potential interethnic tensions on economic processes (Alesina et al., 2002). Similarly, the likely conflicts caused by a lack of ethnic homogeneity limit the ability to reserves accumulation; GMT – GMT index of central bank independence has been updated for more countries (Crowe and Meade, 2008). Given the theoretical debate, it is assumed that it should be inversely related to the dependent variable. 6. EMPIRICAL RESULTS AND DISCUSSION Using simple multy-factor regression model we obtain the results that in general confirms the general logic of introduced assumptions. However, some empirical results are not in line with predictions mostly due to statistical quality of relations. Nevertheless, as it would be shown, such unexpected results could be considered as not contradictive with institutional approach explored in the paper. Table 2 presents regression analysis results. Based on data analysis results in table 2 the following generalizations can be made. Based on data analysis results in table 2 the following generalizations can be made. There is the undeniable fact that the ability to reserves hoarding is very sensitive to the scale effect. The economy of scale variable has a theoretically predictable sign in all specifications. Moreover, all specifications have a statistically significant links except one. Theoretically, this pattern can be explained by the fact that the economy of scale indirectly reflects a specific reduction in the costs of sterilization and other negative manifestations of reserves hypertrophy. This conclusion does not mean that the ability to accumulate reserves is determined solely by the level of economic development. Rather, regression analysis results indicate that the excess reserves are more sensitive to the level of economic development in the sense that it provides a stable and significant amount of net currency inflows. This means that only countries that are politically and macroeconomically resilient can accumulate large reserves. Taking into account that different countries from political regime side form the group, such results also indicate that in different political regimes different factors constitute macroeconomic and political loop of stability that positively affect reserves accumulations. Financial openness demonstrated a theoretically predicted inverse relationship with the dependent variable, although its statistical significance is insufficient. However, indicator of commodity wealth has not shown a direct relation, as could be expected. With the exception of one specification, the statistical significance of this variable is also insufficient in all cases. It can be assumed that structural factors are generally, important in explaining a country’s ability to reserves hoarding but correcting for political and economic features. This is clearly seen in the case of credit depth. It has the correct direction of relation in all specifications with the exception of two, but not all statistically significant. Specification (3) is interesting. 222 Does political regime matter for abnormal hoarding of international exchange reserves? 6. EMPIRICAL RESULTS AND DISCUSSION Viktor Koziuk Table 2 Regression analysis results (1) (2) (3) (4) (5) (6) (7) (8) (9) Ln GDP per capita 1,36798 (2,8846) 0,00467 1,07927 (2,4946) 0,01388 1,20756 (2,1827) 0,03128 0,37911 (0,7105) 0,47941 1,51766 (3,0591) 0,00276 1,42689 (2,951) 0,00386 1,46304 (3,1031) 0,00241 0,6351 (1,5267) 0,13066 1,30922 (3,16823) 0,001961 FO -2,04849 (-1,3688) 0,17368 -1,71269 (-1,2811) 0,20247 -1,01332 (-0,5845) 0,56012 -0,3709 (-0,273) 0,7857 -1,8434 (-1,221) 0,22451 -2,1805 (-1,395) 0,16565 -1,772 (-1,192) 0,23571 -0,0638 (-0,050) 0,96024 -1,72607 (-1,1669) 0,24562 Ln FD -1,11198 (-1,4323) 0,15472 -0,30585 (-0,4407) 0,66016 0,17165 (0,19252) 0,84771 0,0147 (0,0198) 0,98424 -1,1332 (-1,459) 0,14714 -0,9942 (-1,237) 0,21872 -1,3364 (-1,721) 0,0879 Ln Com -0,90184 (-1,3909) 0,16691 -0,9329 (-1,597) 0,11411 -0,8206 (-1,257) 0,21149 -1,0349 (-1,531) 0,12849 -1,0494 (-1,625) 0,10684 -1,1002 (-2,131) 0,03607 -0,93862 (-1,509) 0,133901 ECI -0,42011 (-0,5443) 0,58741 DI -0,90949 (-3,2749) 0,00139 -0,73942 (-2,8239) 0,00550 -0,85014 (-2,6031) 0,01058 -0,3121 (-0,799) 0,42631 -0,6883 (-1,957) 0,05273 -0,8971 (-3,182) 0,00189 -0,7329 (-2,531) 0,01273 -0,1985 (-0,611) 0,54265 -0,73078 (-2,527) 0,01284 RL -5,9875 (-1,117) 0,2672 FE 0,29636 (1,0249) 0,30752 EF 1,51139 (0,7549) 0,45189 GMT -4,5379 (-1,917) 0,05778 GMT* RofL -10,204 (-3,384) 0,00109 GMT* ExRF -1,40767 (-2,3893) 0,018494 Intercept 14,00840 (3,3389) 0,001131 6,68796 (3,3966) 0,000909 5,02533 (1,6713) 0,09764 10,2204 (2,4622) 0,01593 10,2469 (1,8386) 0,06856 13,3754 (3,0923) 0,00249 16,7764 (3,8196) 0,00022 9,9503 (4,0031) 0,00014 12,1308 (3,7887) 0,00024 R2 0,1488 0,1013 0,1071 0,107 0,1565 0,1522 0,1752 0,2041 0,1738 F Statistics F(5,116) = 4,0559 F(4,128) =3,6053 F(5,105) =2,5198 F(6,81) = 1,619 F(6,115) = 3,557 F(6,114) = 3,412 F(6,115) = 4,07 F(5,82) = 4,205 F(5,116) = 4,900 Source: Calculated using the package STATISTICA. The use of the Economic Complexity Index as an indicator of resource richness affects changes in sign in case of the credit depth indicator. This may indicate that the more diversified the economy, the more advanced the banking system needs to be “protected” by reserves. Such an interpretation is close to the phenomenon of financial mercantilism, as pointed by Aizenman and Lee (2008). The same can be seen in specification (4) with indicator of the rule of law. Countries with low quality of institutions that 223 Journal of International Studies Vol.14, No.2, 2021 seek to deepen the financial system are increasing demand for reserves. However, in these specific cases, the statistical links significance is insufficient. Nevertheless, this does not exclude the importance for some countries of the motive for the accumulation associated with the policy of financial protectionism. 6. EMPIRICAL RESULTS AND DISCUSSION Simultaneuosly, such findings rather support discussion with results presented in Leblang and Pepinsky (2008) and Son (2020). Low significance of structure of economic factor in contradiction to export- interest groups can influence the political agenda before reserves accumulation. seek to deepen the financial system are increasing demand for reserves. However, in these specific cases, the statistical links significance is insufficient. Nevertheless, this does not exclude the importance for some countries of the motive for the accumulation associated with the policy of financial protectionism. Simultaneuosly, such findings rather support discussion with results presented in Leblang and Pepinsky (2008) and Son (2020). Low significance of structure of economic factor in contradiction to export- interest groups can influence the political agenda before reserves accumulation. The political regime is a fundamental factor that determines the ability to hoard reserves in large amount. The democracy index is in inverse relation with dependent variable in all specifications, and this relationship is statistically significant except for the specification (8). The ability of autocracies to reserves hoarding, however, is not explained solely by the commodity factor, as might be expected. Most likely, in favor to autocracies play a certain set of factors, namely: insensitivity of reserves to the political business cycle; implicit guarantees of a stable macroeconomic performance; ability to build counter-cyclical fiscal mechanisms without costs for budget negotiations; a selective approach to investor rights guarantees; the need for reserves to offset the risks of balance of payments deterioration as a result of “payment for loyalty” policy. It is likely, that institutional factors of potential inability to reserves accumulation like ethnic fragmentation or the competing elite groups are offset by the ability of autocracies to make the rational macroeconomic policies. This conclusion does not deny the fact that often autocracies in heterogeneous societies tend to seek rent and expropriate political competitors. Political vulnerability to coups and the inability to pursue prudent macroeconomic policy could be precondition for social discontent. Autocratic regimes that are focused on a particular kind of social inclusion are most interested in reserves hypertrophy because it is their kind of success in offsetting all other institutional imperfections. The rule of law has not directly demonstrated a significant impact on the ability to reserves hoarding. There is an inverse relationship between the rule of law index and the dependent variable, but it is not statistically significant. 6. EMPIRICAL RESULTS AND DISCUSSION However, the more interesting fact is that the inclusion of the RL variable dramatically degrades the statistical significance of all others. Specification (4) is the only one in which no variable, despite the correct signs, reached statistical significance. This phenomenon can be explained by more complex forms of interaction between, for example, the political regime and the rule of law, or between the latter and the ability to convert a net currency inflows into economic development. Central bank independence is also key factor to understanding the propensity to reserves accumulation. As can be seen from specification (7), independent central banks are indeed less likely to hold significant reserves. The GMT variable is very close to statistical significance. Even more eloquent is the actual independence expressed by the interaction of the GMT and Rule of Law variables. In the case of specification (8), strong inverse relationship is observed which reduces the statistical significance of economic development and political regimes, but enhances significance of strong inverse relationship between reserves accumulation and resource wealth. This can be explained by the fact that, in the case of actual independence, central banks focused on price stability are not prone to reserve hypertrophy, regardless of the level of economic development, export structure and the nature of the political regime. This is precisely the indirect manifestation of the rule of law. The actual independence of the central bank is the starting point for the development of institutional trajectories, within which the role of alternative to exchange reserves channels of macroeconomic adjustment is increased. Specification (9) confirms this view. Independent central banks pursuing a policy of actual exchange rate floating do not show a tendency to reserves hoarding. The strong political consensus around such practices stems from the ability to create an institutional basis for alternative adjustment to shocks. It also indirectly confirms the tendency of autocracies to have more dependent central banks, which are far from the policy of exchange rate flexibility. 224 Does political regime matter for abnormal hoarding of international exchange reserves? Viktor Koziuk The regression analysis results confirm that political-economic factors and institutional parameters are important in explaining options for countries to choose the mechanisms of macroeconomic adjustment. The priority of international exchange reserves hoarding is determined not only by the optimal exchange rate flexibility with restrictions in the form of structural features of the economy. 6. EMPIRICAL RESULTS AND DISCUSSION This priority also reflects how stable the political and economic consensus is on alternative adjustment options. Accumulation priority is updated with institutional constraints that cannot be overcome within a given political structure. In other words, this is a case where the stability of the political regime relies on a variant of macroeconomic adjustment that does not require institutional changes that tend to compromise the stability of such a regime. Autocracies create the best set of political and economic preconditions for reserves hypertrophy. Whereas democracies allow for the emergence of such institutional trajectories, which create opportunities for alternative macroeconomic adjustment options that do not conflict with the nature of the democratic regime. Central bank independence, exchange rate flexibility and less pronounced dependence on resource wealth are key factors in why democracies are less likely to hoard reserves. The political business cycle and the sensitivity of reserves to it is more an explanation of why reserves change in volumes rather than why they hoard in large amount. Results extrapolation of emerging markets suggests that even if the current level of reserves is lower than optimal, further accumulation of them cannot substitute the needs to radically improve the quality of institutions as a precondition for the development of the financial system and better adaptability to exchange rate fluctuations. REFERENCES Aizenman, J. (2007). Large hoarding of international reserves and the emerging global economic architecture. NBER Working paper, 13277, 3-19. https://doi.org/10.3386/w13277 Aizenman, J., Cheung, Y., & Ito, H. (2015). International reserves before and after the global crisis: is there no end of hoarding?. Journal of International Money and Finance, 52, 102-126. doi:10.1016/j.jimonfin.2014.11.015 Aizenman J., Chinn M., & Ito H. (2013). “Impossible Trinity” Hypothesis in an Era of Global Imbalances: Measurement and Testing. Review of International Economics, 21(3), 447-458. https://doi.org/10.1111/roie.12047 Aizenman J., Jinjarak Y., & Park D. (2015). Financial development and output growth in developing Asia and Latin America: a comparative sectoral analysis. NBER Working Paper, 20917, 1-37. doi:10.3386/w20917 Aizenman J., Lee J. (2008). Financial versus monetary mercantilism: long-run view of large international hoarding. The World Economy, 31(5), 1-22. https://doi.org/10.1111/j.1467-9701.2008.01095.x Aizenman J., Marion N. (2004). international reserve holding with sov Journal, 17, 370-400. https://doi.org/10.2139/ssrn.1288138 Aizenman J., Marion N. (2004). international reserve holding with sovereign risk and costly tax collection. Economic Journal, 17, 370-400. https://doi.org/10.2139/ssrn.1288138 Aizenman J., & Riera-Crichton D. (2006). Real exchange rate and international reserves in the era of growing financial and trade integration. NBER Working Paper, 12363, 1-36. https://doi.org/10.3386/w12363 Alesina A., Devleeschauwer A., & Easterly W. (2002). Fractionalization. NBER Working Paper, 9411, 1-72. https://doi.org/10.2139/ssrn.319762 Aliyev R. (2013). Procyclicality in resource-rich countries. Ministry of Finance of the Czech Republic Working Paper, 1/2013, 1-24. https://www.mfcr.cz/en/about-ministry/research/2013/fiscal-policy-procyclicality-in- resource-36571 Aliyev R. (2012). Monetary policy in resource-rich developing economies. CERGE-EI Working Papers, 466, 1-49. https://www.cerge-ei.cz/pdf/wp/Wp466.pdf iyev R. (2014). Determinants of the choice of exchange rate regime in resource-rich countries. CERGE-EI W Papers, 527, 1-36. https://dx.doi.org/10.2139/ssrn.2856953 rezki R., Brückner M. (2012).Commodity windfalls, polarization, and net foreign assets: Panel data evidence o voracity effect, Journal of International Economics, Elsevier, 86(2), 318-326. doi: 10.1016/j.jinteco.2011.11.00 voracity effect, Journal of International Economics, Elsevier, 86(2), 318-326. doi: 10.1016/j.jinteco.2011.11.001 Auty R. (2001). The political economy of resource-driven growth. European Economic Review, 45, 839-46. https://econpapers.repec.org/RePEc:eee:eecrev:v:45:y:2001:i:4-6:p:839-846 Auty R. (2001). The political economy of resource-driven growth. European Economic Review, 45, 839-46. https://econpapers.repec.org/RePEc:eee:eecrev:v:45:y:2001:i:4-6:p:839-846 BIS (2019). Monetary policy framework in EMEs: Inflation targeting, the exchange rate and financial stability. BIS Annual Economic Report, Chapter II, 31-53. https://www.bis.org/publ/arpdf/ar2019e2.pdf BIS (2019). Monetary policy framework in EMEs: Inflation targeting, the exchange rate and financial stability. BIS Annual Economic Report Chapter II 31 53 https://www bis org/publ/arpdf/ar2019e2 pdf Choi W. G., Sharma S., Stromqvist M. (2007). CONCLUSIONS For emerging markets, the demand for reserves should not be considered as a compensator for the insufficient quality of institutions that impede the deepening of the financial system. CONCLUSIONS Traditional theoretical approaches explain the reserves hypertrophy in the categories of monetary mercantilism, self-insurance, competitive hoarding, etc. In political and economic analysis, the focus is on which of the political regimes better reflects the ability to accumulate external assets. The vulnerability of democracies to the political business cycle and the inability of independent central banks to make such accumulation are most often argued. The limitation of either the political business cycle approach or the preferences of the central bank is that they ignore several fundamental points: the status of the central bank is not an institutionally given, but is one of the institutional choices; macroeconomic adjustment channels alternative to the reserves accumulation can be developed with the relevant political and economic choice to build institutional preconditions for financial development and less painful exchange rate fluctuations; the structural characteristics of the economy and society can play an important role in choosing the optimal macro- financial stability design as they determine the profile of political and economic consensus on the central bank priorities; the political environment determines the degree of non-acceptance of the fiscal burden of excess reserves sterilization, etc. It is empirically confirmed that the propensity for reserve hypertrophy is determined by the level of economic development, the autocracy type of the political regime, and the choice of less dependent central banks in favor of lower exchange rate flexibility. At the same time, the actual independence of the monetary authorities and the monetary policy’s focus on price stability and exchange rate flexibility reduce the demand for excess reserves. The direct influence of the rule of law, ethnic heterogeneity and group fragmentation of society does not play a significant role if the autocratic regime is capable of pursuing a prudent macroeconomic policy, which is also a manifestation of the net currency inflows and, accordingly, the reserves accumulation. In democracies, there is a tendency for strong demand on reserves accumulation where a certain level of financial depth has already been reached that requires “protection” from institutional imperfections. At the same time, opting for an independent central bank and a flexible exchange rate regime requires that democracies focus on institutional changes that enhance financial 225 Journal of International Studies Vol.14, No.2, 2021 development. For emerging markets, the demand for reserves should not be considered as a compensator for the insufficient quality of institutions that impede the deepening of the financial system. development. REFERENCES Capital flows, financial integration, and international reserves holding: the recent experience of emerging market and advanced economies. IMF Working Paper, WP/07/151, 1-36. doi.org/10.5089/9781451867152.001 Cihak M., Demirguc-Kunt A., Feyen E., Levine R. (2012). Benchmarking financial development around the world. World Bank Policy Research Working Paper, 6175, 1-32. https://doi.org/10.1596/1813-9450-6175 Claessens S., & Laeven L. (2003). Financial Development, Property Rights, and Growth. Journal of Finance, 2436. https://doi.org/10.1046/j.1540-6261.2003.00610.x Crowe C., & Meade E. (2008). Central Banks Independence and Transparency: Evolution and Effectiveness. IMF Working Paper, WP/08/119, 1-30. https://doi.org/10.5089/9781451869798.001 Disyatat P. (2003). Rationalizing Asia’s Foreign Reserves Build-up. Bank of Thailand Economic Monitor, Dec., 1-8. https://econpapers.repec.org/RePEc:eee:asieco:v:14:y:2003:i:3:p:389-418 p p p p g y p Dooley M., Folkerts-Landau D., Garber P. (2003). An essay on the revived bretton woods system. NBER Working Paper, 9971, 1-48. doi: 10.3386/w9971 Dooley, M., Folkerts-Landau, D., & Garber, P. (2004). The revived Bretton Woods system: the effects of periphery intervention and reserve management on interest rates and exchange rates center countries. NBER Working Paper, 10332, 1-52. doi:10.3386/w10332 226 Does political regime matter for abnormal hoarding of international exchange reserves? Viktor Koziuk Dreher A., Vaubel R. (2009). Foreign exchange reserves and the political business cycle: a panel data analysis. Journal of International Money and Finance, 28(5), 755-775. doi:10.1016/j.jimonfin.2008.12.007 Durda, C. B., Mendoza, E., & Terrones, M. (2007). Precautionary demand for foreign assets in sudden stop economies: an assessment of the new merchantilism. IMF Working Paper, WP/07/146, 1-54. https://doi.org/10.5089/9781451867107.001 European Central Bank. (2006). The accumulation of foreign reserves. ECB Occasional Paper, 43, 3-73. https://www.ecb.europa.eu/pub/pdf/scpops/ecbocp43.pdf Gosselin, M.-A., & Parent, N. (2005). An Empirical Analysis of Foreign Exchange Reserves in Emerging Asia. Bank of Canada Working Paper, 2005-38, 1-40. https://www.bankofcanada.ca/wp-content/uploads/2012/01/fsr- 1205-gosselin.pdf g p Han I. (2006). The Political Economy of Reserve Accumulation in the Developing World. The Korean Journal of International Relations. 46(5), 231-248. doi:10.14731/kjis.2006.12.46.5.231 Hielscher K., Markwardt G. (2012). The Role of Political Institutions for the Effectiveness of Central Bank Independence. European Journal of Political Economy, 28(3), 286-301. doi:10.1016/j.ejpoleco.2011.08.004 Jager, K. (2016). The Role of Regime Type in the Political Economy of Foreign Reserves Accumulation. European Journal of Political Economy, 44, 79-96. doi:10.1016/j.ejpoleco.2016.05.001 Jager, K. (2016). The Role of Regime Type in the Political Economy of Foreign Journal of Political Economy, 44, 79-96. doi:10.1016/j.ejpoleco.2016.05.001 Koziuk, V. (2016). Financial development, international exchange reserves and political regimes in commodity economies. Economic theory, 3, 82-102. (in Ukrainian) https://doi.org/10.15407/etet2016.03.082 Koziuk, V. (2016). Independence of Central Banks in Commodity Economies. REFERENCES Gerald of National Bank of Ukraine, 235, 6-25. https://doi.org/10.26531/vnbu2016.235.006 Koziuk, V. (2018). Price stability and inflation targeting in commodity economies: macroeconomics vs political economy. Gerald of National Bank of Ukraine, 244, 4-25. https://doi.org/10.26531/vnbu2018.244.01 Kurronen, S. (2012). Financial Sector in Resource-Dependent Economies. BOFIT Discussion Papers,6, 1-35. https://doi.org/10.2139/ssrn.2027444 Leblang, D., & Pepinsky, Th. (2008). To Have or to Hoard? The Political Economy of International Reserves. Paper Presented at the 2008 Annual Meetings of the American Political Science Association, Boston, M.A. https://courses.cit.cornell.edu/tp253/docs/ToHaveAndToHoard.pdf McGrath, L. (2016). Insuring Against Past Perils: The Politics of Post-Currency Crisis Foreign Exchange Reserves Accumulation. Political Science Research and Methods, 5(3), 427-446. https://doi.org/10.1017/psrm.2016.9 Nurbayev, D. (2017). The Rule of Law, Central Bank Independence and Price Stability. Journal of Institutional Economics, 14(4), 659-687. https://doi.org/10.1017/S1744137417000261 Obstfeld M., Shambaugh J., Taylor A. (2008). Financial Stability, the Trilemma, and International Reserves. NBER Working Paper, 14217, 1-48. https://doi.org/10.3386/w14217 Rodrik D. (2006). The Social Cost of Foreign Exchange Reserves. NBER Working Paper, 11952, 1-24. https://doi.org/10.3386/w11952 Sanusi, K. A., Meyer, D. F., & Hassan, A. S. (2019). An Investigation of the Determinants of Foreign Exchange Reserves in Southern Africa Countries. Journal of International Studies, 12(2), 201-212. https://doi.org/10.14254/2071-8330.2019/12-2/12 B. (2020). Democracy and Reserves. Foreign Policy Analysis, 16(3), 417-437. https://doi.org/10.1093/fpa/orz02 Son, B. (2020). Democracy and Reserves. Foreign Policy Analysis, 16(3), 417-437. https://doi.org/10.1093/fpa/orz020 Turner Ph., & Mohanty, M. (2006). Foreign Exchange Reserve Accumulation in Emerging Markets: What Are the Domestic Implications? BIS Quarterly Review, Sept., 39-52. https://www.bis.org/publ/qtrpdf/r_qt0609f.htm olds, O., & Sondergaard, L. (2007). Reserve Accumulation: Objective or By-Product? ECB Occasional Paper Wijnholds, O., & Sondergaard, L. (2007). Reserve Accumulation: Objective or By-Product? ECB Occasional Paper, 73, 3-45 http://citeseerx ist psu edu/viewdoc/download?doi=10 1 1 165 6007&rep=rep1&type=pdf Wijnholds, O., & Sondergaard, L. (2007). Reserve Accumulation: Objective or By-Product? ECB Occasional 3-45. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.165.6007&rep=rep1&type=pdf p p p p yp p Wills, S., & Van der Ploeg, R. (2014). Why do so many oil exporters peg their currency? Foreign reserves as a de- facto sovereign wealth fund. Paper Presented at the joint RES-SPR Conference on “Macroeconomic Challenges Facing Low-Income Countries”. IMF, Washington, D.C, Jan. 30-31. https://www.imf.org/external/np/seminars/eng/2014/lic/pdf/Ploeg1.pdf 227
https://openalex.org/W4385784589	https://ejuow.uowasit.edu.iq/index.php/ejuow/article/download/346/240	Arabic	null	Investigating the impact of flow rate and moisture content for different concentration of liquid desiccant solution	Mağallaẗ Wāsiṭ li-l-ʿulūm al-handasiyyaẗ	2,023	cc-by	4,540	Abstract تمةةة حا ةة تةةلثير هةة ا المتغيةةرا علةةى معةةام اء نظةةام المجييةة مثةة معةةدك إزالةة، الرطوبةة وويةةاءا، الرطوبةة وويةةاءا الم توى،ال راحي ونمب ال راحا، المعقول ومعام نق الكتل . وضة ة النتةا ا المت صة عليهةا نةا ولمةا زا ترويةز الم لةوك ومعةدك تةددق الهةواء ،األولج زا معدك إزال الرطوب ونمب ال راحا المعقول ومعا م انتقةاك الكتلة . عنةد ال صةوك علةى القيمة ال الية لمعةدك تةددق الهةواء األولةج الةداخ ( 0.18 غو ا / ثاني ) ، تؤ ي الزيا ا دج ترويب مدخ م لوك ولوحيد الكالميوم من0.85 إلى0.95 إلى زيا ا معدك إزالة الرطوبة ، ونمةب ال ةراحا ( المعقول ، ومعام نق الكتل بمقداح1.1 _ 1.65 )غ( ، ا / ثاني0.18 - 0.25 ( ) ، و0.01075 ) - 0.0123) م / ث على،التوالج بينما عند ترويز ( مدخ معين لم لوك المجيف0.95 ) ، تؤ ي زيا ا معدك تددق الهواء األولج الداخ من0.1 غو ا / ثاني إلى0.18 غو ا / ثاني إلى زيا ا معدك إزالة الرطوبةةة ، ونمةةةب ال ةةةراحا الم مو ةةة ، ومعامةةة نقةةة الكتلةةةo ( 0.94 - 1.26 )غةةة( ، ا / ثانيةةة0.24 - 0.26 ( ، ) 0.0038 - 0.011) م / ث علةةةى التوالج. Investigating the impact of flow rate and moisture content for different concentration of liquid desiccant solution Hussain H. Mahdy*, Abdulsalam D. M. Hassen and Mohammed G. Al-Azawy Abstract The desiccant air conditioning system consists of two processes, namely cooling and dehumidification, in which the air temperature and humidity are controlled in order to provide comfortable thermal conditions. A typical system includes a dehumidifier, indirect evaporative cooler, and regenerator. The desiccant is selected depending on its ability to absorb water vapor present in the air. In this study, calcium chloride solution was used as a desiccant for the desiccant solution regeneration process, and a flat plate solar collector was employed. Different variables, such as the primary air flow rate, desiccant flow rate, and the concentration of the desiccant solution, were changed during the experiments. The impact of these variables on the performance parameters of the desiccant system such as moisture removal rate, moisture efficiency, enthalpy efficiency, sensible heat ratio, and the mass transfer coefficient was studied. The obtained results revealed that as the solution concentration and the flow rate of primary air increase, the moisture removal rate, sensible heat ratio, and mass transfer coefficient increase. A particular value of inlet primary air flow rate (0.18Kg/s), an increase in the inlet concentration of calcium chloride solution from 0.85 to 0.95 leads to a rise in moisture removal rate, sensible heat ratio, and mass transfer coefficient of (1.1 _1.65) g/s, (0.18- 0.25), and (0.01075-0.0123) m/s respectively. While at a certain inlet concentration of a desiccant solution (0.95), increasing the inlet primary air flow rate from 0.1Kg/s to 0.18Kg/s leads to an increase in the moisture removal rate, sensible heat ratio, and mass transfer coefficient o (0.94-1.26) g/s, (0.24-0.26), and(0.0038 -0.011) m/s respectively. Received 19-6-2022 Revised 6-November-2022 Accepted 13-November-2022 10.31185/ejuow.Vol11.Iss1.346 Keywords: Desiccant column, dehumidification, regeneration process, calcium chloride solution. Keywords: Desiccant column, dehumidification, regeneration process, calcium chloride solution. الخالصة: يتكون نظام تكييف الهواء الجاف من،عمليتين وهما التبريد وإزالة، الرطوبة حية يةتا الةت كا دةج حرة حةراحا الهةواء والرطوبة مةن رة تودير ظروف حراحي مري . يشتم النظام النموذرج على مزي للرطوب ومبر تبخيري غير مباشر و يضًا مُجد . يتا اختياح مةا ا التجييةف اعتمةا ًا .على قدحتها على امتصاص بخاح الماء المورةو دةج الهةواء دةج هة ا الدحا ة تةا ا ةتخدام م لةوك ولوحيةد الكالمةيوم ومجيةف لةي. عملية تجديةد م لةوك ،المجيف تا ا تخدام مجمع األلواح الشممي الممط . تا تغيير متغير،مختلف مث معدك تددق الهواء،األولةج ومعةدك تةددق المجيةف وترويةز م لةوك ،المجيةةف ثنةةاء التجةةاحم. 1. INTRODUCTION [11] presented a novel approach based on descriptive methods compared the performance of desiccant enhanced evaporative cooling systems; examination of characteristics that may have an impact on desiccant improved evaporative air cooling performance. The measurement findings show that the temperature and concentration of the intake desiccant, working air and desiccant to air mass ratios, as well as the channel length, had a significant impact on the DEH's performance. The aim of this study is to investigate the effect of changing the primary air flow rate and moisture content at different concentrations (95%, 90%, 85%) on desiccant system parameters such as: moisture removal rate, moisture efficiency, enthalpy efficiency, sensible heat ratio and mass transfer coefficient. To achieve this goal, the practical tests were carried out using the apparatus shown in Figure 1. 1. INTRODUCTION Indoor or settings are greatly influenced by humidity. High indoor humidity creates an unsettling and harmful atmosphere [1]. Desiccant dehumidification is an old technology that has been used for both industrial and agricultural uses, such as humidity management in textile mills and post-harvest vegetable drying in supermarkets Hussain H. Mahdi st al [2]. Desiccant technology has been discovered to be a viable replacement for traditional cooling systems. Under hot and humid climatic circumstances, it might provide more thermal comfort [3]. There has been a resurgence of interest in solar-powered air conditioning in recent years. The use of liquid-desiccant cooling systems is one possibility that has been researched [4]. As the push for energy saving continues to develop, Use of such liquid desiccants for dehumidification of HVAC (heating, ventilation, and air conditioning) process air is playing an increasingly important role [5] W.Z. Gao. et al. [6] experientially investigated an integrated liquid desiccant and M- cycle indirect evaporative cooler. The removing capacity of moisture in dehumidifier and sensible heat in M- cycle indirect evaporative cooler can be greatly improved while increasing the liquid desiccant flow rate or inlet concentration. Xianhua Ou et al. [7] experiential investigated the heat and mass transfer performance of a liquid desiccant cooling and dehumidification system form the result, with increased chilled water flow rate, desiccant solution flow rate, and concentration, thermal efficiency and moisture efficiency improve. Min Tu. et al. [8] simulated and analyses a novel liquid desiccant air conditioning system. The simulation of this system is formulated with two packed columns, one for regenerating the weak desiccant solution and the other for the dehumidification of the ambient air. first and second law of thermodynamics hour been used to analyse the effect of five key variables on the performance of the system. High efficiency could be reached if proper values of these variables are selected. M.M Bassuoni [9] analytically studied the modelling and performance of a crossflow air dehumidifier using liquid desiccant. the results show that hot humid air and desiccant concentration have the large impact on the performance of the dehumidifier. X. Cui et al. [10] analysed theoretically a liquid desiccant based indirect evaporative cooling system. The influence of several parameters on the performance of the system were investigated including the working –to-intake air flow ratio, the liquid desiccant film length, dimensionless channel length, and the inlet air conditions. Ali Sohani et al. Wasit Journal of Engineering Sciences 2023 11 (1) 2. DESCRIPTION OF EXPERIMENTAL APPARATUS The apparatus that was used in the experiments of this study is shown in Figure 1. It consists of the following parts:  Cross-flow plate heat and mass exchanger (HME).  Cross-flow plate heat and mass exchanger (HME). Air cirulation and treatment system includes: primary and secondary air streams centrifugal fans, Air heater, steam generator (humidifier), water jets in secondary air streams and air flow regulators.  Air cirulation and treatment system includes: primary and secondary air streams centrifugal fa heater, steam generator (humidifier), water jets in secondary air streams and air flow regulator  Liquid desiccant system includes: desiccant jets in primary air stream, liquid desiccant storage tank, and desiccant pump.  Liquid desiccant system includes: desiccant jets in primary air stream, liquid desiccant storage tank, and desiccant pump. p p  Solar collector for regeneration the weak solution desiccant.  Measuring devices includes: temperature and relative humidity measuring devices, arduino device, water and desiccant solution flow meters, digital thermometer, and velocity measuring device (thermo anemometer).  Measuring devices includes: temperature and relative humidity measuring devices, arduino device, water and desiccant solution flow meters, digital thermometer, and velocity measuring device (thermo anemometer). The heat and mass exchanger are essentially made up of thin aluminium sheets that are organized in a precise way to allow two perpendicular air streams (primary and secondary) to flow between parallel plates in alternate passage ways. The heat and mass exchanger construction are made up of fifty-two aluminium plates with a thickness of 0.5mm and a 5mm gap between them. Each air stream can flow through 26 channels to the heat exchanger's aluminium plate layout. Steel wire mesh of 0.5 mm square corrugations is placed between the plates to provide turbulence and enhance heat and mass transfer. Figure 2 shows the cross flow plate heat and mass exchanger. Wasit Journal of Engineering Sciences 2023 11 (1) pg.76 Hussain H. Mahdi st al (a) Snapshot of experimental test rig. (b) Schematic diagram of experimental test rig. Figure 1: The experimental rig. Cross-flow heat and mass exchanger humidifier Solar collector primary air centrifugal Fan Secondary air fan centrifugal Cross-flow heat and mass exchanger primary air centrifugal Fan Secondary air fan centrifugal Solar collector (b) Schematic diagram of experimental test rig. (a) Snapshot of experimental test rig. (a) Snapshot of experimental test rig. Figure 1: The experimental rig. Figure 2: Cross flow plate heat and mass exchanger. Wasit Journal of Engineering Sciences 2023 11 (1) 2. DESCRIPTION OF EXPERIMENTAL APPARATUS Primary air inlet Secondary air outlet Secondary air inlet Primary air outlet Cross- flow plate heat and mass exchanger Secondary air inlet Cross- flow plate heat and mass exchanger Secondary air outlet Figure 2: Cross flow plate heat and mass exchanger. pg.77 Hussain H. Mahdi st al 3. PERFORMANCE PARAMETER OF THE HEAT AND MASS EXCHANGER Two processes take place simultaneously through the heat and mass exchanger (HME) that used in this study. The first process is to remove moisture from the primary air stream passing through the vertical passages of the plates of this heat exchanger by spraying a calcium chloride solution on the surfaces of these passages while the second process that occurs is the cooling of the primary air stream by means of a secondary air stream that is evaporatively cooled and pass in the horizontal passages of the exchanger. A number of parameters have been used to determine the performance of this heat and mass exchanger, A number of parameters have been used to determine the performance of this heat and mass exchanger, including: The rate of Moisture Removal (𝑀𝑅) rate (Kg/s) from the primary air stream, which can be obtained from the following equation [12]: The rate of Moisture Removal (𝑀𝑅) rate (Kg/s) from the primary air stream, which can be obtained from the following equation [12]: 𝑀𝑅= 𝑚̇ 𝑎(𝑊𝑖−𝑊𝑜 ) (1) (1) Where 𝑚̇ 𝑎the mass flow is rate of air with unit of Kg/s and (𝑊𝑖−𝑊𝑜)is the difference between the moisture content at inlet and outlet with units of Kgv/Kgdryair, respectively. According to the concept of efficiency, the performance of process of removing moisture and heat from the primary air stream in the heat exchanger can be expressed in term of both the moisture efficiency (𝜂𝑚) and the enthalpy efficiency (𝜂ℎ), which can be expressed by the equations 2 and 3 respectively [13]. 𝜂𝑚= (𝑊𝑖−𝑊𝑜 ) (𝑊𝑖−𝑊𝑒 ) (2) 𝜂ℎ= (ℎ𝑖−ℎ𝑜 ) (ℎ𝑖−ℎ𝑒 ) (3) (2) (3) (3) The Sensible Heat Ratio (SHR) represents the required handling capability of the various sections of the air conditioner. It is defined as the ratio between the sensible heat load and the overall heat load of the air. As a result, calculating the SHR according to the inlet and outlet of air supply characteristics can be used to assess the effectiveness of cooling and dehumidifying procedures in lowering humidity and temperature. It can be calculated by using equation 4 [6]: The Sensible Heat Ratio (SHR) represents the required handling capability of the various sections of the air conditioner. It is defined as the ratio between the sensible heat load and the overall heat load of the air. 3. PERFORMANCE PARAMETER OF THE HEAT AND MASS EXCHANGER 𝑊𝑒= 0.626185 ( 𝑝𝑠𝑜𝑙 𝑝𝑎𝑡𝑚+𝑝𝑠𝑜𝑙) (7) (7) In term of equilibrium humiidity ratio the following equation can be used to calculate the mass transfer coefficient (𝐾𝑑) between the process air and calcium chloride solution: In term of equilibrium humiidity ratio the following equation can be used to calculate the mass transfer coefficient (𝐾𝑑) between the process air and calcium chloride solution: 𝐾𝑑= 𝑚𝑎(𝑤𝑖−𝑤𝑜 ) 𝐴𝜌(𝑤𝑖−𝑤𝑒 ) (8) (8) Where 𝐴 is the area (m2) and 𝜌 is the fluid density (kg/m3). Where 𝐴 is the area (m2) and 𝜌 is the fluid density (kg/m3) Wasit Journal of Engineering Sciences 2023 11 (1) 3. PERFORMANCE PARAMETER OF THE HEAT AND MASS EXCHANGER As a result, calculating the SHR according to the inlet and outlet of air supply characteristics can be used to assess the effectiveness of cooling and dehumidifying procedures in lowering humidity and temperature. It can be calculated by using equation 4 [6]: 𝑆𝐻𝑅= 𝐶𝑎(𝑡𝑖−𝑡𝑜 ) (ℎ𝑖−ℎ𝑜 ) (4) (4) Where 𝑡 is the temperature at ℃ and ℎ is the enthalpy at Kj/Kg. Where 𝑡 is the temperature at ℃ and ℎ is the enthalpy at Kj/Kg. Where 𝑡 is the temperature at ℃ and ℎ is the enthalpy at Kj/Kg. The Antoine equation is one of the most widely used formulas for determining the vapor pressure of desiccant solution. It can commonly written as follows [14]: 𝑝𝑠𝑜𝑙= 𝑎 𝑒𝑥𝑝 (𝐴− 𝐵 𝑇+𝐶) (5) (5) Where 𝑝𝑠𝑜𝑙 is the desiccant solution's pressure at Pa, while A, B, and C are constants depending on the liquid desiccant temperature and concentration; and 𝑎 can be expressed as: Where 𝑝𝑠𝑜𝑙 is the desiccant solution's pressure at Pa, while A, B, and C are constants depending on the liquid desiccant temperature and concentration; and 𝑎 can be expressed as: 𝑎= 𝑎0+𝑎1𝑥𝑚+𝑎2𝑥𝑚 2 +𝑎3𝑥𝑚 3 +𝑎4𝑥𝑚 4 (6) (6) In which 𝑥𝑚 represents the concentration of the solution; for calcium chloride solution used as a liquid desiccant in this study the coefficient in the Antoine equation take the following values which illustrated in table 1. In which 𝑥𝑚 represents the concentration of the solution; for calcium chloride solution used as a liquid desiccant in this study the coefficient in the Antoine equation take the following values which illustrated in table 1. Table 1: Antoine s equation constant for calcium chloride solution [15]. Table 1: Antoine s equation constant for calcium chloride solution [15]. 𝒂𝟎 𝒂𝟏 𝒂𝟐 𝒂𝟑 𝒂𝟒 A B C cacl2 0.9899 -8.67E-3 -3.68E-2 4.624E-3 -1.67E-4 18.3036 3816.44 46.13 l of Engineering Sciences 2023 11 (1) pg.78 Wasit Journal of Engineering Sciences 2023 11 (1) pg.78 Hussain H. Mahdi st al The equilibrium humidity ratio (𝑊𝑒) of the moisture air in contact with calcium chloride solution can be written for an ideal gas as follows (in terms of atmospheric pressure 𝑝𝑎𝑡𝑚 and desiccant solution 𝑝𝑠𝑜𝑙; see equation 7) [14]. 4. RESULTS AND DISCUSSION Figure 4: Effect of air moisture content on moisture efficiency. Figure 5: Effect of air moisture content on enthalpy efficiency. Figure 6: Effect of air moisture content on SHR. Figure 7: Effect of moisture content on the mass transfer coefficient. Figure 8 illustrates the variations in moisture removal rate as a function of the primary air flow rate. The moisture removal rate increases with increasing air flow rate at the three different concentrations. Referring to the equation 1, as the value of the air mass flow rate rises, the value of the moisture removal rate (M R) rises as well. The difference in the moisture content (𝑊𝑖−𝑊𝑜) in the same equation can be falls when the air flow velocity increases. Higher velocity of process air means a shorter contact time between process air and liquid desiccant. The influence of the airflow rate on enthalpy efficiency and moisture efficiency are shown respectively in Figures 𝐦𝐚𝐢,𝐩 ̇ = 𝟎. 𝟏𝟖𝐊𝐠/𝐬 Enthalpy efficiency 𝒎𝐚𝐢,𝐩 ̇ = 𝟎. 𝟏𝟖𝐊𝐠/𝐬 𝐦𝐚𝐢,𝐩 ̇ = 𝟎. 𝟏𝟖𝐊𝐠/𝐬 𝐦𝐚𝐢,𝐩 ̇ = 𝟎. 𝟏𝟖𝐊𝐠/𝐬 𝐦𝐚𝐢,𝐩 ̇ = 𝟎. 𝟏𝟖𝐊𝐠/𝐬 Hussain H. Mahdi st al Figure 3: Effect of air moisture content on moisture removal. 𝐦𝐚𝐢,𝐩 ̇ = 𝟎. 𝟏𝟖𝐊𝐠/𝐬 𝒎𝐚𝐢,𝐩 ̇ = 𝟎. 𝟏𝟖𝐊𝐠/𝐬 Figure 4: Effect of air moisture content on moisture efficiency. Figure 3: Effect of air moisture content on moisture removal. Figure 4: Effect of air moisture content on moisture efficiency. Figure 5: Effect of air moisture content on enthalpy efficiency. Enthalpy efficiency 𝐦𝐚𝐢,𝐩 ̇ = 𝟎. 𝟏𝟖𝐊𝐠/𝐬 𝐦𝐚𝐢,𝐩 ̇ = 𝟎. 𝟏𝟖𝐊𝐠/𝐬 Enthalpy efficiency Figure 6: Effect of air moisture content on SHR. Figure 5: Effect of air moisture content on enthalpy efficiency. Figure 7: Effect of moisture content on the mass transfer coefficient. 𝐦𝐚𝐢,𝐩 ̇ = 𝟎. 𝟏𝟖𝐊𝐠/𝐬 Figure 7: Effect of moisture content on the mass transfer coefficient. 𝐦𝐚𝐢,𝐩 ̇ = 𝟎. 𝟏𝟖𝐊𝐠/𝐬 Figure 7: Effect of moisture content on the mass transfer coefficient. Figure 8 illustrates the variations in moisture removal rate as a function of the primary air flow rate. The moisture removal rate increases with increasing air flow rate at the three different concentrations. Referring to the equation 1, as the value of the air mass flow rate rises, the value of the moisture removal rate (M R) rises as well. 4. RESULTS AND DISCUSSION Practical experiments were carried out under constant conditions at the entry section of the heat and mass exchanger (HME), including primary initial air temperature of 40 ℃ , and the flow rate of calcium chloride solution is 3 L/min. Figure 3 depicts the effect of the primary air inlet humidity ratio on moisture removal rate at three different values of calcium chloride solution concentrations (0.85, 0.9, and 0.95). The moisture removal rate increases as the air inlet humidity ratio increase. In fact, increasing the air inlet humidity ratio will lead to an increase in the driving force, which in turn increased the mass transfer potential inside the crossflow plate heat and exchanger. As a result, there will be an increase in the moisture removal rate at all concentrations. Figures 4 and 5 present the impact of a moisture content variation on the moisture efficiency and enthalpy efficiency, respectively. The change in the moisture content leads, as is illustrated in the figures, to slight changes in the values of these efficiencies. At a specific value of the moisture content, it is noted that the values of these efficiencies increase with the increase in the concentration of the solution. The effect of air humidity ratio on the SHR is illustrated in Figure 6. The desiccant's humidification capability can be modified by increasing the air humidity ratio due to the higher surface vapor pressure differential between air and desiccant. Therefore, the inlet primary air entering the crossflow plate heat and mass exchanger had a negative effect on the cooling capacity, because of a higher temperature of the supplied air. Thus, the SHR decreases quickly due to the increase of the latent heat fraction in the whole air handling process. Increased moisture content from 0.026 Kg/Kg to 0.035 Kg/Kg led to a decrease in the SHR from 0.26 to 0.20 for 0.95 desiccant solution concentration. Figure 7 shows the mass transfer coefficient under various air moisture contents. The heat exchanger mass transfer coefficient will decrease as the moisture content of the air rises. This behaviour is predicted since a rise in the air humidity ratio raises the water vapor pressure in the air, resulting in a reduction in the mass transfer coefficient. pg.79 Wasit Journal of Engineering Sciences 2023 11 (1) Hussain H. Mahdi st al Figure 3: Effect of air moisture content on moisture removal. 4. RESULTS AND DISCUSSION The difference in the moisture content (𝑊𝑖−𝑊𝑜) in the same equation can be falls when the air flow velocity increases. Higher velocity of process air means a shorter contact time between process air and liquid desiccant. The influence of the airflow rate on enthalpy efficiency and moisture efficiency are shown respectively in Figures 9 and 10. An increase in the rate of air mass flow results in decrease in the humidity ratio and enthalpy differences pg.80 Wasit Journal of Engineering Sciences 2023 11 (1) Hussain H. Mahdi st al between inlet air and outlet air. So, both efficiencies are decreased because of the moisture reduction and enthalpy reduction decrease with the increase in air flow rate. between inlet air and outlet air. So, both efficiencies are decreased because of the moisture reduction and enthalpy reduction decrease with the increase in air flow rate. Figure 11 reveals the effect of the airflow rate on sensible heat ratio (SHR) for three desiccant solution concentrations (0.85, 0.9 and 0.95). From this figure, it can be notice the sensible heat ratio (SHR) increases with increasing air flow rate for all concentrations. The latent heat change was significantly less than that of the sensible heat. The higher airflow rate leads to the decrease of enthalpy reduction and therefore increases the SHR. A raised airflow rate from 0.1 Kg/s to 0.18 Kg/s leading to increase SHR from 0.24 to 0.26 for concentration 0.95. The effect of air velocity on temperature can decrease or increase depending on the contact time but at the same time, the temperature difference increases due to turbulence and therefore increased SHR. Mass transfer coefficient under various air flow rates during dehumidification process and at three different values of the desiccant solution concentrations (0.85, 0.9 and 0.95) is shown in Figure 12. When the air flow rate changes from 0.1 Kg/s to 0.18 Kg/s the mass transfer coefficients increase almost linearly from 0.0037 m/s to 0.0115 m/s at a solution concentration of 0.95. The decrease in heat and mass transfer resistances with increasing air velocity can be considered as the explanation for this case. It is also seen from the figure that at a specific value of the air flow rate, the increase in the solution concentration values leads to an increase in the values of the mass transfer coefficient. Figure 8: Effect of air mass flow rate on moisture removal. Wasit Journal of Engineering Sciences 2023 11 (1) 4. RESULTS AND DISCUSSION Figure 9: Effect of air mass flow rate on moisture efficiency. Figure 10: Effect of air mass flow rate on enthalpy efficiency. Figure 11: Effect of air mass flow rate on SHR. 𝐖𝒂𝒊,𝒑= 𝟎. 𝟎𝟎𝟐𝟗 𝑲𝒈𝒗/𝑲𝒈𝒅𝒓𝒚 𝒂𝒊𝒓 𝐖𝒂𝒊,𝒑= 𝟎. 𝟎𝟎𝟐𝟗 𝑲𝒈𝒗/𝑲𝒈𝒅𝒓𝒚 𝒂𝒊𝒓 𝐖𝒂𝒊,𝒑= 𝟎. 𝟎𝟎𝟐𝟗 𝑲𝒈𝒗/𝑲𝒈𝒅𝒓𝒚 𝒂𝒊𝒓 𝐖𝒂𝒊,𝒑= 𝟎. 𝟎𝟎𝟐𝟗 𝑲𝒈𝒗/𝑲𝒈𝒅𝒓𝒚 𝒂𝒊𝒓 𝐖𝒂𝒊,𝒑= 𝟎. 𝟎𝟎𝟐𝟗 𝑲𝒈𝒗/𝑲𝒈𝒅𝒓𝒚 𝒂𝒊𝒓 𝐖𝒂𝒊,𝒑= 𝟎. 𝟎𝟎𝟐𝟗 𝑲𝒈𝒗/𝑲𝒈𝒅𝒓𝒚 𝒂𝒊𝒓 Figure 8: Effect of air mass flow rate on moisture removal. Figure 8: Effect of air mass flow rate on moisture removal. Figure 9: Effect of air mass flow rate on moisture efficiency. Figure 9: Effect of air mass flow rate on moisture efficiency. Figure 10: Effect of air mass flow rate on enthalpy efficiency. 𝐖𝒂𝒊,𝒑= 𝟎. 𝟎𝟎𝟐𝟗 𝑲𝒈𝒗/𝑲𝒈𝒅𝒓𝒚 𝒂𝒊𝒓 𝐖𝒂𝒊,𝒑= 𝟎. 𝟎𝟎𝟐𝟗 𝑲𝒈𝒗/𝑲𝒈𝒅𝒓𝒚 𝒂𝒊𝒓 Figure 11: Effect of air mass flow rate on SHR. Figure 10: Effect of air mass flow rate on enthalpy efficiency. Wasit Journal of Engineering Sciences 2023 11 (1) pg.81 Hussain H. Mahdi st al Figure 12: Effect of air mass flow rate on the mass transfer coefficient. 𝐖𝒂𝒊,𝒑= 𝟎. 𝟎𝟎𝟐𝟗 𝑲𝒈𝒗/𝑲𝒈𝒅𝒓𝒚 𝒂𝒊𝒓 𝐖𝒂𝒊,𝒑= 𝟎. 𝟎𝟎𝟐𝟗 𝑲𝒈𝒗/𝑲𝒈𝒅𝒓𝒚 𝒂𝒊𝒓 Figure 12: Effect of air mass flow rate on the mass transfer coefficient. 5. CONCLUSION  The rate of moisture removal (M.R) increases with an increase in both the moisture contact of the primary air stream and its mass flow rates  The rate of moisture removal (M.R) increases with an increase in both the moisture contact of the primary air stream and its mass flow rates  Both the moisture efficiency and the enthalpy efficiency change slightly with the change in the moisture content of the primary air stream; as the mass flow rate of this air stream increases, their efficiencies drop.  Both the moisture efficiency and the enthalpy efficiency change slightly with the change in the moisture content of the primary air stream; as the mass flow rate of this air stream increases, their efficiencies drop.  The sensible heat ratio (SHR) decreases with the increase in the moisture content of the primary air stream, while it rises in synchrony with the mass flow rate of this stream.  The same behaviour in the above point applies to the mass transfer coefficient; that is, it decreases with the increase in moisture content and increase with the increase in the rate of air flow.  At specific values of moisture content and air flow rate, the values of all the above mentions performance perimeters increase with the increase in the concentration of the used calcium chloride solution.  At specific values of moisture content and air flow rate, the values of all the above mentions performance perimeters increase with the increase in the concentration of the used calcium chloride solution. Wasit Journal of Engineering Sciences 2023 11 (1) References 1. Elhelw M. (2016). Performance Evaluation for Solar Liquid Desiccant Air Dehumidification System. Alexandria Eng J;55(2): 933–940. 1. Elhelw M. (2016). Performance Evaluation for Solar Liquid Desiccant Air Dehumidification System. Alexandria Eng J;55(2): 933–940. 2. Jradi M and Riffat S. (2014). Energy Performance of an Innovative Liquid Desiccant Dehumidification System with a Counter-Flow Heat and Mass Exchanger Using Potassium Formate. Renew Bioresour;2(1): 5. 3. Rafique MM, Rehman S, Alhems LM, et al. (2018). A Liquid Desiccant Enhanced Two Stage Evaporative Cooling System-Development and Performance Evaluation of a Test Rig. Energies;11(72): 1–17. 4. Mesquita LCS, Harrison SJ and Thomey D. (2006). Modeling of Heat and Mass Transfer in Parallel Plate Liquid-Desiccant Dehumidifiers. Sol Energy;80(11): 1475–1482. 5. Ware III JD. (2013). A Low-Cost-Solar Liquid Desiccant System for Residential Cooling. m Sc Thesis, Florida State Univ. 6. Gao WZ, Cheng YP, Jiang AG, et al. (2015). Experimental Investigation on Integrated Liquid Desiccant - Indirect Evaporative Air Cooling System Utilizing the Maisotesenko - Cycle. Appl Therm Eng;88: 288–296. 7. Ou X, Cai W, He X, et al. (2018). Experimental Investigations on Heat and Mass Transfer Performances of a Liquid Desiccant Cooling and Dehumidification System. Appl Energy;220(March): 164–175. Wasit Journal of Engineering Sciences 2023 11 (1) pg.82 pg.82 Wasit Journal of Engineering Sciences 2023 11 (1) Hussain H. Mahdi st al 8. Tu M, Ren CQ, Zhang LA, et al. (2009). Simulation and Analysis of a Novel Liquid Desiccant Air-Conditioning System. Appl Therm Eng;29(11–12): 2417–2425. 8. 9. Bassuoni MM. (2014). Analytical Modeling and Performance Study of a Cross Flow Air Dehumidifier Using Liquid Desiccant. Adv Mater Res;875–877: 1205–1213. 10. Amer O, Boukhanouf R and Ibrahim HG. (2015). A Review of Evaporative Cooling Technologies. Int J Environ Sci Dev;6(2): 111–117. 11. Lai L, Wang X, Kefayati G, et al. (2021). Evaporative Cooling Integrated with Solid Desiccant Systems: A Review. Energies;14(18): 1–23. 12. Rafiqu MM, Gandhidasan P and Bahaidarah HM. Emerging Energy Efficient Thermally Driven HVAC Technology: Liquid Desiccant Enhanced Evaporative Air Conditioning. In: Desiccant Heating, Ventilating, and Air-Conditioning Systems. Springer, Singapore. pp. 229--256. 13. Elsarrag E. (2007). Moisture Removal Rate for Air Dehumidification by Triethylene Glycol in a Structured Packed Column. Energy Convers Manag;48(1): 327–332. 14. Saman WY and Alizadeh S. (2002). An Experimental Study of a Cross-Flow Type Plate Heat Exchanger for Dehumidification/Cooling. Sol Energy;73(1): 59–71. pg.83
https://openalex.org/W3011744684	https://europepmc.org/articles/pmc7176856?pdf=render	English	null	Screening and functional prediction of differentially expressed circRNAs in proliferative human aortic smooth muscle cells	Journal of Cellular and Molecular Medicine	2,020	cc-by	7,855	K E Y W O R D S circular RNAs, coding potential, HASMCs, microRNA sponge, proliferation Screening and functional prediction of differentially expressed circRNAs in proliferative human aortic smooth muscle cells Wei Chen1,2 \| Jiajie Lin1 \| Bin Li1 \| Shanhu Cao1 \| Huanhuan Li1 \| Jianzhi Zhao1 \| Kun Liu1 \| Yiming Li1 \| Yang Li1 \| Shaoguang Sun1 1Department of Biochemistry and Molecular Biology, Key Laboratory of Medical Biotechnology of Hebei Province, Institute of Medicine and Health, Hebei Medical University, Shijiazhuang, China 2Stem Cell Translational Research Center, Tongji Hospital, Tongji University School of Medicine, Shanghai, China O R I G I N A L A R T I C L E O R I G I N A L A R T I C L E Received: 30 August 2019 \| Revised: 9 January 2020 \| Accepted: 31 January 2020 Received: 30 August 2019 \| Revised: 9 January 2020 \| Accepted: 31 January 2020 Received: 30 August 2019 \| Revised: 9 January 2020 \| Accepted: 31 January 2020 DOI: 10.1111/jcmm.15150 Abstract Abstract Vascular smooth muscle cell (VSMC) proliferation is the pathological base of vascular remodelling diseases. Circular RNAs (circRNAs) are important regulators involved in various biological processes. However, the function of circRNAs in VSMC prolifera- tion regulation remains largely unknown. This study was conducted to identify the key differentially expressed circRNAs (DEcircRNAs) and predict their functions in human aortic smooth muscle cell (HASMC) proliferation. To achieve this, DEcircRNAs between proliferative and quiescent HASMCs were detected using a microarray, fol- lowed by quantitative real-time RT-PCR validation. A DEcircRNA-miRNA-DEmRNA network was constructed, and functional annotation was performed using Gene Ontology (GO) and KEGG pathway analysis. The function of hsa_circ_0002579 in HASMC proliferation was analysed by Western blot. The functional annotation of the DEcircRNA-miRNA-DEmRNA network indicated that the four DEcircRNAs might play roles in the TGF-β receptor signalling pathway, Ras signalling pathway, AMPK signalling pathway and Wnt signalling pathway. Twenty-seven DEcircRNAs with cod- ing potential were screened. Hsa_circ_0002579 might be a pro-proliferation factor of HASMC. Overall, our study identified the key DEcircRNAs between proliferative and quiescent HASMCs, which might provide new important clues for exploring the functions of circRNAs in vascular remodelling diseases. Correspondence Shaoguang Sun, Key Laboratory of Medical Biotechnology of Hebei Province, Department of Biochemistry and Molecular Biology, Institute of Medicine and Health, Hebei Medical University, Shijiazhuang, Hebei, China. Funding information g National Natural Science Foundation of China, Grant/Award Number: 81670273 and 81200215; Natural Science Foundation of Hebei Province, Grant/Award Number: H2019206150; Program of International S&T Cooperation of Hebei Province, Grant/ Award Number: 14393001D 2.2 \| RNA labelling and array hybridization and 3′ tails, which make them resistant to degradation by RNA exonuclease and much more stable than linear RNAs.6 Recently, studies suggested that circRNAs might be involved in various bio- logical processes via different functional models, including binding microRNAs (miRNAs) as competitive endogenous RNAs (ceRNAs),7,8 interacting with RNA binding proteins (RBPs)9,10 or coding proteins/ peptides.11-14 Sample labelling and array hybridization were performed accord- ing to the Agilent One-Color Microarray-Based Gene Expression Analysis protocol (Agilent Technologies) with minor modifications. Briefly, total RNA was digested with RNase R (Epicentre) to remove linear RNAs and enrich circular RNAs. Then, each sample was ampli- fied and transcribed into fluorescent cRNAs along the entire length of the transcripts without 3′ bias utilizing a random priming method. The labelled cRNAs were purified by the RNeasy Mini Kit (QIAGEN). The concentration and specific activity of the labelled cRNAs (pmol Cy3/μg cRNA) were measured by a NanoDrop ND-1000 (Thermo Fisher Scientific). Each labelled cRNA (1 μg) was fragmented by add- ing 5 μL 10 × blocking agent and 1 μL 25 × fragmentation buffer be- fore being heated at 60°C for 30 minutes, followed by the addition of 25 μL 2 × GE Hybridization buffer. Hybridization solution (50 μL) was dispensed into the gasket slide and assembled on the circRNA expression microarray (8x15K, Arraystar) slide. The slides were in- cubated for 17 hours at 65°C in an Agilent Hybridization Oven. The hybridized arrays were washed, fixed and scanned using the Agilent DNA Microarray Scanner (G2505C). The analysis was conducted by Kangchen Bio-tech. However, little is known about the roles of circRNAs in vascular biology. The first study showed that circular antisense non-coding RNA in the INK4 locus (cANRIL) influences the polycomb group (PcG)-mediated repression of the human INK4a/ARF locus, which is associated with atherosclerosis risk.15 CircACTA2 competitively binds miR-548f-5p, which suppresses the expression of α-SMA in VSMCs, and is involved in hypertension.16 Our group demonstrated that circ-Sirt1 controls NF-κB activation via a sequence-specific in- teraction with p65 and enhancement of SIRT1 expression by spong- ing miR-132/212 in the inflammatory phenotypic transformation of VSMCs.17 Liu et al18 found that the expression of circRNA-ZNF609, a sponge for miR-615-5p, has a negative correlation with the veloc- ity of vascular endothelial cell migration and tube formation and is associated with vascular endothelial dysfunction. 2.2 \| RNA labelling and array hybridization Besides, other evi- dence has indicated that the aberrant expression of circRNAs usually accompanies by vascular dysfunction.19,20 Agilent Feature Extraction software (version 11.0.1.1) was used to analyse the acquired array images. Quantile normalization and subsequent data processing were performed using the GeneSpring GX v11.5.1 software package (Agilent Technologies). After quantile normalization of the raw data, circRNAs with at least one out of two samples having flags in Present or Marginal (‘All Targets Value’) were chosen for further data analysis. DEcircRNAs were identified through fold change filtering. Heat map and hierarchical clustering were performed using Agilent GeneSpring GX software (version 11.5.1). This study aimed to screen key differentially expressed circRNAs (DEcircRNAs) in proliferative and quiescent human aortic smooth muscle cells (HASMCs). A total of 134 DEcircRNAs were identi- fied, including 66 up-regulated and 68 down-regulated circRNAs. Seventy-seven DEcircRNAs were predicted to be bound with Argonaute (AGO), and five of them were selected to construct the DEcircRNA-miRNA-DEmRNA crosstalk network. Gene Ontology (GO) and pathway analyses suggested that hsa_circ_0002579, hsa_ circ_0004872, hsa_circ_0006371 and hsa_circ_0040705 mainly participated in the regulation of cell growth pathways. Out of 134 DEcircRNAs, 27 DEcircRNAs with coding potential were screened. Knockdown of hsa_circ_0002579 could inhibit proliferation of HASMCs. Our study provides new clues for exploring the regulation functions of circRNAs in HASMC proliferation and vascular remod- elling diseases. 2.1 \| Cell culture Human aortic smooth muscle cells (ScienCell) were cultured in Smooth Muscle Cell Medium (SMCM, ScienCell) with 2% foetal bo- vine serum (FBS) (ScienCell), 1% 100 × Smooth Muscle Cell Growth Supplement (ScienCell) and 1% 100 × Penicillin/Streptomycin Solution (ScienCell) in a humidified incubator at 37°C containing an atmosphere of 95% air and 5% CO2. Quiescent HASMCs were induced by FBS starvation for 24 hours, and proliferative HASMCs were obtained from quiescent HASMCs after treatment with 10 ng/ mL PDGF-BB (R&D Systems Inc) for 24 hours. 1 \| INTRODUCTION The regulation of VSMC proliferation has been considered a key event due to its major implications for the prevention of pathological vascular conditions.4 1 Cardiovascular diseases (CVD), including vascular remodelling dis- eases, are the leading cause of mortality among humans throughout the world. Abnormal proliferation of vascular smooth muscle cells (VSMCs) is a common feature of many vascular remodelling diseases, including atherosclerosis,1 hypertension2 and vascular aneurysms.3 Non-coding RNAs (ncRNAs) form the dominant product of eu- karyotic transcription, comprising over 73% of the human genome.5 Circular RNAs (circRNAs) are a special novel type of endogenous ncRNAs, forming covalently closed-loop structures without 5′ caps 4762 \| J Cell Mol Med. 2020;24:4762–4772. wileyonlinelibrary.com/journal/jcmm This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2020 The Authors. Journal of Cellular and Molecular Medicine published by Foundation for Cellular and Molecular Medicine and John Wiley & Sons Ltd Wei Chen, Jiajie Lin and Bin Li equally contributed to this study. p y Cellular and Molecular Medicine published by Foundation for Cellular and Molecular Medicine and John Wiley & Sons Ltd J Cell Mol Med. 2020;24:4762–4772. \| 4763 4763 CHEN et al. 2.3 \| Microarray analysis An Arraystar Human circRNA Microarray was designed for the global profiling of human circRNAs. Approximately 5816 circRNAs could be detected by the microarrays. Microarray analysis was performed by Kangchen Bio-tech. The microarray data discussed in this paper have been deposited in NCBI Gene Expression Omnibus and are accessi- ble with the GEO Series accession number GSE77278 (https://www. ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE77278). siRNA-control was as follows: 5′-UUCUCCGAACGUGUCACGUTT-3′ and 5′-ACGUGACACGUUCGGAGAATT-3′. SiRNA-hsa_circ_0002579 or siRNA-control was diluted by DEPC water to form a solution of 20 μM for 30 minutes at 4°C. According to the protocol of the HiperFect Transfection Kit (QIAGEN), HiperFect transfection reagent or siRNA was diluted in a serum-free medium for 5 minutes at room temperature and then HiperFect transfection reagent and siRNA were equal-ratio mixed for 30 minutes at room temperature to form a polyplex. The polyplex was finally adjusted to a proper volume by the same medium and then added into HASMCs. Following 6 hours of transfection in a humidified incubator at 37°C containing an atmos- phere of 95% air, 5% CO2, each well containing HASMCs was replaced with the same volume of fresh complete SMCM. Strand Kit (TaKaRa). The expression of 13 DEcircRNAs was tested by qRT-PCR using SYBR Green assays of SuperReal PreMix Plus (TaKaRa), and the primers used for validating up- or down-regulated circRNAs are displayed in Table S1. The qRT-PCR conditions were as follows: a denaturation step of 3 minutes at 95°C, followed by 40 cycles of 30 seconds at 93°C, 30 seconds at 55°C (adjusted with the Tm of different circRNAs), 20 seconds at 72°C and a final step of 5 minutes at 72°C. All samples in this study were normalized to the internal control β-actin. The comparative CT (2−ΔΔCT) method was used to calculate the fold change of circRNA expression levels, and Student's t test was used to test its statistical significance. All primer pairs for the detection of circRNAs were designed against the circRNA-specific back-splice sites (Table S1). 2.8 \| Western blot Human aortic smooth muscle cells after siRNA-treated for 48 hours were harvested, and total protein was extracted using RIPA buffer (Solarbio) and 1 mM PMSF (Solarbio). Protein concentrations were determined using the Bradford method (Solarbio). Equal amounts of protein samples were separated by SDS-PAGE and transferred onto a PVDF membrane (Merck) after electroblotting. The membrane was blocked with 5% non-fat milk in PBS for 2 hours and then was incubated with the primary antibodies overnight at 4°C: HMGA2 (1:1000, Proteintech), PCNA (1:1,000, Wanleibio), SM22α (1:1,000, CUSABIO) and GAPDH (1:1,000, Wanleibio). After three 5-minutes washes with TBS-tween, the membrane was then incubated with goat anti-rabbit IgG secondary antibody or rat antimouse second- ary antibody (dilution at a 1:2000, Wanleibio) at 37°C for 2 hours. The membranes were visualized by ChemiDoc™ MP Imaging System (BIO-RAD). Five AGO-bound DEcircRNAs (hsa_circ_0000069, hsa_ circ_0002579, hsa_circ_0004872, hsa_circ_0006371 and hsa_circ_0040705) were selected to construct the interaction network. DEmRNAs were obtained from NCBI Gene Expression Omnibus GSE77279 (https://www.ncbi.nlm.nih.gov/geo/query /acc.cgi?acc=GSE77279). Shared miRNAs in DEcircRNAs and DEmRNAs were predicted using Arraystar's homemade miRNA tar- get prediction software based on TargetScan21 and MiRanda.22 The DEcircRNA-miRNA-DEmRNA interaction network was illustrated using Cytoscape_v3.7.1. 2.9 \| Statistical analysis All data were analysed using SPSS 21.0 software (IBM). The mean ± SD and independent-samples t test were used in the statisti- cal analysis. A value of P < .05 was considered significant. 3.1 \| Analysis of circRNA expression profiling The microarray was applied to detect circRNAs in proliferative and quiescent HASMCs. For the distribution of the datasets, there were no distinct differences among the samples (Figure 1A). The circR- NAs that above the top green line and below the bottom green line in the scatter plot (Figure 1B), the red point in the volcano plot (Figure 1C), and the hierarchical clustering (Figure 1D) represent the DEcircRNAs with statistical significance. Among the detectable circRNAs, there were 134 DEcircRNAs in total, including 68 with 2.4 \| Quantitative real-time PCR validation Total RNA was isolated from proliferative and quiescent HASMCs using TRIzol reagent (Thermo Fisher Scientific), and its quantity and quality were examined by a NanoDrop ND-1000 (Thermo Fisher Scientific) and 1% agarose gel electrophoresis. Then, total RNA was reverse-transcribed using reverse transcriptase with random prim- ers according to the manufacturer's instructions of the M-MLV First CHEN et al. 4764 CHEN et al. 2.6 \| Gene ontology and pathway analyses Gene ontology and pathway analyses were performed using the standard enrichment computation method, which was used to de- termine the potential roles of DEcircRNAs in cellular activities. Based on the online software GeneCodis (http://genecodis.cnb. csic.es/analysis), GO classification and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment were conducted for the target genes that might be regulated by these five DEcircRNAs as miRNA sponges in the proliferative and quiescent HASMCs. Gene ontology and pathway analyses are effective methods to uncover the underlying biological function in response to abnormally ex- pressed genes and proteins.23 This analysis was used to determine the biological pathways that had significant enrichment of differen- tially expressed targeted genes. 2.7 \| SiRNA transfection SiRNA-hsa_circ_0002579 and siRNA-control were designed and purchased from GenePharma, China. The sequence of siRNA-hsa_ circ_0002579 was as follows: 5′-UCCUGUUCAUGGGCUCCAUTT-3′ and 5′-AUGGAGCCCAUGAACAGGATT-3′ and that of 4765 \| 4765 wn-regulated expression These 134 DEcircRNAs e exonic, 11% were in- What is more, we also of the circRNAs of 134 with conservation in hu- proliferative HASMCs compared to quiescent HASMCs, except hsa_ circ_0023406; and in Figure 2B, hsa_circ_0009792, hsa_circ_0007422, hsa_circ_0004872, hsa_circ_0040705, hsa_circ_0002720 and hsa_ circ_0001304 were up-regulated in proliferative HASMCs compared to quiescent HASMCs. CHEN et al. \| FI G U R E 1 Expression profile of circRNAs in proliferative HASMCs compared to quiescent HASMCs. A, Box plots of circRNAs show the distribution of intensities from all samples. B, Scatter plot showing circRNA expression variation between proliferative HASMCs and quiescent HASMCs. The values of the X and Y axes in the scatter plots were normalized signal values of the samples (log2 scaled). The green lines are fold change lines, with the default fold change value being 2.0. CircRNAs above the top green line and below the bottom green line had more than 2.0-fold changes between the proliferative HASMCs and quiescent HASMCs. C, Volcano plot showing circRNA differential expression in proliferative HASMCs and quiescent HASMCs. The vertical lines correspond to 2.0-fold up and down, respectively, and the horizontal line represents a P-value of .05. The red point in the plot represents the differentially expressed RNAs with statistical significance. D, Hierarchical clustering was performed to show differential circRNA expression profiling in proliferative HASMCs and quiescent HASMCs. Cluster analysis arranged samples into groups based on their expression levels, which allowed us to hypothesize the relationships among samples. ‘Red’ denotes high relative expression, and ‘green’ denotes low relative expression. E, The exonic circRNAs account for the majority of the differential circRNAs after being classified proliferative HASMCs compared to quiescent HASMCs, except hsa_ circ_0023406; and in Figure 2B, hsa_circ_0009792, hsa_circ_0007422, hsa_circ_0004872, hsa_circ_0040705, hsa_circ_0002720 and hsa_ circ_0001304 were up-regulated in proliferative HASMCs compared to quiescent HASMCs. up-regulated expression and 66 with down-regulated expression with a >2-fold change and a P-value < .05. These 134 DEcircRNAs were classified into three types: 84% were exonic, 11% were in- tronic and 5% were intragenic (Figure 1E). What is more, we also used circbank24 to analyse the conservation of the circRNAs of 134 DEcircRNAs and obtained 41 DEcircRNAs with conservation in hu- mans and mice (Table S2). 3.2 \| Validation by qRT-PCR As many as 58 062 circRNAs from the circBase database were found to have AGO-bound regions,25 indicating that circRNAs might func- tion as miRNA sponges. Among the 134 DEcircRNAs, 77 DEcircRNAs were predicted to bind with AGO (Figure 3A, Table S3), and five AOG-bound DEcircRNAs (hsa_circ_0000069, hsa_circ_0002579, hsa_circ_0004872, hsa_circ_0006371 and hsa_circ_0040705) were To validate the data accuracy of DEcircRNAs in the microarray, we ran- domly selected six up-regulated and seven down-regulated circRNAs and detected them by qRT-PCR. As shown in Figure 2A, hsa_circ_0083756, hsa_circ_0007888, hsa_circ_0006677, hsa_circ_0007146, hsa_ circ_0009065 and hsa_circ_0057072 were down-regulated in 4766 \| FI G U R E 2 Validation of the DEcircRNAs. The expression levels of 13 DEcircRNAs were validated in the samples of HASMCs by qRT-PCR and normalized to the internal reference gene β-actin. A, Apart from hsa_circ_0023406, six verified down-regulated circRNAs that were consistent with the microarray results. B, Six verified up-regulated circRNAs that were consistent with the microarray results. All experiments were replicated three times. The presented values are the mean ± SEM. P < .05 CHEN et al. \| AGO-bound DEcircRNAs in proliferative and quiescent HASMCs (Figure 3C). The GO analysis results showed that DEmRNAs mainly participated in some significant biological processes and cellular components, such as the TGF-β receptor signalling pathway, nega- tive regulation of cell growth and zinc finger. Among those signifi- cantly enriched GO terms, the TGF-β receptor signalling pathway (biological process: 0007179) and negative regulation of cell growth (biological process: 0030308) were noteworthy. TGFBR3 shared the same MREs with hsa_circ_0002579, and ACVR1B shared the same MREs with hsa_circ_0002579 and hsa_circ_0006371, which were within those two terms. According to the KEGG pathway analysis, the Ras signal- ling pathway (KEGG: cfa04014), TNF signalling pathway (KEGG: cfa04668), AMPK signalling pathway (KEGG: cfa04152), Wnt sig- nalling pathway (KEGG: cfa04310) and MAPK signalling pathway (KEGG: cfa04010) were significantly enriched pathways of inter- est (Figure 3C). RASAL2, GAB2, SHC3 and ABL2 were enriched in the Ras signalling pathway. RASAL2 and GAB2 were targets of hsa_circ_0004872 and hsa_circ_0002579, respectively, and these two circRNAs regulated the gene ABL2. SHC3 was the target of hsa_circ_0006371. 3.5 Through the circRNADb,26 we screened 27 DEcircRNAs with inter- nal ribosomal entry sites (IRES) and open reading frames (ORF), sug- gesting that they might code proteins (Table 1). 3.6 \| Comparison of expression levels between DEcircRNAs and corresponding linear mRNAs The main two categories of DEcircRNAs were exonic and intronic. No matter in which portion, there were less than half of the cor- responding mRNAs whose expression was changing along with the differential expression of the circRNAs. In the part of exonic circRNAs, there were 62 circRNAs irrelated with their host genes, 26 co-expression (co-up or co-down) and 23 inverse-expression (circRNAs downward while corresponding mRNAs upward or cir- cRNAs upward while corresponding mRNAs downward) (Figure 4). Only a small moiety of the DEcircRNAs derived from introns and half of them have nothing to do with the corresponding mRNAs (Figure 4). 3.2 \| Validation by qRT-PCR Two genes were enriched in the TNF signalling pathway, including the hsa_circ_0004872-targeted gene CASP10 and the hsa_circ_0002579-targeted gene TNFRSF1B; the AMPK signalling pathway including the hsa_circ_0006371-targeted gene PFKFB4 and the hsa_circ_0002579-targeted gene EEF2K; the Wnt signalling pathway, including the hsa_circ_0006371-targeted gene DVL3 and the hsa_circ_0040705-targeted gene SFRP1; and the MAPK signalling pathway, including the hsa_circ_0040705-tar- geted gene ELK4 and the hsa_circ_0006371-targeted gene MAP3K13. FI G U R E 2 Validation of the DEcircRNAs. The expression levels of 13 DEcircRNAs were validated in the samples of HASMCs by qRT-PCR and normalized to the internal reference gene β-actin. A, Apart from hsa_circ_0023406, six verified down-regulated circRNAs that were consistent with the microarray results. B, Six verified up-regulated circRNAs that were consistent with the microarray results. All experiments were replicated three times. The presented values are the mean ± SEM. P < .05 selected to construct the DEcircRNA-miRNA-DEmRNA crosstalk network (Figure 3B). From the interaction network information, we noted that hsa_circ_0002579 shared the miRNA response elements (MREs) of hsa-miR-let-7 with TGFBR3, ACVR1B and GAB2. Hsa_ circ_0004872 had the MREs of hsa-miR-424-5p, which were also observed in RASAL2 and ABL2. The DEcircRNA hsa_circ_0006371 shared the MREs of hsa-miR-424-3p, hsa-miR-24-3p and hsa-miR- 1298-5p with MAP3K13, shared the MREs of hsa-miR-24-3p with ACVR1B, shared the MREs of hsa-miR-424-5p with ABL2 and shared the MREs of hsa-miR-24-3p and hsa-miR-424-5p with DVL3. Hsa_circ_0040705 shared the MREs of hsa-miR-204-5p with ELK4. Overall, hsa_circ_0000069, hsa_circ_0002579, hsa_circ_0004872, hsa_circ_0006371 and hsa_circ_0040705 were predicted to com- pete with 2, 35, 16, 28 and 10 DEmRNAs for sponging miRNAs, respectively. 3.4 \| Functional annotation of DEmRNAs targeted by DEcircRNAs Gene ontology and KEGG pathway analyses were employed to anno- tate the functions of DEmRNAs that shared the same MREs with five 4767 CHEN et al. \| 4767 CHEN et al. FI G U R E 3 DEcircRNA-miRNA-DEmRNA crosstalk network. A, Venn diagram showing that 77 DEcircRNAs were AGO-bound circRNAs. B, Based on the data of DEcircRNAs and relevant mRNAs in the microarray results mentioned above, 5 AGO-bound DEcircRNAs (hsa_ circ_0000069, hsa_circ_0002579, hsa_circ_0004872, hsa_circ_0006371 and hsa_circ_0040705) were selected to construct the DEcircRNA miRNA-DEmRNA crosstalk network. Red circles: DEcircRNAs. Cyan squares: DEmRNAs. Yellow rhombus: DEmiRNAs. C, Gene Ontology (GO) d th l i f DE RNA t t d b fi AGO b d DE i RNA Th th GO t f th DE RNA \| 4767 EN et al. FI G U R E 3 DEcircRNA-miRNA-DEmRNA crosstalk network. A, Venn diagram showing that 77 DEcircRNAs were AGO-bound circRNAs. B, Based on the data of DEcircRNAs and relevant mRNAs in the microarray results mentioned above, 5 AGO-bound DEcircRNAs (hsa_ circ_0000069, hsa_circ_0002579, hsa_circ_0004872, hsa_circ_0006371 and hsa_circ_0040705) were selected to construct the DEcircRNA- miRNA-DEmRNA crosstalk network. Red circles: DEcircRNAs. Cyan squares: DEmRNAs. Yellow rhombus: DEmiRNAs. C, Gene Ontology (GO) and pathway analysis of DEmRNAs targeted by five AGO-bound DEcircRNAs. The three orange ones are GO terms for the DEmRNAs. The five blue ones are pathway terms for the DEmRNAs 4768 CHEN et al. CHEN et al. 3.4 \| Functional annotation of DEmRNAs targeted by DEcircRNAs E 1 DEcircRNAs with coding potential NA Chrom Strand Gene symbol Genomic length Spliced length Start position End position Protein length circ_0088688 chr9 + STXBP1 5195 469 436 2r + 22 174 aa circ_0108663 chr18 − EPB41L3 27 345 1744 224 1r + 30 516 aa circ_0138079 chr9 + SPTAN1 5882 1102 1098 2r + 51 385 aa irc_0102717 chr14 + NRXN3 5765 398 27 1r + 33 134 aa irc_0001384 chr3 − BDH1 23 847 452 44 1r + 29 145 aa irc_0000504 chr13 − TUBGCP3 11 045 751 73 1r + 29 235 aa irc_0008462 chr21 − SYNJ1 7820 665 96 1r + 66 211 aa irc_0008512 chr17 − ULK2 9209 944 889 2r + 65 354 aa irc_0008449 chr1 − GDAP2 12 231 460 95 1r + 54 139 aa irc_0139157 chr9 + NTRK2 88 275 728 117 1r + 120 243 aa irc_0092831 chr10 − UPF2 22 385 959 73 1r + 67 317 aa irc_0079509 chr7 − SNX13 60 957 1624 1563 2r + 51 578 aa irc_0108090 chr18 − OSBPL1A 3267 416 19 1r + 46 147 aa irc_0065149 chr3 − SETD2 5526 854 794 2r + 330 414 aa irc_0004619 chr1 − FAF1 89 334 377 341 2r + 21 144 aa irc_0003598 chr15 + PDIA3 1783 238 223 2r + 61 104 aa irc_0125090 chr4 + ANK2 34 768 1859 12 1r + 15 620 aa irc_0006188 chr1 − GDAP2 6743 383 95 1r + 23 103 aa irc_0049629 chr19 − FARSA 3554 301 150 2r + 21 157 aa irc_0002092 chr2 − NCKAP1 927 448 14 1r + 27 153 aa irc_0057262 chr2 − NCKAP1 18 580 706 14 1r + 27 239 aa irc_0065140 chr3 − SETD2 24 681 1241 312 1r + 330 419 aa irc_0121581 chr3 − GSK3B 23 374 325 152 2r + 83 193 aa irc_0119391 chr2 + LRRFIP1 20 974 904 127 1r + 110 295 aa irc_0108320 chr18 + DTNA 15 767 397 2 1r + 9 134 aa irc_0069853 chr4 + POLR2B 9056 716 3 1r + 69 260 aa irc_0110127 chr1 + AGL 36 809 2870 4 1r + 46 970 aa viation: DEcircRNAs, differentially expressed circRNAs. 4769 CHEN et al. 3.7 \| Knockdown of hsa_circ_0002579 inhibited the proliferation of HASMCs 3.7 \| Knockdown of hsa_circ_0002579 inhibited the proliferation of HASMCs SiRNA for hsa_circ_0002579 was designed, which could decrease 65% hsa_circ_0002579 levels (data not shown). Compared to con- trol, knockdown of hsa_circ_0002579 reduced the expression level of HMGA2 and PCNA and increased the expression level of SM22a (Figure 5B). From the data obtained by microarray, we found that hsa_ circ_0002579 was up-regulated in the proliferative HASMCs. Gene ontology and pathway analyses showed that 35 DEmRNAs co-expressed with hsa_circ_0002579 (Figure 3B) were enriched in the TGF-β receptor signalling pathway, Ras signalling pathway and AMPK signalling pathway (Figure 3C). Interestingly, high mo- bility group AT-hook 2 (HMGA2) is one of the DEmRNAs that co- expressed with hsa_circ_0002579, 30 shared miRNAs were found between hsa_circ_0002579 and HMGA2 3′-UTR (Figure 5A), in- dicating that hsa_circ_0002579 might target HMGA2 as ceRNA. Taken together, these data showed that knockdown of hsa_ circ_0002579 could inhibit the proliferation of HASMCs. There was a hint that hsa_circ_0002579 could be a promising ceRNA, and HMGA2 was one of its downstream targets. FI G U R E 5 Knockdown of hsa_ circ_0002579 inhibited the proliferation of HASMCs. A, Thirty shared miRNAs between hsa_circ_0002579 and HMGA2 3′-UTR. B, HASMCs were treated with siRNA-hsa_circ_0002579/siRNA-Control for 48 h. Cell lysates were extracted and subjected to Western blotting using specific antibodies against HMGA2, PCNA and SM22α. GAPDH was used as equal loading control. The illustrated results were representative of three independent experiments 4 \| DISCUSSION CircRNAs, mainly produced by precursor mRNA backsplicing of exons, have recently been predicted as a novel class of gene ex- pression potential regulatory factors. CircRNAs often show cell- type-specific, tissue-specific and spatiotemporal-specific patterns, including the cardiovascular system,27 and multitudinous stud- ies suggested that circRNAs are associated with CVD.15,16,18-20 However, direct evidence about circRNAs in VSMCs in terms of proliferation regulation is still absent. Since the ceRNA hypothesis was proposed,28 emerging evidence has indicated that circRNAs could act as miRNA sponges to regulate the stability or translation of mRNAs. In this study, we aimed to explore the roles of circRNAs acting as ceRNAs to mediate the proliferation of VSMCs. Here, we performed microarray analysis to identify 134 DEcircRNAs in prolif- erative HASMCs compared to quiescent HASMCs, and we predicted 77 AGO-bound DEcircRNAs that were promising ceRNA molecules. And hsa_circ_0000069, hsa_circ_0002579, hsa_circ_0004872, hsa_ circ_0006371 and hsa_circ_0040705 were selected to construct the DEcircRNA-miRNA-DEmRNA interaction network to predict their functions. FI G U R E 4 Comparison of expression levels of exonic and intronic DEcircRNAs to corresponding linear RNAs in proliferative HASMCs and quiescent HASMCs. Arrows represent the change of expression level. Upward arrows stand for the up-expression; downward arrows stand for down-expression, and the transverse lines stand for the invariable expression. A, Five categories, including co-expression (up-regulation or down-regulation), inverse-expression (circRNAs downward and mRNA upward or circRNAs upward and mRNA downward) and irrelevant expression, were based on the exonic DEcircRNAs and the corresponding mRNAs. B, There are only three categories in the intronic circRNAs. The relevant and irrelevant expressions are fifty-fifty FI G U R E 4 Comparison of expression levels of exonic and intronic DEcircRNAs to corresponding linear RNAs in proliferative FI G U R E 4 Comparison of expression levels of exonic and intronic DEcircRNAs to corresponding linear RNAs in proliferative HASMCs and quiescent HASMCs. Arrows represent the change of expression level. Upward arrows stand for the up-expression; downward arrows stand for down-expression, and the transverse lines stand for the invariable expression. A, Five categories, including co-expression (up-regulation or down-regulation), inverse-expression (circRNAs downward and mRNA upward or circRNAs upward and mRNA downward) and irrelevant expression, were based on the exonic DEcircRNAs and the corresponding mRNAs. B, There are only three categories in the intronic circRNAs. However, the target miRNAs may not necessarily be affected by low levels of circRNAs.45 Besides, in experimental verification, it is difficult to simulate the expression of miRNAs and ceRNAs in the physiological state through exogenous miRNAs and ceRNAs.46 Furthermore, previous findings have shown that some circRNAs were unable to function as miRNA sponges,25 which suggested that circRNAs may exert their effects through other mechanisms, such as forming complexes with proteins9,10 directly or translating into pro- teins/peptides.11-14 Interestingly, for the translation of circRNAs, out of 134 DEcircRNAs, we screened 27 circRNAs with IRES and ORF, indicating the coding potential of these circRNAs (Table 1). In one previous study, the results showed that circRNAs were typically generated at the cost of canonical mRNA isoforms and even more abundant than the linear counterparts.47 Identifying hundreds of head-to-tail junction reads in published data sets of chroma- tin-bound (nascent) RNA from fly heads,48 circRNAs were consid- ered as being generated co-transcriptionally and could compete with linear splicing mutually.49 However, some studies beg to differ. The expression of some circRNAs is found to be independent of related linear isoforms.50,51 The specifically overexpressed circRNAs in the multiple system atrophy brain have been determined, and the ex- pression levels of linear transcripts are not significantly altered and thus do not follow the pattern of their circular counterparts.52 With the alteration of circRNAs, only a small percentage of the corre- sponding mRNAs also make a change, which is consistent with that only a few circRNAs show co-regulation with their host genes, and circRNAs exhibited changes independent of the cognate mRNA.53 In our study, we also failed to find the clear-cut relationship between circRNAs and the cognate linear transcripts. And we expected that the production of most circRNAs is regulated, which might be indis- pensable in the mediation of gene expression.54 Moreover, circRNAs generated from one locus could be several and different. Thus, the relationship of homology circRNAs, as well as respective functions, required further exploration. DVL3 shared the MREs of hsa-miR-24-3p and hsa-miR-424-5p with hsa_circ_0006371 in HASMCs. One signalling modulator, Rspo3, of the Wnt signalling pathway was identified as a potential regulator of coronary stem proliferation.37 Additionally, the Wnt signalling pathway was shown to be involved in the proliferation and migration of pulmonary arteriolar smooth muscle cells.38 Consequently, we conjecture that hsa_circ_0006371 would be a positive conductor of Wnt signals isolating hsa-miR-24-3p and hsa-miR-424-5p to regulate HASMC proliferation. hsa_circ_0040705. This implied that hsa_circ_0006371 and hsa_ circ_0040705 might synergize the proliferation of HASMCs by the MAPK/JNK signalling pathway. the expression of DEcircRNA-sponged miRNAs was suppressed when these circRNAs were up-regulated in proliferative HASMCs. Here, we found that hsa_circ_0002579 shared the same MREs as hsa-let-7a-5p with GAB2, TGFBR3 and ACVR1B. GAB2 was anno- tated in the Ras signalling pathway, an important pathway involved in cell proliferation. Thus, hsa_circ_0002579 might be a promising mediator of HASMC proliferation via involvement in the Ras signal- ling pathway. The well-known TGF-β signalling pathway is involved in many cellular processes of VSMCs, including cell growth. TGFBR3 is the most abundant TGF-β type II receptor, and the impact of TGFBR3 on TGF-β signalling depends on the circumstances.34 Our results suggested that hsa_circ_0002579 might be involved in the TGF-β signalling pathway to regulate HASMC proliferation through sponging hsa-let-7a-5p. ACVR1B, one type I receptor in the TGF-β signalling pathway, was expected to share MREs of hsa-let-7a-5p with hsa_circ_0002579 and share MREs of hsa-miR-24-3p with hsa_ circ_0006371. MiR-24-3p was shown to participate in the TGF-β sig- nalling pathway and Wnt signalling pathway to regulate the growth of VSMCs.35,36 Hence, hsa_circ_0002579 and hsa_circ_0006371 may synergistically regulate HASMC proliferation through the TGF-β sig- nalling pathway. In terms of the Wnt signalling pathway, we showed that DVL3 shares strong similarity with DSH, functioning as a princi- pal component of the Wnt signalling pathway and governing several cellular processes, including cell proliferation and differentiation.37 DVL3 shared the MREs of hsa-miR-24-3p and hsa-miR-424-5p with hsa_circ_0006371 in HASMCs. One signalling modulator, Rspo3, of the Wnt signalling pathway was identified as a potential regulator of coronary stem proliferation.37 Additionally, the Wnt signalling pathway was shown to be involved in the proliferation and migration of pulmonary arteriolar smooth muscle cells.38 Consequently, we conjecture that hsa_circ_0006371 would be a positive conductor of Wnt signals isolating hsa-miR-24-3p and hsa-miR-424-5p to regulate HASMC proliferation. 4 \| DISCUSSION The relevant and irrelevant expressions are fifty-fifty FI G U R E 4 Comparison of expression levels of exonic and intronic DEcircRNAs to corresponding linear RNAs in proliferative HASMCs and quiescent HASMCs. Arrows represent the change of expression level. Upward arrows stand for the up-expression; downward arrows stand for down-expression, and the transverse lines stand for the invariable expression. A, Five categories, including co-expression (up-regulation or down-regulation), inverse-expression (circRNAs downward and mRNA upward or circRNAs upward and mRNA downward) and irrelevant expression, were based on the exonic DEcircRNAs and the corresponding mRNAs. B, There are only three categories in the intronic circRNAs. The relevant and irrelevant expressions are fifty-fifty FI G U R E 4 Comparison of expression levels of exonic and intronic DEcircRNAs to corresponding linear RNAs in proliferative HASMCs and quiescent HASMCs. Arrows represent the change of expression level. Upward arrows stand for the up-expression; downward arrows stand for down-expression, and the transverse lines stand for the invariable expression. A, Five categories, including co-expression (up-regulation or down-regulation), inverse-expression (circRNAs downward and mRNA upward or circRNAs upward and mRNA downward) and irrelevant expression, were based on the exonic DEcircRNAs and the corresponding mRNAs. B, There are only three categories in the intronic circRNAs. The relevant and irrelevant expressions are fifty-fifty According to recent significant research, many miRNAs, includ- ing miR-24-3p,29 miR-424-5p,30 miR-1298-5p,31 miR-204-5p32 and let-7a-5p,33 are involved in VSMC proliferation and down-regulated in proliferative VSMCs. In the HASMCs, the MREs of the miRNAs mentioned above were predicted in DEcircRNAs. We expected that FI G U R E 5 Knockdown of hsa_ circ_0002579 inhibited the proliferation of HASMCs. A, Thirty shared miRNAs between hsa_circ_0002579 and HMGA2 3′-UTR. B, HASMCs were treated with siRNA-hsa_circ_0002579/siRNA-Control for 48 h. Cell lysates were extracted and subjected to Western blotting using specific antibodies against HMGA2, PCNA and SM22α. GAPDH was used as equal loading control. The illustrated results were representative of three independent experiments FI G U R E 5 Knockdown of hsa_ circ_0002579 inhibited the proliferation of HASMCs. A, Thirty shared miRNAs between hsa_circ_0002579 and HMGA2 3′-UTR. B, HASMCs were treated with siRNA-hsa_circ_0002579/siRNA-Control for 48 h. Cell lysates were extracted and subjected to Western blotting using specific antibodies against HMGA2, PCNA and SM22α. GAPDH was used as equal loading control. The illustrated results were representative of three independent experiments 4770 CHEN et al. ACKNOWLEDGEMENTS 10. Abdelmohsen K, Panda AC, Munk R, et al. Identification of HuR tar- get circular RNAs uncovers suppression of PABPN1 translation by CircPABPN1. RNA Biol. 2017;14:361-369. This work was supported by grants from the National Natural Science Foundation of China (81670273 and 81200215 to SS), the Natural Science Foundation of Hebei Province (H2019206150 to SS) and the Program of International S&T Cooperation of Hebei Province (14393001D to SS). 11. Pamudurti NR, Bartok O, Jens M, et al. Translation of CircRNAs. Mol Cell. 2017;66(9–21):e7. 12. Wesselhoeft RA, Kowalski PS, Anderson DG. Engineering circu- lar RNA for potent and stable translation in eukaryotic cells. Nat Commun. 2018;9:2629. 13. Zhang M, Zhao K, Xu X, et al. A peptide encoded by circular form of LINC-PINT suppresses oncogenic transcriptional elongation in glioblastoma. Nat Commun. 2018;9:4475. CONFLICT OF INTEREST The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. 14. Liang WC, Wong CW, Liang PP, et al. Translation of the circular RNA circbeta-catenin promotes liver cancer cell growth through activa- tion of the Wnt pathway. Genome Biol. 2019;20:84. 15. Burd CE, Jeck WR, Liu Y, Sanoff HK, Wang Z, Sharpless NE. Expression of linear and novel circular forms of an INK4/ARF- associated non-coding RNA correlates with atherosclerosis risk. PLoS Genet. 2010;6:e1001233. ETHICAL APPROVAL The human aortic smooth muscle cells (HASMCs) we used in this study were commercial cells purchased from ScienCell. 18. Liu C, Yao MD, Li CP, et al. Silencing of circular RNA-ZNF609 ameliorates vascular endothelial dysfunction. Theranostics. 2017;7:2863-2877. However, further experimental work remains to be done to validate those forecasting results. 3. Wang S, Zhang X, Yuan Y, et al. BRG1 expression is increased in thoracic aortic aneurysms and regulates proliferation and apoptosis of vascular smooth muscle cells through the long non-coding RNA HIF1A-AS1 in vitro. Eur J Cardiothorac Surg. 2015;47:439-446. In our study, we investigated the DEcircRNAs that were found in the proliferative and quiescent conditions of HASMCs to help elucidate their function. Next, the DEcircRNA-miRNA-DEmRNA crosstalk network was constructed for further investigations. Then, we analysed the coding potential of DEcircRNA and the ef- fect of DEcircRNA on the corresponding linear RNA. Finally, hsa_ circ_0002579, one of the DEcircRNAs, might target HMGA2 to promote the proliferation of HASMCs by sponging miRNAs, which may be a new therapeutic target for VSMC proliferative vascular diseases. These findings provide novel insights for understanding the functions of circRNAs in HASMC proliferation and vascular remodelling diseases. 4. Zheng B, Bernier M, Zhang XH, et al. miR-200c-SUMOylated KLF4 feedback loop acts as a switch in transcriptional programs that con- trol VSMC proliferation. J Mol Cell Cardiol. 2015;82:201-212. 5. Djebali S, Davis CA, Merkel A, et al. Landscape of transcription in human cells. Nature. 2012;489:101. 6. Chen LL, Yang L. Regulation of circRNA biogenesis. RNA Biol. 2015;12:381-388. 7. Hansen TB, Jensen TI, Clausen BH, et al. Natural RNA circles func- tion as efficient microRNA sponges. Nature. 2013;495:384-388. 8. Zheng Q, Bao C, Guo W, et al. Circular RNA profiling reveals an abundant circHIPK3 that regulates cell growth by sponging multiple miRNAs. Nat Commun. 2016;7:11215. 9. Du WW, Yang W, Liu E, Yang Z, Dhaliwal P, Yang BB. Foxo3 circular RNA retards cell cycle progression via forming ternary complexes with p21 and CDK2. Nucleic Acids Res. 2016;44:2846-2858. In vitro, the overexpression of miR-424-5p inhibited VSMC pro- liferation,30 which implies that miR-424-5p might be induced as a signal to counteract the proliferation. The MREs of hsa-miR-424-5p were predicted in hsa_circ_0006371 and hsa_circ_0004872, which were also predicted in the DEmRNA ABL2. ABL2 was annotated in the Ras signalling pathway. Some researchers have identified ABL2 as the target of many regulators to facilitate the proliferation process.39 Here, we proposed that hsa_circ_0006371 and hsa_circ_0004872 could sponge hsa-miR-424-5p to up-regulate ABL2 to be involved in the proliferation of HASMCs via the Ras signalling pathway. Moreover, hsa_circ_0006371 shared the same MREs of hsa-miR- 24-3p, hsa-miR-424-5p and hsa-miR-1298-5p with MAP3K13 (namely LZK). The JNK signalling pathway was downstream of MAP3K13. A specific role of miR-1298 in regulating VSMC prolifera- tion was demonstrated.31 The MAPK/JNK pathway also might regu- late VSMC proliferation.40 In addition, another DEmRNA annotated in the MAPK pathway was ELK4, an ERK-regulated cofactor of the SRF transcription factor, which has been shown to act as a target of the ERK, JNK and p38 MAPK families.41-44 Based on the network, we predicted that ELK4 shared the MREs of hsa-miR-204-5p with By DEcircRNAs-DEmRNAs co-expression subnetwork analysis, we found has_circ_0002579 was co-expression with 35 DEmRNAs in the proliferative HASMCs (Figure 3B), and 30 miRNAs shared by hsa_circ_0002579 and HMGA2 3′-UTR (Figure 5A). Studies have shown that HMGA2 is overexpressed in cancer cells55,56 and promotes angiogenesis.57 And miR-4500,58 miR-98-5p59 and miR-485-5p60 have been reported to directly target the 3′-UTR of HMGA2 in other cell-type studies. These cases profile the bright prospects of hsa_circ_0002579, which acts as a ceRNA molecule interacting with miRNAs to regulate the proliferation of HASMCs by targeting HMGA2. We found that hsa_circ_0002579 was up-reg- ulated in the proliferative HASMCs (Figure 1), and knockdown of hsa_circ_0002579 could decrease the expression level of HMGA2 and PCNA (Figure 5B). We speculated that hsa_circ_0002579 might cross-interact with miRNAs to increase the expression of HMGA2, ultimately positively regulating the proliferation of VSMCs. CHEN et al. 4771 ORCID Wei Chen https://orcid.org/0000-0002-8375-518X Jiajie Lin https://orcid.org/0000-0003-3338-6264 Shaoguang Sun https://orcid.org/0000-0002-2827-0948 Wei Chen https://orcid.org/0000-0002-8375-518X Jiajie Lin https://orcid.org/0000-0003-3338-6264 Shaoguang Sun https://orcid.org/0000-0002-2827-0948 21. Pasquinelli AE. MicroRNAs and their targets: recognition, reg- ulation and an emerging reciprocal relationship. Nat Rev Genet. 2012;13:271-282. Shaoguang Sun https://orcid.org/0000-0002-2827-0948 22. Enright AJ, John B, Gaul U, Tuschl T, Sander C, Marks DS. MicroRNA targets in Drosophila. Genome Biol. 2003;5:R1. AUTHOR CONTRIBUTIONS SS conceived and designed this study. WC, JL, BL, SC, YL, YL, KL and SS performed the bioinformatics analysis. WC, JL, BL, HL and JZ per- formed the experiments. WC, JL, SC and SS wrote the manuscript. All authors read and approved the final manuscript. 16. Sun Y, Yang Z, Zheng B, et al. A novel regulatory mechanism of smooth muscle alpha-actin expression by NRG-1/circACTA2/miR- 548f-5p axis. Circ Res. 2017;121:628-635. 17. Kong P, Yu Y, Wang L, et al. circ-Sirt1 controls NF-kappaB activa- tion via sequence-specific interaction and enhancement of SIRT1 expression by binding to miR-132/212 in vascular smooth muscle cells. Nucleic Acids Res. 2019;47:3580-3593. DATA AVAILABILITY STATEMENT 19. Holdt LM, Stahringer A, Sass K, et al. Circular non-coding RNA ANRIL modulates ribosomal RNA maturation and atherosclerosis in humans. Nat Commun. 2016;7:12429. Raw sequence reads have been deposited in the GEO database (GSE77278 and GSE77279). 20. Wang K, Long B, Liu F, et al. A circular RNA protects the heart from pathological hypertrophy and heart failure by targeting miR-223. Eur Heart J. 2016;37:2602-2611. 46. Thomson DW, Dinger ME. Endogenous microRNA sponges: evi- dence and controversy. Nat Rev Genet. 2016;17:272-283. 25. Militello G, Weirick T, John D, Doring C, Dimmeler S, Uchida S. Screening and validation of lncRNAs and circRNAs as miRNA sponges. Brief Bioinform. 2017;18:780-788. 47. Salzman J, Gawad C, Wang PL, Lacayo N, Brown PO. Circular RNAs are the predominant transcript isoform from hundreds of human genes in diverse cell types. PLoS One. 2012;7:e30733. 26. Chen X, Han P, Zhou T, Guo X, Song X, Li Y. circRNADb: a com- prehensive database for human circular RNAs with protein-coding annotations. Sci Rep. 2016;6:34985. 48. Rodriguez J, Tang CH, Khodor YL, Vodala S, Menet JS, Rosbash M. Nascent-Seq analysis of Drosophila cycling gene expression. Proc Natl Acad Sci USA. 2013;110:E275-E284. 27. Tan WL, Lim BT, Anene-Nzelu CG, et al. A landscape of circular RNA expression in the human heart. Cardiovasc Res. 2017;113:298-309. 49. Ashwal-Fluss R, Meyer M, Pamudurti NR, et al. circRNA biogenesis competes with pre-mRNA splicing. Mol Cell. 2014;56:55-66. 28. Salmena L, Poliseno L, Tay Y, Kats L, Pandolfi PP. A ceRNA hy- pothesis: the Rosetta stone of a hidden RNA language? Cell. 2011;146(3):353–358. 50. Salzman J, Chen RE, Olsen MN, Wang PL, Brown PO. Cell- type specific features of circular RNA expression. PLoS Genet. 2013;9:e1003777. 29. Chan MC, Hilyard AC, Wu C, et al. Molecular basis for antagonism between PDGF and the TGFbeta family of signalling pathways by control of miR-24 expression. EMBO J. 2010;29:559-573. 51. Rybak-Wolf A, Stottmeister C, Glazar P, et al. Circular RNAs in the mammalian brain are highly abundant, conserved, and dynamically expressed. Mol Cell. 2015;58:870-885. 30. Merlet E, Atassi F, Motiani RK, et al. miR-424/322 regulates vascu- lar smooth muscle cell phenotype and neointimal formation in the rat. Cardiovasc Res. 2013;98:458-468. 52. Chen BJ, Mills JD, Takenaka K, Bliim N, Halliday GM, Janitz M. Characterization of circular RNAs landscape in multiple system at- rophy brain. J Neurochem. 2016;139:485-496. 31. Hu W, Wang M, Yin H, et al. MicroRNA-1298 is regulated by DNA methylation and affects vascular smooth muscle cell function by targeting connexin 43. Cardiovasc Res. 2015;107:534-545. 53. You X, Vlatkovic I, Babic A, et al. Neural circular RNAs are derived from synaptic genes and regulated by development and plasticity. Nat Neurosci. 2015;18:603-610. 32. Courboulin A, Paulin R, Giguere NJ, et al. Role for miR-204 in human pulmonary arterial hypertension. J Exp Med. 2011;208:535-548. 54. Li Z, Huang C, Bao C, et al. Exon-intron circular RNAs regulate tran- scription in the nucleus. Nat Struct Mol Biol. 2015;22:256-264. 33. Cao H, Hu X, Zhang Q, et al. Upregulation of let-7a inhibits vascular smooth muscle cell proliferation in vitro and in vein graft intimal hyperplasia in rats. J Surg Res. 2014;192:223-233. 55. Chang HY, Ye SP, Pan SL, et al. Overexpression of miR-194 re- verses HMGA2-driven signatures in colorectal cancer. Theranostics. 2017;7:3889-3900. 34. Heldin CH, Moustakas A. Signaling receptors for TGF-beta family members. Cold Spring Harb Perspect Biol. 2016;8:a022053. 56. Wang Y, Hu L, Wang J, et al. HMGA2 promotes intestinal tumori- genesis by facilitating MDM2-mediated ubiquitination and degra- dation of p53. J Pathol. 2018;246:508-518. 35. Yang J, Chen L, Ding J, et al. MicroRNA-24 inhibits high glucose-in- duced vascular smooth muscle cell proliferation and migration by targeting HMGB1. Gene. 2016;586:268-273. 36. Yang J, Fan Z, Yang J, Ding J, Yang C, Chen L. MicroRNA-24 at- tenuates neointimal hyperplasia in the diabetic rat carotid artery injury model by inhibiting Wnt4 signaling pathway. Int J Mol Sci. 2016;17:765. 57. Sakata J, Hirosue A, Yoshida R, et al. HMGA2 contributes to distant metastasis and poor prognosis by promoting angiogenesis in oral squamous cell carcinoma. Int J Mol Sci. 2019;20:2473. 58. Yu FY, Tu Y, Deng Y, et al. MiR-4500 is epigenetically downregu- lated in colorectal cancer and functions as a novel tumor suppres- sor by regulating HMGA2. Cancer Biol Ther. 2016;17:1149-1157. 37. Sharma M, Castro-Piedras I, Simmons GE Jr, Pruitt K. Dishevelled: a masterful conductor of complex Wnt signals. Cell Signal. 2018;47:52-64. 59. Guo Z, He C, Yang F, Qin L, Lu X, Wu J. Long non-coding RNA- NEAT1, a sponge for miR-98-5p, promotes expression of oncogene HMGA2 in prostate cancer. Biosci Rep. 2019;39:BSR20190635. 38. Luo L, Hong X, Diao B, Chen S, Hei M. Sulfur dioxide attenuates hypoxia-induced pulmonary arteriolar remodeling via Dkk1/Wnt signaling pathway. Biomed Pharmacother. 2018;106:692-698. 60. Chen Z, Li Q, Wang S, Zhang J. miR-485-5p inhibits bladder cancer metastasis by targeting HMGA2. Int J Mol Med. 2015;36:1136-1142. 39. Kazi JU, Rupar K, Marhall A, et al. ABL2 suppresses FLT3-ITD- induced cell proliferation through negative regulation of AKT sig- naling. Oncotarget. 2017;8:12194-12202. 40. Gan J, Li P, Wang Z, et al. Rosuvastatin suppresses platelet-derived growth factor-BB-induced vascular smooth muscle cell prolifera- tion and migration via the MAPK signaling pathway. Exp Ther Med. 2013;6:899-903. 4772 \| 4772 CHEN et al. REFERENCES 23. Ramanan VK, Shen L, Moore JH, Saykin AJ. Pathway analysis of genomic data: concepts, methods, and prospects for future devel- opment. Trends Genet. 2012;28:323-332. 1. Bennett MR, Sinha S, Owens GK. Vascular smooth muscle cells in atherosclerosis. Circ Res. 2016;118:692-702. 2. Yu ZL, Wang JN, Wu XH, et al. Tanshinone IIA prevents rat basilar artery smooth muscle cells proliferation by inactivation of PDK1 during the development of hypertension. J Cardiovasc Pharmacol Ther. 2015;20:563-571. 24. Liu M, Wang Q, Shen J, Yang BB, Ding X. Circbank: a comprehen- sive database for circRNA with standard nomenclature. RNA Biol. 2019;16:899-905. SUPPORTING INFORMATION Additional supporting information may be found online in the Supporting Information section. 41. Janknecht R, Ernst WH, Nordheim A. SAP1a is a nuclear target of signaling cascades involving ERKs. Oncogene. 1995;10:1209-1216. 42. Price MA, Rogers AE, Treisman R. Comparative analysis of the ter- nary complex factors Elk-1, SAP-1a and SAP-2 (ERP/NET). EMBO J. 1995;14:2589-2601. How to cite this article: Chen W, Lin J, Li B, et al. Screening and functional prediction of differentially expressed circRNAs in proliferative human aortic smooth muscle cells. J Cell Mol Med. 2020;24:4762–4772. https://doi.org/10.1111/jcmm.15150 43. Janknecht R, Hunter T. Activation of the Sap-1a transcription factor by the c-Jun N-terminal kinase (JNK) mitogen-activated protein ki- nase. J Biol Chem. 1997;272:4219-4224. 44. Janknecht R, Hunter T. Convergence of MAP kinase pathways on the ternary complex factor Sap-1a. EMBO J. 1997;16:1620-1627. 45. Denzler R, Agarwal V, Stefano J, Bartel DP, Stoffel M. Assessing the ceRNA hypothesis with quantitative measurements of miRNA and target abundance. Mol Cell. 2014;54:766-776.
https://openalex.org/W1908300887	https://www.notulaebiologicae.ro/index.php/nsb/article/download/5593/8299	English	null	An Assessment of the Accumulation Potential of Pb, Zn and Cd by Avicennia marina (Forssk.) Vierh. in Vamleshwar Mangroves, Gujarat, India	Notulae Scientia Biologicae	2,011	cc-by	3,609	Abstract The study deals with the accumulation of Pb, Zn and Cd in an important mangrove species, Avicennia marina (Forssk.) Vierh., in the Vamleshwar mangrove ecosystem, near Narmada estuary, West coast of Gujarat, India with height differences of 0.5, 1.5, 2.5 meters and carried out under field conditions during October, 2009. The site was located on 21°30´11.55´´ N latitude and 72°43´53.68´´ E longitude. Mangrove receives heavy metal pollution from upstream areas of Narmada estuary and highly populated settlements. However, little is known about the capacity of mangrove plants to take up and store heavy metals in them. Water, sediment and plant parts such as roots, stems and leaves were analyzed for finding the trace metal accumulation of different height groups by Inductive Coupled Plasma Analyser (ICPA). Amount of the content of metals found in the water, sediment and plant parts were in the order of Pb>Zn>Cd. The average contents of heavy metals in the waters were 57.83 mg l-1 for Pb, 3.89 83 mg l-1 for Zn and 0.42 mg l-1 for Cd. It was observed that the average contents of Pb (73.6 mg l-1), Zn (8.1 mg l-1) and Cd (0.73 mg l-1) in the sediments were below the critical soil concentrations. The concentrations of heavy metals in different parts of Avicennia marina were in the order Roots>stem>leaf except for Cd, but Cd found higher in leaf. The ranges of the content of heavy metals in plants were 18.5-102.2 mg l-1 for Pb, 3.5-19.5 mg l-1 for Zn and 0.2-4.1 mg l-1 for Cd. The concentrations of all heavy metals in Avicennia marina except Pb were falling within the normal range and were much more in the plants have the highest height. The present study has shown the potential of Avicennia marina as a phytoremediation species for selected heavy metals in many mangrove ecosystems. nia marina (Forssk.) Vierh., heavy metal accumulation, mangrove plant parts, sediments, Vamleshwar mangrove Keywords: Avicennia marina (Forssk.) Vierh., heavy metal accumulation, mangrove plant parts, sediments, Vamleshwar mangrove ecosystem ticulate matter, which eventually settle down and become incorporated into sediments. Surface sediment is the most important reservoir or sink of metals and other pollut- ants in aquatic environments. Sediment-bound pollutants can be taken up by rooted aquatic macrophytes and other aquatic organisms. Abstract Water pollution around the mining and smelting areas has been clearly demonstrated by the high concentrations of heavy metals in water, sediments, and aquatic organisms (Cardwell et al., 2002; Demirezen and Aksoy, 2004; Samecka-Cymerman and Kempers, 2004). In recent years, there has been an ever-increasing interest in the study of metal-accumulating plants for en- vironmental remediation application, termed as ‘phytore- mediation’. One method of phytoremediation is phytoex- traction, which uses metal-accumulating plants to remove pollutants from contaminated soils by concentrating them in the harvestable aboveground parts (Salt et al., 1998). For wastewater containing high concentrations of heavy metals, plant roots can also be used to absorb and adsorb An Assessment of the Accumulation Potential of Pb, Zn and Cd by Avicennia marina (Forssk.) Vierh. in Vamleshwar Mangroves, Gujarat, India Isaiah J. NIRMAL KUMAR1, Poliyaparambil Ravi SAJISH1, Rita NIRMAL KUMAR2, George BASIL1, Viyol SHAILENDRA1 Isaiah J. NIRMAL KUMAR1, Poliyaparambil Ravi SAJISH1, Rita NIRMAL KUMAR2, George BASIL1, Viyol SHAILENDRA1 1Institute of Science and Technology for Advanced Studies and Research (ISTAR), P. G. Department of Environmental Science and Technology, Vallabh Vidya Nagar, Gujarat-388120, India; istares2005@yahoo.com 2N. V. Patel College of Pure and Applied Sciences, Department of Biological Science and Environmental Science, Vallabh Vidya Nagar, Gujarat-388 120, India; nvpas2009@gmail.com Received 11 December 2010; accepted 03 February 2011 Available online at www.notulaebiologicae.ro Not Sci Biol, 2011, 3(1):36-40 Print ISSN 2067-3205; Electronic 2067-3264 Available online at www.notulaebiologicae.ro Not Sci Biol, 2011, 3(1):36-40 Print ISSN 2067-3205; Electronic 2067-3264 Notulae Scientia Biologicae Introduction Mangrove ecosystems are highly productive and play a vital role as a major primary producer within estuarine systems. The uniqueness of A. marina root system serves as a habitat and nursery area for many juvenile fish and crustaceans, which have both direct and indirect socio- economic significance and are of great importance to many scientific studies. They also provide erosion miti- gation and stabilization for adjacent coastal landforms (Harty, 1997). Environmental pollution with toxic metals has been attracting considerable public attention over the past few decades. One of the major sources of heavy metal pollution is the mining and smelting of metalliferous ores (Li and Thornton, 2001). Besides the contamination from the weathering and leaching processes of mine tailings, untreated mine drainage also contributes large amounts of heavy metals to nearby streams and rivers. Heavy metals released into aquatic systems are generally bound to par- Nirmal Kurmar, J. I. et al. / Not Sci Biol, 2011, 3(1):36-40 37 Analysis of plant parts, water and sediment for heavy metals Fig. 1. Location of the study area 37 pollutants, i.e., rhizofiltration. Aquatic plants can absorb bulk mineral salts either from the sediments via the root system, from the water phase by the leaves, or from both sources (Rai et al., 2003).i ( ) In a plant-soil system, strong absorption and fixation of heavy metals by soil can easily cause residual accumulation in the soil, resulting in over-absorption of heavy metals by growing plants (Lian et al., 1999; Ravikumar et al., 2007). These plant products are harmful to the health of humans (Defew et al., 2005). The uptake of heavy metals by plants are passive, and its translocation from roots to other plant organs is generally low (Baker and Walker, 1990; De Lac- erda and Abrao, 1986; Nirmal Kumar et al., 2008). Many studies have been carried out on various plants to deter- mine its heavy metal accumulation capability in differ- ent aquatic and forest environments and vegetable crops (Nirmal Kumar et al., 2006, 2007, 2009; Silva et al., 2006; Yu et al., 2007; ). There are several studies on heavy metal contamination in mangrove sediments and their effects on organisms but little is known about heavy metals uptake by mangrove plants (Seng et al., 1987; Ismail and Asmah, 1992). Therefore, an attempt was made to understand the metal phytoremediation potential of A. Introduction marina in a mangrove ecosystem at Vamleshwar, Gujarat, India, that would benefit in the wise-management and utilization of this ecosystem. The present study presents data on the concentrations of three metals (Pb, Zn and Cd) in water, sediment, and in leaves, stems and roots of A. marina. Fig. 1. Location of the study area Fig. 1. Location of the study area Analysis of plant parts, water and sediment for heavy metalsh Analysis of plant parts, water and sediment for heavy metalsh Analysis of plant parts, water and sediment for heavy metalsh The plant parts of roots, stems and leaves samples were dried at 60°C for two days to achieve constant weight, homogenised and grounded to a fine powder. Samples then stored in plastic vials with labels and kept in desicca- tors. The samples of water, sediment and plant-parts were chemically examined for detection for Pb, Zn and Cd. Ac- curately 1 g of dry powder of each sample was weighed, and digested with con. HNO3, H2SO4 and H2O2 (2:6:6) as prescribed by Saison et al. (2004). Towards the end of the digestion, the flasks were brought to near dryness. The solutions were made to 20 ml each in measuring cylinder with double distilled water and examined for Pb, Zn and Cd by Inductive Coupled Plasma Analyser (ICPA) (Per- kin-Elmer ICP Optima 3300 RL, U.S.A) at Sophisticated Instrumentation Centre for Applied Research and Testing (SICART), Vallabh Vidyanagar, Gujarat, India. Mean val- ues of triplicate of each sample of the water, sediment and plant samples were calculated and considered. Sampling site S l p g Sampling was conducted at Vamleshwar A. marina plantation, which is located 45km to west of Ankleshwar, Gujarat, India and facing the Arabian Sea at a latitude of 21°30´11.55´´ N and longitude of 72°46´53.68´´ E. The study area has various natural resources, vast array of bio- logical diversity and coastal and riverine fishing activities (Fig. 1). The plant parts of different height groups viz., 0.5 m, 1.5 m and 2.5 m were collected in three replicates from the study site (n=10) during October 2009. The roots were carefully removed from the sediments while the leaves and stems were collected using a pair of scissors. The samples were immediately wrapped with an aluminium foil and la- beled. All the samples were kept in a cooler box with ice at 4°C for storage during transportation and brought to the laboratory (Prica et al., 2007). After reaching the laborato- ry, root, stem and leaves samples were washed with double distilled water before storage. Water and sediment samples were collected from same sites where plant samples were collected. Soon after collection, the water samples were filtered through 0.45 µm millimeter filter and preserved in plastic bottles by the addition of a few drops of nitric acid. Sediment samples were preserved in air-dry plastic bags. The samples were labeled carefully and brought to the laboratory for further analysis. Statistical analysis l Statistical analysis Pearson correlation coefficient analysis was done be- tween metal-pairs in plants to check if differences exist between different metal combinations in either root, stem or leaf system. The products of the correlation coefficient (r) were evaluated as follows: (r) were evaluated as follows: • 0-0.3: No correlation; • 0.3-0.5: Low correlation; • 0.5-0.7 Medium correlation; • 0.7-0.9 High correlation; • 0.9-1.0 Very high correlation. Further the comparison of the concentration of an ele- ment in an aquatic organism with that of the same element in the water in which the organism lives was made. This is the ratio between the concentration of the element in Nirmal Kurmar, J. I. et al. / Not Sci Biol, 2011, 3(1):36-40 38 Fig. 2A. Fig. 2B. Fig. 2C. Fig. (2A-2C). The mean concentrations of Pb, Zn and Cd in the roots, stems, leaves, water and sediments of A. marina (n=3) plantation of different height groups viz., 0.5 M, 1.5 M and 2.5 M the organism and that of in the water, which is known as Concentration Factor (De Bortoli et al., 1968). Further the data was analyzed using ANOVA to detect if any significant differences in means of each heavy metal exist between mangrove sediments, roots, stems and leaves and of different heights. Significance value was set at 5%. Results and discussion Heavy metal accumulation in the mangrove sediments, water and mangrove plant partsh g p p The average concentration of Pb (73.6 mg l-1) was higher in the sediments followed by Zn (8.1 mg l-1), and Cd (0.73 mg l-1) (Fig. 2A-2C). like sediment the concen- tration of Pb (57.83 mg l-1) in water was greater than the other two elements (Zn-3.89 mg l-1and Cd-0.42 mg l-1) (Fig. 2). However, when compared to the critical soil con- centrations, the total concentrations for all selected heavy metals present in the sediments were below the critical soil concentration values (Alloway, 1990) (Tab. 1). Mean- while, the comparison of the total concentrations of the selected heavy metals in the sediments with those from other countries showed that the concentration of Cd in this study was greater than those measured in Buloh River and Khatib Bongsu River at Singapore by Cuong et al. (2005), Brisbane River, Australia by Mackey et al. (1992) and Mai Po river, Hong Kong by Ong Che (1999). The values of the ratio between element concentrations in the sediments and those in the water were low for Pb (1.27) where as that of Zn was observed high (2.08) (Tab. 2). Fig. 2A. Fig. 2A. Fig. 2B. Fig. 2B. Fig. 2B. Fig. 2B. Fig. 2C.h Heavy metal concentrations were detected in the man- grove parts prove that Pb, Zn and Cd were distributed in all three parts of A. marina. From the observations, A. marina accumulated high concentration of heavy metals in the root system compared to the other parts of plant except Cd which found high in leaves followed by root and stem. Greater accumulation was observed in the plant parts of tallest plant group i.e. 2.5 m. The average content of Pb in leaf, stem, and root were 48.64, 21.94 and 91.07 mg l-1 while for Zn 11.24, 5.78 and 15.53 mg l-1 whereas for Cd 2.40, 0.77 and 1.70 mg l-1 respectively (Fig. 2) The average content of the elements in the plants declined ac- cording to this sequence: Pb>Zn>Cd. Fig. (2A-2C). The mean concentrations of Pb, Zn and Cd in the roots, stems, leaves, water and sediments of A. marina (n=3) plantation of different height groups viz., 0.5 M, 1.5 M and 2.5 M Fig. (2A-2C). The mean concentrations of Pb, Zn and Cd in the roots, stems, leaves, water and sediments of A. Results and discussion marina (n=3) plantation of different height groups viz., 0.5 M, 1.5 M and 2.5 M Tab. 1. The total concentrations of heavy metals (mg L-1) in mangrove sediments of Vamleshwar mangrove ecosystem (present study) relative to other countries. Critical soil concentrations follow Alloway (1990) Tab. 1. The total concentrations of heavy metals (mg L-1) in mangrove sediments of Vamleshwar mangrove ecosystem (present study) relative to other countries. Critical soil concentrations follow Alloway (1990) Pb Zn Cd References Vamleshwar Mangrove Ecosystem, Gujarat, India 73.6±4.72 8.1±2.94 0.73±0.153 Present study Singapore (Buloh River) 12.28±5.18 51.24±39.97 0.181±0.189 Cuong et al. (2005) Singapore (Khatib Bongsu River) 30.98±6.16 120.23±13.90 0.266±0.171 Cuong et al. (2005) Australia (Brisbane River) 20.1-81.9 40.8-144.0 <0.1-1.9 Mackey et al. (1992) Hong Kong (Mai Po) 161.6-219.8 277.2-321.2 0.5-0.6 Ong Chee (1999) Critical Soil Concentrations 100-400 70-400 3-8 Alloway (1990) Nirmal Kurmar, J. I. et al. / Not Sci Biol, 2011, 3(1):36-40 39 Tab. 2. Elemental concentration ratios between sediments and water The ANOVA also showed marked distinction of the heavy metals (Pb, Zn and Cd) concentrations between mangrove sediments and the roots of A. marina (Pb, p=0.01; Zn, p=0.005; Cd, p=0.01). h Sediment (mg l-1) Water (mg l-1) Sediment/Water Pb 73.6 57.83 1.27 Zn 8.1 3.89 2.08 Cd 0.73 0.42 1.74 p p p The study showed the transport mechanism of metals from sediment to plant and their accumulation in various parts of the plant. The transport mechanism and accumu- lation pattern of heavy metals can be elaborated as follows: sediment>root system>stem system>leaf system. The critical ranges of heavy metals in plants were 30- 300 mg l-1 for Pb, 100-400 mg l-1 for Zn and 10-30 mg l-1 for Cd. The concentrations of all heavy metals in A. ma- rina except Pb were within the normal range in the present study. Conclusions This study has shown that A. marina possesses the ca- pacity to take up selected heavy metals via its roots and storing in its leaves without any sign of injury. As it grows more aged, its capability of accumulating heavy metals is also increasing much fold. This suggests the potential of A. marina as a phytoremediation species for many mangrove ecosystems. The luxuriant growth of A. marina in com- parison to other mangrove species is evident to its adapt- ability even under polluted conditions. Acknowledgements Th h h The authors are highly thankful to Ministry of Envi- ronment and Forests (MoEF), New Delhi for financial support for the project. Tab. 3. Concentration factors calculated for different parts of A. marina and elements Statistical analysis Th The concentration factor (C. F) for all three heavy metals and that of different components of A. marina also were evaluated (Tab. 3). The mean C. F. value of elements in the different parts declined according to this sequence: Cd> Zn > Pb. The mean C. F. for the various elements calculated for root is rather high followed by leaf and stem (Tab. 3). The concentration distribution in plant parts like leaves, stems and roots may vary depending on the con- centration of heavy metals in the sediment, height of the plants, the types of heavy metals and also the tolerance of the species and its parts towards the heavy metals. The Tab. 3. Concentration factors calculated for different parts of References A comparison of sediment quality results with acid volatile sulfide (AVS) and simultaneously extracted metals (SEM) ratio in Vojvodina (Serbia) sediments. Science of the Total Environment 389:235-44. Harty C (1997). Mangrove in New South Wales and Victoria. Vista Publication, Melbourne. Ismail A, Asmah MIN (1992). Copper, zinc, lead and cadmium in intertidal molluscs and sediment off Seberang Perai coastline, Malaysia. 4th Princess Chulabhorn International Science Congress, Bangkok, Thailand. Rai UN, Tripathi RD, Vajpayee P, Oandey N, Ali MB, Gupta DK (2003). Cadmium accumulation and its phytoxicity in Potamogeton pectinatus L. (Potamogetonaceae). Bulletin of Environmental Contamination and Toxicolog 70:566–575. Li XD, Thornton I (2001). Chemical partitioning of trace and major elements in soils contaminated by mining and smelting activities. Applied Geochemistry 16:1693-1706. Ravikumar S, Prakash Williams G, Shanthy S, Anitha Anantha Gracelin N, Babu S, Parimala PS (2007). Effect of heavy metals (Hg and Zn) on the growth and phosphate solubilising activity in halophilic phosphobacteria isolated from Manakudi mangrove. Journal of Environmental Biology 28:109-114. Lian Y, Xu J, Lin P, Meguro S, Kawachi S (1999). Five heavy metals in propagules of ten mangrove species of China. Journal of Wood Science 45:18-24. Mackey AP, Hodgkinson M, Nardella R (1992). Nutrient levels and heavy metals in mangrove sediments from the Brisbane River, Australia. Marine Pollution Bulletin 24:418-20. Saison C, Schwartz C, Morel JL (2004). Hyperaccumulation of metals by Thlaspi caerulescens as affected by root development and Cd-Zn/Ca-Mg interactions. International Journal of Phytoremediation 6(1):49-61. Nirmal Kumar JI, Soni H, Nirmal Kumar R (2006). Biomonitoring of selected freshwater macrophytes to assess the lake trace element contamination: A case study of Nal Sarovar Bird Sanctuary, Gujarat, India. Journal of Limnology 65(1):9-16. Salt DE, Smith RD, Raskin I (1998). Phytoremediation. Annual Review of Plant Physiology and Plant Molecular Biology 49:643-648. Samecka-Cymerman A, Kempers AJ (2004). Toxic metals in aquatic plants surviving in surface water polluted by copper mining industry. Ecotoxicology and Environmental Safety 59:64-69. Nirmal Kumar JI, Soni H, Nimal Kumar R (2007). Characterization of Heavy Metals in Vegetables Using Inductive Coupled Plasma Analyzer (ICPA). Journal of Applied Science and Environmental Management 11(3):75- 79. Seng CE, Lim PPE, Ang TT (1987). Heavy metal concentrations in coastal sea water and sediment of Prai Industrial Estate, Penang, Malaysia. Mar Pollut Bull 18:611-612. Nirmal Kumar JI, Soni H, Nimal Kumar R, Bhatt I (2008). References Pb Zn Cd Mean Leaf 0.84 2.89 5.71 3.15 Stem 0.38 1.49 1.83 1.23 Root 1.57 4.0 4.05 3.21 Mean 0.93 2.79 3.86 SD 0.60 1.26 1.95 Alloway BJ (1990). Heavy metals in soils. John Wiley and Sons, Inc, New York. Baker AJ, Walker PI (1990). Ecophysiology of metal uptake by tolerant plants, p. 155-178. In: Shaw AJ (Eds.). Heavy metal tolerance in plants evolutionary aspects. CRC, Press, Boca Raton, FL. Cardwell AJ, Hawker DW, Greenway M (2002). Metal accumulation in aquatic macrophytes from southeast Queensland, Australia. Chemosphere 48:653-663. present study is well corroborated with works by De Lac- erda andAbrao (1986); Baker and Walker (1990).hfi present study is well corroborated with works by De Lac- erda andAbrao (1986); Baker and Walker (1990).hfi Cuong DT, Bayen S, Wurl O, Subramanian K, Wong KKS, Sivasothi N, Obbard JP (2005). Heavy metal contamination in mangrove habitats of Singapore. Marine Pollution Bulletin 50:1713-1744. The output of Pearson correlation coefficient (r) analy- sis on combinations of different metal-pairs which are present together in either roots, stems or leaves of Avicen- nia marina showed that a high +ve correlation (r=0.6-0.9) between all the metal-pairs analyzed (Tab. 4) except in the root system. De Bortoli M, Gaglione P, Malvicini A, Polvani A (1968). Concentration factors for strontium and caesium in fish of the lakes in the region of Varese. Giorn Fisica Sanit Radioprot 12:324-331. Tab. 4. Pearson correlation coefficient between concentrations of heavy metal pairs in root, stem and leaf systems of A. marina Analysis metal-pair Root system Stem system Leaf system Zn × Pb 0.655 0.844 0.635 Cd × Pb -0.982 0.990 0.980 Zn × Cd -0.500 0.911 0.775 high correlation (r=0.6-0.9) De Lacerda LD, Abrao JJ (1986). Heavy metal accumulation by mangrove and saltmarsh intertidal sediments. Marine Pollution Bulletin 17:246-250. Tab. 4. Pearson correlation coefficient between concentrations of heavy metal pairs in root, stem and leaf systems of A. marina Defew LH, Mair JM, Guzman HM (2005). An assessment of metal contamination in mangrove sediments and leaves from Punta Mala Bay, Pacific Panama. Marine Pollution Bulletin 50:547-552. Demirezen D, Aksoy A (2004). Accumulation of heavy metals in Typha angustifolia (L.) and Potamogeton pectinatus (L.) Nirmal Kurmar, J. I. et al. / Not Sci Biol, 2011, 3(1):36-40 40 living in Sultan Marsh (Kayseri, Turkey). Chemosphere 56:685-696. Prica M, Dalmacija B, Roncevic S, Krcmar D, Becelic M (2007). References Biomonitoring of trace elements in some selected aquatic macrophytes with reference to lake contamination: A case study of Pariyej community reserve, Gujarat, India. Asian Journal of Microbiology Biotechnology and Environmental Science 10(4):803-810. Silva CAR, de Silva AP, Oliveira SR (2006). Concentration, stock and transport rate of heavy metals in a tropical red mangrove, Natal, Brazil. Marine Chemistry 99:2-11. Yu KF, Kamber BS, Lawrence MG, Greig A, Zhao JX (2007). High precision analysis on annual variation of heavy metals, lead isotopes and rare earth elements in mangrove tree rings by inductively coupled plasma mass spectrometry. Nuclear Instruments and Methods in Physics Research, Section B-Beam interactions with materials and atoms 255:399- 408. Nirmal Kumar JI, Soni H, Nimar Kumal R, Bhatt I (2009). Hyper accumulation and mobility of heavy metals in vegetable crops in India. The Journal of Agriculture and Environment 10:29-38. Ong Che RG (1999). Concentration of seven heavy metals in sediment and mangrove root samples from Mai Po, Hong Kong. Marine Pollution Bulletin 39:269-79.
https://openalex.org/W4220894509	https://zenodo.org/records/5930586/files/Joint_Migration_Policy_Roundtable__2__Handbook.pdf	English	null	JOINT HANDBOOK ON MIGRANTS' INTEGRATION: A SERVICE DESIGN AND SUPPLY PERSPECTIVE	Zenodo (CERN European Organization for Nuclear Research)	2,022	cc-by	29,803	(JANUARY 2022) These projects have received funding from the European Union’s Horizon 2020 research and innovation programme under the grant agreements No 822380, 822215, 822601, 822717, 870930 and 870980. These projects have received funding from the European Union’s Horizon 2020 research and innovation programme under the grant agreements No 822380, 822215, 822601, 822717, 870930 and 870980. These projects have received funding from the European Union’s Horizon 2020 research and innovation programme under the grant agreements No 822380, 822215, 822601, 822717, 870930 and 870980. Disclaimer All opinions and positions contained in this Joint Handbook on Migrants’ Integration are those of the projects concerned and reflect only the authors’ views and not those of the European Commission. 2 2 Table of Contents Executive summary ................................................................................................................................ 3 Definitions .............................................................................................................................................. 4 Abbreviations ......................................................................................................................................... 5 1. Introduction ........................................................................................................................................ 7 1.1 Aim of this Handbook ................................................................................................................... 7 1.2 The 2020 Whitepaper ................................................................................................................... 8 2. Service design and social integration ............................................................................................... 10 2.1. Can service design contribute to social integration? ............................................................... 10 2.2. From lessons learned to acknowledged pitfalls and policy recommendations ....................... 11 CO-DESIGN AND CONVERGENCE .................................................................................................. 11 COMMUNITY ORIENTED APPROACHES ......................................................................................... 13 CO-CREATION, SOCIAL DESIGN AND TRANSFORMATIVE IMPACTS ............................................... 15 HUMAN RIGHTS LITERACY IN SERVICE DESIGN AND SUPPLY ........................................................ 17 PRIVACY AND DATA PROTECTION ................................................................................................. 22 PERSONALIZATION OF INTELLIGENT TECHNOLOGIES FOR THIRD COUNTRY NATIONALS ............ 26 3. Sustainability of solutions and social integration ........................................................................... 28 3.1. Digital solutions for social integration: what are the challenges for sustainability? .............. 28 3.2. From lessons learned to acknowledged pitfalls and policy recommendations ....................... 29 BOTTLENECKS IN MIGRANTS’ INTEGRATION PROCEDURES .......................................................... 29 IMPACTS OF SOCIO-TECHNICAL SOLUTIONS FOR MIGRANTS’ INTEGRATION .............................. 31 DEVELOPING A VIABLE UPTAKE STRATEGY ................................................................................... 33 SUSTAINABILITY OF THE NOVEL ICT SOLUTIONS ........................................................................... 35 PSYCHOSOCIAL DIMENSIONS OF AI SUPPORTED SERVICES .......................................................... 39 ASSESSING SERVICE ACCESSIBILITY: TOWARDS A NEW STANDARD ............................................. 41 2 2 Executive summary On October 21 and 22, 2021, the six projects named MIICT, REBUILD, NADINE, MICADO, easyRights, and WELCOME, all funded under the European Union’s Horizon 2020 research and innovation program, organized their second Joint Migration Policy Roundtable. This Handbook reflects on the two thematic lines common to all the six projects, namely: co-creation, participatory design and multi-stakeholder collaboration on the one hand; ICT or AI based tools and novel digital services for supporting integration on the other. Its contents are the result of the discussion among Roundtable participants, which was articulated in two thematic sessions: the first taking stock of the co- creational dimension to frame the wider issue of service co-design for social integration; the second reflecting on the sustainability conditions of newly developed ICT or AI enabled tools and services. The twin aim was to identify pitfalls of the current approaches to migrant and refugee integration as carried out at local level, and to formulate policy recommendations on how to improve the migrant and refugee integration process within the EU. In this sense, the Handbook aspires to provide a strategic toolbox to support reflections in the development of future migration policies at the EU and national levels. After a brief Introduction reiterating the global aim of the document and briefly reporting about the proceedings of the first Joint Migration Policy Roundtable held in October 2020 by the same six projects, the remainder of this Handbook text is structured in two main sections, each mirroring a thematic session of the second Roundtable as explained, both co-authored by individual consortium members of the six aforementioned projects. Section 2 deals with the overarching question: Can service design contribute to social integration? This topic is dealt with according to the following structure: ● Co-design and convergence (responsible project MICADO) ● Co-creation, social design and transformative impacts (responsible project REBUILD) ● Co-creation, social design and transformative impacts (responsible project REBUILD) ● Human rights literacy in service design and supply (responsible project easyRights) ● Human rights literacy in service design and supply (responsible project easyRights) ● Privacy and data protection (responsible project MIICT) ● Personalization of intelligent technologies for TCNs (responsible project WELCOME). Definitions Migrants. This Handbook uses migrants as an umbrella term and adopts the inclusivist approach to defining migrants as described in the glossary of Migration and Home Affairs of the European Commission1. Therefore, migrants in this document are mobile persons that reside outside the territory of the State of which they are nationals or citizens of, irrespective of the causes, voluntary or involuntary, and the means, regular or irregular, of this change of residence. This includes individuals who fit into any of these categories: migrant, refugee, or asylum seeker. Third-Country Nationals (TCNs). According to EU Law, a TCN is any person who is not a citizen of the European Union within the meaning of Art. 20(1) of the TFEU (Treaty on the Functioning of the European Union) and who is not a person enjoying the European Union right to free movement, as defined in Art. 2(5) of the Regulation (EU) 2016/399 (Schengen Borders Code)2. Service design. This term is used to refer to the practice of arranging the actors, infrastructure, communication, and material elements to constitute or improve the quality of service. A service design methodology is characterized by attention to the needs of service users and the capabilities of service providers. In this Handbook, service design tools and methods are commonly established within co-design sessions that allow the study and consideration of the different needs and requirements of involved stakeholders and share the design process with them as experts in the area, to guarantee the sustainability, adoption, and relevance of the proposed services. Community-oriented approach. The term highlights the role that local communities play in the integration process as well as their responsibility to make integration a smooth process. In this approach, both local and migrant communities are relevant to guarantee the existence of a social landscape of integration. Social design. The term identifies the socially oriented design aiming to understand societal problems and support positive societal change. Social design can impact the quality of social relations, through the activation of real transformations and changes. Through the design of objects, social design intervenes in society-at-large and on the relationships between the people who are part of it. Human rights literacy. 1 https://ec.europa.eu/home-affairs/pages/glossary/migrant_en 2 2 https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32016R0399 1 https://ec.europa.eu/home-affairs/pages/glossary/migrant_en 2 https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32016R0399 1 https://ec.europa.eu/home-affairs/pages/glossary/migrant_en 2 https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32016R0399 Executive summary Section 3 examines the challenges for sustainability of digital solutions for social integration, taking on the following perspectives: ● Bottlenecks in migrants' integration procedures (responsible project NADINE) ● Impacts of socio-technical solutions for migrants’ integration (responsible project REBUILD) ● Psychosocial dimensions of AI-supported services (responsible project WELCOME) ● Assessing service accessibility: towards a new standard (responsible project easyRig Key lessons learned on the various issues are collected and interpreted in terms of acknowledged pitfalls and evidence-based policy recommendations, representing the “actionable” contents of this Handbook. 3 Definitions The term does not only involve human rights related to education or specific professional skills, but also the empowerment of migrants, public officials, members of institutions, civil society, and NGOs in their respective communication, bearing in mind the innate human dignity of any human being. 1.1 Aim of this Handbook In the Member States of the European Union (EU) and Associated Countries, the integration of migrants and refugees remains a challenge for both public authorities and local communities3. The integration of already established migrants - especially the next generations - as well as “newcomers” such as refugees and asylum seekers, constitutes a serious challenge for our societies and at the same time a unique opportunity to improve our societies in many respects. Obstacles and pitfalls are nonetheless in place, which need to be kept in due consideration. MIICT, Re-Build, NADINE, MICADO, WELCOME, and easyRights are six Horizon 2020 co-funded innovation actions (IA), which work towards the development of ICT-enabled solutions and toolkits for the implementation of novel inclusion policies and integration support services in collaboration between public administration, private sector and civil society. The six projects apply participatory co-creation methodologies in the development of ICT solutions and follow a collaborative approach involving not only migrants, but also public authorities and other key stakeholders for leveraging the tools and services4. In fact, efficient and effective integration of migrants and refugees can only be achieved by clearly understanding the individual situation and the social embedment of the problem. It needs the involvement of key stakeholders such as researchers, public authorities, and civil society organizations as well as the migrants and refugees themselves. On October 14-16, 2020, the six EU-funded projects organized a Joint Migration Policy Roundtable, structured in three sessions, each focusing on one of their common pillars, namely: co-creation and participatory design; multi-stakeholder collaboration; ICT tools and digital services for supporting integration. These common thematic pillars were described in a consolidated Whitepaper, which was meant to be a strategic input for improving integration and shaping future migration policies in the EU (see Joint Migration Policy Whitepaper, 2020, p.8 ff.)5. Following this successful experience, on October 21-22, 2021, the six EU projects organized a Second Joint Migration Policy Roundtable focused on the contribution of service design to integration and the sustainability conditions for the newly developed digital solutions in support of migrants. Two thematic sessions were foreseen in this new Roundtable: the first taking stock of the co-creational dimension to frame the wider issue of service design for social integration; the second reflecting on the sustainability of newly developed ICT or AI enabled tools and services. 3 https://cordis.europa.eu/programme/id/H2020_DT-MIGRATION-06-2018-2019 4 For more information of all six projects funded under the same program see here: https://cordis.europa.eu/search?q=contenttype%3D%27project%27%20AND%20programme%2Fcode%3D%27DT- MIGRATION-06-2018-2019%27andp=1andnum=10andsrt=/project/contentUpdateDate:decreasing 5 https://digital-strategy.ec.europa.eu/en/news/migration-whitepaper-new-approach-digital-services-migrants 5 5 6 6 3 https://cordis.europa.eu/programme/id/H2020_DT-MIGRATION-06-2018-2019 This Handbook aims to be an executive document to feed policy decisions and discussions in relation to two operational questions and some associated issues: ● Can service design contribute to social integration? ● Can service design contribute to social integration? ○ Co-design and convergence (responsible project MICADO) ○ Community-oriented approaches (responsible project NADINE) ○ Co-creation, social design and transformative impacts (responsible project REBUILD ○ Human rights literacy in service design and supply (responsible project easyRights) ○ Privacy and data protection (responsible project MIICT) ○ Personalization of intelligent technologies for TCNs (responsible project WELCOME). ● Digital solutions for social integration: what are the challenges for sustainability? ○ Personalization of intelligent technologies for TCNs (responsible project WELCOME). ● Digital solutions for social integration: what are the challenges for sustainability? ○ Bottlenecks in migrants' integration procedures (responsible project NADINE) ○ Impacts of socio-technical solutions for migrants’ integration (responsible project REBUILD) ○ Developing a viable uptake strategy (responsible project MICADO) ○ Sustainability of the novel ICT solutions (responsible project MIICT) ○ Psychosocial dimensions of AI-supported services (responsible project WELCOME) ○ Assessing service accessibility: towards a new standard (responsible project easyRights). g g ( p p j ) ● Digital solutions for social integration: what are the challenges for sustainability? ○ Bottlenecks in migrants' integration procedures (responsible project NADINE) ○ Impacts of socio-technical solutions for migrants’ integration (responsible project REBUILD) ○ Developing a viable uptake strategy (responsible project MICADO) ○ Sustainability of the novel ICT solutions (responsible project MIICT) ● Digital solutions for social integration: what are the challenges for sustainability? ○ Bottlenecks in migrants' integration procedures (responsible project NADINE) ○ Impacts of socio-technical solutions for migrants’ integration (responsible project REBUILD) ○ Developing a viable uptake strategy (responsible project MICADO) ○ Sustainability of the novel ICT solutions (responsible project MIICT) ○ Psychosocial dimensions of AI-supported services (responsible project WELCOME) ○ Assessing service accessibility: towards a new standard (responsible project easyRights). Key lessons learned on the above issues are collected and interpreted in terms of acknowledged pitfalls and evidence-based policy recommendations representing the “actionable” contents of this Handbook. 6 https://digital-strategy.ec.europa.eu/en/news/migration-whitepaper-new-approach-digital-services-migrants 1.1 Aim of this Handbook The two sessions had the common goal of discussing the aforementioned topics and reporting about the key issues emerged in the six EU projects: a number of findings and reflections are presented in this Handbook in the form of recognized pitfalls and policy recommendations nurtured by the lessons learned throughout the implementation of the six projects. 7 7 1.2 The 2020 Whitepaper In December 2020 the Joint Migration Policy Whitepaper was published6 as a strategic input for improving integration and shaping the future migration policy of the EU. It focused on ICT-enabled solutions, participatory design and co-creation of those solutions, as well as end-to-end collaboration between the various parties involved in migrant integration in European societies. The Whitepaper was largely disseminated among the competent EU and national institutions and researchers in the domain of “migration and migrants’ integration”. It inspired those Institutions and served as a useful source of information for programs and policy papers at the EU level. One of its most important contributions was the fact that it was used as an important tool at an EU level during the preparation of the “Action plan on integration and inclusion” that was published at the end of 2020. The Whitepaper was created to provide to the key stakeholders and the national administrations in the EU Member States, practical input for future national policies. Descriptions of field experiences were effective in widening the perspectives of inclusion policies by public administration and civil society. The six Migration06 projects highlighted how they worked to increase perceived effectiveness, users’ acceptance, and the adoption potential of the services innovated or digitally transformed. The 2020 Whitepaper also provided specific policy recommendations at the end of each session that were meant to pave the way for further discussions on how to address the challenges of migrant integration through ICT-enabled solutions. Emphasis was put on understanding migrants’ lives as a prerequisite for defining the scope of their active participation in co-creation and participatory 8 design sessions, on adopting a wider perspective before addressing specific issues and challen and on recognizing the necessity of co-design for trust-building, removing barriers to participa first, and then pushing participatory design ahead. The crucial benefit of multi-stakeholder collaboration when using ICT to support the integratio migrants was also highlighted. Its value does not only come from being an effective support to public decision-making process but also to the delivery of digital services to migrants. Moreover, importance of reiterated public funding to ensure the long-term sustainability of digital platfo facilitating collaboration among stakeholders was reinforced. Publishing success stories of migr in mainstream media was recommended to ensure smooth and multi-stakeholder integra processes. The last of the three sessions focused on ICT tools and digital services to support integration. 1.2 The 2020 Whitepaper Providing ad education by public authorities and to promote information through ICT” by national laws were the last Finally, the Joint Migration Policy Roundtable 2020 co mainly addressing how the EC could leverage the inno roll-out. These general recommendations can be summ activities of funded projects; 2) work with practitioners i authorities; 4) boost the uptake of novel solutions; 5) pr bring the projects to broad stakeholder attention. 1.2 The 2020 Whitepaper Mig integration policies require data, statistics, transparent and comparable indicators at the regi and local levels. Learning from the challenges in the field, the six projects highlighted the potential for overcoming this problem of implementing ICT solutions to support the delivery of pu services aimed at integration. To boost the digital transformation of public administration an support migrants through innovative ICT tools, the Whitepaper recommended to the EC encourage central governments, local authorities, and other public bodies to guarantee free ac to tools data obviously in coherence with relevant national and EU laws and regulations. Also boost the integration of migrants, it was suggested to host communities to act as major stakehol in future research and innovation activities. Providing adequate financial resources to support di education by public authorities and to promote “interculturalism and multilingual mig information through ICT” by national laws were the last two recommendations in that session. Finally, the Joint Migration Policy Roundtable 2020 concluded with six general recommendat mainly addressing how the EC could leverage the innovations and eventually support their fu roll-out. These general recommendations can be summarized as follows: 1) support the outre activities of funded projects; 2) work with practitioners in the field; 3) help the engagement of pu authorities; 4) boost the uptake of novel solutions; 5) present the migration projects more widely bring the projects to broad stakeholder attention. design sessions, on adopting a wider perspective befo and on recognizing the necessity of co-design for trust first, and then pushing participatory design ahead. The crucial benefit of multi-stakeholder collaboration w migrants was also highlighted. Its value does not only c public decision-making process but also to the delivery o importance of reiterated public funding to ensure the facilitating collaboration among stakeholders was reinfo in mainstream media was recommended to ensure processes. The last of the three sessions focused on ICT tools and d integration policies require data, statistics, transparent and local levels. Learning from the challenges in the potential for overcoming this problem of implementing services aimed at integration. To boost the digital tran support migrants through innovative ICT tools, the encourage central governments, local authorities, and o to tools data obviously in coherence with relevant nat boost the integration of migrants, it was suggested to ho in future research and innovation activities. 7 UN General Assembly Resolution 70/1, Transforming Our World: The 2030 Agenda for Sustainable Development (25 September 2015) UN Doc A/RES/70/1, SDG 1.4, at 15. 8 Hesselman, M., De Wolf, A. H., and Toebes, B. C. (Eds.). (2020). Socio-Economic Human Rights in Essential Public Services Provision. Routledge. 9 https://www.universal-rights.org/urg-policy-reports/human-rights-in-the-digital-age-making-digital- technology-work-for- human-rights/ 2. Service design and social integration 2. Service design and social integration 2.1. Can service design contribute to social integration? design sessions, on adopting a wider perspective before addressing specific issues and challenges, and on recognizing the necessity of co-design for trust-building, removing barriers to participation first, and then pushing participatory design ahead. design sessions, on adopting a wider perspective before addressing specific issues and challenges, and on recognizing the necessity of co-design for trust-building, removing barriers to participation first, and then pushing participatory design ahead. The crucial benefit of multi-stakeholder collaboration when using ICT to support the integration of migrants was also highlighted. Its value does not only come from being an effective support to the public decision-making process but also to the delivery of digital services to migrants. Moreover, the importance of reiterated public funding to ensure the long-term sustainability of digital platforms facilitating collaboration among stakeholders was reinforced. Publishing success stories of migrants in mainstream media was recommended to ensure smooth and multi-stakeholder integration processes. The last of the three sessions focused on ICT tools and digital services to support integration. Migrant integration policies require data, statistics, transparent and comparable indicators at the regional and local levels. Learning from the challenges in the field, the six projects highlighted the high potential for overcoming this problem of implementing ICT solutions to support the delivery of public services aimed at integration. To boost the digital transformation of public administration and to support migrants through innovative ICT tools, the Whitepaper recommended to the EC to encourage central governments, local authorities, and other public bodies to guarantee free access to tools data obviously in coherence with relevant national and EU laws and regulations. Also, to boost the integration of migrants, it was suggested to host communities to act as major stakeholders in future research and innovation activities. Providing adequate financial resources to support digital education by public authorities and to promote “interculturalism and multilingual migrant information through ICT” by national laws were the last two recommendations in that session. Finally, the Joint Migration Policy Roundtable 2020 concluded with six general recommendations mainly addressing how the EC could leverage the innovations and eventually support their future roll-out. These general recommendations can be summarized as follows: 1) support the outreach activities of funded projects; 2) work with practitioners in the field; 3) help the engagement of public authorities; 4) boost the uptake of novel solutions; 5) present the migration projects more widely; 6) bring the projects to broad stakeholder attention. 9 9 10 This chapter has been mainly authored by the partners of the MICADO project funded under the Horizon2020 program with Grant Agreement number 822717. ICT represents only one, although very relevant, dimension of complex service design processes that more and more require the involvement of all relevant actors (migrants, relevant stakeholders including public authorities, researchers, citizens, and local communities, civil society actors, refugee organizations, and migrant associations, research institutions, and IT companies) with the aim to facilitate migrants’ integration and develop solutions to ease becoming EU citizens. Focusing on service design and on the role of ICT for developing better and more effective services for migrants according to their actual needs can be considered one of the integration strategies deployed by single countries, regions, or cities. As many studies and policy documents including the 2020 Whitepaper already highlighted, the adoption of a co-creation approach, and in general, the involvement of local communities and migrants as active participants can play a crucial role in developing solutions for migrants, as it increases the capacity of service providers to cope with the needs of all. Through the co-creative process, various co-design tools and workshops focusing on service design activities, and through combining co-creation and AI technologies, access to public service for migrants can be significantly facilitated. Moreover, designing service is a complex task that can play a relevant role in easing the integration process if it is carried out while being aware of the implications and impacts on migrants. What are the lessons learned from the six Migration 06 projects? What are the key pitfalls and the suggestions we can consider when developing new services or making policies to design and adopt new innovative solutions? 2.1. Can service design contribute to social integration? Access to essential public services in a reliable, affordable, and adequate manner is a key condition for living a life of human dignity and well-being. It also lies at the core of fostering healthy, inclusive, and sustainable societies as highlighted by the United Nations’ Sustainable Development Goals: [...] by 2030 ensure that all men and women, particularly the poor and the vulnerable, have equal rights to economic resources, as well as access to basic services, ownership, and control over land and other forms of property, inheritance, natural resources, appropriate new technology, and financial services including microfinance7. Services as systems fulfilling public needs can be analysed from different perspectives. Among them, essential public services such as water supply, healthcare, or different types of civic or administrative services need to meet certain quality and universality standards to truly serve the whole public and the public interest8. Inequalities related to ethnic and migration backgrounds are evident, such as socioeconomic and spatial segregation (including labour, school, and housing ones), little use of welfare resources compared to autochthonous populations, modest upward mobility processes, deficient recognition of migrant skills and competencies in the hosting societies, discrimination in access to opportunities, concealed racism, etc. Frequently, it is very confusing and difficult for migrants to understand how to “navigate” in a host country, the offered opportunities, or how to overcome obstacles in accessing crucial resources and basic services. Recent advancements in technological development and services digitization represent an opportunity to enable migrants to better understand the context in which they are in. As highlighted in the report at the seventh Glion Human Rights Dialogue (Glion VII), an important benefit of digital technology such as artificial intelligence and algorithmic decision making is to improve public service provision, with positive implications for, inter alia, the right to health and the right to an adequate standard of living9. Furthermore, ICT solutions can also quicken and ease integration processes towards more equal access to services as well as provide flexibilization and adaptation to the needs of migrants that are characterized by an increasing differentiation connected to their diverse backgrounds, skills, patterns, and experiences. Integration goals, at the same time, are challenging ICT development and representing an opportunity toward policy innovation thanks to the experimentation of emerging technological solutions in environments and domains hardly represented in the ICT market demand. 10 An ICT tool that supports migrants’ integration can tackle different domains of the integration process. These domains can include assistance in access to the labour market, information about public health services, guides to education and training opportunities, or provide advice on how to apply for public housing, just to give a few examples. The challenge is hence to decide which part of the integration journey a tool should focus on. By co-creating a tool with migrants and service providers, this decision can be taken to best fit the local needs and existing service infrastructure. The service design needs to be based on the IT requirements that were agreed upon. These requirements must be measurable i.e. they need to have performance indicators that are aligned with the overall goal of the service and its specific functions. A robust service design approach needs to be built on co-creation and convergence activities. While key stakeholders and end-users are involved in the co-creation process to ensure that the services meet stakeholders’ demands, the feedback from users’ needs to be analysed and translated into requirements for software development in subsequent convergence workshops. In these convergence meetings, technical partners can discuss software development strategies with partners who work closely with the target groups and understand their needs. The output of the convergence activity is a plan for a unified, broad-based, adaptive solution that addresses crucial local demands and can be implemented by technical partners in a satisfactory way. There are different methods that can be applied to gather relevant input for service design. One option is to collect input from migrants and service providers and create user stories. User stories and collected requirements such as descriptions of functionalities and features are analysed, assessed, and categorized in terms of their criticality and data availability. In the next step, researchers, technical developers, IT experts, and software programmers developed a concept for a minimal viable product based on user requirement analysis. As a result of that activity, a universal solution with selected functions can be devised that is relevant to all the local needs. The Minimal Viable Product then can be tested in the piloting cities, which allows for further (local) iterations of the general solution tailored to specific demands. Use cases, application scenarios, user stories can be summarized into a logbook for development and uptake. CO-DESIGN AND CONVERGENCE10 CO-DESIGN AND CONVERGENCE10 As a substantial human-centred activity of the development process, service design needs to ensure that a product or a service is developed based on the needs of its users and allows for high-quality interaction between service providers and users. Furthermore, service design needs to take into consideration the adequate use of available resources, such as human resources, infrastructure, communication, and material components of a service. A successful service design approach improves the quality and efficiency of service from the perspective of users and providers. When designing services for migrants’ integration a wide range of stakeholders needs to be considered in the process. This includes migrants, service providers but also any other stakeholder who will be directly or indirectly affected by the service. Key target groups can comprise public administrations (local authorities in charge of migrants' reception), civil society organizations and non-governmental organizations that offer services to migrants, IT providers that manage, maintain, and administer the technical infrastructure used by service organizations, and migrants (asylum seekers and refugees). The knowledge and experience of each target group can feed into the development process to create a minimal viable product or service 11 11 This chapter has been mainly authored by the partners of the NADINE project funded under the Horizon2020 program with Grant Agreement number 822601. methodology needs to be in place to translate stakeholders’ inputs into tangible requirements for IT development. ● Parallel piloting in different locations (cities) demands a well-orchestrated pilot and testing choreography. Pitfalls ● Managing the diversity of languages appears a major issue not only in the practical application but also in the technical development of the ICT tools. ● If local requirements and needs in pilot locations differ in a substantial way, a general (singular) IT solution- or one-size-fits-all solution, might not be feasible Policy Recommendations ● ICT tools that are developed for migrants should consider their language capabilities and provide content in languages they can master. Provide funds for the development of basic technologies and methods that enable the effective translation of contents in ICT systems for migrant integration. ● Support existing language and translation technologies and propagate their application in future funded projects. ● Propagate standardization of migrant information and data in order to support exchange between organizations on local as well as on (cross-)national level. methodology needs to be in place to translate stakeholders’ inputs into tangible requirements for IT development. ● Parallel piloting in different locations (cities) demands a well-orchestrated pilot and testi choreography. ● ICT tools that are developed for migrants should consider their language capabilities and provide content in languages they can master. Provide funds for the development of basic technologies and methods that enable the effective translation of contents in ICT systems for migrant integration. ● Support existing language and translation technologies and propagate their application in future funded projects. ● Propagate standardization of migrant information and data in order to support exchange between organizations on local as well as on (cross-)national level. Collaborative discussions and feedback processes can bring invaluable ideas and suggestions in response to specific local user demands. These processes are particularly relevant for the development of apps that are easy to use. If designed carefully, even a complex tool that helps migrants to navigate local bureaucracies can offer an intuitive user experience. One option is to use guided processes that explain complex administrative processes step-by-step. It has to be noted that while co-creation approaches deliver vital insights into user needs and preferences, not all suggestions by all stakeholders can be implemented. Some functions, such as translation of content into multiple languages, cannot be fully implemented due to a lack of resources or because of technical limitations. Here, the challenge is to provide solutions that are compromising between user demands and technical development, that all partners can live with. Key Lessons Learned ● Involving relevant stakeholder groups from the very beginning of the development is important for successful co-creation processes and paves the way for uptake of services. ● When starting co-creation activities with multiple stakeholders, a robust convergence ● Involving relevant stakeholder groups from the very beginning of the development is important for successful co-creation processes and paves the way for uptake of services. ● When starting co-creation activities with multiple stakeholders, a robust convergence 12 organizations fill in the gaps local government leaves in the process and provide (more personalized) guidance and support. To this end, local communities can determine the quality and effectiveness of the integration process. organizations fill in the gaps local government leaves in the process and provide (more personalized) guidance and support. To this end, local communities can determine the quality and effectiveness of the integration process. The role of migrant communities in the process is often neglected, however, with their feedback, the process can become smoother and ensure long-term integration. This includes asking for and considering their feedback on the integration process and considering their wants and needs, as well as their qualifications. External factors may also determine whether the process will be successful or not. For instance, it has been found that in case a recently arrived migrant is reunited with their family in their host country, the migrant enters the labour market quicker leading to faster employment attachment12. Additionally, through community-building initiatives bringing together the host community and the migrant community, resources can be unlocked, and a stronger community can be built. These initiatives can tackle the most basic needs of migrants, such as finding adequate housing, which fosters the autonomy of migrants, or they can be more niche, such as cultural exchange between the host and migrant communities. As a result of a community-oriented approach, the migrants’ lives, as well as their economic and social standing, will improve. Furthermore, the social connections within local communities will be enhanced which will ensure faster and better integration for more migrants. This approach also creates increased empowerment for migrants and the local communities, resulting in more resilient communities. Besides considering migrants’ feedback on the process, by finding the overlaps and the gaps in the services provided by local authorities and civil society organizations, a less complex integration system can be created that better corresponds to migrants’ needs. Key Lessons Learned ● Considering migrants’ needs, qualifications, and feedback on the integration process can improve and facilitate the process better. ● A multi-stakeholder approach to integration fills in the gaps and, consequently, a simpler process can be achieved. ● Through community-building initiatives, more resilient communities can be built inclusive of the host and migrant communities. ● The integration can be achieved more successfully if the process of integrating is considered as an equal action from both sides, the host and migrant communities. 12 European University Institute (2019), Working Paper on ‘Refugee integration in Europe since the ‘crisis’’, for more information see https://cadmus.eui.eu/handle/1814/62446. COMMUNITY ORIENTED APPROACHES11 An often-neglected aspect of good migrant integration practices is the involvement of local communities and migrants as active participants in the process. While ICT-enabled solutions alleviate some of the burden migrants face when entering and integrating into their host country, becoming a part of the community brings additional stability to migrants’ lives. Displacement leads to disruption in most, if not all, aspects of migrants’ lives. A community-oriented approach takes into account the experiences of migrants, their individual needs and struggles, and tackles these challenges through the provision of personalized support. This approach also highlights the role local communities play in the integration process as well as their responsibility to ensure a smoother process. As a result of this approach, both individuals and communities are mobilized and supported, community structures are re-established, and the individual rights of migrants are protected and respected. The role local communities play in the integration process, as well as the local context itself, determine the integration conditions of migrants. In fact, a multi-stakeholder approach, involving local authorities and civil society organizations, ensures processes that correspond to the needs of migrants. Local authorities are generally the first point of contact for recently arrived migrants. They provide essential services and have a responsibility to guarantee a safe environment and promote long-term integration. On the other hand, complementing the work of local authorities, civil society 13 organizations fill in the gaps local government leaves in the process and provide (more personalized) guidance and support. To this end, local communities can determine the quality and effectiveness of the integration process. ● Societal and cultural context, public opinion, media, and local culture play a significant role in supporting community-based organizations and initiatives. Pitfalls ● Long and complex bureaucratic procedures. ● Overlap between services, leading to the increased administrative burden ● Lack of Institutional readiness in a host community for wide integration and social inclusion. 12 European University Institute (2019), Working Paper on ‘Refugee integration in Europe since the ‘crisis’’, for more information see https://cadmus.eui.eu/handle/1814/62446. Key Lessons Learned ● Considering migrants’ needs, qualifications, and feedback on the integration process can improve and facilitate the process better. ● A multi-stakeholder approach to integration fills in the gaps and, consequently, a simpler process can be achieved. ● Through community-building initiatives, more resilient communities can be built inclusive of the host and migrant communities. ● The integration can be achieved more successfully if the process of integrating is considered as an equal action from both sides, the host and migrant communities. ● Societal and cultural context, public opinion, media, and local culture play a significant role in supporting community-based organizations and initiatives. Pitfalls ● Long and complex bureaucratic procedures. ● Overlap between services, leading to the increased administrative burden ● Lack of Institutional readiness in a host community for wide integration and social inclusion. 12 European University Institute (2019), Working Paper on ‘Refugee integration in Europe since the ‘crisis’’, for more information see https://cadmus.eui.eu/handle/1814/62446. 13 This chapter has been mainly authored by the partners of the REBUILD project funded under the Horizon2020 program with Grant Agreement number 822215. 14 Moretti, M. (2019). Design for migration: Nuove pratiche di design verso una società più inclusiva. Roots-Routes, 9(29), 1- 9. 15 Papanek, V., and Fuller, R. B. (1972). Design for the real world (p. 22). London: Thames and Hudson. Brock, B. (1977). Ästhetik als Vermittlung. Arbeitsbiographien eines Generalisten, hg. von K. Fohrbeck. Köln. Policy Recommendations ● Considering migrants’ feedback on the integration process would allow for the creation of an integration system that addresses its existing issues from the point of view of migrants, allowing for a more robust integration process that focuses on the needs of migrants. ● Considering migrants’ feedback on the integration process would allow for the creation of an integration system that addresses its existing issues from the point of view of migrants, allowing for a more robust integration process that focuses on the needs of migrants. ● There needs to be stronger support for community-building initiatives involving different approaches and activities that bring the host and the migrant communities closer together. These initiatives may range from sharing cultural experiences to providing accommodation to migrants. Community-building practices enhance the individuals’ sense of community, lead to better and more efficient integration, and result in a more resilient community including both the host and migrant communities. ● Synergies between stakeholders must be strengthened to fill in the gaps in the integration process and create a less complex system that is easier to navigate. Such a system, built also with input from the migrants, will alleviate some of the burdens on the migrants, resulting in faster and more sustainable integration. ● The complex division and regulation of competencies between the national government, the states, the regions, and the cities must be considered to reduce the bureaucratic complexity and navigation. ● Enabling local residents to benefit from their interaction with migrants can generate larger opportunities for integration and widen welcoming attitudes and behaviours. Key Lessons Learned ● Considering migrants’ needs, qualifications, and feedback on the integration process can improve and facilitate the process better. 12 European University Institute (2019), Working Paper on ‘Refugee integration in Europe since the ‘crisis’’, for more information see https://cadmus.eui.eu/handle/1814/62446. 14 14 15 Papanek, V., and Fuller, R. B. (1972). Design for the real world (p. 22). London: Thames and Hudson. Brock, B. (1977). Ästhetik als Vermittlung. Arbeitsbiographien eines Generalisten, hg. von K. Fohrbeck. Köln ● Potential conflict and social tensions can arise in the community-led development initiatives. participatory and collective design movement in the past, digital transformation and the service society are on the agenda of today's co-creation practitioners and researchers. participatory and collective design movement in the past, digital transformation and the service society are on the agenda of today's co-creation practitioners and researchers. In the co-creation and co-production definitions theoretical and political issues arise, among which: how does the service work? Who is the owner of knowledge of service provision? Who is accountable for the way the service works? Understandability, ownership, and accountability are key issues of service systems for social integration and can only be designed with political purposes in mind: is co-creation looking for paternalistic rather than fraternalist approaches? Which kind of democratization is co-production really implementing? And of course, what's the ultimate goal of co- creation action? Is it actually meant to be transformative? To be effectively transformative, co-creation lets the users share their own life experiences and actively inform design solutions. This knowledge is ultimately elaborated by designers and technologists into the logic and purpose of AI-based solutions. At the same time the co-creation team might value the other way around: not only to target novel users’ practices by means of design and technology and to imagine how different the migrants’ and refugees’ life might be, but also how interacting with migrants and refugees themselves could foster generative processes towards novel research paradigms and methods. By means of continuous iteration and experimentation with interactive technology both the fine-tuning of the technology and the transformative action of the participatory design happens. Continuous iterations between ideas and concepts, and between mock-ups and prototypes experimentation, allows also to empower literacy-at-large in all the participants, at linguistic, cultural, and technological levels. Literacy at large is conceived as a prerequisite for human contact, social integration, and service development and also represents an objective for removing barriers and integrating diverse user requirements. On the contrary, unbalanced or improper literacy might have drawbacks as weak awareness of duties and rights and thus decreases the possibility for all the participants to co- production to define novel opportunities to innovate social integration. Key Lessons Learned ● Language literacy is the prerequisite for co-creation processes and requires to be investigated with creative co-design solutions such as visual languages and visual codes. CO-CREATION, SOCIAL DESIGN AND TRANSFORMATIVE IMPACTS13 Co-creation is widely investigated in the context of the social function of design that leads to the creation of artifacts capable of acting on the relationships, structures, and dynamics within western societies. These are digital super-objects that open debates, guide reflections, tackle dominant narratives14. In particular, the social design approach supports migrants and refugees in life rebuilding and facilitation of access to public services in the host countries. In this perspective, it is possible to investigate the link between the empowerment of local communities and the design and implementation of AI-based solutions and, consequently, to debate the actual transformative impact of co-creation and participatory design. Social design is focusing on the relation between technology and design and on how to transform work practices since the 50s, trying to challenge the status quo, by means of both novel design of service provision and service fruition strategies15. Though workplaces and computer systems were at the heart of the Scandinavian 15 participatory and collective design movement in the past, digital transformation and the service society are on the agenda of today's co-creation practitioners and researchers. 16 This chapter has been mainly authored by the partners of the easyRights project funded under the Horizon2020 program with Grant Agreement number 870980. 17 Even when formal citizenship rights are acknowledged, discrepancies and inequalities between individuals with and without a migrant origin persist. 18 Vienna Declaration and Programme of Action (1993). Vienna. Retrieved from https://www.ohchr.org/EN/ProfessionalInterest/Pages/Vienna.aspx this could only be done via a transparent proposition and clarity of objectives and values. Pitfalls ● Projects and initiatives focusing on migrants’ integration are often designed and implemented with a silo-like approach, undermining the overall effectiveness of any solution, and their actual mid-and long-term sustainability. Policy Recommendations ● Promote a multi-stakeholder approach involving governments, the private sector, the local community, civil society, public authorities, migrant communities. ● The experimentation cannot be done only with separate groups but should involve the entire actors’ ecosystem. Having all the sectors in society with diverse expectations and needs, solutions for integration might be valuable for scaling up and diversifying accordingly, in order to treat everyone equally. ● Large and extensive experiments would also require observing value, attitudes, and behaviour changes over time. ● Adopt a co-regulation approach by experimenting with a broad range of co-design tools to foster migrant integration and migrants’ active participation in policy design. HUMAN RIGHTS LITERACY IN SERVICE DESIGN AND SUPPLY16 Accessing, understanding, and exercising rights are central factors for successful inclusion processes of migrants in the host societies. In this perspective, at least two different instantiations of the broad concept of rights are relevant: (i) irrespective of the type of migration, residence status, and citizenship, migrants are entitled to the recognition and adherence to basic human rights that are this could only be done via a transparent proposition and clarity of objectives and values. ● Projects and initiatives focusing on migrants’ integration are often designed and implemented with a silo-like approach, undermining the overall effectiveness of any solution, and their actual mid-and long-term sustainability. ● Promote a multi-stakeholder approach involving governments, the private sector, the local community, civil society, public authorities, migrant communities. ● The experimentation cannot be done only with separate groups but should involve the entire actors’ ecosystem. Having all the sectors in society with diverse expectations and needs, solutions for integration might be valuable for scaling up and diversifying accordingly, in order to treat everyone equally. ● Large and extensive experiments would also require observing value, attitudes, and behaviour changes over time. ● Adopt a co-regulation approach by experimenting with a broad range of co-design tools to foster migrant integration and migrants’ active participation in policy design. The adoption of visual grammar based on pictograms or on the mixed text and graphic representations allows even illiterate people to understand the context, get aware of the mission of the team, and exchange their ideas. ● In relation to language literacy, sharing the same conceptual models is fundamental for guaranteeing mutual understanding. ● Guidance and mediation might be used in co-creation approaches. The role of the introductory tutorial, video explanation, and interactive prototype ought to be explored for training all the non-designers participants with co-design methods. ● Co-design allows to collect and identify the perspective of a multitude of stakeholders, to build shared decisions, and to find the way to involve at the same time local populations and migrants, and refugees as novel citizens. ● Furthermore, co-design seeks integration instead of separation and does that by means of cooperation between migrants and local communities. ● Cooperation also means to eliminate the distrust and foster acceptance of the society a 16 because of that, there is a crucial topic at all societal levels: literacy on human rights. This was clear since human rights have been recognized as special and different from others. It is not by chance that the Preamble to the Universal Declaration of Human Rights (UDHR) exhorts "every individual and every organ of society" to "strive by teaching and education to promote respect for these rights and freedoms". The concern was since then that human rights recognition could have a transformative impact on people’s lives and on the building of just and robust democracies. Moreover, human rights must be embedded in social practices and give birth, day by day, to new generative social practices19. Institutions are required to ensure that human rights are respected but also individuals are, in this perspective, to be fully involved in the creation and development of human rights-based culture in order to empower them as citizens. As a matter of fact, human rights literacy is a crucial “citizenship tool”. Knowing them strengths both citizens and institutions. It strengthens the work of organizations that operate in the name and on behalf of the public authorities and it strengthens the ability of individuals to interface with the State at all levels. Although human rights are frequently cited in the day-by-day conversations, they are quite under- researched, also by people that should know them. Students of law and international relations or political science may study human rights in a university setting, but most people, especially those operating on the field and interacting with im(migrants), receive no education, formally or informally, about human rights. Even at the European level, the public administration staff are not sufficiently aware of the impact on human rights of their policies. Human rights literacy has been studied and investigated largely in countries experiencing gross and persistent violations of human rights; they often design human rights education programs and other measures to combat human rights violations 20. A South African research project titled “Human Rights Literacy: A quest for meaning “explored the student-teachers’ conceptualization ontology of human rights and meaning of human rights values; it concluded that “human rights literacy, as understandings of dignity, equality and freedom as both rights (legal claims) and values (normative action) within horizontal and vertical applications, aid rights-based education towards transformative action and an open and democratic society”21 . 19 Ventura, C. A., C. Rubens, M. S. Gutier, I. Mendes (2016). Alternatives for the enforcement of the right to health in Brazil. Nursing Ethics, 23(3): 318–327. Ventura, C. A., I. A. Mendes, S. de Godoy, L. Fumincelli, M. C. Souza and V. D. Souza (2019). Perceptions of brazilian nursing faculty members regarding literacy of human rights related to health in nursing undergraduate programs. BMC International Health and Human Rights, 27: 19-27. https://doi.org/10.1186/s12914-019-0213-7 20 Becker, Anne; Annamagriet de Wet and Willie van Vollenhoven (2015). Human rights literacy: Moving towards rights- based education and transformative action through understandings of dignity, equality and freedom. South African Journal of Education, 35(2): 1-12. Hamel, R.E. (1994). Indigenous education in Latin America: policies and legal framework. In Skutnabb-Kangas, Tove and Robert Philipson (Eds.) Linguistic Human Rights. Overcoming linguistic discrimination, Mounton de Gruyter: 271-288. Roux, C.D. and Du Preez, P. (2015). Human right literacy: a quest for meaning. DOI: 10.13140/RG.2.1.2310.4160 21 Becker, Anne; Annamagriet de Wet and Willie van Vollenhoven (2015). Human rights literacy: Moving towards rights-based education and transformative action through understandings of dignity, equality and freedom. South African Journal of Education, 35(2): 1-12. Quotation from page 1. 21 Becker, Anne; Annamagriet de Wet and Willie van Vollenhoven (2015). Human rights literacy: Moving towards rights-based education and transformative action through understandings of dignity, equality and freedom. South African Journal of Education, 35(2): 1-12. Quotation from page 1. tura, C. A., C. Rubens, M. S. Gutier, I. Mendes (2016). Alternatives for the enforcement of the right to health in Braz ng Ethics, 23(3): 318–327. HUMAN RIGHTS LITERACY IN SERVICE DESIGN AND SUPPLY16 Accessing, understanding, and exercising rights are central factors for successful inclusion processes of migrants in the host societies. In this perspective, at least two different instantiations of the broad concept of rights are relevant: (i) irrespective of the type of migration, residence status, and citizenship, migrants are entitled to the recognition and adherence to basic human rights that are universally acknowledged in all free and democratic countries; (ii) the rights of “new citizens”17- including the right to access welfare services and to exercise civil rights - which are formally granted to regular migrants under the general framework of integration, but depend, in many cases, on the fulfilment of a complex set of rules and conditions. In this “game” human rights literacy plays a crucial role. At the Vienna Conference on Human Rights in 1993, the UN secretary-general stated that human rights constitute a “common language of humanity” (World Conference on Human Rights, Vienna, 1993)18. They are at the very core of what “humanity” profoundly means. Using this metaphor leads us to say that human rights need to be fully known and understood, and not only declared in laws, treaties and regulations in order to effectively permeate our societies, governments, policies, and actions. People have to know a language to use it and the more they use it, the more they learn it, in a continuous positive loop. So, people need to learn the language of human rights to fully use it and 17 because of that, there is a crucial topic at all societal levels: literacy on human rights. This was clear since human rights have been recognized as special and different from others. 22 Regulation (EU) 2021/2303 - formally signed by the co-legislators on 15 December 2021 - concerning the establishment of a European Union Agency for Asylum (EUAA). 23 See: https://www.ohchr.org/EN/Issues/Education/Training/WPHRE/SecondPhase/Pages/SecondphaseEvaluation.aspx Pitfalls Some recurrent pitfalls are to be considered at the level of service design and provision in relation to human rights literacy and to the key scope of assuring human rights respect at all levels of the services creation process. Awareness and empowerment ● Migrants are predominantly weakly aware of being the owner of inalienable human rights; also, when they are fully aware they retract from exercising them as they often feel not Key Lessons Learned Although in 2005 the World Programme for Human Rights Education of the United Nations was launched to tackle the need for specific training on this issue and human rights literacy is considered as a fundamental asset for different subjects, be they individuals or organizations, human rights education programs with a focus on vulnerable groups and public services have been implemented only to a limited extent. In several countries23 training activities on human rights have been addressed exclusively to police forces with national or international competence. However, this kind of training is only compulsory for missions in sensitive areas or are otherwise taken up on a voluntary basis. Although in some cases training plans include personnel dealing with vulnerable groups (children, women, minorities, indigenous peoples, etc.) or being involved in activities challenging the gender perspectives, the absence of quantitative reports on trained staff suggests a weak commitment and implementation in this respect. Of the known measures, only a small proportion of institutions and persons that are actual contact points for migrants, for accessing public services, and for service delivery are covered. Still, it is evident that service provision requires a full awareness of involved human rights as well as of risks associated with denied rights once they are undermined by obstacles in services access. If we consider the plurality of actors that are indirectly and directly relevant for the design of and for the delivery of/access to public services, a complex landscape of actors unfolds consisting of multiple stakeholders from different areas of policy, administration, non-profit organizations, public service sectors and individual networks and communities of migrants. Some recurrent pitfalls are to be considered at the level of service design and provision in relation to human rights literacy and to the key scope of assuring human rights respect at all levels of the services creation process. public policy are people’s empowerment and compliance with international human rights standards. Both aspects are guided by the central element of human rights that is human dignity. In this sense, freedom —in the form of self-determination— is one of the key aspects in the development of the concept of human dignity. Likewise, international standards carry a set of rights that appeal to superior, higher needs of human beings, beyond the satisfaction of very basic needs, i.e., freedom, equality, peace, etc. The very recent EU Regulation 2021/230322, approved after the Policy Roundtable, introduces important obligations for member states as per human rights literacy that already covers some of the policy recommendations provided below and considered relevant to be included at national, regional, and local levels. regional, and local levels. Key Lessons Learned Although in 2005 the World Programme for Human Rights Education of the United Nations was launched to tackle the need for specific training on this issue and human rights literacy is considered as a fundamental asset for different subjects, be they individuals or organizations, human rights education programs with a focus on vulnerable groups and public services have been implemented only to a limited extent. In several countries23 training activities on human rights have been addressed exclusively to police forces with national or international competence. However, this kind of training is only compulsory for missions in sensitive areas or are otherwise taken up on a voluntary basis. Although in some cases training plans include personnel dealing with vulnerable groups (children, women, minorities, indigenous peoples, etc.) or being involved in activities challenging the gender perspectives, the absence of quantitative reports on trained staff suggests a weak commitment and implementation in this respect. Of the known measures, only a small proportion of institutions and persons that are actual contact points for migrants, for accessing public services, and for service delivery are covered. Still, it is evident that service provision requires a full awareness of involved human rights as well as of risks associated with denied rights once they are undermined by obstacles in services access. If we consider the plurality of actors that are indirectly and directly relevant for the design of and for the delivery of/access to public services, a complex landscape of actors unfolds consisting of multiple stakeholders from different areas of policy, administration, non-profit organizations, public service sectors and individual networks and communities of migrants. Human rights emerge from everyday social practices, and their articulation can also establish new and innovative practices. Therefore, the term “human rights literacy” involves not only human rights related to education or specific professional skills, but also the empowerment of migrants, institutions, public officials, civil society and NGOs in their respective communication, bearing in mind the innate human dignity of man. Two of the main characteristics of a human rights-focused 18 Pitfalls Some recurrent pitfalls are to be considered at the level of service design and provision in relation to human rights literacy and to the key scope of assuring human rights respect at all levels of the services creation process. Awareness and empowerment ● Migrants are predominantly weakly aware of being the owner of inalienable human rights; also, when they are fully aware they retract from exercising them as they often feel not 22 Regulation (EU) 2021/2303 - formally signed by the co-legislators on 15 December 2021 - concerning the establishment of a European Union Agency for Asylum (EUAA). 23 See: https://www.ohchr.org/EN/Issues/Education/Training/WPHRE/SecondPhase/Pages/SecondphaseEvaluation.aspx Key Lessons Learned Although in 2005 the World Programme for Human Rights Education of the United Nations was launched to tackle the need for specific training on this issue and human rights literacy is considered as a fundamental asset for different subjects, be they individuals or organizations, human rights education programs with a focus on vulnerable groups and public services have been implemented only to a limited extent. In several countries23 training activities on human rights have been addressed exclusively to police forces with national or international competence. However, this kind of training is only compulsory for missions in sensitive areas or are otherwise taken up on a voluntary basis. Although in some cases training plans include personnel dealing with vulnerable groups (children, women, minorities, indigenous peoples, etc.) or being involved in activities challenging the gender perspectives, the absence of quantitative reports on trained staff suggests a weak commitment and implementation in this respect. Of the known measures, only a small proportion of institutions and persons that are actual contact points for migrants, for accessing public services, and for service delivery are covered. Still, it is evident that service provision requires a full awareness of involved human rights as well as of risks associated with denied rights once they are undermined by obstacles in services access. If we consider the plurality of actors that are indirectly and directly relevant for the design of and for the delivery of/access to public services, a complex landscape of actors unfolds consisting of multiple stakeholders from different areas of policy, administration, non-profit organizations, public service sectors and individual networks and communities of migrants. Education and training ● In public administrations, literacy on human rights is mostly developed in the field, in the sense that operators, whether they are top management or front office staff, had minimum opportunities to participate in training courses on human rights or issues related to migrants. ● Training on the field is very often developed by operators thanks to roles played in public administrations or to immersive voluntary experiences that allowed them to develop a relevant understanding and sensitiveness towards migrants’ problems; no structured training introduces big uncertainty in the service response across Europe especially in relation to those services and procedures that define the way migrants get potentially integrated. ● The lack of systematic training also worsens the already existing conditions of scarce alignment of laws and regulations across the EU deepening migrants’ uncertainties and fragility and exposing them to the risk of basic rights violation. ● Legal implications of critical knowledge on, and understanding of, human rights may be hard to evaluate and consider. Nevertheless, they are crucial and require specialized skills and competencies at very different levels of services supply side with special attention to the roles played by translators and cultural mediators supporting migrants in the interaction with public authorities. ● Training on human rights is important to build safeguarding policies in organizations and must be tailored to different contexts and trainees. Public authorities and private actors frequently don’t have the same educational needs and gaps. ● Quite often local operators don’t master basic language skills to work with migrants and serve them. being in the conditions for rights reclamation. ● Many services operators are fully aware that services do have impacts on human rights and also can list most of the negative implications consequent to the failing of the service journey completion but still, they almost uniquely gain these competencies thanks to every day “fieldwork!” which makes their knowledge and preparedness incomplete and inadequate in any case but especially from the point of view of legal implications. ● Many services operators are fully aware that services do have impacts on human rights and also can list most of the negative implications consequent to the failing of the service journey completion but still, they almost uniquely gain these competencies thanks to every day “fieldwork!” which makes their knowledge and preparedness incomplete and inadequate in any case but especially from the point of view of legal implications. ● Human rights violations often occur because of the discretionary practices, especially of street-level bureaucrats and/or operators involved in providing services to migrants. ● Human rights violations often occur because of the discretionary practices, especially of street-level bureaucrats and/or operators involved in providing services to migrants. ● Local language knowledge (by migrants and professionals dealing with them) can be an obstacle in acknowledging basic rights. ● Local language knowledge (by migrants and professionals dealing with them) can be an obstacle in acknowledging basic rights. ● Migrants frequently must rely on “mediators” (be they professional ones like lawyers or informal ones) to enforce their basic rights. ● Migrants frequently must rely on “mediators” (be they professional ones like lawyers or informal ones) to enforce their basic rights. Awareness and empowerment ● Migrants are predominantly weakly aware of being the owner of inalienable human rights; also, when they are fully aware they retract from exercising them as they often feel not 22 Regulation (EU) 2021/2303 - formally signed by the co-legislators on 15 December 2021 - concerning the establishment of a European Union Agency for Asylum (EUAA). 19 Policy Recommendations Key obstacles for migrants to exercise their rights is the complexity of bureaucratic procedures; this is particularly challenging if migrants do not have any previous knowledge or experience with such administrative systems, and even more so if they have no or limited literacy in the script and language of the host country. Access to welfare services, especially for these groups, often requires face-to-face interactions and a strong, not discretionary, street-level bureaucratic system. As a consequence, improving the knowledge of basic human rights is a precondition for a more inclusive society. Policy measures are necessary to be developed targeting two different, rather complementary goals: i) deepening the human rights literacy of actors involved in service design and provision; ii) embedding human rights-oriented solutions in ICT enabling services, especially those to be used by migrants. ● Service interfaces (digital or not) fail in making available clear services pathways; they also require too many steps/clicks, do not usually have multiple language options including migrants’ mother-tongue language; finally, service usability is not assured to people owning only basic digital skills. ● Service interfaces (digital or not) fail in making available clear services pathways; they also require too many steps/clicks, do not usually have multiple language options including migrants’ mother-tongue language; finally, service usability is not assured to people owning only basic digital skills. ● The language barrier is widely recognized as an impediment to the completion of service journey worsening migrants’ vulnerability and undermining their basic rights; nevertheless, the offer of multiple language functions is rarely considered as a key point of attention in service design and supply at the levels of ICT, documental and staff provision. ● While it is widely shared that ICT solutions can facilitate access to information and increase the transparency and accessibility of the activities of public institutions and governments, they can be misused for disinformation, data falsification and manipulation so implying the necessity to consider that new technologies generate demands for a new generation of human rights that arise the need of specific and innovative research and professional practices. Policy Recommendations Service design and supply ● Long timing in accessing services and completing integration procedures is not clearly understood as a form of infringement of basic human rights by civil servants; this specific topic is, on the contrary, identified as one of the most evident forms of human rights negation by many actors providing services and support to migrants. ● Services are not explicitly ”human-rights oriented” as well as they are not “children” or “gender” oriented. 20 24 This chapter has been mainly authored by the partners of the MIICT project funded under the Horizon 2020 program with Grant Agreement number 822380 25 European Union, Charter of Fundamental Rights of the European Union, 26 October 2012, 2012/C 326/02. Available at: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A12012P%2FTXT. 26 Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation), OJ L 119, 4.5.2016. Available at: https://eur- lex.europa.eu/eli/reg/2016/679/oj. Embedding human rights-oriented solutions in ICT enabled services Embedding human rights-oriented solutions in ICT enabled services ● design ICT-enabled solutions with user-friendly interfaces which inform and reassure migrants about their rights and obligations; ● design ICT-enabled solutions with user-friendly interfaces which inform and reassure migrants about their rights and obligations; ● make ICT interfaces simple and usable by simple basic digital skills; ● make service ICT interfaces available in multiple languages (migrants’ native language included) available in both vocal and textual modes; ● supply through ICT solutions service pathways clear and easy to understand; ● develop a framework for digital rights that encompasses all platform stakeholders, including - but not limited to- designers, users, and shareholders. Finally, policies should generally target specific human rights issues when specifying and fixing service quality standards: ● access and completion of service-related procedures must be short and should not affect integration processes like those related to housing, job search, family creation, schooling, security, or to health provisions; ● services’ pathways have to be understood easily thanks to a minimum provision of clear and accessible information provided before and during the interaction with services themselves; ● homogeneity of service procedures and access modes across EU countries and cities is a way to reduce the discretion of service operators and guarantee equal access to human rights. ● develop an adequate and compulsory training framework on human rights for public servants at national, regional, and local level. Deepening human rights literacy ● make human rights literacy a strategic cultural, educational, and policy goal for the integration of migrants in Europe; ● provide human rights-related training to all operators and organizations acting in origin countries to make migrants capable to exercise their rights; ● guarantee training on human rights for key actors in service provision for them to gain competences on consequences of human rights infringement and abilities to navigate human rights as drivers for solving procedural issues; ● make legal implications related to human rights infringement a focal point of training on human rights; ● reduce the aleatory of competencies developed through fieldwork experience as a unique form of provision of a human rights literacy; ● support the development of human rights-based training with adequate economi resources but also with organizational ones; ● control should be adequately granted of the compliance of the practices of National Public Administrations and of National Law Enforcement Authorities (the Police) to the fundamental rights, as defined in the EU Charter on Fundamental Rights; 21 complied with in every instance of personal data processing, especially in the context of ICT-enabled systems and solutions. Article 6 clarifies the applicable grounds for processing and the rights of the data subjects are covered under Chapter III of the Regulation. Article 25 establishes the principles of data protection by design and by default as overarching and which must be duly considered as early as the planning of the development of ICT-enabled systems. complied with in every instance of personal data processing, especially in the context of ICT-enabled systems and solutions. Article 6 clarifies the applicable grounds for processing and the rights of the data subjects are covered under Chapter III of the Regulation. Article 25 establishes the principles of data protection by design and by default as overarching and which must be duly considered as early as the planning of the development of ICT-enabled systems. It is also important to take the following proposals into account, as they pose important implications for further strengthening privacy and data protection across the EU in the near future: ● The Data Governance Act (DGA)27, introduced as part of the European Data Strategy, proposes to strengthen the data sharing between the data subjects and any users of that data through secure data intermediaries/data-sharing service providers. In addition, through the implementation of additional measures that are intended to ensure full compliance with the GDPR in respect of personal data protection, the proposal aims to establish better control overall for the data subjects. ● The proposal for a Digital Services Act (DSA)28, as part of the package including the Digital Markets Act, aims to improve transparency and accountability for online platforms by introducing due diligence obligations for them. It proposes additional measures to be taken by Very Large Online Platforms, such as conducting risk assessments. Also, the interplay between the DSA and the GDPR would be of key importance for privacy and data protection considerations and, while the DSA would uphold the GDPR principles it also elaborates further on certain aspects, such as clarifying what is perceived as “transparent”. Thus, the DSA may be considered as building on the GDPR even further and introducing even stronger protection for data subjects. PRIVACY AND DATA PROTECTION24 Data protection is a human right. On a European level, it is enshrined in Article 7 (“Respect for private and family life”) and Article 8 (“Protection of personal data”) of the Charter of Fundamental Rights of the European Union25. Furthermore, both privacy and security of personal data are extensively protected by EU Regulation 2016/679, also known as the General Data Protection Regulation (“GDPR”)26. The GDPR has a wide-reaching application and applies in all cases of processing of “personal data'' which it defines as “any information relating to an identified or identifiable natural person” that includes “name, an identification number, location data, an online identifier or to one or more factors specific to the physical, physiological, genetic, mental, economic, cultural or social identity of that natural person” (Article 4(1) of the GDPR). Furthermore, it also covers all EU organizations regardless of whether the data itself is being processed outside of the EU, as well as non-EU organizations that, however, process the personal data of EU citizens (Article 3 of the GDPR). Article 5 covers the principles of processing which are the cornerstones that must be 22 Data Protection Impact Assessment The carrying out of a Data Protection Impact Assessment (DPIA) when developing ICT environments to be used by migrants appears especially useful in terms of monitoring whether the developed platform or tool is compliant with all data protection considerations. Also, ensuring that the DPIA is observed along lite and updated in terms of the risks encountered and corresponding measures to mitigate/eliminate the risks that are in place, is considered a good practice. The regular provision of an updated DPIA makes sure that any new challenges encountered from the technology developed are handled in a timely manner. Key Lessons learned32 The key lessons learned from a legal and ethical point of view focus on three interconnected dimensions while developing and implementing ICT-enabled systems, as well as their testing and use by human participants. These include legal, human rights, and data protection considerations respectively, and must be duly incorporated at each step of the development process. Such a comprehensive and all-encompassing approach was found to work quite well in ensuring that all safeguards are put properly in place to provide adequate and sufficient protection of the data subjects’ rights, especially in cases where a vulnerable group of people, such as migrants, is concerned. In terms of data protection33 more specifically, the following aspects were focused upon: Importance of implementing data protection by default and by design; ● Following the six principles of the GDPR, namely that: ○ the data processing must be on a lawful basis, fair and transparent; ○ purpose limitation of the processing must be observed; ○ data minimization must be ensured; ○ the processing must be accurate; ○ storage limitation must be defined; ○ integrity and confidentiality must be followed; ○ the data controller and data processor must have accountability in respect to all principles listed above. ○ the data controller and data processor must have accountability in respect to all principles listed above. Data Protection Impact Assessment 32 Similar challenges regarding ethics, privacy and data protection are equally experienced when creating digital tools. One example of how these challenges were overcome is provided below from the MIICT project. This can provide a blueprint for future initiatives wishing to embark on research using co-creative methods to create ICT solutions to migrant integration. 33 These data protection protocols are from the IMMERSE platform created in the MIICT project. They can serve as a blueprint for other researchers working in the area to ensure that full data protection considerations are put in place when creating ICTs to aid migrant integration. communications through introducing clearly defined rules in this respect and thus complementing the provisions of the GDPR. Such a comprehensive and all-encompassing approach was found to work quite well in ensuring that all safeguards are put properly in place to provide adequate and sufficient protection of the data subjects’ rights, especially in cases where a vulnerable group of people, such as migrants, is concerned. In terms of data protection33 more specifically, the following aspects were focused upon: ● Importance of implementing data protection by default and by design; ● Following the six principles of the GDPR, namely that: ○ the data processing must be on a lawful basis, fair and transparent; ○ purpose limitation of the processing must be observed; ○ data minimization must be ensured; ○ the processing must be accurate; ○ storage limitation must be defined; ○ integrity and confidentiality must be followed; ○ the data controller and data processor must have accountability in respect to all principles listed above. Data Protection Impact Assessment The carrying out of a Data Protection Impact Assessment (DPIA) when developing ICT environments to be used by migrants appears especially useful in terms of monitoring whether the developed platform or tool is compliant with all data protection considerations. Also, ensuring that the DPIA is observed along lite and updated in terms of the risks encountered and corresponding measures to mitigate/eliminate the risks that are in place, is considered a good practice. The regular provision of an updated DPIA makes sure that any new challenges encountered from the technology developed are handled in a timely manner. Joint Collaboration and Cooperation Strict collaboration and cooperation among all the actors involved in the ICT development also including those whose having a role in the ICT product exploitation is valuable, in order to develop an overall picture for all implications that the Platform has. This requires working closely and l l ith th d l th t ll t i l ti t th t h i l d l t d ● Another pending proposal, the Artificial Intelligence Act (AIA)29, envisages a complete ban on certain AI systems which manipulate vulnerable individuals, are used for social scoring or run real-time remote biometric identification systems in publicly accessible spaces for law enforcement. It also defines and provides for the regulation of high-risk AI systems and also lays down strengthened transparency requirements. ● Last but not least, the proposal for an EU ePrivacy Regulation30 which will repeal the current e-Privacy Directive31, proposes to enhance the privacy protection of electronic ● Last but not least, the proposal for an EU ePrivacy Regulation30 which will repeal the current e-Privacy Directive31, proposes to enhance the privacy protection of electronic 27 Proposal for a Regulation of the European Parliament and of the Council on European data governance (Data Governance Act), COM/2020/767 final. Available at: https://eur-lex europa eu/legal-content/EN/TXT/?uri=CELEX%3A52020PC0767 28 Proposal for a Regulation of the European Parliament and of the Council on a Single Market For Digital Services (Digital Services Act) and amending Directive 2000/31/EC, COM/2020/825 final. Available at: https://eur-lex.europa.eu/legal- content/en/TXT/?uri=COM%3A2020%3A825%3AFIN 29 Proposal for a Regulation of the European Parliament and of the Council laying down harmonized rules on artificial intelligence (artificial intelligence act) and amending certain union legislative acts, COM/2021/206 final. Available at: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52021PC0206 . 30 Proposal for a Regulation of the European Parliament and of the Council concerning the respect for private life and the protection of personal data in electronic communications and repealing Directive 2002/58/EC (Regulation on Privacy and Electronic Communications) COM/2017/010 final - 2017/03 (COD). Available at: https://eur-lex.europa.eu/legal- content/EN/TXT/?uri=CELEX%3A52017PC0010. 31 Directive 2002/58/EC of the European Parliament and of the Council of 12 July 2002 concerning the processing of personal data and the protection of privacy in the electronic communications sector (Directive on privacy and electronic communications), OJ L 201, 31.7.2002, p. 37–47. Available at: https://eur-lex.europa.eu/legal- content/EN/ALL/?uri=celex%3A32002L0058. 23 communications through introducing clearly defined rules in this respect and thus complementing the provisions of the GDPR. Key Lessons learned32 The key lessons learned from a legal and ethical point of view focus on three interconnected dimensions while developing and implementing ICT-enabled systems, as well as their testing and use by human participants. These include legal, human rights, and data protection considerations respectively, and must be duly incorporated at each step of the development process. Common pitfalls in terms of the development of ICT-enabled systems include: ● Not having sufficient measures in place to ensure that the required express and informed consent has been obtained from the data subject prior to the data processing. ● Reactive and not proactive actions by the data controllers in terms of existing risks to the personal data of data subjects. Data controllers must anticipate risks before they take place and not wait until they have already materialized. ● Services offered based on various individual characteristics, such as age, gender, and nationality which may increase the risk of discrimination. ● Excessive processing of personal data that is not necessary to be provided for the purposes of using the ICT-enabled system and the services offered through it. ● The personal data provided is used beyond the stated means and purposes and the technical and organizational measures in place to ensure the security of the data are not reviewed regularly as the external threat environment and technology changes. ● Relevant information is not adequately communicated to the data subjects due to language difficulties which affect the right to information. ● The “once only” principle for the processing of personal data by public bodies is not adhered to. The recommendations below have been developed from a legal and ethical standpoint to ensure that a full data protection framework is built in ICT-enabled systems for migrants' integration: The recommendations below have been developed from a legal and ethical standpoint to ensure that a full data protection framework is built in ICT-enabled systems for migrants' integration: ● ICT-enabled systems must process data without causing any negative effects to the users ● Only the data that is necessary for the provision of the service must be collected, following a pre-defined scenario and applying the “once only” principle where data must be collected only once at the different stages and where such information is relevant. The same information is not to be asked again when it comes to the provision of another service requiring the same or parts of already input information. ● Data controllers must periodically revise the collected information in order to keep it up to date. ● Users must be informed about their right to withdraw consent through ICT-enabled systems and must have the opportunity to exercise that right through it. The same information is not to be asked again when it comes to the provision of another service requiring the same or parts of already input information. ● Data controllers must periodically revise the collected information in order to keep it up to date. ● Users must be informed about their right to withdraw consent through ICT-enabled systems and must have the opportunity to exercise that right through it transparent, as well as with the actors interacting with migrants in the pilot countries so that the practical perspective is more easily captured, understood, and duly taken into account. Indeed, such a joint effort provides for a comprehensive data protection framework to be developed and also for the bridging of any existing gaps in expertise between legal and technical experts–that is to say, from a legal point of view to understand the risks and considerations in terms of technical aspects and from a technical point of view to then adequately implement these considerations into practice. transparent, as well as with the actors interacting with migrants in the pilot countries so that the practical perspective is more easily captured, understood, and duly taken into account. Indeed, such a joint effort provides for a comprehensive data protection framework to be developed and transparent, as well as with the actors interacting with migrants in the pilot countries so that the practical perspective is more easily captured, understood, and duly taken into account. Indeed, such a joint effort provides for a comprehensive data protection framework to be developed and also for the bridging of any existing gaps in expertise between legal and technical experts–that is to say, from a legal point of view to understand the risks and considerations in terms of technical aspects and from a technical point of view to then adequately implement these considerations into practice. such a joint effort provides for a comprehensive data protection framework to be developed and also for the bridging of any existing gaps in expertise between legal and technical experts–that is to say, from a legal point of view to understand the risks and considerations in terms of technical aspects and from a technical point of view to then adequately implement these considerations into practice. Joint Collaboration and Cooperation Strict collaboration and cooperation among all the actors involved in the ICT development also including those whose having a role in the ICT product exploitation is valuable, in order to develop an overall picture for all implications that the Platform has. This requires working closely and regularly with the developers so that all aspects in relation to the technical development are made 32 Similar challenges regarding ethics, privacy and data protection are equally experienced when creating digital tools. One example of how these challenges were overcome is provided below from the MIICT project. This can provide a blueprint for future initiatives wishing to embark on research using co-creative methods to create ICT solutions to migrant integration. 33 These data protection protocols are from the IMMERSE platform created in the MIICT project. They can serve as a blueprint for other researchers working in the area to ensure that full data protection considerations are put in place when creating ICTs to aid migrant integration. 24 Policy Recommendations transparent, as well as with the actors interacting with migrants in the pilot countries so that the practical perspective is more easily captured, understood, and duly taken into account. Indeed, such a joint effort provides for a comprehensive data protection framework to be developed and also for the bridging of any existing gaps in expertise between legal and technical experts–that is to say, from a legal point of view to understand the risks and considerations in terms of technical aspects and from a technical point of view to then adequately implement these considerations into practice. Pitfalls Common pitfalls in terms of the development of ICT-enabled systems include: ● Not having sufficient measures in place to ensure that the required express and informed consent has been obtained from the data subject prior to the data processing. ● Reactive and not proactive actions by the data controllers in terms of existing risks to the personal data of data subjects. Data controllers must anticipate risks before they take place and not wait until they have already materialized. ● Services offered based on various individual characteristics, such as age, gender, and nationality which may increase the risk of discrimination. ● Excessive processing of personal data that is not necessary to be provided for the purposes of using the ICT-enabled system and the services offered through it. ● The personal data provided is used beyond the stated means and purposes and the technical and organizational measures in place to ensure the security of the data are not reviewed regularly as the external threat environment and technology changes. ● Relevant information is not adequately communicated to the data subjects due to language difficulties which affect the right to information. ● The “once only” principle for the processing of personal data by public bodies is not adhered to. Policy Recommendations The recommendations below have been developed from a legal and ethical standpoint to ensure that a full data protection framework is built in ICT-enabled systems for migrants' integration: ● ICT-enabled systems must process data without causing any negative effects to the users. ● Only the data that is necessary for the provision of the service must be collected, following a pre-defined scenario and applying the “once only” principle where data must be collected only once at the different stages and where such information is relevant. ● The controller shall consult the supervisory authority prior to processing when the processing would result in a high risk for the rights and freedoms of users. ● Data controllers must appoint Data Protection Officers (DPO). ● Data controllers must appoint Data Protection Officers (DPO). ● Users must be informed through the system of the processing activities taking place. ● Users must be informed through the system of the processing activities taking place. ● ICT-enabled systems must provide information to migrants about their rights in the processing of personal data and the ways to exercise them, following a layered approach. ● ICT-enabled systems must provide information to migrants about their rights in the processing of personal data and the ways to exercise them, following a layered approach. ● The system must inform users for how long the personal data will be stored or at least the criteria of how this period is determined. ● The system must inform users for how long the personal data will be stored or at least the criteria of how this period is determined. ● The presented information through the system must be free of charge. ● All information must be presented through the system in a clear and understandable wa ● The ICT-enabled systems should enable their users to exercise their rights as data subje through the system. ● All collected data must go through pseudonymization and encryption processes for bet data security. ● The system must ensure the availability, confidentiality, integrity of the process information. ● In cases of data breaches, the system must be able to restore access to personal data. ● Establishment of data protection policy and taking steps to make sure that the policy is implemented. ● Regular security checks to be carried out to evaluate whether the taken technical and organizational measures are up to date. ● The ICT-enabled system needs to be interoperable with relevant e-governance syste and eID schemes. ● When the provided services include automated decision-making or profiling the system must provide information to the users about the nature and method of that processing. 34 This chapter has been mainly authored by the partners of the WELCOME project funded under the Horizon2020 program with Grant Agreement number 870930. ● ICT-enabled systems must automatically inform data controllers, migrants, and d protection authorities about cases of data breaches without delay. ● Data controllers should conduct data protection impact assessments whenever a new service is planned to be integrated into the ICT-enabled system. 25 ● Support targeted Natural Language Processing (NLP) applications that cover the analysis and synthesis of languages spoken by TCNs (it is not enough to focus on machine translation as it is not enough to focus on selected majority languages). ● TCNs do not form one homogeneous group; they have diverse educational, cultural, and social backgrounds. ● TCNs do not form one homogeneous group; they have diverse educational, cultural, and social backgrounds. ● Each TCN has its own communication skills, its own unique experiences, its own needs and values, and its individual capacities and capabilities. ● If we furthermore take into account that TCNs, as a rule, do not speak well the language of the host country and are not familiar with its social and cultural norms, let alone the bureaucratic procedures, the need for personalized assistance of each individual becomes obvious. Given that personalized assistance of TCNs by authorities and civil society sector is heavily conditioned by the availability of human and financial resources, intelligent technologies in terms of conversational agents (CAs) that assume the role of a personal assistant can be designed and provide support to the public administrations and civil society in the whole process. Common pitfalls can be observed on both extremes of the personalization of the TCN-targeted services: Common pitfalls can be observed on both extremes of the personalization of the TCN-targeted services: ● Intent to acquire and process sensitive personal data (such as, e.g., health conditions) of the TCNs. This might result in optimized assistance, but implies significant critical issues related to trust, data protection, and also technical complexity. ● Intent to generalize the services across such features as age, gender, country of origin, etc. This simplifies the technical realization, but does not meet the personalization needs of the TCNs. ● Intent to cover too many services and thus be too ambitious for the given time and financial frame (which will necessarily lead to superficial support in all of them). ● Support empirical studies that identify the concrete needs of TCNs during their reception and integration in the countries of the European Union and align these needs with concrete ICT-based services. ● Support the development of service-specific CA solutions (i.e., coaching of TCNs how to prepare and how to act during a job interview); they will be more beneficial than generic solutions. ERSONALIZATION OF INTELLIGENT TECHNOLOGIES FOR THIRD COUNTRY NATIONALS34 Third Country Nationals (TCNs) who arrive in Europe are required to go through complex registration procedures, have to learn the language of the host country, find employment, etc. Very often, they are overburdened and need more background information, some guidance (or coaching) and training. However, it must be taken into account that the needs depend on the educational, cultural, and social background of each TCN. Therefore, personalized support is needed. On the other hand, over the last decade, significant advances have been achieved with respect to intelligent ICT-based technologies, and, more notably, with respect to personal assistants or conversational agents. This calls for the use and further development of these technologies in the context of TCN reception and integration in the EU. ● Third Country Nationals (TCNs) are expected to complete a number of multiple-sta reception and integration procedures, in which some of the stages presuppose t 34 This chapter has been mainly authored by the partners of the WELCOME project funded under the Horizon2020 program with Grant Agreement number 870930. 26 successful completion of the preceding stages; often, the documentation they have to deal with is more complex (and specific) than documentation migrants moving from one European country to another have to handle; successful completion of the preceding stages; often, the documentation they have to deal with is more complex (and specific) than documentation migrants moving from one European country to another have to handle; ● TCNs do not form one homogeneous group; they have diverse educational, cultural, and social backgrounds. ● Each TCN has its own communication skills, its own unique experiences, its own needs and values, and its individual capacities and capabilities. ● If we furthermore take into account that TCNs, as a rule, do not speak well the language of the host country and are not familiar with its social and cultural norms, let alone the bureaucratic procedures, the need for personalized assistance of each individual becomes obvious. Given that personalized assistance of TCNs by authorities and civil society sector is heavily conditioned by the availability of human and financial resources, intelligent technologies in terms of conversational agents (CAs) that assume the role of a personal assistant can be designed and provide support to the public administrations and civil society in the whole process. Pitfalls Common pitfalls can be observed on both extremes of the personalization of the TCN-targeted services: ● Intent to acquire and process sensitive personal data (such as, e.g., health conditions) of the TCNs. This might result in optimized assistance, but implies significant critical issues related to trust, data protection, and also technical complexity. ● Intent to generalize the services across such features as age, gender, country of origin, etc. This simplifies the technical realization, but does not meet the personalization needs of the TCNs. ● Intent to cover too many services and thus be too ambitious for the given time and financial frame (which will necessarily lead to superficial support in all of them). Policy Recommendations ● Support empirical studies that identify the concrete needs of TCNs during their reception and integration in the countries of the European Union and align these needs with concrete ICT-based services. ● Support the development of service-specific CA solutions (i.e., coaching of TCNs how to prepare and how to act during a job interview); they will be more beneficial than generic solutions. 1. Digital solutions for social integration: what are the challenges for sustainabilit The sustainability of ICT tools that facilitate migrants’ integration is inherently dependent on a multitude factors. These are both positive and negative, ranging from the design of the tool and the information and services provided therein, to the acceptance and uptake in the communities where such tools are introduced and implemented. Sustainability is also linked to the national, regional, and local context. Delivering new forms of innovation for facilitating migrants’ integration and strengthening the social integration through digital solutions carries out different challenges and poses several questions such as how to ensure access to innovation outcomes to end-users as well as the usability of the solution for migrants and also for the service providers. How to assess the impacts of socio-technical solutions for migrants’ integration and guarantee the inclusion of privacy, legal and technical considerations as a way to foster sustainability? Is the technology involved compatible and easy to integrate with the existing service components for example municipality ecosystems? And eventually to what extent does the solution address the issue of long-term sustainability? This session presents interim findings from work in progress at the six Migration 06 projects and discusses policy-related issues on sustainability and the adoption of solutions and tools of social integration. As a result of the implementation of inclusion policy, ICT-enabled solutions and toolkits may have elevated the task of public administration, local authorities to deliver and manage better and customized services. Facilitating communication and access to services through ICT solutions helps improve migrants’ autonomy and inclusion. Moreover, it helps to define efficient integration tools and implementation at the local level for sustainable inclusion of migrants and effective public perception of migration. The contribution of the six projects on this topic demonstrates the reusability or scalability of developed solutions at the European level and their potential to ensure long-term sustainability as well as take-up by identified users. This section lays out a number of key lessons learned on how to approach sustainability of digital solutions and enumerates challenges and pitfalls to overcome when managing migration, welcoming, orientation, and integration using technological means from the different angles of social integration and challenges for institutional sustainability. ● Support targeted Natural Language Processing (NLP) applications that cover the analysis and synthesis of languages spoken by TCNs (it is not enough to focus on machine translation as it is not enough to focus on selected majority languages). ● Support targeted Natural Language Processing (NLP) applications that cover the analysis and synthesis of languages spoken by TCNs (it is not enough to focus on machine translation as it is not enough to focus on selected majority languages). 27 27 3. Sustainability of solutions and social integration 3.1. Digital solutions for social integration: what are the challenges for sustainability? BOTTLENECKS IN MIGRANTS’ INTEGRATION PROCEDURES35 Migrants’ integration has been and still is a challenge across the EU. This is proved by the EU-level documents, which attempted to harmonize the various approaches followed and offer guidelines in the integration process, and often remained unenforced for very long time36. The crucial stages of integration policy enforcement can be identified as the arrival and application for asylum status, the verification of information, interviews and decisions, access to housing, and labour market integration. The stages, highlighted here for their strategic role in the integration process, are complemented by other steps, most of which are interconnected in many ways. The approaches to identified key stages of integration differ across the Member States, creating various bottlenecks and heterogeneities across the EU. Nevertheless, the integration process should be considered as a whole as opposed to standalone stages or approaches in order to improve and enhance the efficiency of integration. Decision and appeal Evaluating asylum applications is a national competence, although guidelines are available at the EU level. As a result, national legislation prescribes the maximum length of the process, however, delays occur quite often, especially when large amounts of applications are filed at the same time. Although a short procedure may seem appealing and cost-efficient, the quality of the process should not be neglected, as a higher quality process reduces the number of appeals filed after a decision has been made. The time needed for reviewing the appeals and the specificities of the process are a national competence as well. The evaluation of appeals often lasts longer in the Member States where judicial authorities are in charge of the procedure. The delays applicants face during the decision and appeal (if applicable) stages negatively affect their prospects in other key aspects of the integration process, including access to housing and to the labour market. Arrival and application for the asylum status According to the Asylum Procedures Directive37, the application process should be completed within six months, or nine months in specific cases, following the submission of the application. However, in practice, delays are common in many Member States. For instance, in some countries, the average waiting period to even get their status registered can last as long as 12 months. Delays can be due to a lack of harmonization across the EU regarding the conditions that asylum seekers must fulfil in order to gain access to the labour market, a lack of administrative capacity to process large amounts of applications, and an increased administrative burden on the employers hiring asylum applicants. 3.2. From lessons learned to acknowledged pitfalls and policy recommendations 3.2. From lessons learned to acknowledged pitfalls and policy recommendations 35 This chapter has been mainly authored by the partners of the NADINE project funded under the Horizon2020 program with Grant Agreement number 822601. 36 See as examples: the‘Common European Asylum System’ https://ec.europa.eu/home-affairs/what-we- do/policies/asylum_en European Commission, (2020) as well as the ‘Action plan on Integration and Inclusion 2021-2027’. https://ec.europa.eu/migrant-integration/news/the-ec-presents-its-eu-action-plan-on-integration-and-inclusion-2021-2027 37 Directive 2013/32/EU of the European Parliament and of the Council of 26 June 2013 on common procedures for granting and withdrawing international protection. Available at: https://eur-lex.europa.eu/legal- content/en/ALL/?uri=celex%3A32013L0032 1. Digital solutions for social integration: what are the challenges for sustainabilit Detected problems include surviving bottlenecks in migrants' integration procedures, insufficient analysis of the impacts of migrants’ integration (and of the costs of non-integration), limited consideration of the psychosocial dimensions of migrant profiles during the development of ICT or AI-supported integration services, and the lack of an assessment standard for service accessibility. We also propose a number of policy recommendations that may assist in overcoming some of those hindering factors as well as improving the known favourable conditions to a wider uptake of the novel ICT solutions. 28 35 This chapter has been mainly authored by the partners of the NADINE project funded under the Horizon2020 program with Grant Agreement number 822601. housing options and overflowing reception centres, migrants often face inadequate living conditions. The bottlenecks related to housing also include a shortage of affordable housing in the area and the limited budget capacities of cities. Labour market Gaining access to the labour market is a decisive factor in migrant integration, as it is connected to other aspects of integration, such as access to language training, housing, etc. The shorter the waiting period, the better will be the economic integration of migrants. On the other hand, longer waiting periods negatively affect migrants’ motivations, and a lack of long-term permits negatively affects the prospects of finding a stable job as well. Another aspect connected to access to the labour market is the insufficient capability of speaking the local language, which is also connected to accessing social networks and building social capital. Although the bottlenecks connected to the different aspects of the integration process are varied, ICT tools can mitigate these issues, for instance by providing comprehensive information in a user- friendly format in many languages on what documents are needed, or providing information on migrants’ rights, and giving them quick access to local networks to enable social integration. Key Lessons Learned ● Approaches to migrant integration are different across the EU. For instance, there is a lack of harmonization on the conditions asylum seekers must fulfil to gain access to the labour market. ● The main bottlenecks connected with each key aspect of the integration process are varied and can differ between the Member States. Pitfalls ● Delays in key aspects of integration, including registration as asylum seekers and gaining access to the labour market. ● Long administrative delays negatively affect the integration of migrants and their prospects. ● Lack of access to information for migrants. ● Difficulties and conflicts may arise in the process of integration. Policy Recommendations ● Through exploring digital solutions, multiple bottlenecks can be alleviated or even eliminated. ICT-enabled solutions, complemented by additional changes and more harmonized approaches across the EU could, for instance, shorten administrative delays which is one of the biggest hindrances in migrant integration. ● The lack of availability and accessibility of information is an obstacle for most migrants, however, ICT tools could provide more accessible information. This would reduce some of the burdens on migrants and lead the way to a smoother integration process. Key Lessons Learned ● Approaches to migrant integration are different across the EU. For instance, there is a lack of harmonization on the conditions asylum seekers must fulfil to gain access to the labour market. Housing Housing is another key aspect of the integration process, upon which other steps in the integration process rely, such as access to education, employment, and healthcare. However, due to limited Housing is another key aspect of the integration process, upon which other steps in the integration process rely, such as access to education, employment, and healthcare. However, due to limited 35 This chapter has been mainly authored by the partners of the NADINE project funded under the Horizon2020 program with Grant Agreement number 822601. 29 29 housing options and overflowing reception centres, migrants often face inadequate living conditions. The bottlenecks related to housing also include a shortage of affordable housing in the area and the limited budget capacities of cities. IMPACTS OF SOCIO-TECHNICAL SOLUTIONS FOR MIGRANTS’ INTEGRATION38 Assessing the impact of socio-technical solutions such as those developed by the six projects present at the roundtable can be of great help in fostering outputs’ sustainability. Indeed, by showing in qualitative and quantitative terms the benefits of the solutions developed, the chance to find further (financial) support for their improvement and upscaling increases considerably. Certainly, assessing the impact of socio-technical innovations in the field of migrant inclusion also represents a challenge as no standardized approach exists yet. Talking about impact means answering questions like “What is the difference a project makes?” and “For whom?”. Impact assessment is a way to assure transparency and increase accountability but is also a way of increasing reflexivity within a project or within an organization: it is a tool for learning and improving development activities and outputs. More importantly in this context, impact assessment can also be a tool helping innovators better communicate the benefits of the solutions developed to potential investors, up-takers, and other stakeholders. The areas considered while assessing the impact of socio-technical solutions for migrant integration are the social, economic, political, and technological ones. Each dimension is articulated in several sub-dimensions which include, but are not limited to: impact on education and human capital; impact on health; impact on social capital; impact on employment and economic empowerment; impact on service efficiency; impact on digitalization and technology usage; impact on trust in local institutions and impact on policies. In order to map all the potential impacts of socio-technical projects and take into consideration all outputs, it is important to consider the impact on different stakeholders, mainly: migrants and refugees, Public Administrations (PAs), and other actors acting as Local Service Providers (LSPs) and the whole society. Clearly, the final expected impact of socio-technical projects, as those presented at the Policy Roundtable this document is the output of, is the inclusion of migrants in the host society, being this a multifaceted and still debated concept. They are expected to increase migrants’ integration by improving access to and increasing the effectiveness of already-existing services. 38 This chapter has been mainly authored by the partners of the REBUILD project funded under the Horizon2020 program with Grant Agreement number 822215. Labour market ● The main bottlenecks connected with each key aspect of the integration process are varied and can differ between the Member States. Pitfalls ● Delays in key aspects of integration, including registration as asylum seekers and gaining access to the labour market. ● Long administrative delays negatively affect the integration of migrants and their prospects. ● Lack of access to information for migrants. ● Difficulties and conflicts may arise in the process of integration. Policy Recommendations ● Through exploring digital solutions, multiple bottlenecks can be alleviated or even eliminated. ICT-enabled solutions, complemented by additional changes and more harmonized approaches across the EU could, for instance, shorten administrative delays which is one of the biggest hindrances in migrant integration. ● The lack of availability and accessibility of information is an obstacle for most migrants, however, ICT tools could provide more accessible information. This would reduce some of the burdens on migrants and lead the way to a smoother integration process. ● Approaches to migrant integration are different across the EU. For instance, there is a lack of harmonization on the conditions asylum seekers must fulfil to gain access to the labour market. ● Difficulties and conflicts may arise in the process of integration. ● Through exploring digital solutions, multiple bottlenecks can be alleviated or even eliminated. ICT-enabled solutions, complemented by additional changes and more harmonized approaches across the EU could, for instance, shorten administrative delays which is one of the biggest hindrances in migrant integration. ● The lack of availability and accessibility of information is an obstacle for most migrants, however, ICT tools could provide more accessible information. This would reduce some of the burdens on migrants and lead the way to a smoother integration process. ● The lack of availability and accessibility of information is an obstacle for most migrants, however, ICT tools could provide more accessible information. This would reduce some of the burdens on migrants and lead the way to a smoother integration process. 30 30 GRANTS’ INTEGRATION38 s those developed by the six utputs’ sustainability. Indeed olutions developed, the chanc g increases considerably. Cer migrant inclusion also repres What is the difference a proje ansparency and increase acco within an organization: it is a More importantly in this er communicate the benefits takeholders. io-technical solutions for mig nes. Each dimension is articu o: impact on education and employment and economic nd technology usage; impact ● Mixing qualitative and quantitative methods in assessing impacts helps in better communicating to different stakeholders and potential up-takers: for some the economic aspect is crucial while for others more soft, social, and story-telling-based descriptions are better welcomed. Pitfalls ● It is still difficult to find comparable data on migrant integration across EU countries. Moreover, within the same country, regional and local data are often missing or not comparable. The lack of standardized approaches in monitoring the integration phenomenon lead also to the lack of historical series of data that are crucial in mapping the impact of policies and investments. ● Often, when data are available, they are not provided as open and FAIR data, this is a shortcoming in terms of re-use of available data and research progress. Key Lessons Learned ● Considering different stakeholders, in different countries when assessing the impacts of socio-technical solutions for migrant integration is really important as the differences among contexts are high, and customized solutions emerged as the most valuable in terms of adoption and long-term sustainability. ● Mixing qualitative and quantitative methods in assessing impacts helps in better communicating to different stakeholders and potential up-takers: for some the economic aspect is crucial while for others more soft, social, and story-telling-based descriptions are better welcomed. Pitfalls ● It is still difficult to find comparable data on migrant integration across EU countries. Moreover, within the same country, regional and local data are often missing or not comparable. The lack of standardized approaches in monitoring the integration phenomenon lead also to the lack of historical series of data that are crucial in mapping the impact of policies and investments. ● Often, when data are available, they are not provided as open and FAIR data, this is a shortcoming in terms of re-use of available data and research progress. ● Considering different stakeholders, in different countries when assessing the impacts of socio-technical solutions for migrant integration is really important as the differences among contexts are high, and customized solutions emerged as the most valuable in terms of adoption and long-term sustainability. ● Mixing qualitative and quantitative methods in assessing impacts helps in better communicating to different stakeholders and potential up-takers: for some the economic aspect is crucial while for others more soft, social, and story-telling-based descriptions are better welcomed. ● It is still difficult to find comparable data on migrant integration across EU countries. ○ economic impact: efficiency; internal working routine and work processes; accountability and transparency ○ political impact: policies; institutions; ○ technological impact: services’ digitalisation; access to information. By gathering data during development and testing activities, it is possible to describe in qualitative and quantitative terms the impact on the above mentioned dimensions. Data can be of help in better describing the value generated by outputs and their potential if scaled up from pilots and proofs of concept to actual large-scale implementation. In parallel to this analysis, it is also useful to consider the cost of non-integration of migrants for the host society so as to be able to show - in financial terms - the opportunity space for investing in socio-technical solutions for their integration. In doing so, an in-depth analysis of the different and complementary costs that the EU Member States incur as a consequence of integration failure has been carried out, including, e.g., cost of irregular migrants’ imprisonment, cost of unemployment benefits, cost of migrants’ healthcare, cost of low educational performance for migrant students, cost of discrimination and cost of the reception. The analysis shows that the costs for a migrant that is not properly integrated into the host society can be as high as 52,000 Euro per year for a single person. An important disclaimer on the above figure is needed: the analysis conducted maps the cost of non-integration with reference to the Italian case and can be used only as a reference for reflection and further analysis on the situation in other EU countries. The assessment of costs of non-integration deserves further analysis, but it clearly shows how investments in integration - including investments in socio-technical tools - can be of great relevance for the host countries and can result, even, in a cost-saving for governments. The assessment of costs of non-integration deserves further analysis, but it clearly shows how investments in integration - including investments in socio-technical tools - can be of great relevance for the host countries and can result, even, in a cost-saving for governments. Key Lessons Learned Key Lessons Learned ● Considering different stakeholders, in different countries when assessing the impacts of socio-technical solutions for migrant integration is really important as the differences among contexts are high, and customized solutions emerged as the most valuable in terms of adoption and long-term sustainability. IMPACTS OF SOCIO-TECHNICAL SOLUTIONS FOR MIGRANTS’ INTEGRATION38 The areas of impact and dimensions considered for assessing the impact on migrants and on PAs and LSPs are: ● Impacts on migrants: ● Impacts on migrants: ● Impacts on migrants: ○ social impact: education and human capital; health; citizenship and social capital ○ economic impact: employment; access to bank and financial services; economic empowerment; ○ political impact: digital democracy; trust in host country institutions; civic participation; ○ technological impact: digital literacy; technology usage; information asymmetries. ● Impacts on LPSs and PAs: ○ social impact: workforce human capital; collaboration and networking; workforce working condition; 38 This chapter has been mainly authored by the partners of the REBUILD project funded under the Horizon2020 program with Grant Agreement number 822215. 31 31 Policy Recommendations ● To support the long-term sustainability of projects like the ones presented in this roundtable it would be advisable to have dedicated programs at the national and at EU level for supporting the research-to-market phase of the innovation produced. Innovation actions, even when they proved to deliver effective solutions, need further economic investments for their outputs to reach the market and be further customized to final users’ needs. ● To support the long-term sustainability of projects like the ones presented in this roundtable it would be advisable to have dedicated programs at the national and at EU level for supporting the research-to-market phase of the innovation produced. Innovation actions, even when they proved to deliver effective solutions, need further economic investments for their outputs to reach the market and be further customized to final users’ needs. ● PAs should be better supported in collecting data on the status of migrant integration: assessing the impact of project results and of their outputs can help prioritizing this goal, which would also provide useful information for better navigating the multitude of projects and initiatives active across EU countries. ● Besides this, a program of technology transfer, with ad hoc experts, should be offered to policymakers to support them in taking decisions on socio-technical solutions adoption. 39 This chapter has been mainly authored by the partners of the MICADO project funded under the Horizon2020 program with Grant Agreement number 822717. Moreover, within the same country, regional and local data are often missing or not comparable. The lack of standardized approaches in monitoring the integration phenomenon lead also to the lack of historical series of data that are crucial in mapping the impact of policies and investments. ● Often, when data are available, they are not provided as open and FAIR data, this is a shortcoming in terms of re-use of available data and research progress. 32 To ensure these three levels of sustainability, specific processes and activities need to be prepared. The critical question is who takes over the management of each level. One possibility is to delegate the responsibility to one institution. However, this strategy does depend on the institutional networking capacity of the ICT developer in terms of personal resources, technical infrastructure, and expertise. Relevant is to refer to the different actors who were already involved in the co-creation process. Software developers may form a central caretaker hub - a technical core team that ensures the long-term open-source development of the software solution. Delegation of responsibilities for sustainable uptake also needs to be linked to sustainable business planning and a strategy to secure funding for continuous development, maintenance, and development of the IT tools. A comprehensive guide to the uptake of IT solutions should be developed which includes a summary of the purpose of the uptake activities, a roadmap for local activities, and a clear management structure that outlines the responsibilities of each stakeholder (partner) involved in the uptake process. The guide needs to present the general plan for uptake and sustainability and provide input on how these strategies can be adapted and implemented in different local contexts. Finally, the guide needs to include strategic mid-term and long-term goals of an uptake strategy. Key Lessons Learned ● A generic uptake and sustainability plan is needed that outlines uptake demands and requirements for usage, operations, and technical development. ● Generic uptake plans need to be modified and translated into local uptake plans that meet contextual conditions. ● End users and uptake entities should be involved in the planning of uptake strategies from the very beginning of the ICT development work. Pitfalls ● Without the constant monitoring of (administrative) changes on the organization level of institutions and organizations being the potential uptakers of the solution, uptake activities may miss the target. ● Due to institutional changes, previously stated uptake commitments may be taken back, or substantially altered. ● While the planning for a sustainable uptake strategy happens during the project life-cycle, uptake activities are rarely continued after a project end. Policy Recommendations ● Financial support may be useful to target the uptake of ICT solutions that have already proven to be effective in integration of migrants. ● Relevant to the uptake may be more a capillary information sharing of success stories and solutions. DEVELOPING A VIABLE UPTAKE STRATEGY39 An effective uptake strategy needs to assess the key social, economic, and technical environments in which the IT solution will be operating. Potential risks and hurdles to uptake need to be evaluated and addressed. Sustainability principles guiding responsible use of resources and limiting the environmental burden need to be part of this strategy. An uptake and sustainability strategy or plan should offer: ● general measures and guidelines, which may apply to any location in which the technical solution will be taken up; ● specific strategies for local uptake that take into consideration local demands, existing IT infrastructure and processes, management structures, and responsibilities of different institutions in the domain of service provision. ICT solutions need to address three levels of sustainability, i.e., the sustainability of usage, technical maintenance, and sustainable development. ICT solutions need to address three levels of sustainability, i.e., the sustainability of usage, technical maintenance, and sustainable development. 1) The end-users should be trained and get familiarized with the solution, so they can use it independently and incorporate the usage into their daily routine. 2) The technical operability and the continued administration of content presented on the IT tool need to be managed in a sustainable manner. The responsibility of continued administration and maintenance needs to be assigned to an agency (the organization that provides services to migrants and/or an external IT provider). 3) ICT solutions are living products and require sustainable development. Software development needs to keep abreast of technical innovations and changes in the environment in which the software is embedded. 3) ICT solutions are living products and require sustainable development. Software development needs to keep abreast of technical innovations and changes in the environment in which the software is embedded. 33 ● Promotion of networked experiments of ICT solutions among different communities and institutions may be part of a targeted uptake strategy. ● The European Commission may consider the provision of supportive financial frameworks that enable future usage, operations, and technical development beyond projects’ runtime (e.g. operation funds). Key Lessons Learned ● A generic uptake and sustainability plan is needed that outlines uptake demands and requirements for usage, operations, and technical development. ● Generic uptake plans need to be modified and translated into local uptake plans that meet contextual conditions. ● End users and uptake entities should be involved in the planning of uptake strategies from the very beginning of the ICT development work. Pitfalls ● Without the constant monitoring of (administrative) changes on the organization level of institutions and organizations being the potential uptakers of the solution, uptake activities may miss the target. ● Due to institutional changes, previously stated uptake commitments may be taken back, or substantially altered. ● While the planning for a sustainable uptake strategy happens during the project life-cycle, uptake activities are rarely continued after a project end. Policy Recommendations ● Financial support may be useful to target the uptake of ICT solutions that have already proven to be effective in integration of migrants. ● Relevant to the uptake may be more a capillary information sharing of success stories and solutions. ● Promotion of networked experiments of ICT solutions among different communities and institutions may be part of a targeted uptake strategy. ● The European Commission may consider the provision of supportive financial frameworks that enable future usage, operations, and technical development beyond projects’ runtime (e.g. operation funds). Key Lessons Learned ● A generic uptake and sustainability plan is needed that outlines uptake demands and requirements for usage, operations, and technical development. ● Generic uptake plans need to be modified and translated into local uptake plans that meet contextual conditions. ● End users and uptake entities should be involved in the planning of uptake strategies from the very beginning of the ICT development work. Pitfalls ● Without the constant monitoring of (administrative) changes on the organization level of institutions and organizations being the potential uptakers of the solution, uptake activities may miss the target. ● Due to institutional changes, previously stated uptake commitments may be taken back, or substantially altered. SUSTAINABILITY OF THE NOVEL ICT SOLUTIONS40 40 This chapter has been mainly authored by the partners of the MIICT project funded under the Horizon 2020 program with Grant Agreement 822380 41 Farbenblum, B., Berg, L. and Kintominas, A. (2018). Transformative Technology for Migrant Workers: Opportunities, Challenges, and Risks. New York: Open Society Foundations. 42 Bock, A. K., Bontoux, L., Nascimento, S., and Szczepanikova, A. (2016). The future of the EU collaborative economy. JRC Science for Policy report. 43 Virketyte, A.; Wiklund, I. (2019). The Creative Process for Digital Social Innovation in the Context of Migrant Integration [PhD dissertation]. Jönköping: Jönköping University. 44 Farbenblum, B., Berg, L. and Kintominas, A. (2018). Transformative Technology for Migrant Workers: Opportunities, Challenges, and Risks. New York: Open Society Foundations. Virketyte, A.; Wiklund, I. (2019). Op. Cit.. Teberga, P. M. F., and Oliva, F. L. (2018). Identification, analysis and treatment of risks in the introduction of new technologies by start-ups. Benchmarking, 25(5), 1363–1381. SUSTAINABILITY OF THE NOVEL ICT SOLUTIONS40 Novel ICT solutions that deploy matchmaking tools linking migrants and service providers together need to be made sustainable according to many aspects. First, delivered information must be up to date, reliable, credible and officially endorsed by public sector organizations. This is a crucial requirement, challenged by the twin facts that information is not static and the user needs change across time. Also, translations of relevant documents and information need to be sensitive to dialects and constantly maintained and available in real-time. This implies that continual monitoring of the platform needs to be undertaken by stakeholder groups including the migrants themselves, who can give valuable feedback on the utility of the information contained in it. Another important aspect for developers to consider is accessibility. Any ICT social innovation tool is to be designed for accessibility and not to deepen the digital divide that may occur on grounds such as educational level, language skills, gender, age, previous exposure to digital devices, and income41, whereby ordinarily marginalized groups are excluded42. ICT solutions, therefore, need to be accompanied by training opportunities and the provision of free WIFI to users. Another issue concerns value. A well-documented social innovation risk is failure to provide value to the targeted audiences43, especially in the long term, where user retention and the expansion and/or narrowing down of the services offered start playing a crucial role in determining whether a technological tool survives the test over time or becomes obsolete. One more aspect for policymakers to consider is ethics, as migrants are vulnerable and can be at risk of exploitation, so the administrator of a platform must veto all links and information. Another factor enhancing sustainability obviously comes from securing long-term funding for the maintenance, expansion, and implementation of later-stage learning outcomes in relation to social innovations of the ICT type44. Key Lessons Learned Accessibility At the primary level, technological solutions must ensure that they are accessible to their targeted end-users. A robust, but appropriate technical infrastructure is necessary, such that it is responsive Key Lessons Learned Accessibility At the primary level, technological solutions must ensure that they are accessible to their targeted end-users. A robust, but appropriate technical infrastructure is necessary, such that it is responsive Key Lessons Learned rketyte, A.; Wiklund, I. (2019). Op. Cit.. berga, P. M. F., and Oliva, F. L. (2018). Identification, analysis and treatment of risks in the introduction of new chnologies by start-ups. Benchmarking, 25(5), 1363–1381. A well-documented social innovation risk is failure to provide value to the targeted audiences43, especially in the long term, where user retention and the expansion and/o narrowing down of the services offered start playing a crucial role in determining whether technological tool survives the test over time or becomes obsolete. One more aspect for policymakers to consider is ethics, as migrants are vulnerable and can be at ris of exploitation, so the administrator of a platform must veto all links and information. Another factor enhancing sustainability obviously comes from securing long-term funding for the maintenance, expansion, and implementation of later-stage learning outcomes in relation to socia innovations of the ICT type44. Key Lessons Learned Accessibility At the primary level, technological solutions must ensure that they are accessible to their targeted end-users. A robust, but appropriate technical infrastructure is necessary, such that it is responsive 40 This chapter has been mainly authored by the partners of the MIICT project funded under the Horizon 2020 program h Virketyte, A.; Wiklund, I. (2019). Op. Cit.. ● While the planning for a sustainable uptake strategy happens during the project life-cycle, uptake activities are rarely continued after a project end. Policy Recommendations ● Financial support may be useful to target the uptake of ICT solutions that have already proven to be effective in integration of migrants. ● Relevant to the uptake may be more a capillary information sharing of success stories and solutions. ● Promotion of networked experiments of ICT solutions among different communities and institutions may be part of a targeted uptake strategy. ● The European Commission may consider the provision of supportive financial frameworks that enable future usage, operations, and technical development beyond projects’ runtime (e.g. operation funds). ● Generic uptake plans need to be modified and translated into local uptake plans that meet contextual conditions. ● End users and uptake entities should be involved in the planning of uptake strategies from the very beginning of the ICT development work. 34 SUSTAINABILITY OF THE NOVEL ICT SOLUTIONS40 Novel ICT solutions that deploy matchmaking tools linking migrants and service providers togethe need to be made sustainable according to many aspects. First, delivered information must be up to date, reliable, credible and officially endorsed by publi sector organizations. This is a crucial requirement, challenged by the twin facts that information i not static and the user needs change across time. Also, translations of relevant documents and information need to be sensitive to dialects and constantly maintained and available in real-time This implies that continual monitoring of the platform needs to be undertaken by stakeholder group including the migrants themselves, who can give valuable feedback on the utility of the informatio contained in it. Another important aspect for developers to consider is accessibility. Any ICT social innovation tool i to be designed for accessibility and not to deepen the digital divide that may occur on grounds such as educational level, language skills, gender, age, previous exposure to digital devices, and income41 whereby ordinarily marginalized groups are excluded42. ICT solutions, therefore, need to be accompanied by training opportunities and the provision of free WIFI to users. Another issue concerns value. 45 Nedelcu, M. and Soysüren, I. (2020). Precarious Migrants, Migration Regimes and Digital Technologies: The Empowerment-Control Nexus. Journal of Ethnic and Migration Studies. https://www.tandfonline.com/doi/full/10.1080/1369183X.2020.1796263 Farbenblum, B., Berg, L. and Kintominas, A. (2018). Transformative Technology for Migrant Workers: Opportunities, Challenges, and Risks. New York: Open Society Foundations. https://www.opensocietyfoundations.org/publications/transformative-technology-migrant-workers-opportunities- challenges-and-risks Gelb, S. and Krishnan, A. (2018). Technology, Migration and the 2030 Agenda for Sustainable Development. London: Overseas Development Institute. https://odi.org/en/publications/technology-migration-and-the-2030-agenda-for-sustainable-development/ Responsiveness Also, at the foundational level stands the challenge of responsiveness, understood here as the capability to reflect the current needs and wants of end-users, and the potential to continue doing so in the future, as those needs and wants change and evolve. It is vital for an ICT tool to offer a wide spectrum of relevant information and services that bring practical value to end-users and are of everyday utility. The creation of a one-stop-shop approach is important in enabling migrants and service providers to save time and be more independent and proactive in their own integration journeys. to different devices and operating systems, as well as a user interface design (UID) that can accommodate both desktop and mobile versions without great divergences. Accessibility also relies on the user base being able to access a device, through which to make use of the particular ICT tool, and being capable of operating this device at least on its basic functionalities. One of the key lessons learned is to ensure that within the different pilot locations support sessions are offered in order to ensure that specific groups - such as those that are illiterate and the visually impaired - are able to use the ICT tool. Another lesson is to ensure that the platform is available on mobile phones and mobile apps since they are acknowledged as the preferred medium for accessing information and services by migrant users45. This would ensure that migrants can access vital services whilst on the move. to different devices and operating systems, as well as a user interface design (UID) that can accommodate both desktop and mobile versions without great divergences. Accessibility also relies on the user base being able to access a device, through which to make use of the particular ICT tool, and being capable of operating this device at least on its basic functionalities. to different devices and operating systems, as well as a user interface design (UID) that can accommodate both desktop and mobile versions without great divergences. Accessibility also relies on the user base being able to access a device, through which to make use of the particular ICT tool, and being capable of operating this device at least on its basic functionalities. One of the key lessons learned is to ensure that within the different pilot locations support sessions are offered in order to ensure that specific groups - such as those that are illiterate and the visually impaired - are able to use the ICT tool. Another lesson is to ensure that the platform is available on mobile phones and mobile apps since they are acknowledged as the preferred medium for accessing information and services by migrant users45. This would ensure that migrants can access vital services whilst on the move. Accessibility 35 Reliability Reliability is intrinsically tied to the credibility of content. The challenge of verification, fact- checking, and ensuring security and trust must be foundational to the design and development of any ICT platform. Due to the highly sensitive nature of information additional care must be taken to validate that information and check that it is truthful and corresponds to official sources, and services offered are genuine. An additional concern arises from the fact that migrant end-users constitute a vulnerable group and do stand at a heightened risk of exploitation and abuse. All privacy and data protection risks need to be examined to ensure migrants are not put at risk. Opening the platform to the public or up-scaling for commercial use would mandatorily entail the allocation of resources for the verification, validation, and fact-checking of information, services, and users (e.g. service providers). 46 Codagnone, C. and Kluzer, S. (2011). ICT for the Social and Economic Integration of Migrants into Europe. JRC Scientific and Technical reports. Luxembourg: Publications Office of the European Union. We recommend that the EC strongly encourages governments, local governance structures, and public bodies to take ownership of novel ICT solutions that have proven successful in the field and have garnered support from non-governmental organizations and service providers, working with migrants, as well as migrants themselves, after the end of project activities, should those solutions be project-based. In supporting and commissioning ICT tools for migration and integration, governments should ensure consultation with and feedback from non-governmental organizations working with migrants, as well as migrant communities and sub-groups within those communities. Good working relationships and stable collaborations with Internet providers and mobile service providers are also of utmost importance. The sustainability of existing ICT tools relies upon the ongoing and long-term value creation for migrant end-users, regardless of how their needs and preferences change in time. Hence, ensuring funding and human resources necessary for the provision of this value is of crucial importance. Digital education and inclusion It is highly recommended that the digital education of migrants and the breaking down of the Marginalized groups Some ICTs may lead to the exclusion of marginalized groups such as women, the elderly, unaccompanied migrants, and migrants who are illiterate or visually impaired. The creation of local hubs or information points could enable migrants who are less comfortable with using ICT tools to access information and services, build their digital skills, and acquire new qualifications. This would enable those migrants who fear being tracked by the authorities to use the ICT solution with the support of trained staff. The fear of digital surveillance is a concern that can jeopardize the success of any ICT solution. Take-up and ownership Project-based social innovations carry the inherent characteristics of being time- and resource- constrained and being capable of reaching a finite number of beneficiaries before activities come to an end46. This is why the post-project take-up and the assumption of ownership over any innovation remain challenges worth addressing. Policy Recommendations Engagement and ownership The sustainability of existing ICT tools relies upon the ongoing and long-term value creation for migrant end-users, regardless of how their needs and preferences change in time. Hence, ensuring funding and human resources necessary for the provision of this value is of crucial importance. needs and preferences, and must be provided with different services. One important issue related to this aspect of tailoring is the need to provide adequate languages and translations that take into consideration the local context and region. Another important factor is to enable the ICT solution to add new modules and functionalities, whilst keeping services and information updated and modifiable in the various migrant languages. needs and preferences, and must be provided with different services. One important issue related to this aspect of tailoring is the need to provide adequate languages and translations that take into consideration the local context and region. Another important factor is to enable the ICT solution to add new modules and functionalities, whilst keeping services and information updated and modifiable in the various migrant languages. consideration the local context and region. Another important factor is to enable the ICT solution to add new modules and functionalities, whilst keeping services and information updated and modifiable in the various migrant languages. Marginalized groups Some ICTs may lead to the exclusion of marginalized groups such as women, the elderly, unaccompanied migrants, and migrants who are illiterate or visually impaired. The creation of local hubs or information points could enable migrants who are less comfortable with using ICT tools to access information and services, build their digital skills, and acquire new qualifications. This would enable those migrants who fear being tracked by the authorities to use the ICT solution with the support of trained staff. The fear of digital surveillance is a concern that can jeopardize the success of any ICT solution. Take-up and ownership Project-based social innovations carry the inherent characteristics of being time- and resource- constrained and being capable of reaching a finite number of beneficiaries before activities come to an end46. This is why the post-project take-up and the assumption of ownership over any innovation remain challenges worth addressing. Policy Recommendations Engagement and ownership Multi-stakeholder engagement in all stages of ideation, creation, and implementation helps reap the fruits of continuous learning from a spectrum of points of view. The encouragement of cross- organizational communication in the formulation of migration and integration policies at any level is to be done paying due attention to meeting migrants’ needs through ICT tools, as well as conventional methods, especially in the COVID-19 and post-pandemic realities. Pitfalls Tailored service provision In the early stages of co-creation, the need for customization arose hand-in-hand with the realization that end-users in different pilot countries act in different environments, have different 36 36 Engagement and ownership Engagement and ownership Multi-stakeholder engagement in all stages of ideation, creation, and implementation helps reap the fruits of continuous learning from a spectrum of points of view. The encouragement of cross- organizational communication in the formulation of migration and integration policies at any level is to be done paying due attention to meeting migrants’ needs through ICT tools, as well as conventional methods, especially in the COVID-19 and post-pandemic realities. We recommend that the EC strongly encourages governments, local governance structures, and public bodies to take ownership of novel ICT solutions that have proven successful in the field and have garnered support from non-governmental organizations and service providers, working with migrants, as well as migrants themselves, after the end of project activities, should those solutions be project-based. In supporting and commissioning ICT tools for migration and integration, governments should ensure consultation with and feedback from non-governmental organizations working with migrants, as well as migrant communities and sub-groups within those communities. Good working relationships and stable collaborations with Internet providers and mobile service providers are also of utmost importance. The sustainability of existing ICT tools relies upon the ongoing and long-term value creation for migrant end-users, regardless of how their needs and preferences change in time. Hence, ensuring funding and human resources necessary for the provision of this value is of crucial importance. Multi-stakeholder engagement in all stages of ideation, creation, and implementation helps reap the fruits of continuous learning from a spectrum of points of view. The encouragement of cross- organizational communication in the formulation of migration and integration policies at any level is to be done paying due attention to meeting migrants’ needs through ICT tools, as well as conventional methods, especially in the COVID-19 and post-pandemic realities. We recommend that the EC strongly encourages governments, local governance structures, and public bodies to take ownership of novel ICT solutions that have proven successful in the field and have garnered support from non-governmental organizations and service providers, working with migrants, as well as migrants themselves, after the end of project activities, should those solutions be project-based. In supporting and commissioning ICT tools for migration and integration, governments should ensure consultation with and feedback from non-governmental organizations working with migrants, as well as migrant communities and sub-groups within those communities. Good working relationships and stable collaborations with Internet providers and mobile service providers are also of utmost importance. Building confidence and trust It is recommended that the EC urges member states to refrain from unjustified and indiscriminate violations of freedom of communication and the right to privacy of correspondence of migrants, which extend to their online activities as well. On the other hand, data security in an environment of increasingly frequent cyberattacks against online public infrastructure is a matter of growing concern. Protection of personal data mechanisms and adequate and effective remedies in cases of breaches ought to be established across the EU with a particular emphasis on vulnerable groups and users with lower awareness of online threats and risks. It is important that end-users have trust and confidence when accessing and utilizing ICT tools, not only from the viewpoint of being able to easily complete tasks and resolve issues (which is readily addressed by digital education), but also of having the perception and knowledge that their data and interactions are kept confidential and secure. This would reduce the scepticism and raise the motivation in benefiting from digital solutions for migration and integration. Investing in and fostering (through targeted policy and legislation) stable progress in the area of personal data protection, cybersecurity, and cyber rights is recommended to achieve the goal of ICT tools’ sustainability. digital barriers that many migrants face are recognized as common goals among all EU Member States. The process of digitalization should not remain divorced from the issues of social inequality, marginalization, and exclusion that vulnerable groups such as migrants often encounter in their daily lives. Hence, the EC must continue to promote and build support for a rights-based understanding of ICT infrastructure and the Internet, emphasizing the value of equal treatment and non-discrimination. Furthermore, it is recommended that the EC supports governments both methodologically and financially in providing such education and locally countering the digital barriers while offering fair and adequate opportunities for NGOs working with migrants to benefit from renewing or expanding their existing ICT infrastructure, where this is needed, so that they can provide access to ICT tools for migration and integration to their beneficiaries on their premises. Finally, it is suggested that the EC considers targeting policies and measures that engage specific sub-groups of migrants who are at higher risk of being marginalized and excluded from accessing and benefiting from ICT tools, such as illiterate women, the elderly, those living in extreme material deprivation, and those with low or no education. Gender mainstreaming and acknowledgment of intersectionality are vital in this respect. Support for governments and NGOs that work with such groups is essential in overcoming the socio-economic barriers before their full- fledged inclusion and successful integration. Without continuous capacity-building and skill- development among both migrants and professionals working with migrants, the sustainability of novel ICT tools for migration and integration could be compromised. Digital education and inclusion highly recommended that the digital education of migrants and the breaking down of the 37 the end of a project-based life stage is not solely a question of stakeholder willingness and support, but also of available finance. Hence, it is recommended that the EC and the member states act to financially secure the long-term exploitation of successful digital social innovations either through designated funding or amendments in national budgets that free up resources to be pooled in this direction. Importantly, funding is the single defining factor behind many of the good practices and lessons learned presented here – for instance, migrant digital education and bridging the digital barrier are highly dependent upon the availability of human and financial resources to implement these activities. Therefore, governments ought to make room for this set of integration measures in their general migration budgets, while NGOs should be given ample opportunity to apply for and gain the funding that allows them to carry out tasks related to migration and integration via digital tools. 48 World Health Organisation (2018) Mental health promotion and mental health care in refugees and migrants’ Technical guidance. https://www.euro.who.int/__data/assets/pdf_file/0004/386563/mental-health-eng.pdf?ua=1 47 This chapter has been mainly authored by the partners of the WELCOME project funded under the Horizon2020 program with Grant Agreement number 870930. Funding for sustainability Of all the challenges and pitfalls discussed in this section of the Handbook, lack of funding for scaling, bringing to maturity, and implementing changes in technology on the basis of later-stage learning remains the most considerable constraint to the longevity and sustainability of platforms. Funding is also crucial for ensuring that comprehensive and reliable test and end-user feedback can be collected and analysed in a consistent manner over time. Thus, the assumption of ownership over a given technological tool for migration and integration at 38 47 This chapter has been mainly authored by the partners of the WELCOME project funded under the Horizon2020 program with Grant Agreement number 870930. 48 World Health Organisation (2018) Mental health promotion and mental health care in refugees and migrants’ Technical guidance. https://www.euro.who.int/__data/assets/pdf_file/0004/386563/mental-health-eng.pdf?ua=1 • Developing ICT solutions from scratch and trying to implement them in a vacuum, without considering the differences in each country might end up in a non-applicable system. It has been observed that host countries may be unable to support TCNs as their systems are not ready, and remain unprepared, especially after the change in the migration flows of the period after 2015. Key Lessons Learned Key lessons learned in terms of the development of ICT systems include: ● Asking and involving the target groups (migrants, asylum seekers, and refugees) from the beginning of the design of any process and within all phases of the implementation of a project until its evaluation and not only in pilot phases. ● Harmonizing the language between technology and human sciences. ● Taking into consideration the several dimensions of a TCN’s experience when discussing how to automate integration procedures and design/develop specific tools for their support. Pitfalls Common pitfalls in terms of the development of ICT systems include: • Not taking into account the specific characteristics of the target groups when trying to match them with needs. Target group’s needs must be recorded correctly in order for the services offered to be adapted to various characteristics, such as age, gender, religion and language, to avoid increasing the risk of discrimination. • Developing ICT solutions from scratch and trying to implement them in a vacuum, without considering the differences in each country might end up in a non-applicable system. It has been observed that host countries may be unable to support TCNs as their systems are not ready, and remain unprepared, especially after the change in the migration flows of the period after 2015. Key lessons learned in terms of the development of ICT systems include: ● Asking and involving the target groups (migrants, asylum seekers, and refugees) from the beginning of the design of any process and within all phases of the implementation of a project until its evaluation and not only in pilot phases. ● Harmonizing the language between technology and human sciences. ● Taking into consideration the several dimensions of a TCN’s experience when discussing how to automate integration procedures and design/develop specific tools for their support. reception, integration, and coaching of TCNs in Europe. However, the vast majority of these technologies are realized as a “one-size-fits-all” solution, which means that they assume a predetermined average profile of their users with respect to their needs, background, and capacity. Even if we assume that all TCNs need to go through the same procedures and thus require assistance on the same matters, their needs are different because of the large variation of their social, cultural, and educational backgrounds. Furthermore, forced migration backgrounds often imply high levels of psychological distress, deep personal traumatic experiences, often dangerous journeys full of privation, and a profound disillusion upon arrival in the host country, which may affect the mental and psychological equilibrium of the individuals concerned. In particular, through the process of adaptation to a new environment, and because of their past experiences of discrimination, migrants are more prone to psychological stress than the average, which across time may lead to symptoms of depression or emotional burden as they strive to adapt to the receiving context conditions. All the six projects involved in this Handbook preparation faced similar challenges related to the psychosocial status of TCNs when designing AI-supported integration services. The following table can provide a blueprint for future projects on how these challenges can be taken into account during the preparation of ICT and AI-based tools for TCN reception and integration in the EU. Key Lessons Learned Key lessons learned in terms of the development of ICT systems include: ● Asking and involving the target groups (migrants, asylum seekers, and refugees) from the beginning of the design of any process and within all phases of the implementation of a project until its evaluation and not only in pilot phases. ● Harmonizing the language between technology and human sciences. ● Taking into consideration the several dimensions of a TCN’s experience when discussing how to automate integration procedures and design/develop specific tools for their support. Pitfalls Common pitfalls in terms of the development of ICT systems include: • Not taking into account the specific characteristics of the target groups when trying to match them with needs. Target group’s needs must be recorded correctly in order for the services offered to be adapted to various characteristics, such as age, gender, religion and language, to avoid increasing the risk of discrimination. PSYCHOSOCIAL DIMENSIONS OF AI SUPPORTED SERVICES47 Leaving one’s home country and migrating to a different one is often complex and stressful, related to events before departure, during travel and transit, and after arrival. Refugees in particular have often been exposed to stressful events such as wars and other forms of armed conflict, persecution, discrimination or natural disasters48. Upon arrival in the host country, third-country nationals have to cope with the need to adapt to a new society, often with a new language, different climate, and majority culture. Their minority status puts migrants and refugees in a disadvantaged position, prone to discrimination and increase vulnerability. These difficult processes and experiences often raise concerns about the fate of their families, regardless of whether the families travelled with them or stayed in the country of origin or transit. Additional challenges may involve learning the language of the host country, searching for employment, getting acquainted with the health and social services, managing the bureaucratic requirements for obtaining permission to reside in their new community. A key issue is the fact that there is uncertainty about how long they can stay in the country and that creates a lot of stress and difficulties. The threat of being homeless, deported, detained, and/or suffering from long waiting times until their status is finally determined is a common problem among many MS. The number and complexity of the necessary administrative procedures are considerably higher than those of the usual procedures that the locals need to follow routinely, and if we consider that even locals often feel overstrained, so do Third Country Nationals (TCNs), at a much higher rate. The role of social and governmental services and civil society has been fundamental for the successful assistance of TCNs, but there is a need for more support. In view of the fact that intelligent human-computer interaction technologies and more recently, Artificial Intelligence (AI) have become an important element in a number of different applications, including, e.g., information delivery, education, and coaching, simulation and training, entertainment, and media, these technologies appear to be an ideal instrument to support the 39 reception, integration, and coaching of TCNs in Europe. However, the vast majority of these technologies are realized as a “one-size-fits-all” solution, which means that they assume a predetermined average profile of their users with respect to their needs, background, and capacity. PSYCHOSOCIAL DIMENSIONS OF AI SUPPORTED SERVICES47 Even if we assume that all TCNs need to go through the same procedures and thus require assistance on the same matters, their needs are different because of the large variation of their social, cultural, and educational backgrounds. Furthermore, forced migration backgrounds often imply high levels of psychological distress, deep personal traumatic experiences, often dangerous journeys full of privation, and a profound disillusion upon arrival in the host country, which may affect the mental and psychological equilibrium of the individuals concerned. In particular, through the process of adaptation to a new environment, and because of their past experiences of discrimination, migrants are more prone to psychological stress than the average, which across time may lead to symptoms of depression or emotional burden as they strive to adapt to the receiving context conditions. All the six projects involved in this Handbook preparation faced similar challenges related to the psychosocial status of TCNs when designing AI-supported integration services. The following table can provide a blueprint for future projects on how these challenges can be taken into account during the preparation of ICT and AI-based tools for TCN reception and integration in the EU. Key Lessons Learned Key lessons learned in terms of the development of ICT systems include: ● Asking and involving the target groups (migrants, asylum seekers, and refugees) from the beginning of the design of any process and within all phases of the implementation of a project until its evaluation and not only in pilot phases. ● Harmonizing the language between technology and human sciences. ● Taking into consideration the several dimensions of a TCN’s experience when discussing how to automate integration procedures and design/develop specific tools for their support. Pitfalls Common pitfalls in terms of the development of ICT systems include: • Not taking into account the specific characteristics of the target groups when trying to match them with needs. Target group’s needs must be recorded correctly in order for the services offered to be adapted to various characteristics, such as age, gender, religion and language, to avoid increasing the risk of discrimination. • Developing ICT solutions from scratch and trying to implement them in a vacuum, without Policy Recommendations ● It is highly important to consider the socio-psychological dimensions (that are taken into account by professionals) during the interaction with TCNs. The analysis of psychosocial backgrounds is of primary importance in the context of AI-supported services: it reveals a large variety of social and cultural backgrounds, different (very often traumatic) life experiences, expectations, etc. Moreover, each TCN has his/her own communication skills, own traumata, needs and values, formal or informal educational background as well as individual capacities and capabilities. ● AI-supported services cannot be a “one-size-fits-all”. To ensure adequate service provision, there is a need to identify and differentiate the most distinctive psychosocial features and take them into account in technology development and implementation. 49 This chapter has been mainly authored by the partners of the easyRights project funded under the Horizon2020 program with Grant Agreement number 870980. 50 https://nws.eurocities.eu/MediaShell/media/3rd_Integrating_Cities_Report_October_2018_FINAL.pdf 51 ● There is a special issue in the area of integration in the employment domain: an appropriate match of the individual's profile with his/her own special qualitative characteristics is rarely made. This automatically results in a profiling that is poor or negative for the TCN. Common pitfalls in terms of the development of ICT systems include: Common pitfalls in terms of the development of ICT systems include: • Not taking into account the specific characteristics of the target groups when trying to match them with needs. Target group’s needs must be recorded correctly in order for the services offered to be adapted to various characteristics, such as age, gender, religion and language, to avoid increasing the risk of discrimination. • Developing ICT solutions from scratch and trying to implement them in a vacuum, without considering the differences in each country might end up in a non-applicable system. It has been observed that host countries may be unable to support TCNs as their systems are not ready, and remain unprepared, especially after the change in the migration flows of the period after 2015. 40 https://www.eupan.eu/wpcontent/uploads/2019/02/2018_1_BG_Citizens_Charters_and_Public_Service_Delivery_Standar ds.pdf 52 F. Adrodegari, D. Bonetti and N. Saccani (2017). A Framework to assess the level of Customer Centricity in Manufacturing Companies. Paper presented at the RESER Conference, Bilbao, Spain. 53 Hayes B. E. (2013). TCE - Total Customer Experience, Building business through customer-centric measurement and analytics. Lexington, Kentucky: Business Over Broadway. 54 Manning, H. and Bodine, K. (2012), Outside in. The Power of Putting Customers at the Center of Your Business , Houghton Mifflin Harcourt, Boston, MA. 55 Adrodegari et al. (2017). Op. Cit. 55 Adrodegari et al. (2017). Op. Cit. ● providing timely, high standard and easily accessible services; and, ● being accountable when people (citizens/users) become part of service provision. ASSESSING SERVICE ACCESSIBILITY: TOWARDS A NEW STANDARD49 Formal complaint proc out as well as contingent processes covering the cases where standards are not met. In general “all countries consider that it would be worthwhile to explore further benefits/possibilities of developing some form of common non-binding public se standards on an EU level, but there are different views on the scope of the needed coop need to be agreed upon. Still, most countries point towards the development of comm for quality of public service delivery across the EU, which will not be binding for the M but will offer them a benchmark and guidelines when adopting their national standa 2018, p.5). Generally, the common principles for quality service delivery are related to ● listening to and understanding the customer; ● designing and implementing policies and services that meet citizens’ expectatio ● providing timely, high standard and easily accessible services; and, ● being accountable when people (citizens/users) become part of service provisio User centricity is generally acknowledged as a key concept in service (re)design for bo and public sector innovators. In the private (for profit) sector, the equivalent con Centricity is Customer Centricity. This posits that the customer must ‘come first’, i should put all its efforts into troubleshooting its customers and creating a product or really of value for them52. Following this train of logic, the analytical unit for a assessment of C stomer Centricit is not a specific good or ser ice b t the compan implementation varies significantly across countries, due to the different legal systems in place. So while some charters are enforced by law, others are more informal commitments and/or statements of principle. Also, several countries report challenging experiences in measuring service compliance with the target levels set out in the Charters, so they either use the feedback received from users or rely on general commitments to improve quality, without the recourse to specific monitoring and evaluation mechanisms. The contents of these Charters across countries present a high degree of similarity (although with some differentiations) for the commitments made: on-time delivery targets, a positive attitude of public sector workers, non-discriminating actions, transparent and accountable behaviours, inclusiveness and openness of government etc. Access to public services is promised in a multi-modal fashion (e.g. by post, online, by phone and in-person). Mechanisms are put in place for the general public to provide feedback on public service delivery and there are intentions to use this feedback to improve public service development in the future. Formal complaint procedures are set out as well as contingent processes covering the cases where standards are not met. In general “all countries consider that it would be worthwhile to explore further the potential benefits/possibilities of developing some form of common non-binding public service delivery standards on an EU level, but there are different views on the scope of the needed cooperation, which need to be agreed upon. Still, most countries point towards the development of common principles for quality of public service delivery across the EU, which will not be binding for the Member States, but will offer them a benchmark and guidelines when adopting their national standards” (EUPAN, 2018, p.5). Generally, the common principles for quality service delivery are related to Generally, the common principles for quality service delivery are related to ● listening to and understanding the customer; ● designing and implementing policies and services that meet citizens’ expectations; ● providing timely, high standard and easily accessible services; and, ASSESSING SERVICE ACCESSIBILITY: TOWARDS A NEW STANDARD49 The quality of integration policies varies widely across European countries. In spite of the standards set out by EU and international law, citizens still experience disparities within and between Member States and cities, and such disparities worsen migrants’ conditions as they are not treated fairly and equally wherever they apply for support. In a report made by Eurocities in 201850 it was clearly stated that most cities have performed (and are still performing) significant efforts to promote citizenship and equal access to rights and services by the migrants through implementing dedicated measures, such as: in some cases, setting up new bodies in charge of implementing inclusive policies or introducing monitoring mechanisms against discrimination; in other cases, adopting new guidelines and plans for equality; or investing in accessible services offered in many languages and tailored to the specific needs of migrants and refugees. EUPAN, the informal network of the European Commission and Directors General responsible for public administration in the EU Member States, has largely promoted the adoption of standards for quality management in the public sector, including for the measurement of satisfaction in service usage as a proxy for compliance; these initiatives are certainly relevant, especially given the positive impacts that the use of standards has generated in both the public and the private sectors. Citizens’ charters have been largely used since the 1990s and recently confirmed (see the EUPAN network report of 201851) as a good way to develop or adopt standards, or quasi-standards. Their 41 implementation varies significantly across countries, due to the different legal system while some charters are enforced by law, others are more informal commitments and/ of principle. Also, several countries report challenging experiences in measuring servi with the target levels set out in the Charters, so they either use the feedback received rely on general commitments to improve quality, without the recourse to specific m evaluation mechanisms. The contents of these Charters across countries present a h similarity (although with some differentiations) for the commitments made: on-time de a positive attitude of public sector workers, non-discriminating actions, transparent an behaviours, inclusiveness and openness of government etc. Access to public services is multi-modal fashion (e.g. by post, online, by phone and in-person). Mechanisms are p the general public to provide feedback on public service delivery and there are intentio feedback to improve public service development in the future. 57 For more info https://ec.europa.eu/newsroom/document.cfm?doc_id=47559 56 See e.g. M. Wiśniewski (2001) Using SERVQUAL to assess customer satisfaction with public sector services, In Managing Service Quality, 11(6): 350-388. ● being accountable when people (citizens/users) become part of service provision. ● being accountable when people (citizens/users) become part of service provision. User centricity is generally acknowledged as a key concept in service (re)design for both the private and public sector innovators. In the private (for profit) sector, the equivalent concept to User Centricity is Customer Centricity. This posits that the customer must ‘come first’, i.e. a company should put all its efforts into troubleshooting its customers and creating a product or service that is really of value for them52. Following this train of logic, the analytical unit for a hypothetical assessment of Customer Centricity is not a specific good or service, but the company that produces it. For instance, a customer centric provider should integrate customer experience programs throughout the organization itself53 or use feedback collection methods and tools as well as customer satisfaction metrics to ground its decision-making54. Indeed, as noted by some scholars55 “the 42 57 For more info https://ec.europa.eu/newsroom/document.cfm?doc_id=47559 • Digital Interaction • Digital Interaction o To have the option to digitally interact with their administrations • Digital Interaction o To have the option to digitally interact with their administrations • Accessibility, security, availability and usability o That the services are made more accessible (including findable) and secure and can be used by all in a non-discriminatory manner, with appropriate assistance available upon need. o That the principles of universal design have been applied to the setting up of the services and that the websites are simple to read and easy to understand. o That the authenticity of digital public services is secured and can be recognised in a clear and consistent manner. transition from a product centric approach to a customer centric one is very complex and requires a transformation process that covers all the organizational aspects of the company”. This is a very interesting research and practical domain, but obviously not transferable as such to the public sector, which is eminently subject to the rule of law in the daily management of its current affairs. In fact, while the collection of inputs from government service users in the form of satisfaction metrics has been customary since long, particularly in association with the implementation of quality standards56, it remains true that a government body or agency demonstrating their propensity to listen to the opinions of citizens and/or businesses cannot be considered user-centric only because of that. With the 2017 Tallinn Declaration on eGovernment57, 32 Member States narrowed the focus on 8 basic principles (aka standards), the fulfilment of which in a context of digital public service delivery or existing service digitization should ensure that citizens and businesses are genuinely placed at the centre of the newly built or redesigned public services. These 8 principles are reported below. When interacting with public administrations and using digital public services, citizens and businesses should expect: When interacting with public administrations and using digital public services, citizens and businesses should expect: • Protection of personal data and privacy • Protection of personal data and privacy o That the handling of personal data respects the general data protection regulation and privacy requirements in the EU and national levels, when applicable informing citizens about the use and storage of their personal data and allowing citizens to access and ask for the correction and deletion of personal data, where appropriate. f p p y o That the handling of personal data respects the general data protection regulation and privacy requirements in the EU and national levels, when applicable informing citizens about the use and storage of their personal data and allowing citizens to access and ask for the correction and deletion of personal data, where appropriate. • Redress and complaint mechanisms • Redress and complaint mechanisms o That redress mechanisms are available online and that citizens and business have access to complaint procedures online, while also in other available channel(s) of their choice. These 8 standards add further dimensions to Principle #6 (of 12) of the European Interoperability Framework (EIF), which generically underlines the importance of analysing, understanding and reflecting users’ needs and requirements during the design and development of public services in accordance with their expectations. As a matter of fact, Online Availability and Online Usability have been considered for quite a long time the two key components of User Centricity. The former refers to the existence of an electronic channel for public services either through a portal or standalone. The latter looks at the overall user experience of support, help and feedback functionalities, ease and speed of use. This is the situation created at the EU level with the eGovernment Benchmark Framework 2012-201558. More recently, and notably after the publication in April 2016, of the eGovernment Action Plan for 2016-2020, Mobile Friendliness (and accessibility) was also added, to reflect the fact that while the use of mobile Internet has become widespread, only 1 in 3 European public websites at the time was designed to be ‘mobile friendly’59. 58 https://ec.europa.eu/digital-agenda/sites/digital- agenda/files/eGovernment%20Benchmarking%20method%20paper%20published%20version_0.pdf 59 https://ec.europa.eu/digital-single-market/en/news/eu-egovernment-report-2016-shows-online-public-services- improved-unevenly • Reduction of the administrative burden • Reduction of the administrative burden o That public administrations make efforts to reduce the administrative burden on citizens and businesses, namely by optimizing and/or creating digital processes and services where relevant and possible, and by offering personalised and pro-active services. o Not to be asked to provide the same information to public services more than once, in due respect of data protection rules and regulations. • Digital delivery of public services o That public services can as much as possible and appropriate, especially upon request of the user, be fully handled online, including the provision of any evidence required to obtain a right or fulfil obligations. o That the status of service delivery can be checked online where relevant. • Citizen engagement o That digital means are used to empower citizens and businesses to voice the views, allowing policy makers to collect new ideas, involve citizens more in the creation of public services and provide better digital public services. • Incentives for digital service use ncentives for digital service use o The barriers to use digital public services should be effectively removed, including by extending and promoting the benefits of, for example, higher confidence, speed, effectivity and reduced costs to individuals who are able to use them. 56 See e.g. M. Wiśniewski (2001) Using SERVQUAL to assess customer satisfaction with public sector services, In Managing Service Quality, 11(6): 350-388. 43 g / / g p p p _ p 59 https://ec.europa.eu/digital-single-market/en/news/eu-egovernment-report-2016-shows-online-public-services- improved-unevenly It would provide a common basis for the harmonization of public services to migrants and refugees across Europe, under the perspectives of quality of communication and effectiveness of information delivery. It should embed at least the following three characteristics: coverage, coherence, and outreach. Coverage The assessment standard should differentiate the various dimensions in which a service is articulated and provide a series of reference points (levels or steps) by which progress in accessing that service can be calibrated; also, it should be able to sense whether public services are designed to consider many possible use cases in different and complex situations including those that often occur when migrants are involved. ● Shared understanding and guidelines on citizens’ needs and expectations may improve the efficiency of the public sector across Europe and increase transparency and accountability so widening citizens’ knowledge about, and confidence in, public services. ● Existing service quality standards have been introduced and used discontinuously across Europe and developed with scarce knowledge of the users’ experience: accessibility should take into account the users’ experience for it to become an integration measure. ● Existing service quality standards have been introduced and used discontinuously across Europe and developed with scarce knowledge of the users’ experience: accessibility should take into account the users’ experience for it to become an integration measure. ● New standards are required that are more focused on User Centricity and Interoperability of these services. Linguistic offer in service provision is crucial but does not represent the unique obstacle in service access. Discretionary in the interpretation of service accessibility and the ways to guarantee it across service sectors and across countries introduces disparities and inequalities and represents a significant obstacle to migrants’ integration. Bureaucratic complexity of service procedures represents an infringement to information rights and makes service access difficult and, more critical, open to failures. Differentiation in accessibility modes improves service quality but at the same time widens the bureaucratic complexity of the service thus challenging its accessibility. Compliance with basic human rights and human rights literacy represents a critical challenge in service provision especially when services do have deep impacts on citizenship and cascade effects on other services and rights. Coverage The assessment standard should differentiate the various dimensions in which a service is articulated and provide a series of reference points (levels or steps) by which progress in accessing that service can be calibrated; also, it should be able to sense whether public services are designed t id ibl i diff t d l it ti i l di th th t ft ● Shared understanding and guidelines on citizens’ needs and expectations may improve the efficiency of the public sector across Europe and increase transparency and accountability so widening citizens’ knowledge about, and confidence in, public services. ● Existing service quality standards have been introduced and used discontinuously across Europe and developed with scarce knowledge of the users’ experience: accessibility should take into account the users’ experience for it to become an integration measure. ● New standards are required that are more focused on User Centricity and Interoperability of these services. Pitfalls Linguistic offer in service provision is crucial but does not represent the unique obstacle in service access. Discretionary in the interpretation of service accessibility and the ways to guarantee it across service sectors and across countries introduces disparities and inequalities and represents a significant obstacle to migrants’ integration. Bureaucratic complexity of service procedures represents an infringement to information rights and makes service access difficult and, more critical, open to failures. Differentiation in accessibility modes improves service quality but at the same time widens the bureaucratic complexity of the service thus challenging its accessibility. Compliance with basic human rights and human rights literacy represents a critical challenge in service provision especially when services do have deep impacts on citizenship and cascade effects on other services and rights. Policy Recommendations Improving the accessibility of public services is surely crucial to migrants’ integration in Europe; it also represents an opportunity to homogenize European citizenship and reduce the administrative stress of the public sector. The creation of a standard to assess service accessibility would reduce disparities and unequal legal implications consequent to eventual failures in administrative service and procedure accomplishment. In compliance with the User Centricity principle, such a standard could and should be a collection of minimum requirements for enforcing the informational rights of foreign migrants and refugees entering the European Union legally. Mainstreaming a user-centric approach to the evaluation of the quality of public services for migrants can and should constitute the next immediate challenge for EU policy making: this would enable at the same time, the harmonization of basic service delivery irrespective of the place of fruition and a concrete progress towards User Centricity and Interoperability of this special category of government services - differentiating themselves not because of a fundamental heterogeneity but only due to the particular nature of served beneficiaries. ● Although accessibility has been up to now considered one among other service quality issues, it is evident that it represents a relevant goal for migrant’s integration. The way to the adoption of a quality standard for public services, and more specifically for public service accessibility, is paved and the challenging perspective of integration is making it an urgent issue across EU countries. ● The absence of a shared accessibility standard makes the accessibility itself a deep differentiated quality of services creating disparities and inequalities across Europe. ● Differences in service laws and in-service governance models across EU countries make the adoption of a service accessibility standard a complex challenge: considerations are needed for the level of decentralization of service provision. ● Differences in service laws and in-service governance models across EU countries make the adoption of a service accessibility standard a complex challenge: considerations are needed for the level of decentralization of service provision. 44 44 ● Shared understanding and guidelines on citizens’ needs and expectations may improve the efficiency of the public sector across Europe and increase transparency and accountability so widening citizens’ knowledge about, and confidence in, public services. ● Existing service quality standards have been introduced and used discontinuously across Europe and developed with scarce knowledge of the users’ experience: accessibility should take into account the users’ experience for it to become an integration measure. ● New standards are required that are more focused on User Centricity and Interoperability of these services. Pitfalls Linguistic offer in service provision is crucial but does not represent the unique obstacle in service access. Discretionary in the interpretation of service accessibility and the ways to guarantee it across service sectors and across countries introduces disparities and inequalities and represents a significant obstacle to migrants’ integration. Bureaucratic complexity of service procedures represents an infringement to information rights and makes service access difficult and, more critical, open to failures. Differentiation in accessibility modes improves service quality but at the same time widens the bureaucratic complexity of the service thus challenging its accessibility. Compliance with basic human rights and human rights literacy represents a critical challenge in service provision especially when services do have deep impacts on citizenship and cascade effects on other services and rights. Policy Recommendations Improving the accessibility of public services is surely crucial to migrants’ integration in Europe; it also represents an opportunity to homogenize European citizenship and reduce the administrative stress of the public sector. The creation of a standard to assess service accessibility would reduce disparities and unequal legal implications consequent to eventual failures in administrative service and procedure accomplishment. In compliance with the User Centricity principle, such a standard could and should be a collection of minimum requirements for enforcing the informational rights of foreign migrants and refugees entering the European Union legally. It would provide a common basis for the harmonization of public services to migrants and refugees across Europe, under the perspectives of quality of communication and effectiveness of information delivery. It should embed at least the following three characteristics: coverage, coherence, and outreach. Coherence As services are interfaces between citizens and their rights, the assessment standard should be able to detect the internal contradictions among service delivery process components, which may establish conditions for possible discriminations to guarantee equal service access (and therefore access to rights) to all beneficiaries. Coverage The assessment standard should differentiate the various dimensions in which a service is articulated and provide a series of reference points (levels or steps) by which progress in accessing that service can be calibrated; also, it should be able to sense whether public services are designed to consider many possible use cases in different and complex situations including those that often occur when migrants are involved. 45 Outreach The assessment standard should verify that a service scope, rules, and procedures are clearly formulated, promptly available, and readily comprehensible to its prospective users; all the beneficiaries, migrants included, should be able to complete their service journeys in autonomy (without external support) avoiding the risk to fail due to lack of understanding. 46 The Migration06 cluster Editor’s contact: info@designscapes.eu These projects have received funding from the European Union’s Horizon 2020 research and innovation programme under the grant agreements No 822380, 822215, 822601, 822717, 870930 and 870980. The Migration06 cluster Editor’s contact: info@designscapes.eu Editor’s contact: info@designscapes.eu These projects have received funding from the European Union’s Horizon 2020 research and innovation programme under the grant agreements No 822380, 822215, 822601, 822717, 870930 and 870980. These projects have received funding from the European Union’s Horizon 2020 research and innovation programme under the grant agreements No 822380, 822215, 822601, 822717, 870930 and 870980.
https://openalex.org/W4389288438	https://liksprava.com/index.php/journal/article/download/1002/907	Russian	null	Влияние различных вариантов комбинированной фармакотерапии больных с хронической сердечной недостаточностью на показатели гемограммы	Lìkarsʹka sprava/Lìkarsʹka sprava	2,013	cc-by	3,067	НА ДОПОМОГУ ПРАКТИЧНОМУ ЛІКАРЕВІ НА ДОПОМОГУ ПРАКТИЧНОМУ ЛІКАРЕВІ УДК 616.12–008.46–002.2–059:615.3 Поступила 03.04.201 В. В. КУТОВОЙ, В. К. КАЗИМИРКО, Л. Н. ИВАНИЦКАЯ, Т. С. СИЛАНТЬЕВА, А. Г. ДУБКОВА (Киев) ВЛИЯНИЕ РАЗЛИЧНЫХ ВАРИАНТОВ КОМБИНИРОВАННОЙ ФАРМАКОТЕРАПИИ БОЛЬНЫХ С ХРОНИЧЕСКОЙ СЕРДЕЧНОЙ НЕДОСТАТОЧНОСТЬЮ НА ПОКАЗАТЕЛИ ГЕМОГРАММЫ Национальная медицинская академия последипломного образования им. П. Л. Шупика <liliya_2808@mail.ru> Поступила 03.04.2013 Поступила 03.04.2013 62 ISSN 0049-6804. «Лік. справа=Врачеб. дело». 2013, № 7 НА ДОПОМОГУ ПРАКТИЧНОМУ ЛІКАРЕВІ © Â. Â. Êóòîâîé, Â. Ê. Êàçèìèðêî, Ë. Í. Èâàíèöêàÿ, Ò. Ñ. Ñèëàíòüåâà, À.Ã. Äóáêîâà, 2013 В. В. Кутовой и др. Влияние различных вариантов комбинированной фармакотерапии ХСН 63 Цель исследования – выявить и клинически оценить влияние различных вариантов фармакотерапии больных с ХСН на выраженность анемии на основа- нии изучения уровня эритроцитов (эр.), гемоглобина (Hb), цветового показателя (ЦП), гематокритного числа (Ht) и коррелятивных взаимосвязей между ними и показателями, характеризующими функциональное состояние сердца, – конечно- диастолическим объёмом (КДО), фракцией выброса (ФВ), индексом массы мио- карда левого желудочка (ИММ ЛЖ), функциональным классом (ФК) ХСН, дис- танцией 6-минутного теста (Д6МТ). у ( ) Материалы и методы. В исследование вошло 66 человек в возрасте от 62 до 88 лет (29 мужчин и 37 женщин) с эссенциальной артериальной гипертензией (АГ) III степени и ишемической болезнью сердца (ИБС), ХСН IIА–IIБ стадии II–IV функционального класса (ФК) по NYHA и фракцией выброса левого же- лудочка (ФВ ЛЖ) < 45 %. Стадию ХСН и ФК определяли в соответствии с ре- комендациями Украинской ассоциации кардиологов по диагностике, лечению и профилактике сердечной недостаточности у взрослых [1]. В процессе диагности- ки использовали такие диагностические методы, как эхокардиографию (эхоКГ), рентгенографию грудной клетки, электрокардиограмму (ЭКГ) в 12 отведениях, холтеровское мониторирование ЭКГ (по показаниям), стандартный спектр лабо- раторных исследований, в частности общий анализ крови, количество эритроци- тов, лейкоцитов, тромбоцитов, гематокритное число, скорость оседания эритро- цитов – СОЭ. Рассчитывали уровень гемоглобина, коэффициенты корреляции между показателями, для статистической обработки результатов исследования использовали пакет статистических программ «STATISTICA for Windows 5.1». р р Результаты и их обсуждение. Результаты изучения влияния различных ва- риантов фармакотерапии больных с ХСН на уровень показателей гемограммы приведены в табл. 1. Таблица 1. Влияние различных вариантов фармакотерапии больных с хронической сердечной недостаточностью на показатели гемограммы (М ± m) Показатель Уровень средних показателей до и через 12 мес лечения І группа – ИАПФ + БАБ + Д (n = 34) ІІ группа – БРАII + БАБ + Д (n = 14) ІІІ группа – ИАПФ + БРАII + БАБ + Д(n = 18) Эр., · 1012 в 1 л до лечения 3,92 ± 0,24 3,95 ± 0,25 РІ > 0,5 3,89 ± 0,23 РІ > 0,5; РІІІ > 0,5 через 12 мес 4,12 ± 0,12 Р > 0,2 4,07 ± 0,25 Р > 0,5 РІ > 0,5 4,38 ± 0,13 Р > 0,05 РІ > 0,1; РІІ > 0,2 Эр. ВЛИЯНИЕ РАЗЛИЧНЫХ ВАРИАНТОВ КОМБИНИРОВАННОЙ ФАРМАКОТЕРАПИИ БОЛЬНЫХ С ХРОНИЧЕСКОЙ СЕРДЕЧНОЙ НЕДОСТАТОЧНОСТЬЮ НА ПОКАЗАТЕЛИ ГЕМОГРАММЫ Установлено, что различные варианты комбинированной фармакотерапии больных с хронической сердечной недостаточностью (ХСН) сопровождаются повышенным со- держанием эритроцитов, гемоглобина и цветового показателя, тенденцией к увеличению гематокритного числа. При этом наиболее выраженные достоверные сдвиги отмечены у больных, принимавших комбинацию ингибиторов ангиотензинпревращающего фермен- та, блокаторов ангиотензина ІІ, β-адреноблокаторов и диуретического средства. Выявлена прямая положительная коррелятивная связь между уровнем гемоглобина и дистанцией 6-минутного теста, обратная – между уровнем гемоглобина и конечно- диастолическим объёмом левого желудочка, функциональным классом (ФК) ХСН. Ко- личество эритроцитов в крови обратно коррелирует с ФК ХСН, прямо положительно – с дистанцией 6-минутного теста. Ключевые слова: хроническая сердечная недостаточность, фармакотерапия, гемо- грамма. Вступление. В последние годы установлено, что в патофизиологии хрони- ческой сердечной недостаточности (ХСН) провоспалительные цитокины при- водят к нарушению метаболизма железа, эритропоэза, развитию анемии [2–5], являющейся независимым фактором риска и неблагоприятного прогноза при ХСН. Исходя из этих данных, актуально выяснение характера влияния со- временной терапии больных с ХСН, включающей комбинацию ингибиторов ангиотензинпревращающего фермента (АПФ), блокаторов ангиотензина II (БРАII), β-адреноблокаторов (БАБ) и диуретиков (Д) [1], на показатели гемо- граммы. © Â. Â. Êóòîâîé, Â. Ê. Êàçèìèðêî, Ë. Í. Èâàíèöêàÿ, Ò. Ñ. Ñèëàíòüåâà, À.Ã. Äóáêîâà, 2013 В. В. Кутовой и др. Влияние различных вариантов комбинированной фармакотерапии ХСН (мужчины), ·1012 в 1 л до лечения 4,00 ± 0,19 3,96 ± 0,21 РІ > 0,5 3,98 ± 0,24 РІ > 0,5; РІІ > 0,5 через 12 мес 4,40 ± 0,14 Р > 0,05 4,05 ± 0,26 Р > 0,5 РІ > 0,2 4,50 ± 0,10 Р < 0,05 РІ > 0,5; РІІ > 0,1 Эр. (женщины), ·1012 в 1 л до лечения 3,87 ± 0,18 3,90 ± 0,21 РІ > 0,5 3,89 ± 0,22 РІ > 0,5; РІІ > 0,5 через 12 мес 3,85 ± 0,11 Р > 0,5 4,09 ± 0,24 Р > 0,5 РІ > 0,2 4,26 ± 0,16 Р > 0,1 РІ > 0,05; РІІ > 0,5 Hb, г/л до лечения 113,98 ± 3,67 114,20 ± 3,91 РІ > 0,5 113,82 ± 3,84 РІ > 0,5; РІІ > 0,5 через 12 мес 124,01 ± 4,43 Р < 0,05 124,70 ± 3,26 Р < 0,05 РІ > 0,5 132,46 ± 3,10 Р < 0,001 РІ > 0,1; РІІ > 0,05 Таблица 1. Влияние различных вариантов фармакотерапии больных с хронической сердечной недостаточностью на показатели гемограммы (М ± m) ISSN 0049-6804. «Лік. справа=Врачеб. дело». 2013, № 7 НА ДОПОМОГУ ПРАКТИЧНОМУ ЛІКАРЕВІ 64 Показатель Уровень средних показателей до и через 12 мес лечения І группа – ИАПФ + БАБ + Д (n = 34) ІІ группа – БРАII + БАБ + Д (n = 14) ІІІ группа – ИАПФ + БРАII + БАБ + Д(n = 18) Hb (мужчины), г/л до лечения 115,93 ± 4,17 116,28 ± 4,14 РІ > 0,5 115,86 ± 4,31 РІ > 0,5; РІІ > 0,5 через 12 мес 130,78 ± 3,98 Р < 0,01 129,28 ± 4,32 Р < 0,05 РІ > 0,5 138,33 ± 2,78 Р < 0,01 РІ > 0,1; РІІ > 0,05 Hb (женщины), г/л до лечения 110,84 ± 3,62 111,28 ± 3,52 РІ > 0,5 110,95 ± 2,46 РІ > 0,5; РІІ > 0,5 через 12 мес 117,25 ± 4,96 Р > 0,2 120,12 ± 2,68 Р < 0,05 РІ > 0,5 126,60 ± 3,44 Р < 0,001 РІ > 0,1; РІІ > 0,1 ЦП, ед. В. В. Кутовой и др. Влияние различных вариантов комбинированной фармакотерапии ХСН до лечепия 0,79 ± 0,04 0,77 ± 0,04 РІ > 0,5 0,78 ± 0,03 РІ > 0,5; РІІ > 0,5 через 12 мес 0,89 ± 0,01 Р < 0,02 0,88 ± 0,03 Р < 0,02 РІ > 0,5 0,86 ± 0,02 Р < 0,05 РІ > 0,1; РІІ > 0,5 Ht, % до лечения 34,84 ± 1,51 34,89 ± 1,69 РІ > 0,5 35,2 ± 1,7 РІ > 0,5; РІІ > 0,5 через 12 мес 36,52 ± 1,87 Р > 0,2 37,50 ± 2,68 Р > 0,2 РІ > 0,5 39,66 ± 2,23 Р > 0,5 РІ > 0,2; РІІ > 0,5 Ht (мужчины), % до лечения 36,12 ± 1,59 35,97 ± 1,66 РІ > 0,5 36,33 ± 1,58 РІ > 0,5; РІІ > 0,5 через 12 мес 38,53 ± 1,19 Р > 0,2 37,82 ± 2,80 Р > 0,5 РІ > 0,5 42,71 ± 1,92 Р < 0,02 РІ > 0,05; РІІ > 0,1 Ht (женщины), % до лечения 32,97 ± 1,63 33,28 ± 1,50 РІ > 0,5 33,44 ± 1,65 РІ > 0,5; РІІ > 0,5 через 12 мес 34,48 ± 1,76 Р > 0,5 37,16 ± 2,59 Р > 0,1 РІ > 0,2 36,60 ± 2,53 Р > 0,2 РІ > 0,5; РІІ > 0,5 Окончание табл. 1 Как видно из табл. 1, через 12 мес после лечения средние величины количества эритроцитов в группах (в том числе у мужчин и женщин) достоверно не отлича- лись как от исходных, так и между собой (в норме содержание эритроцитов у мужчин – (4–5) · 1012 в 1 л, у женщин – (3,7–4,5) · 1012 в 1 л). Через год диапазон колебаний индивидуальных показателей в І группе у женщин составил (2,28– 4,54) · 1012 в 1 л, у мужчин – (3,56–5,35)·1012 в 1 л, во ІІ – соответственно (3,5– 5,29) · 1012 в 1 л и (3,54–5,54) · 1012 в 1 л, в ІІІ – (3,19–5,1) · 1012 в 1 л и (4,18– 5,02) · 1012 в 1 л. Процент женщин в трёх группах, содержание эритроцитов у которых было ниже 3,7 · 1012 в 1 л, составил соответственно 30,43; 27,3 и 9,09. У мужчин І и ІІ групп показатели ниже 4,0·1012 в 1 л отмечены соответственно у 56,52 и 25 %, в ІІІ группе таких пациентов не было. В. В. Кутовой и др. Влияние различных вариантов комбинированной фармакотерапии ХСН Как у мужчин, так и у женщин терапия ХСН сопровождалась уменьшением количества пациентов с низким со- держанием эритроцитов в крови, увеличением среднего показателя по группе. Через год лечения в трёх группах средние результаты у мужчин превысили ис- 65 В. В. Кутовой и др. Влияние различных вариантов комбинированной фармакотерапии ХСН ходные соответственно на 10; 2,27 и 13,06 %. У женщин во ІІ и ІІІ группах по- казатели были выше по сравнению с контролем на 6,1; 4,87 и 9,51 %. Между собой средние данные в группах достоверно не различались. Более выраженные изменения, по сравнению с содержанием эритроцитов, отмечены в содержании гемоглобина (в норме уровень гемоглобина у мужчин составил 130–160 г/л, у женщин – 120–140 г/л). Терапия больных с ХСН сопро- вождалась увеличением средних значений гемоглобина: показатели пациентов трёх групп через год превышали контрольные соответственно на 8,8; 9,19 (Р < 0,05) и 16,38 % (Р < 0,001). У мужчин всех трёх групп через год средние величины со- держания гемоглобина превышали исходные данные соответственно на 12,81 (Р < 0,02); 11,18 (Р < 0,05) и 19,39 % (Р < 0,01). Положительная тенденция от- мечена и у женщин: во всех трёх группах средние результаты через год лечения были выше исходных соответственно на 5,78 (Р > 0,2), 7,94 (Р < 0,05) и 14,1 % (Р < 0,01). Анализ индивидуальных данных показал, что уровень гемоглобина ниже 120 г/л через год терапии наблюдался у 33,33 % пациентов І группы, у 40,9 % – ІІ и у 18,2 % – ІІІ. Все варианты терапии больных с ХСН сопрово- ждались уменьшением количества пациентов с низким (менее 120 г/л) содержа- нием гемоглобина в крови, увеличением средних значений в группах. Наиболее выраженные достоверные сдвиги отмечены в ІІІ группе (комбинация ИАПФ + БРАII + БАБ + Д). б С ( Лечение больных с ХСН сопровождалось достоверным увеличением ЦП (нор- ма 0,85–1,1): средняя величина в І группе через год терапии превысила исходную на 12,66 % (Р < 0,02), во ІІ – на 14,28 % (Р < 0,02), в ІІІ – на 10,26 % (Р < 0,05). ЦП ниже 0,85 сохранялся у 15,18 % больных І группы, у 35,71 % – ІІ и у 15 % – ІІІ. Таким образом, фармакотерапия больных с ХСН сопровождалась выраженным уменьшением количества пациентов со сниженным (< 0,85) значением ЦП. Наи- меньшее количество таких больных отмечено в І и ІІІ группах. ру Под влиянием 12-месячной терапии (см. табл. В. В. Кутовой и др. Влияние различных вариантов комбинированной фармакотерапии ХСН 1) тенденцию к увеличению имели средние результаты гематокритного числа (норма у мужчин 0,4–0,52 %, у женщин 0,36–0,42 %). Так, исходный показатель в І группе повысился на 4,82 %, во ІІ – на 7,48 %, в ІІІ – на 12,67 %. Такая же динамика отмечалась в подгруппах мужчин и женщин. Наиболее высокий уровень гематокритного числа (42,71 % ± 1,92 %) наблюдался через год у мужчин ІІІ группы – выше исходного на 17,56 % (Р < 0,05). Анализ индивидуальных данных показал, что, несмотря на 12-месячное лечение, величина гематокритного числа менее 36 имела место у 67,96 % больных І группы, у 50 % – ІІ и у 43,75 % – ІІІ. Нами не выявлено прямой коррелятивной зависимости между показателями гематокритного числа и гемоглобином (r = +0,005), а также гематокритного числа и содержанием сывороточного железа (r = +0,01). Мы изучали коррелятивные связи между показателями гемограммы, с одной стороны, функциональными показателями и данными эхоКГ – с другой (табл. 2). Таблица 2. Коэффициент корреляции между показателями гемограммы и некоторыми функциональными показателями и данными эхокардиографии Таблица 2. Коэффициент корреляции между показателями гемограммы и некоторыми функциональными показателями и данными эхокардиографии Показатель гемо- граммы Показатели функциональные и эхокардиограммы, коэффициент корреляции, r ИММ ЛЖ, г/м2 КДО, мл ФВ, % ФК ХСН Д6МТ, м Гемоглобин, г/л –0,076 Р > 0,5 –0,321 Р < 0,05 +0,180 Р > 0,1 –0,760 Р < 0,01 +0,701 Р < 0,01 Эритроциты, ·1012 в 1 л +0,073 Р > 0,5 –0,018 Р > 0,5 –0,057 Р > 0,5 –0,628 Р < 0,01 +0,441 Р < 0,05 Как видно из табл. 2, выявлена прямая положительная коррелятивная связь между уровнем гемоглобина и дистанцией 6-минутного теста (Д6МТ), обратная – между уровнем гемоглобина и КДО ЛЖ, ФК ХСН. Количество эритроцитов в крови обратно коррелирует с ФК ХСН, прямо положительно – с дистанцией 6-минутного теста. Таблица 2. Коэффициент корреляции между показателями гемограммы и некоторыми функциональными показателями и данными эхокардиографии Показатель гемо- граммы Показатели функциональные и эхокардиограммы, коэффициент корреляции, r ИММ ЛЖ, г/м2 КДО, мл ФВ, % ФК ХСН Д6МТ, м Гемоглобин, г/л –0,076 Р > 0,5 –0,321 Р < 0,05 +0,180 Р > 0,1 –0,760 Р < 0,01 +0,701 Р < 0,01 Эритроциты, ·1012 в 1 л +0,073 Р > 0,5 –0,018 Р > 0,5 –0,057 Р > 0,5 –0,628 Р < 0,01 +0,441 Р < 0,05 Как видно из табл. ВПЛИВ РІЗНИХ ВАРІАНТІВ КОМБІНОВАНОЇ ФАРМАКОТЕРАПІЇ ХВОРИХ З ХРОНІЧНОЮ СЕРЦЕВОЮ НЕДОСТАТНІСТЮ НА ПОКАЗНИКИ ГЕМОГРАМИ В. В. Кутовий, В. К. Казимирко, Л. М. Іваницька, Т. С. Силантьєва, А. Г. Дубкова (Київ) Встановлено, що різні варіанти комбінованної фармакотерапії хворих з хронічною сер- цевою недостатністю (ХСН) супроводжуються збільшенням вмісту еритроцитів, гемоглобіну в крові, кольорового показника, тенденцією до підвищення гематокритного числа. При цьому найбільш виражені достовірні зрушення відмічено у пацієнтів, які приймали комбінацію інгібіторів ангіотензинперетворюючого ферменту, блокаторів ангіотензину ІІ, β-адреноблока- торів і діуретичного засобу. Виявлено прямий позитивний корелятивний зв’язок між рівнем гемоглобіну та дистанцією 6-хвилинного тесту, обернений – між рівнем гемоглобіну та кінцево- діастолічним об’ємом лівого шлуночка, функціональним класом (ФК) ХСН. Кількість ери- троцитів в крові обернено корелює з ФК ХСН, прямо позитивно – з дистанцією 6-хвилинного тесту. Ключові слова: хронічна серцева недостатність, фармакотерапія, гемограма. С п и с о к л и т е р а т у р ы 1. Воронков Л. Г., Амосова К. М., Багрій А. Е. та ін. Рекомендації Асоціації кардіологів України з лікування хронічної серцевої недостатності у дорослих (перегляд 2011) // Серцева недостатність. – 2011. – № 1. – С. 101–116. 2. Дзяк Г. В., Василенко А. М., Шейко С. О., Василенко В. А. Кардіоренальний анемічний син- дром у хворих з ХСН // Там само. – 2010. – № 3. – С. 10–17. 3. Bolger A. P., Bartlett F. R., Penston H. S. et al. Intravenous iron alone for the treatment of anemia in patients with chronic heart failure // J. Am. Coll. Cardiol. – 2006. – Vol. 48, N 6. – P. 1225– 1227. 4. Crichton R. R., Danielson B. G., Geisser P. Iron therapy with special emphasis on intravenous adminis- tration / 3–rd ed. Bremen: UNI-Med. Verlag A. G. – 2006. – 96 p. / g p 5. Iversen P. O., Woldbaek P. R., Tonnessen T. et al. Decreased hematopoiesis in bone marrow of mice with congestive heart failure // Am. J. Physiol. Regul. Integr. Соmр. Physiol. – 2002. – Vol. 282. – P. 166–172. 5. Iversen P. O., Woldbaek P. R., Tonnessen T. et al. Decreased hematopoiesis in bone marrow of mice with congestive heart failure // Am. J. Physiol. Regul. Integr. Соmр. Physiol. – 2002. – Vol. 282. – P. 166–172. ВПЛИВ РІЗНИХ ВАРІАНТІВ КОМБІНОВАНОЇ ФАРМАКОТЕРАПІЇ ХВОРИХ З ХРОНІЧНОЮ СЕРЦЕВОЮ НЕДОСТАТНІСТЮ НА ПОКАЗНИКИ ГЕМОГРАМИ В. В. Кутовой и др. Влияние различных вариантов комбинированной фармакотерапии ХСН 2, выявлена прямая положительная коррелятивная связь между уровнем гемоглобина и дистанцией 6-минутного теста (Д6МТ), обратная – между уровнем гемоглобина и КДО ЛЖ, ФК ХСН. Количество эритроцитов в крови обратно коррелирует с ФК ХСН, прямо положительно – с дистанцией 6-минутного теста. Как видно из табл. 2, выявлена прямая положительная коррелятивная связь между уровнем гемоглобина и дистанцией 6-минутного теста (Д6МТ), обратная – между уровнем гемоглобина и КДО ЛЖ, ФК ХСН. Количество эритроцитов в крови обратно коррелирует с ФК ХСН, прямо положительно – с дистанцией 6-минутного теста. 66 ISSN 0049-6804. «Лік. справа=Врачеб. дело». 2013, № 7 НА ДОПОМОГУ ПРАКТИЧНОМУ ЛІКАРЕВІ Выводы. 1. Комбинированная фармакотерапия больных с ХСН сопровожда- ется увеличением содержания эритроцитов, уменьшением количества пациентов с низким содержанием эритроцитов и гемоглобина (< 120 г/л), со сниженным (< 0,85) уровнем ЦП, а также тенденцией к увеличению гематокритного числа. Наиболее выраженные достоверные сдвиги отмечены в ІІІ группе (комбинация ИАПФ + БРАII + БАБ + Д). 2. Выявлена прямая положительная коррелятивная связь между уровнем гемоглобина и дистанцией 6-минутного теста, обратная – между уровнем гемоглобина и КДО ЛЖ, ФК ХСН. Количество эритроцитов в крови обратно коррелирует с ФК ХСН, прямо положительно – с дистанцией 6-минутного теста. Выводы. 1. Комбинированная фармакотерапия больных с ХСН сопровожда- ется увеличением содержания эритроцитов, уменьшением количества пациентов с низким содержанием эритроцитов и гемоглобина (< 120 г/л), со сниженным Выводы. 1. Комбинированная фармакотерапия больных с ХСН сопровожда- ется увеличением содержания эритроцитов, уменьшением количества пациентов с низким содержанием эритроцитов и гемоглобина (< 120 г/л), со сниженным (< 0,85) уровнем ЦП, а также тенденцией к увеличению гематокритного числа. Наиболее выраженные достоверные сдвиги отмечены в ІІІ группе (комбинация д р р р ц ( 0 / ), (< 0,85) уровнем ЦП, а также тенденцией к увеличению гематокритного числа. Наиболее выраженные достоверные сдвиги отмечены в ІІІ группе (комбинация р р ру ( ИАПФ + БРАII + БАБ + Д). 2. Выявлена прямая положительная коррелятивная связь между уровнем гемоглобина и дистанцией 6-минутного теста, обратная – между уровнем гемоглобина и КДО ЛЖ, ФК ХСН. Количество эритроцитов в крови обратно коррелирует с ФК ХСН, прямо положительно – с дистанцией 6-минутного теста. EFFECT OF PHARMACOTHERAPY OF CHRONIC HEART FAILURE (CHF) ON THE PARAMETERS OF HEMOGRAM . Kazimirko, V. V. Kutovyi L. M. Ivanitska, T. S. Silantieva, A. G. Dubkova (Kiev, Ukraine) P. L. Shupyk National Medical Academy of Postgraduate Education Found that different versions of the combined drug therapy of chronic heart failure accompanied by an increase of erythrocytes, hemoglobin in the blood of patients, color index, tended to increase hematocrit. The most pronounced changes are marked valid in the group of patients receiving the combination of ACE inhibitors + ARB-II + beta-blockers + diuretics. Found to have a direct positive correlation between hemoglobin levels and distance 6-minute test, the back - between hemoglobin and end-diastolic volume of the left ventricle, the functional class of heart failure. The number of red blood cells inversely correlated with functional class of heart failure, a direct positive – with a distance of 6-minute test. Key words: chronic heart failure, pharmacotherapy, hemogram.
https://openalex.org/W2029540615	https://jm.copernicus.org/articles/21/137/2002/jm-21-137-2002.pdf	English	null	A comparison between smaller (&gt;63 μm) and larger (&gt;150 μm) planktonic foraminiferal faunas from the Pleistocene of ODP Site 1073 (Leg 174A), New Jersey margin, NW Atlantic Ocean	Journal of micropalaeontology	2,002	cc-by	8,428	INTRODUCTION juvenile specimens and to reduce the eﬀect of dissolution en- countered with ﬁne (>62 µm) size-fractions. Lu & Keller (1995) recommended using the >106 µm fraction, whilst Martin & Liddell (1989) proposed a ‘sieve method’ whereby multiple size-fractions are analysed for each sample. However, whatever size-fraction is used, in order to achieve reliable comparisons between foraminiferal faunas of diﬀerent workers only the same size-fraction should be considered. Planktonic foraminiferal faunas have been used extensively in reconstructing past oceanographic and climatic conditions. In the modern ocean, their distribution shows a distinct global latitudinal pattern, forming ﬁve major biogeographical provinces, which is closely related to the hydrographic features of oceanic water masses (e.g. Bé & Tolderlund, 1971; Bé, 1977; Bé & Hutson, 1977). The geographical and vertical distribution of planktonic foraminifera is principally related to their prefer- ences for speciﬁc temperatures and food requirements (e.g. Bé & Tolderlund, 1971; Kipp, 1976; Bé, 1977). This relationship has been applied to fossil faunas to understand past distribution and abundance patterns. The aim of this paper is to compare the distribution and palaeoecological signiﬁcance of Pleistocene planktonic foraminiferal faunas in the >63 µm and the >150 µm size- fractions from one site (ODP Site 1073, Leg 174A) in the northwest Atlantic Ocean. A detailed study of the palaeo- climatic and palaeoceanographic history of the Pleistocene of Site 1073 based on planktonic foraminifera in the >150 µm size-fraction is dealt with in a separate paper (Olson & Smart, in press). The use of particular size-fractions in foraminiferal analysis has important implications for the interpretation of modern and past environments. Various authors have suggested the use of speciﬁc size-fractions for the study of foraminifera, including >63 µm, >125 µm, >150 µm and even >250 µm. The CLIMAP (Climate, Long-Range Investigation, Mapping and Prediction) Project members (McIntyre et al., 1976; CLIMAP, 1981, 1984) recommended the use of the >149 µm size-fraction as a standard in palaeoceanographic and palaeoecological studies. This is because the smaller-sized planktonic foraminifera which are diﬃcult to identify and time consuming to count are eliminated from the sample (Imbrie & Kipp, 1971). However, in benthic foraminiferal studies, it has been argued (e.g. Schröder et al., 1987; Sen Gupta et al., 1987) that in order to include the large number of smaller-sized and ecologically signiﬁcant species in the analysis, the >63 µm size-fraction must be used. INTRODUCTION Following the recommendations of the CLIMAP project, many authors have employed the >149/150 µm size-fraction in planktonic foraminiferal studies. Kellogg (1984) argued in favour of using the >149 µm sieve size to avoid taxonomic problems with A comparison between smaller (>63 µm) and larger (>150 µm) planktonic foraminiferal faunas from the Pleistocene of ODP Site 1073 (Leg 174A), New Jersey margin, NW Atlantic Ocean CHRISTOPHER W. SMART Department of Geology, University of Plymouth, Drake Circus, Plymouth PL4 8AA, UK (e-mail: csmart@plymouth.ac.uk) ABSTRACT – Planktonic foraminiferal faunas have been studied from the Pleistocene of ODP Site 1073 (Leg 174A), New Jersey margin, NW Atlantic Ocean and their abundances have been compared in the >63 µm and >150 µm size-fractions from the same samples. Trends in the relative abundance of many species are similar in the two size-fractions, although the general level varies considerably. The mean abundance and ranges of Neogloboquadrina pachyderma (sinistral), N. pachyderma (dextral), Globorotalia inﬂata and Globigerina bulloides are greater in the >150 µm size-fraction compared with the >63 µm size-fraction. Turborotalita quinqueloba, Globigerinita uvula, G. glutinata, G. clarkei, and juvenile species are more abundant in the >63 µm size-fraction than the >150 µm size-fraction. Peaks (c. 60%) in abundance of G. uvula occur in the >63 µm size-fraction only, although the causes of these patterns are unclear. The data suggest that, in general, consistent palaeoclimatic/palaeoceanographic information is achieved by studying planktonic foraminiferal faunas from either size-fraction. However, because particular smaller species are either under-represented or even absent from the larger (>150 µm) size-fraction, the smaller (>63 µm) size-fraction must be included in studies of planktonic foraminifera. Furthermore, studies that involve planktonic foraminifera in the >63 µm size-fraction could provide diﬀerent transfer function estimates for sea surface temperatures in areas where workers have only used larger (>125 µm and >150 µm) size-fractions. J. Micropalaeontol. 21(2): 137–147, December 2002. Journal of Micropalaeontology, 21: 137–147. Journal of Micropalaeontology, 21: 137–147. 0262-821X/02 $15.00 2002 The Micropalaeontological Society Site 1073 Site 1073 (Hole 1073A) was drilled as part of the New Jersey Mid-Atlantic Sea-Level Transect in order to evaluate the eﬀects of ‘Icehouse’ glacial-eustatic change at a passive continental margin characterized by predominantly siliciclastic sedimen- tation (Austin et al., 1998). Site 1073 (39(13.5214#N, 72(16.5461#W; 639 m water depth) is located on the continental slope oﬀshore New Jersey (Fig. 1). A single hole (Hole 1073A) was drilled to a depth of 663.6 mbsf (metres below seaﬂoor). Core recovery at this site was excellent (99.9%) and a thick (519.8 m) Pleistocene sequence (<1.77 Ma) dominated by clays, silty clays, sandy muds and muddy sands was recovered (Austin 137 C. W. Smart Fig. 1. Location map of Site 1073. Bathymetric contours (in metres) are shown as dotted lines. Fig. 1. Location map of Site 1073. Bathymetric contours (in metres) are shown as dotted lines. 0 mbsf to 120 mbsf. Peaks in calcium carbonate are generally correlated with interglacial stages and opal to glacial stages (McHugh & Olson, 2002). et al., 1998). The thick Pleistocene sequence recovered from Site 1073, where planktonic foraminifera are common, provides a unique opportunity to study climatic change in detail in an upper slope setting. Sediment accumulation rates for the Pleistocene at Site 1073 vary from 25 cm ka1 (OIS 4) to 270 cm ka1 with the highest values occurring during glacial periods OIS 2 and 10 (McHugh & Olson, 2002). The high sedimentation rates recorded at Site 1073 are because of its close proximity (c. 20 km) to the Hudson Canyon Shelf Valley System. The stratigraphy of the Pleistocene succession of Site 1073 is summarized in Figure 2. Age control is provided by nannofossil biostratigraphy (Wei, 2001) and oxygen isotope stratigra- phy (McHugh & Olson, 2002). A radiocarbon age of 14,400 55 years was recorded at 0.40 mbsf suggesting that the Holocene is either very thin or missing (McHugh & Olson, 2002). The section from c. 0–30 mbsf is represented by a sandy debris ﬂow (McHugh & Olson, 2002). Sample preparation and data collection Sediment samples were examined for planktonic foraminifera at intervals varying from c. 1–5 m above c. 100 mbsf and at less frequent intervals from c. 100 mbsf to 434.47 mbsf. Each of the 20 cm3 samples were split into two sub-samples of 10 cm3. Each sub-sample was dried, weighed, disaggregated using a 10% sodium hexametaphosphate solution, and then washed over a 63 µm sieve. The washed sub-samples were dried and re-weighed to calculate the percentage of coarse fraction (percentage of sediment >63 µm). For each interval, one of the sub-samples was studied for foraminifera in the >63 µm size-fraction, whilst the other was analysed for foraminifera in the >150 µm size- fraction. For analysis of the >150 µm size-fraction, the >63 µm portion was dry-sieved using a 150 µm sieve. For each of the sub-samples, the fractions were subdivided using a modiﬁed Otto microsplitter until aliquots of approximately 300 plank- tonic foraminiferal specimens were attained for foraminiferal analysis. Specimens were picked and mounted on gummed slides for identiﬁcation and the proportions of the various species were Investigations of multichannel seismic data on the slope and adjacent continental shelf of the NW Atlantic Ocean indicate the presence of four Pleistocene sequence boundaries labelled pp1, pp2, pp3 and pp4 (Mountain et al., 1994; Austin et al., 1998). The positions of sequence boundaries pp1, pp2 and pp3 are shown in Figure 2. The base of the Pleistocene section is marked by a surface of discontinuity (pp4) at 520 mbsf which is just below the section sampled for isotopic analysis (McHugh & Olson, 2002). These boundaries are inferred to reﬂect phases of rapid sea-level rise, drowning of the shelf, and subsequent sediment starvation on the slope. Predominantly oﬀ-lapping reﬂectors beneath the slope reﬂect enhanced oﬀshore sediment transport and margin progradation during periods of falling sea-level (Savrda et al., 2001). The percentages of calcium carbonate (B. Balsam & J. Damuth, unpublished data) and opal (McHugh & Olson, 2002) reveal an inverse correlation that is especially evident from 138 Pleistocene planktonic foraminifera from the NW Atlantic Ocean Pleistocene planktonic foraminifera from the NW Atlantic Ocean Fig. 2. The stratigraphy of Site 1073 based on nannofossil biostratigra- phy (Wei, 2001), radiocarbon ages and oxygen stable isotope stratigra- phy (McHugh & Olson, 2002). Numbers refer to Oxygen Isotope Stages (OIS) with the glacials shaded. Sample preparation and data collection The positions of sequence boundaries pp1, pp2 and pp3 (Austin et al., 1998) are shown with wavy lines. Note E. huxleyi=Emiliania huxleyi, P. lacunosa=Pseudoemiliania lacunosa, F.O., ﬁrst occurrence; L.O., last occurrence; mbsf, metres below seaﬂoor. Globigerinita glutinata. Globigerinita glutinata mainly includes specimens similar to that illustrated in Kennett & Srinivasan (1983, ﬁg. 3, plate 56). Some specimens have a bulla and some specimens have chambers that increase more rapidly in size with the ultimate chamber being much larger than the penultimate chamber. Globigerinita glutinata was distinguished from Glo- bigerina bulloides mainly by the wall structure. Globigerinita glutinata has a smooth to ﬁnely hispid and non-spinose test surface with ﬁne pores whereas G. bulloides shows a hispid, densely perforate surface and usually has a more sugary appear- ance. Some specimens of G. bulloides have a very small aperture but are consistent with the surface texture of G. bulloides. calculated. The association between relative abundances was measured using correlation coeﬃcients. Diﬀerences between relative abundances of diﬀerent size-fractions were tested for signiﬁcance using the Wilcoxon Test for paired data. For a number of samples, the weight of the picked portion was determined in order to calculate the number of foraminifera per gram of dry sediment. The Planktonic Fragmentation Index (PFI) was recorded in the >150 µm size-fraction and was calcu- lated as the percentage of test fragments relative to whole planktonic foraminifera+test fragments. Species diversity was recorded in terms of the alpha index (Fisher et al., 1943) using the graph in Murray (1991). Globigerina bulloides. In this study, G. bulloides includes a few specimens which could be considered in the Globigerina falconensis and G. umbilicata group, although a clear distinction was not apparent in the samples analysed from Site 1073. Globorotalia spp. (other). The Globorotalia spp. (other) includes specimens which could not be clearly identiﬁed because many were juveniles. Sample preparation and data collection The positions of sequence boundaries pp1, pp2 and pp3 (Austin et al., 1998) are shown with wavy lines. Note E. huxleyi=Emiliania huxleyi, P. lacunosa=Pseudoemiliania lacunosa, F.O., ﬁrst occurrence; L.O., last occurrence; mbsf, metres below seaﬂoor. Srinivasan (1983). A faunal reference list is given in the Appendix. A number of speciﬁc taxonomic concepts have been incorporated into this study and details of these are given below. Srinivasan (1983). A faunal reference list is given in the Appendix. A number of speciﬁc taxonomic concepts have been incorporated into this study and details of these are given below. Neogloboquadrina pachyderma and N. dutertrei. Both dextral (d) and sinistral (s) forms of N. pachyderma were recognized and their proportions were calculated. Transitional forms (P–D intergrade of Kipp, 1976) between N. pachyderma (d) and Neogloboquadrina dutertrei could not be consistently distin- guished and were thus combined with N. pachyderma (d). Forms considered to be N. dutertrei were typically relatively large in size, comprised at least ﬁve chambers in the ﬁnal whorl, had a wide umbilicus and a high spire (e.g. Giraudeau, 1993). Forms with four or four-and-a-half chambers and a small umbilicus, with an aperture often bordered by a lip, were assigned to N. pachyderma. Turborotalita quinqueloba. In this study, Turborotalita quinque- loba includes Turborotalita quinqueloba (sensu stricto), Globi- gerina exumbilicata and Globigerinita clarkei. Globigerina exumbilicata could not be consistently distinguished from T. quinqueloba in the samples studied, although it was considered to be rare. It also proved diﬃcult to separate consistently the occasionally reasonably abundant (up to 10%) G. clarkei from T. quinqueloba in the samples studied particularly where preser- vation was poor. It was, therefore, decided to include G. clarkei within the T. quinqueloba group. However, preliminary work carried out by the author has shown that it is possible to distinguish these taxa with reasonable consistency using the SEM. Globigerinita clarkei is consistently smaller than T. quin- queloba, its test is more compressed, chambers are less inﬂated, intercameral sutures are less incised, pores are larger and it has a thicker wall (Rögl & Bolli, 1973; Boltovskoy, 1977, 1991). According to Boltovskoy (1977), G. clarkei should be placed in the genus Globigerinita and this is followed here. Fig. 2. The stratigraphy of Site 1073 based on nannofossil biostratigra- phy (Wei, 2001), radiocarbon ages and oxygen stable isotope stratigra- phy (McHugh & Olson, 2002). Numbers refer to Oxygen Isotope Stages (OIS) with the glacials shaded. RESULTS A total of 53 sediment samples was studied for planktonic foraminifera from ODP Site 1073. Only 31 samples yielded suﬃcient specimens in both size-fractions (>63 µm and >150 µm) for analysis and interpretation. A number of samples were found to contain too few specimens or were barren of foraminifera. Table 1 lists the number of samples studied together with their sub-bottom depths. Seven species dominate the planktonic foraminiferal assem- blage in the >63 µm size-fraction in the studied interval of Site 1073: T. quinqueloba (mean 55.1%), G. uvula (14.5%), N. pachyderma (d) (9.1%), N. pachyderma (s) (8.1%), G. glutinata (6.0%), G. inﬂata (2.2%) and G. bulloides (2.1%). In the >150 µm size-fraction, apart from G. uvula which is rare/absent (0.1%), these species are also signiﬁcant components of the assemblage, although their mean abundances diﬀer: N. pachyderma (d) (34.1%), T. quinqueloba (19.9%), N. pachyderma (s) (18.7%), G. bulloides (11.3%), G. inﬂata (7.7%), and G. glutinata (4.0%). Correlation coeﬃcients and signiﬁcance of diﬀerences (Wilcoxon Test) of species between the two size-fractions are given in Table 2. There is overwhelming evidence of diﬀerences (p= 0.000) between the two size-fractions in the abundances of N. pachyderma (s), N. pachyderma (d), G. bulloides, T. quinque- loba, G. uvula and G. inﬂata, and strong evidence of a diﬀerence (p=0.009) in the abundance of G. glutinata. Neogloboquadrina pachyderma (s), N. pachyderma (d), G. bulloides and G. inﬂata are more abundant in the >150 µm compared with the >63 µm size-fraction whereas, T. quinqueloba, G. uvula and G. glutinata are more abundant in the >63 µm compared with the >150 µm size-fraction (see Figs 3 and 4). A comparison of species abundances between the two size-fractions show that there is strong positive correlation of variations in N. pachyderma (s) (r=0.889, p=0.000), G. bulloides (r=0.801, p=0.000) and N. pachyderma (d) (r=0.705, p=0.000) (Table 2). Weaker positive correlations are also evident in T. quinqueloba (r=0.506, p=0.004), G. uvula (r=0.431, p=0.016) and G. glutinata (r=0.385, p=0.032). There is little/no correlation in the abundance of G. inﬂata (r=0.104, p=0.579) between the two size-fractions. The results of the correlation coeﬃcients and signiﬁcance of diﬀer- ences for the other planktonic foraminifera are less clear due mainly to the fact that these taxa occur rarely (<5%) in the samples studied. Large ﬂuctuations occur in the relative abundance of T. quinqueloba and N. pachyderma (d) throughout the studied interval in both size-fractions (Fig. 3). Taxonomy rare 19H-3, 55-57 171.55 x 20H-3, 56-58 181.02 x 21H-3, 56-58 188.54 x 22H-3, 56-58 198.16 x 23H-3, 56-58 206.36 x 24H-1, 56-58 212.36 v. rare 24H-3, 56-58 215.32 v. rare 25X-3, 56-58 219.26 v. rare 26X-3, 56-58 227.96 v. rare 26X-4, 56-58 229.46 v. rare 28X-3, 56-58 247.06 x 32X-3, 55-57 284.89 v. rare 34X-3, 64-66 303.74 v. rare 36X-3, 56-58 321.96 x 38X-3, 55-57 340.75 x 40X-3, 55-57 359.55 x 46X-3, 56-58 414.86 x 48X-3, 57-59 434.47 x Samples represented by crosses (‘x’) contained suﬃcient numbers of foraminifera and the data are presented in this study. 140 Taxonomy For the most part, planktonic foraminifera were identiﬁed following the suggestions of Saito et al. (1981) and Kennett & 139 C. W. Smart RESULTS A total of 53 sediment samples was studied for planktonic foraminifera from ODP Site 1073. Only 31 samples yielded suﬃcient specimens in both size-fractions (>63 µm and >150 µm) for analysis and interpretation. A number of samples were found to contain too few specimens or were barren of foraminifera. Table 1 lists the number of samples studied together with their sub-bottom depths. Seven species dominate the planktonic foraminiferal assem- blage in the >63 µm size-fraction in the studied interval of Site 1073: T. quinqueloba (mean 55.1%), G. uvula (14.5%), N. pachyderma (d) (9.1%), N. pachyderma (s) (8.1%), G. glutinata (6.0%), G. inﬂata (2.2%) and G. bulloides (2.1%). In the >150 µm size-fraction, apart from G. uvula which is rare/absent (0.1%), these species are also signiﬁcant components of the assemblage, although their mean abundances diﬀer: N. pachyderma (d) (34.1%), T. quinqueloba (19.9%), N. pachyderma (s) (18.7%), G. bulloides (11.3%), G. inﬂata (7.7%), and G. glutinata (4.0%). Graphs showing abundance ﬂuctuations of the seven most abundant species are shown in Figures 3 and 4. These are species that exceed 10% in two or more samples in either size-fractions. On the whole, the trends in the relative abun- dances of many species are similar in the two size-fractions, although the general level varies considerably between the two size-fractions. Correlation coeﬃcients and signiﬁcance of diﬀerences (Wilcoxon Test) of species between the two size-fractions are given in Table 2. There is overwhelming evidence of diﬀerences (p= 0.000) between the two size-fractions in the abundances of N. pachyderma (s), N. pachyderma (d), G. bulloides, T. quinque- loba, G. uvula and G. inﬂata, and strong evidence of a diﬀerence (p=0.009) in the abundance of G. glutinata. Neogloboquadrina pachyderma (s), N. pachyderma (d), G. bulloides and G. inﬂata are more abundant in the >150 µm compared with the >63 µm size-fraction whereas, T. quinqueloba, G. uvula and G. glutinata are more abundant in the >63 µm compared with the >150 µm size-fraction (see Figs 3 and 4). A comparison of species abundances between the two size-fractions show that there is strong positive correlation of variations in N. pachyderma (s) (r=0.889, p=0.000), G. bulloides (r=0.801, p=0.000) and N. pachyderma (d) (r=0.705, p=0.000) (Table 2). Weaker positive correlations are also evident in T. Taxonomy quinqueloba (r=0.506, p=0.004), G. uvula (r=0.431, p=0.016) and G. glutinata (r=0.385, p=0.032). There is little/no correlation in the abundance of G. inﬂata (r=0.104, p=0.579) between the two size-fractions. The results of the correlation coeﬃcients and signiﬁcance of diﬀer- ences for the other planktonic foraminifera are less clear due mainly to the fact that these taxa occur rarely (<5%) in the samples studied. Large ﬂuctuations occur in the relative abundance of T. quinqueloba and N. pachyderma (d) throughout the studied interval in both size-fractions (Fig. 3). The abundance of N. pachyderma (s) shows moderate variation in both size-fractions (Fig. 3). Peaks in abundance of N. pachyderma (s) in the >63 µm size-fraction are consistent with peaks in abundance in the >150 µm size-fraction and occur during Oxygen Isotope Stages (OIS) 2, 4, 12, and 16(?). Fluctuations in the abundance of G. inﬂata are evident particularly in the >150 µm size-fraction, although, for the most part, values rarely exceed 15% (Fig. 4). Higher abundances of G. bulloides are recorded in both size-fractions from the interval 37.09–65.63 mbsf, with peaks occurring in OIS 3 and 4 (Fig. 4). Globigerinita glutinata typically has low abundances (<10%) in both size-fractions throughout the studied interval (Fig. 4). A number of peaks in abundance (up to 63.7%) of G. uvula occur Table 1. Samples studied from ODP Site 1073 (Hole 1073A) Samples Sub-bottom depth Samples Comments (>63 µm) (m) (>150 µm) 1H-1, 55-57 0.55 x 1H-3, 56-58 3.56 x 1H-5, 14-16 6.14 barren 2H-3, 55-57 10.15 barren 2H-5, 55-57 13.15 barren 3H-3, 55-57 19.45 barren 4H-3, 55-57 29.01 barren 4H-5, 55-57 31.95 barren 5H-2, 49-51 37.09 x 5H-3, 116-118 39.26 x 5H-4, 56-58 40.16 x 6H-3, 56-58 47.39 x 6H-4, 55-57 48.88 x 6H-5, 55-57 50.38 x 6H-6, 48-50 51.81 x 7H-1, 55-57 54.65 x 7H-3, 56-58 57.66 x 7H-5, 48-50 60.58 x 8H-1, 55-57 64.15 x 8H-2, 53-55 65.63 x 8H-3, 56-58 67.16 x 8H-4, 54-56 68.64 x 10H-1, 55-57 83.15 v. rare 10H-2, 55-57 84.65 v. rare 10H-3, 55-57 86.15 v. rare 10H-7, 55-57 92.15 v. rare 11H-5, 55-57 98.65 v. rare 12H-2, 55-57 103.65 v. rare 12H-3, 55-57 105.15 v. rare 13H-2, 55-57 113.15 v. rare 14H-3, 56-58 124.16 x 15H-3, 55-57 133.65 x 16H-3, 55-57 143.15 x 17H-3, 55-57 152.65 x 18H-3, 55-57 162.15 v. Pleistocene planktonic foraminifera from the NW Atlantic Ocean Pleistocene planktonic foraminifera from the NW Atlantic Ocean Fig. 3. Relative abundance variations of Turborotalita quinqueloba, Neogloboquadrina pachyderma (d), and Neogloboquadrina pachyderma (s) in the >63 µm size-fraction (dashed lines) and >150 µm size-fraction (solid lines) plotted versus sub-bottom depth (m) at Site 1073. Also shown is the stratigraphy with glacial OIS numbered and shaded (McHugh & Olson, 2002), together with the positions of sequence boundaries pp1, pp2 and pp3 (Austin et al., 1998). Fig. 3. Relative abundance variations of Turborotalita quinqueloba, Neogloboquadrina pachyderma (d), and Neogloboquadrina pachyderma (s) in the >63 µm size-fraction (dashed lines) and >150 µm size-fraction (solid lines) plotted versus sub-bottom depth (m) at Site 1073. Also shown is the stratigraphy with glacial OIS numbered and shaded (McHugh & Olson, 2002), together with the positions of sequence boundaries pp1, pp2 and pp3 (Austin et al., 1998). abundances of T. quinqueloba (>63 µm: r=0.338, p=0.063; >150 µm: r=0.362, p=0.045). abundances of T. quinqueloba (>63 µm: r=0.338, p=0.063; >150 µm: r=0.362, p=0.045). in the >63 µm size-fraction, although this species is rare/absent in the >150 µm size-fraction (Fig. 4). Peaks are recorded during OIS 2, 3, 4, 5, 9, and 15(?). DISCUSSION Si h k Fluctuations in the alpha index are shown in Figure 5. Although the means and ranges of the alpha index are the same in both size-fractions (2.1 and 1.0–3.5 respectively, >63 µm n=31, >150 µm n=28), there is little in the way of correlation between the trends (r=0.204, p=0.297) and there is no evidence of a diﬀerence in the values (p=0.952). Since the work of the CLIMAP project (McIntyre et al., 1976; CLIMAP, 1981, 1984), the >149/150 µm size-fraction has com- monly been taken as a standard in studies of planktonic foraminifera. Most workers are in agreement that the use of the coarser size-fraction avoids taxonomic problems encountered with smaller-sized specimens. However, it has been argued (e.g. Schröder et al., 1987) that in order to include the large number of smaller-sized specimens which may be ecologically important, only the >63 µm size-fraction should be used in the analysis. The use of including smaller planktonic foraminifera (63–150 µm) in palaeoceanographic studies has clearly been shown by a number of workers including Marquard & Clark (1987) and Bauch (1994). In a recent study of the glacial–interglacial palae- oceanography of the polar North Atlantic, Kandiano & Bauch (2002) clearly demonstrated the advantage of studying plank- tonic foraminifera in smaller size-fractions (80–150 µm). Carstens et al. (1997) have demonstrated the value of studying smaller living forms (>63 µm) in net tow samples from the Fram Strait, Arctic Ocean. Their study revealed that signiﬁcant diﬀer- ences exist in the species composition and abundances between the 63 µm mesh and previous studies from the same area which used meshes of 150 µm and 200 µm. They calculated that only 10–30% of the assemblage was recovered in studies using the larger (i.e. >150 µm and >200 µm) mesh sizes compared with the smaller (>63 µm) mesh size. The concentration (number per gram of sediment) of plank- tonic foraminifera is shown in Figure 5. Although data were not calculated for all the studied samples (particularly the >150 µm size-fraction), there is strong evidence (p=0.001) that the number of planktonic foraminifera per gram is higher in the >63 µm compared with the >150 µm size-fraction. In the >63 µm size-fraction, the mean number per gram is 360, with a range of 18–3422 (n=30), compared with the >150 µm size- fraction where the mean is 30 and range is 2–80 (n=13). RESULTS The abundance of N. pachyderma (s) shows moderate variation in both size-fractions (Fig. 3). Peaks in abundance of N. pachyderma (s) in the >63 µm size-fraction are consistent with peaks in abundance in the >150 µm size-fraction and occur during Oxygen Isotope Stages (OIS) 2, 4, 12, and 16(?). Table 1. Samples studied from ODP Site 1073 (Hole 1073A) part, values rarely exceed 15% (Fig. 4). Higher abundances of G. bulloides are recorded in both size-fractions from the interval 37.09–65.63 mbsf, with peaks occurring in OIS 3 and 4 (Fig. 4). Globigerinita glutinata typically has low abundances (<10%) in both size-fractions throughout the studied interval (Fig. 4). A number of peaks in abundance (up to 63.7%) of G. uvula occur Fluctuations in the abundance of G. inﬂata are evident particularly in the >150 µm size-fraction, although, for the most 140 Pleistocene planktonic foraminifera from the NW Atlantic Ocean DISCUSSION Si h k There is a good correlation of peaks in the number of planktonic foraminifera per gram between the two size-fractions (r=0.775, p=0.002). The planktonic fragmentation index (PFI) was calculated for the >150 µm size-fraction and the results are shown in Figure 5. Peaks in the percentage of fragments occur during OIS 2, 9, and 10. The peak in fragments (42.2%) at 152.65 mbsf correlates with the large peak in the total number of planktonic foramini- fera (>63 µm) per gram (3422 g1). In general, increases in % fragments are associated with decreases in the abundance of T. quinqueloba and low % fragments correlate with high 141 C. W. Smart Fig. 4. Relative abundance variations of Globorotalia inﬂata, Globigerina bulloides, Globigerinita glutinata, and Globigerinita uvula in the >63 µm size-fraction (dashed lines) and >150 µm size-fraction (solid lines) plotted versus sub-bottom depth (m) at Site 1073. Also shown is the stratigraphy with glacial OIS numbered and shaded (McHugh & Olson, 2002), together with the positions of sequence boundaries pp1, pp2 and pp3 (Austin et al., 1998). Fig. 4. Relative abundance variations of Globorotalia inﬂata, Globigerina bulloides, Globigerinita glutinata, and Globigerinita uvula in the >63 µm size-fraction (dashed lines) and >150 µm size-fraction (solid lines) plotted versus sub-bottom depth (m) at Site 1073. Also shown is the stratigraphy with glacial OIS numbered and shaded (McHugh & Olson, 2002), together with the positions of sequence boundaries pp1, pp2 and pp3 (Austin et al., 1998). Fig. 4. Relative abundance variations of Globorotalia inﬂata, Globigerina bulloides, Globigerinita glutinata, and Globigerinita uvula in the >63 µm size-fraction (dashed lines) and >150 µm size-fraction (solid lines) plotted versus sub-bottom depth (m) at Site 1073. Also shown is the stratigraphy with glacial OIS numbered and shaded (McHugh & Olson, 2002), together with the positions of sequence boundaries pp1, pp2 and pp3 (Austin et al., 1998). Fig. 4. Relative abundance variations of Globorotalia inﬂata, Globigerina bulloides, Globigerinita glutinata, and Globigerinita uvula in the >63 µm size-fraction (dashed lines) and >150 µm size-fraction (solid lines) plotted versus sub-bottom depth (m) at Site 1073. Also shown is the stratigraphy with glacial OIS numbered and shaded (McHugh & Olson, 2002), together with the positions of sequence boundaries pp1, pp2 and pp3 (Austin et al., 1998). bundance variations of Globorotalia inﬂata, Globiger ction (solid lines) plotted versus sub-bottom depth ( 142 Pleistocene planktonic foraminifera from the NW Atlantic Ocean Table 2. DISCUSSION Si h k Correlation coeﬃcients and signiﬁcance of diﬀerences (Wilcoxon Test) of relative abundances of planktonic foraminifera between the two size-fractions (>63 µm and >150 µm) at Site 1073 Correlation coeﬃcient between size-fractions Signiﬁcance Wilcoxon Test of paired diﬀerences Number (r) (p) (p) (n) N. pachyderma (s) 0.889 0.000 0.000 31 G. bulloides 0.801 0.000 0.000 31 N. pachyderma (d) 0.705 0.000 0.000 31 G. puncticulata 0.681 0.000 0.026 31 G. scitula 0.605 0.000 0.123 31 T. quinqueloba 0.506 0.004 0.000 31 G. ruber 0.494 0.005 0.024 31 G. uvula 0.431 0.016 0.000 31 G. glutinata 0.385 0.032 0.009 31 O. universa 0.175 0.346 0.063 31 G. inﬂata 0.104 0.579 0.000 31 G. pyramidalis 0.053 0.779 0.262 31 Globorotalia spp. (other) 0.002 0.990 0.374 31 G. truncatulinoides 0.091 0.627 0.013 31 G. sacculifer 0.107 0.566 0.314 31 G. cyclostomus 0.145 0.436 0.004 31 N. dutertrei 0.159 0.394 0.001 31 G. hirsuta — — 0.109 31 G. menardii — — 0.317 31 Globigerina bulloides includes G. bulloides, G. falconensis and G. umbilicata. Turborotalita quinqueloba includes T. quinqueloba, G. exumbilicata and G. clarkei. Globigerina bulloides includes G. bulloides, G. falconensis and G. umbilicata. Turborotalita quinqueloba includes T. quinqueloba, G. exumbilicata and G. clarkei. Table 2. Correlation coeﬃcients and signiﬁcance of diﬀerences (Wilcoxon Test) of relative abundances of planktonic foraminifera between the two size-fractions (>63 µm and >150 µm) at Site 1073 Table 2. Correlation coeﬃcients and signiﬁcance of diﬀerences (Wilcoxon Test) of relative abundances of planktonic foraminifera between the two size-fractions (>63 µm and >150 µm) at Site 1073 Fig. 5. Variations in the alpha index and concentration (number per gram of sediment) of planktonic foraminifera in the >63 µm size-fraction (dashed lines) and >150 µm size-fraction (solid lines), and planktonic fragmentation index (PFI) of the >150 µm size-fraction, all plotted versus sub-bottom depth (m) at Site 1073. Also shown is the stratigraphy with glacial OIS numbered and shaded (McHugh & Olson, 2002), together with the positions of sequence boundaries pp1, pp2 and pp3 (Austin et al., 1998). Fig. 5. Variations in the alpha index and concentration (number per gram of sediment) of planktonic foraminifera in the >63 µm size-fraction (dashed lines) and >150 µm size-fraction (solid lines), and planktonic fragmentation index (PFI) of the >150 µm size-fraction, all plotted versus sub-bottom depth (m) at Site 1073. DISCUSSION Si h k Also shown is the stratigraphy with glacial OIS numbered and shaded (McHugh & Olson, 2002), together with the positions of sequence boundaries pp1, pp2 and pp3 (Austin et al., 1998). In a recent study of size distributions of planktonic foramini- fera from the Arabian Sea, Peeters et al. (1999) recommend the use of the >150 µm size-fraction for standard faunal analysis because most species have reached their adult stage of ontogeny in this size-fraction. However, they also suggest that the >125 µm and >250 µm have to be used as well so that a reliable 143 C. W. Smart C. W. Smart estimate of the small and large species respectively can be determined. G. bulloides at Site 1073 during OIS 3 and 4 possibly indicates that increased upwelling and nutrient levels occurred during both interglacial and glacial stages. In this study, particular species are either under-represented or even absent from the larger size-fraction (>150 µm) whilst they may have high abundances in the smaller size-fraction (>63 µm). In addition, higher proportions of juvenile species occur in the smaller size-fraction. The most signiﬁcant species which occurs almost entirely in the >63 µm size-fraction is G. uvula. At certain intervals it comprises up to c. 60% of the total planktonic foraminiferal assemblage in the >63 µm size- fraction. In the >150 µm size-fraction, it has very low abun- dances and is commonly absent. Globigerinita uvula is a small species which is generally considered to be associated with colder/subpolar waters (e.g. Bé & Tolderlund, 1971; Bé & Hutson, 1977) and, together with T. quinqueloba, is relatively more susceptible to dissolution than N. pachyderma (s), G. inﬂata, G. truncatulinoides, G. bulloides, G. glutinata and N. pachyderma (d) (Malmgren, 1983). Although Niebler & Gersonde (1998) noted low numbers of G. uvula in the >125 µm size-fraction from all areas of the Antarctic Circumpolar Front, southern South Atlantic Ocean, they found a distinct maximum in abundance at c. 7 (C representing the average summer surface water temperature at the Subantarctic Front. Recently, Boltovskoy et al. (2000) reported high abundances of the typically ‘cold water’ forms, G. bulloides, T. quinqueloba and G. uvula, from the southwestern Atlantic occurring in water tem- peratures as high as 20–24 (C. Boltovskoy et al. (2000) suggest that this casts doubt on the use of these species as indicators of cool waters. The causes of the abundance peaks of G. DISCUSSION Si h k uvula at Site 1073 are not clear as there appears to be no consistent correlation with the abundances of other taxa, the oxygen stable isotope record, and dissolution reﬂected in the PFI. Turborotalita quinqueloba is signiﬁcantly more abundant in the >63 µm size-fraction than in the >150 µm size-fraction. This is not surprising since it is well known to be a small species which commonly dominates in size-fractions of <150 µm (Kellogg, 1984; Bauch, 1994; Peeters et al., 1999; Kandiano & Bauch, 2002). In a study of size variations in T. quinqueloba from the LGM of the Norwegian–Greenland Sea, Bauch (1994) noted the palaeoceanographic signiﬁcance of T. quinqueloba and found that the main abundances occur in the <149 µm size- fraction. At Site 1073, there is a relationship between PFI values (% fragments) and relative abundances of T. quinqueloba; i.e. increases in PFI are associated with decreases in % T. quinque- loba and decreases in PFI are associated with increases in % T. quinqueloba. It has been suggested by a number of workers (e.g. Thunell, 1975; Le & Shackleton, 1992) that the PFI is a reliable indicator of carbonate dissolution and T. quinqueloba is susceptible to dissolution (e.g. Malmgren, 1983). Thus, the abundance of T. quinqueloba at Site 1073 can be explained, at least in part, by changes in the amount of carbonate dissolution (reﬂected in the PFI). In this study, G. clarkei is included in the T. quinqueloba group. Although readily distinguished under the SEM, these two species proved diﬃcult to consistently separate under a light microscope and were, therefore, grouped together. For the most part, G. clarkei has only been reported rarely in the literature. However, as pointed out by Boltovskoy (1977, 1990, 1991) this is because most foraminiferal workers use coarse size-fractions (>125 µm and >150 µm) and because of its very small size (maximum diameter, 100–150 µm; height 60–80 µm) it will be missed. Furthermore, Boltovskoy (1991) reported extremely high abundances (mean c. 20%, and up to 50%) of G. clarkei from the >63 µm size-fraction of Quaternary sediments recov- ered from Atlantic, Indian and Paciﬁc Ocean DSDP Sites. Although abundances of G. clarkei are not as high from the Pleistocene of ODP Site 1073 (up to 10%), their presence in the >63 µm size-fraction only and not the >150 µm size-fraction is signiﬁcant. DISCUSSION Si h k As might be expected, the mean percentages and percentage ranges of the larger species N. pachyderma (sinistral and dextral), G. bulloides and G. inﬂata are greater in the >150 µm size-fraction than in the >63 µm size-fraction. Neogloboquadrina pachyderma is known to be a cool-water species which is resistant to dissolution (Martinez et al., 1998). Neogloboquadrina pachyderma (s) is a polar species which peaks in colder water temperatures of between 0( and 9(C (Bé & Tolderlund, 1971). Neogloboquadrina pachyderma (d) is a subpolar species which peaks in waters of between 10( and 14(C, but can calcify at temperatures lower than 7–8(C (Bé & Tolderlund, 1971; Reynolds & Thunell, 1985; Wu & Hillaire-Marcel, 1994). At Site 1073, peaks in the abundance of N. pachyderma (s and d) are consistent in both the >63 µm and the >150 µm size-fractions. Peaks in N. pachyderma (s) are associated with 18O increases and reﬂect cooler/glacial conditions. Boltovskoy (1977) concluded that G. clarkei is a eurythermic species found from the subantarctic to equatorial zone and, because of its thick wall, is the most resistant species to dissolution among Recent planktonic foraminifera. In this study, no consistent correlation between abundances of G. clarkei and dissolution indices (e.g. PFI) was found. Clearly, further, more detailed work would be required to investigate the causes of abundance and distribution of G. clarkei in the world’s ocean. Globigerina bulloides is a transitional to polar species and occurs mainly in subpolar areas and upwelling regions (e.g. Thiede, 1975; Hemleben et al., 1989). Globigerina bulloides occurs in water temperatures of between 0(C and 23.3(C, although the optimum temperature for this species is 9–11(C (Bé & Tolderlund, 1971). Recently, Schmidt & Mulitza (2002) suggested a temperature range of 3–19(C and optimum of 10–12(C for G. bulloides in the North Atlantic. Reynolds & Thunell (1985) suggest that the distribution and abundance of G. bulloides may be related to food availability rather than to particular temperature ranges. The peaks in abundance of There does not appear to be a consistent relationship between species diversity (reﬂected by alpha index values) and the relative abundance ﬂuctuations of the planktonic foraminiferal taxa in either size-fraction. However, as might be expected, lowered diversity values correlate with major increases in individual species. A good example of this are the peaks in G. uvula which generally correlate with decreases in alpha index values in the >63 µm size-fraction. ACKNOWLEDGEMENTS I wish to thank all the members of the ODP Leg 174A technical and scientiﬁc party onboard the JOIDES Resolution for their assistance during the cruise. This research used samples pro- vided by the Ocean Drilling Program (ODP). The ODP is sponsored by the US National Science Foundation (NSF) and participating countries (including the UK) under the manage- ment of the Joint Oceanographic Institutions (JOI), Inc. I wish to thank the reviewers, M. Kucera and B. A. Malmgren, whose comments greatly improved the manuscript. I would also like to thank H. C. Olson for her helpful suggestions on an earlier version of the manuscript and D. E. Wright (Plymouth) for his help on statistics. Funding for this research was provided by a NERC/ODP Rapid Response Award. Higher proportions of juvenile species occur in the smaller size-fraction and the mean percentages and abundance ranges of larger species, e.g. N. pachyderma (s), N. pachyderma (d), G. inﬂata and G. bulloides are higher in the >150 µm size-fraction compared with the >63 µm size-fraction. The smaller species, T. quinqueloba, G. uvula, G. glutinata and G. clarkei are more abundant in the >63 µm size-fraction compared with the >150 µm size-fraction. A comparison of the dominant species abundances between the two size-fractions reveals strong positive correlations of variations in N. pachyderma (s) (r=0.889, p=0.000), G. bulloides (r=0.801, p=0.000) and N. pachyderma (d) (r=0.705, p=0.000), and weaker positive correlations in T. quinqueloba (r=0.506, p=0.004), G. uvula (r=0.431, p=0.016) and G. glutinata (r=0.385, p=0.032). There is little/no correlation in the abun- dance of G. inﬂata (r=0.104, p=0.579) between the two size- fractions. Correlation coeﬃcients and signiﬁcance of diﬀerences for other species are less clear because they occur rarely (<5%) at Site 1073. APPENDIX: Faunal reference list i APPENDIX: Faunal reference list Globigerina bulloides d’Orbigny, 1826 APPENDIX: Faunal reference list Globigerina bulloides d’Orbigny, 1826 Globigerina exumbilicata Herman, 1974 Globigerina falconensis Blow, 1959 Globigerina umbilicata Orr & Zaitzeﬀ, 1971 Globigerinita clarkei (Rögl & Bolli)=Globigerina clarkei Rögl & Bolli, 1973 Globigerinita clarkei (Rögl & Bolli)=Globigerina clarkei Rögl & Bolli, 1973 For the most part, consistent palaeoclimatic/palaeoceano- graphic interpretations are achieved from the planktonic foraminiferal abundance ﬂuctuations from the two size- fractions. Peaks in abundance of N. pachyderma (s) are associ- ated with 18O increases and reﬂect cooler/glacial conditions. The peaks in abundance of G. bulloides during OIS 3 and 4 may reﬂect increases in upwelling and nutrients. DISCUSSION Si h k The study of diﬀerent size-fractions has important impli- cations for establishing planktonic foraminiferal transfer 144 Pleistocene planktonic foraminifera from the NW Atlantic Ocean Pleistocene planktonic foraminifera from the NW Atlantic Ocean Pleistocene planktonic foraminifera from the NW Atlantic Ocean solution have an important inﬂuence on the abundance of this dissolution-susceptible species. functions. For example, Niebler & Gersonde (1998) concluded that in the southern South Atlantic Ocean, their analysis of planktonic foraminifera in the >125 µm size-fraction, instead of the >150 µm size-fraction, increased the temperature signal stored in the preserved faunal record. Apart from N. pachy- derma (s), Niebler & Gersonde (1998) found that their study of the >125 µm size-fraction resulted in species such as T. quinque- loba, G. uvula, G. bulloides, N. pachyderma (d) and G. inﬂata contributing considerably to the colder water assemblages de- posited in the Polar Front Zone and the Subantarctic Zone of the Antarctic Circumpolar Front. They noted that previous high-latitude investigations restricted to the >150 µm size- fraction have overlooked this relationship. At Site 1073, the smaller species, T. quinqueloba, G. clarkei and G. uvula, are more signiﬁcant components in the >63 µm size-fraction compared with the >150 µm size-fraction. Thus planktonic foraminiferal studies involving the >63 µm size-fraction could potentially produce diﬀerent transfer function estimates for sea surface temperatures in a given study area compared with studies that utilize the >150 µm size-fraction and even the >125 µm size-fraction. The most signiﬁcant species which occurs almost exclusively in the >63 µm size-fraction is G. uvula (mean abundance 14.5%). At particular intervals, peaks of up to 60% are recorded in its abundance in the >63 µm size-fraction only. In the >150 µm size-fraction G. uvula is rare/absent (mean abundance 0.1%). The small species G. clarkei, which in this study is grouped in T. quinqueloba, occurs in moderate abundance (up to 10%) in the >63 µm size-fraction but is absent from the >150 µm size-fraction. The palaeoenvironmental signiﬁcance of the observed trends in G. uvula and G. clarkei is not fully understood and, therefore, further detailed studies from other areas are required. The results of this study indicate that since particular species are either under-represented or even absent from the >150 µm size-fraction and may have high abundances in the >63 µm size-fraction, it is recommended that the >63 µm size-fraction cannot be ignored in studies of high-latitude planktonic forami- nifera. CONCLUSIONS A comparison of the planktonic foraminiferal faunas between the >63 µm and >150 µm size-fractions from the same samples from the Pleistocene of ODP Site 1073 (Leg 174A), New Jersey margin, NW Atlantic Ocean reveal important diﬀerences. Although the trends in the relative abundances of many species are similar in the two size-fractions, the general level varies considerably between the two size-fractions. DISCUSSION Si h k Furthermore, the study of diﬀerent size-fractions has important implications for establishing planktonic foraminiferal transfer functions. Studies involving smaller (>63 µm) plank- tonic foraminifera could generate diﬀerent transfer function estimates for sea surface temperatures in a given study area compared with studies that only use larger (>150 µm and >125 µm) planktonic foraminifera. REFERENCES Austin, J.A. Jr, Christie-Blick, N. & Malone, M.J. et al. 1998. Proceed- ings of the Ocean Drilling Program, Initial Reports, 174A. College Station, TX (Ocean Drilling Program). Malmgren, B.A. 1983. Ranking of dissolution susceptibility of plank- tonic foraminifera at high latitudes of the South Atlantic Ocean. Marine Micropaleontology, 8: 183–191. Bauch, H.A. 1994. Signiﬁcance of variability in Turborotalita quinque- loba (Natland) test size and abundance for paleoceanographic inter- pretations in the Norwegian–Greenland Sea. Marine Geology, 121: 129–141. Marquard, R.S. & Clark, D.L. 1987. Pleistocene paleoceanographic correlations: Northern Greenland Sea to central Arctic Ocean. Marine Micropaleontology, 12: 325–341. Martin, R.E. & Liddell, W.D. 1989. Relation of counting methods to taphonomic gradients and biofacies zonation of foraminiferal sediment assemblages. Marine Micropaleontology, 15: 67–89. Boltovskoy, E. 1977. Time range, geographic distribution and taxo- nomic situation of Globigerinita clarkei (Rögl & Bolli). Revista Espanola de Micropaleontología, 9: 89–93. Martinez, J.I., Taylor, L., De Deckker, P. & Barrows, T. 1998. Plank- tonic foraminifera from the eastern Indian Ocean: distribution and ecology in relation to the Western Paciﬁc Warm Pool (WPWP). Marine Micropaleontology, 34: 121–151. p p g Boltovskoy, E. 1990. Late Pleistocene–Holocene planktic foraminifera of the western equatorial Paciﬁc. Boreas, 19: 119–125. Boltovskoy, E. 1991. Globigerinita clarkei (Rögl & Bolli) – an unfairly ignored small planktic foraminifer. Boreas, 20: 151–154. ignored small planktic foraminifer. Boreas, 20: 151–154. McHugh, C.M.G. & Olson, H.C. 2002. Pleistocene chronology of continental margin sedimentation: new insights into traditional models, New Jersey. Marine Geology, 186: 389–411. Boltovskoy, E., Boltovskoy, D. & Brandini, F. 2000. Planktonic foraminifera from south-western Atlantic epipelagic waters: abun- dance, distribution and year-to-year variations. Journal of the Marine Biological Association of the United Kingdom, 79: 203–213. McIntyre, A., Kipp, N.G., Bé, A.W.H. et al. 1976. Glacial North Atlantic 18,000 years ago: a CLIMAP reconstruction. In: Cline, R.M. & Hays, J.D. (Eds), Investigation of Late Quaternary paleoceanogra- phy and paleoclimatology. Geological Society of America Memoir, 145: 43–76. Bé, A.W.H. 1977. An ecological, zoogeographical and taxonomic review of Recent planktonic foraminifera. In: Ramsay, A.T.S. (Ed.), Oceanic Micropaleontology, Part 1. Academic Press, London, 1–100. Bé, A.W.H. & Hutson, H. 1977. Ecology of planktonic foraminifera and biogeographic patterns of life and fossil assemblages in the Indian Ocean. Micropaleontology, 23: 369–414. Mountain, G.S., Miller, K.G., Blum, P. et al. 1994. Proceedings of the Ocean Drilling Program, Initial Reports, 150: College Station, TX (Ocean Drilling Program). Bé, A.W.H. & Tolderlund, D.S. 1971. ACKNOWLEDGEMENTS The inverse relation- ship between the abundance of Turborotalita quinqueloba and PFI (% fragments) suggests that variations in carbonate dis- Globigerinita glutinata (Egger)=Globigerina glutinata Egger, 1895 Globigerinita glutinata (Egger)=Globigerina glutinata Egger, 1895 Globigerinita uvula (Ehrenberg)=Pylodexia uvula Ehrenberg, 1861 Globigerinita uvula (Ehrenberg)=Pylodexia uvula Ehrenberg, 1861 Globigerinoides cyclostomus (Galloway & Wissler)=Globigerina cyclostoma Galloway & Wissler, 1927 Globigerinoides cyclostomus (Galloway & Wissler)=Globigerina cyclostoma Galloway & Wissler, 1927 Globigerinoides pyramidalis (van den Broeck)=Globigerina bulloides d’Orbigny var. rubra d’Orbigny subvar. pyramidalis van den Broeck, 1876 145 C. W. Smart Globigerinoides ruber (d’Orbigny)=Globigerina rubra d’Orbigny, 1839 Fisher, R.A., Corbet, A.S. & Williams, C.B. 1943. The relationship between the number of species and the number of individuals in a random sample of an animal population. Journal of Animal Ecology, 12: 42–58. Globigerinoides sacculifer (Brady)=Globigerina sacculifer Brady, 1877 Giraudeau, J. 1993. Planktonic foraminiferal assemblages in surface sediments from the southwest African continental margin. Marine Geology, 110: 47–62. Globorotalia hirsuta (d’Orbigny)=Rotalina hirsuta d’Orbigny, 1839 Hemleben, C., Spindler, M. & Anderson, O.R. 1989. Modern planktonic foraminifera. Springer-Verlag, Berlin. Globorotalia inﬂata (d’Orbigny)=Globigerina inﬂata d’Orbigny, 1839 Imbrie, J. & Kipp, N.G. 1971. A new micropaleontological method for quantitative paleoclimatology: applications to a late Pleistocene Caribbean core. In: Turekian, K.K. (Ed.), The Late Cenozoic Glacial Ages. Yale University Press, Newhaven, CT, 71–131. Globorotalia menardii (Parker, Jones & Brady)=Rotalia menardii Parker, Jones & Brady, 1865 Ages. Yale University Press, Newhaven, CT, 71–131. Globorotalia puncticulata (Deshayes)=Globigerina punctulata d’Orbigny, 1826 nomen nudum Kandiano, E.S. & Bauch, H.A. 2002. Implications of planktic fora- miniferal size fractions for the glacial–interglacial paleoceanography of the polar North Atlantic. Journal of Foraminiferal Research, 32: 245–251. g y Globorotalia scitula (Brady)=Pulvinulina scitula Brady, 1882 Globorotalia trucatulinoides (d’Orbigny)=Rotalia truncatuli- noides d’Orbigny, 1839 Kellogg, T.B. 1984. Paleoclimatic signiﬁcance of subpolar foraminifera in high-latitude marine sediments. Canadian Journal of Earth Sciences, 21: 189–193. Neogloboquadrina dutertrei (d’Orbigny)=Globigerina dutertrei d’Orbigny, 1839 Kennett, J.P. & Srinivasan, M.S. 1983. Neogene planktonic foraminifera: a phylogenetic atlas. Hutchinson Ross Publishing Company, Stroudburg, Pennsylvania. Neogloboquadrina pachyderma (Ehrenberg)=Aristospira pachy- derma Ehrenberg, 1861 Kipp, N.G. 1976. New transfer function for estimating past sea-surface conditions from sea-bed distribution of planktonic foraminiferal assemblages in the North Atlantic. In: Cline, R.M. & Hays, J.D. (Eds), Investigation of Late Quaternary paleoceanography and paleoclimatology. Geological Society of America Memoir, 145: 3–42. h kl b di l i ﬂ i i Orbulina universa d’Orbigny, 1839 Turborotalita quinqueloba (Natland)=Globigerina quinqueloba Natland, 1938 Manuscript received 21 January 2002 Le, J. ACKNOWLEDGEMENTS & Shackleton, N.J. 1992. Carbonate dissolution ﬂuctuations in the western equatorial Paciﬁc during the late Quaternary. Paleoceanography, 7: 21–42. Manuscript accepted 23 July 2002 Lu, G. & Keller, G. 1995. Planktonic foraminiferal turnovers in the subtropical Paciﬁc during the late Pleistocene to early Eocene. Journal of Foraminiferal Research, 25: 97–116. REFERENCES Distribution and ecology of living planktonic foraminifera in surface waters of the Atlantic and Indian oceans. In: Funnell, B.M. & Riebridge, W.R. (Eds), The Micro- paleontology of the oceans. Cambridge University Press, Cambridge, 105–149. Murray, J.W. 1991. Ecology and palaeoecology of benthic foraminifera. Longman, Harlow. Niebler, H.-S. & Gersonde, R. 1998. A planktic foraminiferal transfer function for the southern South Atlantic Ocean. Marine Micropaleontology, 34: 213–234. CLIMAP 1981. Seasonal reconstructions of the Earth’s surface at the last glacial maximum. Geological Society of America Map and Chart Series, MC-36: 1–18. Olson, H.C. & Smart, C.W. in press. Pleistocene climatic history reﬂected in planktonic foraminifera from ODP Site 1073 (Leg 174A), New Jersey margin, NW Atlantic Ocean. Marine Micro- paleontology. CLIMAP 1984. The last interglacial ocean. Quaternary Research, 21: 123–224. Peeters, F., Ivanova, E., Conan, S., Brummer, G.-J., Ganssen, G., Troelstra, S. & van Hinte, J. 1999. A size analysis of planktic foraminifera from the Arabian Sea. Marine Micropaleontology, 36: 31–63. Carstens, J., Hebbeln, D. & Wefer, G. 1997. Distribution of planktic foraminifera at the ice margin in the Arctic (Fram Strait). Marine Micropaleontology, 29: 257–269. 146 Pleistocene planktonic foraminifera from the NW Atlantic Ocean Pleistocene planktonic foraminifera from the NW Atlantic Ocean Schröder, C.J., Scott, D.B. & Medioli, F.S. 1987. Can smaller benthic foraminifera be ignored in paleoenvironmental analyses? Journal of Foraminiferal Research, 17: 101–105. Reynolds, L. & Thunell, R.C. 1985. Seasonal succession of planktonic foraminifera in the subpolar North Paciﬁc. Journal of Foraminiferal Research, 15: 282–301. f Sen Gupta, B.K., Shin, I.C. & Wendler, S.T. 1987. Relevance of specimen size in distribution studies of deep-sea benthic foraminifera. Palaios, 2: 332–338. Rögl, F., Bolli, H.M. 1973. Holocene to Pleistocene planktonic for- aminifera of Leg 15, Site 147 (Cariaco Basin (Trench), Caribbean Sea) and their climatic interpretation. In: Edgar, N.T., Saunders, J.B. et al., (Eds), Initial Reports of the Deep Sea Drilling Project, 15: US Government Printing Oﬃce, Washington, DC, 558–615. Thiede, J. 1975. Distribution of foraminifera in surface waters of a coastal upwelling area. Nature, 253: 712–714. Saito, T., Thompson, P.R. & Breger, D. 1981. Systematic index of Recent and Pleistocene planktonic foraminifera. University of Tokyo Press, Japan. Thunell, R.C. 1975. Optimum indices of calcium carbonate dissolution in deep-sea sediments. Geology, 4: 525–528. Wei, W. 2001. Calcareous nannofossils from the New Jersey conti- nental margin. In: Christie-Blick, N., Austin, J.A. Jr & Malone, M.J. (Eds), Proceedings of the Ocean Drilling Program, Scientiﬁc Results, 174A. College Station, TX (Ocean Drilling Program), 1–16 [Online]. Savrda, C.E., Krawinkel, H., McCarthy, F.M.G., McHugh, C.M.G., Olson, H.C. & Mountain, G. 2001. Ichnofabrics of a Pleistocene slope succession, New Jersey Margin: relations to climate and sea-level dynamics. Palaeogeography, Palaeoclimatology, Palaeoecology, 171: 41–61. Wu, G. & Hillaire-Marcel, C. 1994. Oxygen isotope compositions of sinistral Neogloboquadrina pachyderma tests in surface sediments: North Atlantic Ocean. Geochimica et Cosmochimica Acta, 58: 1303–1312. Schmidt, G.A. & Mulitza, S. 2002. Global calibration of ecological models for planktic foraminifera from coretop carbonate oxygen-18. Marine Micropaleontology, 44: 125–140. 147
https://openalex.org/W4294955082	https://sam.ensam.eu/bitstream/10985/22703/2/I2M_IJLC_2022_LARATTE.pdf	English	null	Midpoint and endpoint characterization factors for mineral resource dissipation: methods and application to 6000 data sets	The international journal of life cycle assessment	2,022	cc-by	17,256	Alexandre Charpentier Poncelet1 · Philippe Loubet1 · Christoph Helbig2,3 · Antoine Beylot4 · Stéphanie Muller4 · Jacques Villeneuve4 · Bertrand Laratte5 · Andrea Thorenz2 · Axel Tuma2 · Guido Sonnemann1 Alexandre Charpentier Poncelet1 · Philippe Loubet1 · Christoph Helbig2,3 · Antoine Beylot4 · Stéphanie Muller4 · Jacques Villeneuve4 · Bertrand Laratte5 · Andrea Thorenz2 · Axel Tuma2 · Guido Sonnemann1 Received: 23 May 2022 / Accepted: 10 August 2022 © The Author(s) 2022 3 Ecological Resource Technology, University of Bayreuth, Universitaetsstr. 30, 95447 Bayreuth, Germany 4 BRGM, 45060 Orléans, France 5 Arts Et Métiers, University of Bordeaux, CNRS, INP, I2M, 33400 Talence, Bordeaux, France Communicated by Matthias Finkbeiner Science Arts & Métiers (SAM) is an open access repository that collects the work of Arts et Métiers Institute of Technology researchers and makes it freely available over the web where possible. This is an author-deposited version published in: https://sam.ensam.eu Handle ID: .http://hdl.handle.net/10985/22703 This document is available under CC BY license To cite this version : Alexandre CHARPENTIER PONCELET, Philippe LOUBET, Christoph HELBIG, Antoine BEYLOT, Stéphanie MULLER, Jacques VILLENEUVE, Bertrand LARATTE, Andrea THORENZ, Axel TUMA, Guido SONNEMANN - Midpoint and endpoint characterization factors for mineral resource dissipation: methods and application to 6000 data sets - The International Journal of Life Cycle Assessment - 2022 Any correspondence concerning this service should be sent to the repository Administrator : scienceouverte@ensam.eu Any correspondence concerning this service should be sent to the repository Administrator : scienceouverte@ensam.eu Any correspondence concerning this service should be sent to the repository Administrator : scienceouverte@ensam.eu The International Journal of Life Cycle Assessment https://doi.org/10.1007/s11367-022-02093-2 LCIA OF IMPACTS ON HUMAN HEALTH AND ECOSYSTEMS Abstract Purpose The accessibility to most metals is crucial to modern societies. In order to move towards more sustainable use of metals, it is relevant to reduce losses along their anthropogenic cycle. To this end, quantifying dissipative flows of mineral resources and assessing their impacts in life cycle assessment (LCA) has been a challenge brought up by various stakeholders in the LCA community. We address this challenge with the extension of previously developed impact assessment methods and evaluating how these updated methods compare to widely used impact assessment methods for mineral resource use. Methods Building on previous works, we extend the coverage of the average dissipation rate (ADR) and lost potential service time (LPST) methods to 61 metals. Midpoint characterization factors are computed using dynamic material flow analysis results, and endpoint characterization factors, by applying the market price of metals as a proxy for their value. We apply these methods to metal resource flows from 6000 market data sets along with the abiotic depletion potential and ReCiPe 2016 methods to anticipate how the assessment of dissipation using the newly developed methods might compare to the latter two widely used ones. Results and discussion The updated midpoint methods enable distinguishing between 61 metals based on their global dissipa- tion patterns once they have been extracted from the ground. The endpoint methods further allow differentiating between the value of metals based on their annual average market prices. Metals with a high price that dissipate quickly have the highest endpoint characterization factors. The application study shows that metals with the largest resource flows are expected to have the most impacts with the midpoint ADR and LPST methods, metals that are relatively more expensive have a greater relative contribution to the endpoint assessment. Conclusion The extended ADR and LPST methods provide new information on the global dissipation patterns of 61 met- als and on the associated potentially lost value for humans. The methods are readily applicable to resource flows in current life cycle inventories. This new information may be complementary to that provided by other impact assessment methods addressing different impact pathways when used in LCA studies. Additional research is needed to improve the characteriza- tion of the value of metals for society and to extend the methods to more resources. * Christoph Helbig christoph.helbig@uni-bayreuth.de Alexandre Charpentier Poncelet alexandre.charpentier.poncelet@gmail.com 1 UMR 5255, Univ. Bordeaux, CNRS, INP, ISM, 33400 Talence, Bordeaux, France 2 Resource Lab, University of Augsburg, Universitaetsstr. 16, 86159 Augsburg, Germany 1 UMR 5255, Univ. Bordeaux, CNRS, INP, ISM, 33400 Talence, Bordeaux, France 2.1 Extending the coverage of the ADR and LPST methods The MaTrace model initially developed by Nakamura et al. (2014) allows quantifying losses of a cohort of extracted met- als to the environment, other material flows (non-functional recycling), and waste disposal facilities (including tailings and slags) over time (Helbig et al. 2020). As in former studies (Beylot et al. 2020a, 2021; Charpentier Poncelet et al. 2021; Helbig et al. 2020), these losses are considered as dissipa- tive flows in this study. With this model, the global anthro- pogenic cycles of metals are studied one at a time, meaning there are no direct links between the material flow models of each metal. Contrastingly, Helbig et al. (2021) analyzed seven major metals simultaneously with MaTrace-multi, which adds another level of information and another level of modeling complexity that would not be possible to handle for 61 metals at this point. Therefore, despite the potential limitations associated with the study of single cycles (e.g., mass-balance discrepancies between different metals used in the same applications), we here focus on the individual cycles of 61 metals as studied by Charpentier Poncelet et al. (2022b). In the latter article, the losses of metals are evalu- ated over time and their average lifetimes in the economy are estimated based on the most recent data possible (most typically between 2010 and 2020). The latter correspond to the average duration over which metals remain in use in the economy after extraction. In this context, it is proposed to develop methods to assess the dissipation of mineral resources and its impacts in life cycle assessment (LCA) (Berger et al. 2020; Beylot et al. 2020b, 2021; van Oers et al. 2020; Zampori and Sala 2017). Charpentier Poncelet et al. (2019) proposed a framework to consider the impacts of the dissipative flows of metals on the area of protection (AoP) natural resources in LCA. Based on that idea, we developed two dissipation-oriented midpoint methods for life cycle impact assessment (LCIA) applicable to extraction flows in the life cycle inventories (LCI) and computed characterization factors (CF) for 18 metals (Charpentier Poncelet et al. 2021). These methods relied on the results from the dynamic material flow analysis (MFA) model of Helbig et al. (2020). The first method is the lost potential service time (LPST), which quantifies the lost opportunity to use metallic elements in the economy due to dissipative flows over time horizons of 25, 100, or 500 years. 2.1 Extending the coverage of the ADR and LPST methods The second one is the average dissipation rate (ADR), which assesses the expected dissipation rates of metals from extrac- tion until their complete dissipation. p p The main objective of this article is to increase the coverage of the LPST and ADR methods based on our extension of the aforementioned dynamic MFA model to 61 metals (Charpentier Poncelet et al. 2022b). We also explore the use of price-based endpoint CFs to represent the loss of value associated with the dissipation of metals. Complementarily, we investigate potential impact assess- ment results using these new methods, and compare them with widely used LCIA methods. To do so, we apply the newly developed CFs to all non-empty market data sets from the ecoinvent database version 3.7.1 (Moreno Ruiz et al. 2020; Wernet et al. 2016) and compare the LCIA results with those for the abiotic depletion potential (ADP) ultimate reserves method (van Oers et al. 2019, 2002), the surplus ore potential (SOP) method (Vieira et al. 2016) Midpoint CFs for 61 metals are derived from the results of that article. The Python code and compiled datasets underlying that study are accessible online (Helbig and Charpentier Poncelet 2022). The overview of developments proposed in this article based on our previous work is pre- sented in Fig. 1. Abstract Keywords Dissipation · Losses · Metals · Mineral resources · Life cycle impact assessment · Circula Communicated by Matthias Finkbeiner 3 Ecological Resource Technology, University of Bayreuth, Universitaetsstr. 30, 95447 Bayreuth, Germany 4 BRGM, 45060 Orléans, France 5 Arts Et Métiers, University of Bordeaux, CNRS, INP, I2M, 33400 Talence, Bordeaux, France 1 UMR 5255, Univ. Bordeaux, CNRS, INP, ISM, 33400 Talence, Bordeaux, France 2 Resource Lab, University of Augsburg, Universitaetsstr. 16, 86159 Augsburg, Germany :(0123 1 23456789) 3 The International Journal of Life Cycle Assessment and the surplus cost potential (SCP) method (Vieira et al. 2016). The latter two methods are included in ReCiPe 2016 (Huijbregts et al. 2017). Materials and methods are detailed in “Sect. 2,” results are presented and analyzed in “Sect. 3,” and conclusions are drawn in “Sect. 4.” 1 Introduction All metals in the periodic table can provide a socio-economic benefit for modern societies (Graedel et al. 2013). Yet, min- eral resources are non-renewable and accessible in constrained quantities (Drielsma et al. 2016; Schulze et al. 2020), driving the will to keep them in the economy for as long as possible through, e.g., recycling or other circular economy strategies (Blomsma and Tennant 2020; European Commission 2020a; Reuter et al. 2019; UNEP 2013). Dissipative flows of mineral resources are flows that become inaccessible for future use (Beylot et al. 2020b; Helbig et al. 2020). They can include flows to the environment, to waste disposal facilities, and flows to materials where their specific physicochemical char- acteristics are no longer positively contributing to the material characteristics (non-functional recycling) (Beylot et al. 2020b; Charpentier Poncelet et al. 2021; Helbig et al. 2020). 2.1.1 Midpoint characterization factors based on dynamic MFA data The midpoint CFs for the ADR and LPST methods are com- puted using the approach described by Charpentier Poncelet et al. (2021). The ADR method allows distinguishing between the relative dissipation rates of metals after extrac- tion in current conditions of consumption and recycling in 1 3 3 The International Journal of Life Cycle Assessment Loss rates & lifetime of 61 metals CFLPST CFLPV AoP Natural resources Endpoint/damage Midpoint Life cycle inventory Previous work (references presented in caption) Further LCIA method developments (this article) Extension of midpoint CFs to 61 metals Dynamic MFA results for 61 metals Method for computing midpoint CFs & computation of CFs for 18 metals CFPVLR Price-based endpoint CFs for 61 metals CFADR Resource extraction from ground Dissipation over time after extraction Inaccessibility for use in technosphere Lost potential to make use of value of inaccessible metals Extraction of metals CF: characterization factor ADR: average dissipation rate LPST: lost potential service time Legend LPV: lost potential value PVLR: potential value loss rate Further LCIA method developments (this article) Extension of midpoint CFs to 61 metals Price-based endpoint CFs for 61 metals Dynamic MFA results for 61 metals Extraction of metals Life cycle inventory Resource extraction from ground Lost potential to make use of value of inaccessible metals Inaccessibility for use in technosphere Dissipation over time after extraction CF: characterization factor ADR: average dissipation rate LPST: lost potential service time method for computing midpoint CFs (Charpentier Poncelet et al. 2021); dynamic MFA results for 18 metals (Helbig et al. 2020); and extended dynamic MFA results (Charpentier Poncelet et al. 2022b) method for computing midpoint CFs (Charpentier Poncelet et al. 2021); dynamic MFA results for 18 metals (Helbig et al. 2020); and extended dynamic MFA results (Charpentier Poncelet et al. 2022b) Fig. 1 Overview of impact pathway and further development of the ADR and LPST methods based on previous work (adapted from Charpentier Poncelet et al. 2021). References to previous works: Fig. 1 Overview of impact pathway and further development of the ADR and LPST methods based on previous work (adapted from Charpentier Poncelet et al. 2021). References to previous works: given the expected dissipation pattern of metal i over a given time horizon of 25, 100, or 500 years, as shown in Eq. (2): the economy. 2.1.2 Endpoint characterization factors based on metal prices y p y Different periods are considered to estimate the price of other metals due to a lack of data. Given that iron ores are directly refined into steel, the price of iron is derived from that of steel. The 10-year period for the latter ranges from 2001 to 2010 because no data are available for subsequent years (Kelly and Matos 2014). Osmium (one of the PGMs), for which no price data is available, is estimated to remain at a constant price of $US400 per ounce (Labbé and Dupuy 2014). The price of scandium is estimated from the price of scandium oxides between 2012 and 2018, which ranged between $US4600 and $US5400 per kilogram (European Commission 2020b). For the latter, we consider an average yearly price of $US5000 per kg from 2012 to 2018 and a scandium content of 65.2% in scandium oxides (Sc2O3). The price of thulium (Tm) is cal- culated based on the price of its oxides in the third semester of 2015 (Bru et al. 2015) and considering a stoichiometric content of 87.4%. Similarly, the prices of the remaining REEs, namely erbium (Er), ytterbium (Yb), holmium (Ho), and lute- tium (Lu), are calculated from the average yearly prices of their oxides between 2009 and 2014 as reported by Stormcrow (2014) and considering their respective stoichiometric contents in oxides (approximately 87.5%). It should be noted that the prices of REEs underwent an important price peak in 2011 due to Chinese bans on exports in the early 2010s (Bru et al. 2015), which has a notable effect on the reported prices for these metals. Such potential limitations associated with using the market price data are discussed in “Sect. 3.1.2.” As the final step of the impact pathway, we evaluate the potential socio-economic impacts due to the dissipation of different mineral resources (cf. Fig. 1). The Joint Research Centre (JRC) of the European Commission suggested that the average prices of resources over a given period can be used as a proxy to reflect the complex utility that resources have for humans and are practical to do so since the data is easily available (Beylot et al. 2020a). The underlying assump- tion of using the prices of metals as an indication of their socio-economic value is that prices reflect at least to some extent the value of metals for society, albeit not perfectly (see discussion in, e.g., Beylot et al. 2.1.2 Endpoint characterization factors based on metal prices 2020a; Ecorys 2012; Henckens et al. 2016; Huppertz et al. 2019; and Watson and Eggert 2021). The most expensive metals are generally used in specialized applications, because their high price does not justify their use in low value-added applications in which less specific or cheaper materials can be used. i For this study, we consider that recent price statistics are most likely to be representative of the value of resources answering the current demand for different applications. We further assume that this value is maintained over time as long as metals remain in the economy. This assumption is likely to be more realistic for the short than the long term for most metals, given that it is impossible to predict the long-term trends in demand for given applications, nor the development of new applications. Whenever possible, the 10-year average prices from 2006 to 2015 are consid- ered. This period is chosen because most price statistics are obtained from US Geological Survey (USGS) statistics (Kelly and Matos 2014) for which 2015 is the most recent year for which data are available for most metals. Details for specific metals are presented below. Price averages and references are provided in the SI. Endpoint CFs are computed by multiplying the midpoint indicators with the average price of metals, allowing to com- pare the relative potential value lost due to the dissipation of different metals over time. This approach differs from the JRC approach proposing to directly characterize dissipative flows identified in the LCI with price-based CFs (Beylot et al. 2020a). Yet, metals retain value for humans for as long as they are in use and quantifying the problem of inaccessibility neces- sitates a time dimension (Dewulf et al. 2021) that is taken into account with the proposed ADR and LPST methods. The price of barium is derived from statistics for barites because barium is almost exclusively consumed in compound forms (Johnson et al. 2017). We considered a barium content of 58.9% in barites (BaSO4) based on its stoichiometric content. Other data are gathered to fill data gaps and cover the plati- num group metals and rare earth elements that are not covered separately in USGS statistics. All of the prices are adjusted to $US1998 to match the reference unit of the USGS price data. Prices for individual platinum group metals (PGMs) except osmium are compiled from Johnson Matthey (2021). 2.1.1 Midpoint characterization factors based on dynamic MFA data The LPST method measures the (relative) lost opportunity to make use of metals over time once dissipated based on these same conditions. The rationale for the LPST method is summarized in the supporting information (SI). The main equations for the ADR and LPST methods are replicated below. the economy. The LPST method measures the (relative) lost opportunity to make use of metals over time once dissipated (2) LPSTi,TH = OSTi,TH −STi,TH (2) where the LPST, OST, and ST are expressed in kg.yr/kg (= yr). The CFs of the ADR method are calculated as the ratio between the ADR of metal i and that of iron (Fe), as shown in Eq. (3): The ADR is calculated as the inverse of the average life- time of metals in the economy, referred to as the total service time (STTOT) in Eq. (1): (3) CFADRi = ADRi∕ADRFe (3) (1) ADRi = 1∕STTOTi. ADRi = 1∕STTOTi. (1) where the CFADR for metal i is expressed in kg Fe-eq./kg. There is no time horizon for the ADR method, as it inte- grates the time function in its computation to provide a yearly rate of dissipation (Charpentier Poncelet et al. 2021). where the STTOT represents the total expected service time of metal i in the economy after extraction and until its com- plete dissipation (expressed in kg.yr/kg = yr). The ADR is expressed in kg/kg.yr = yr−1.f Similarly, those of the LPST method are calculated as shown in Eq. (4): The LPST is calculated as the difference between the opti- mal service time, OST, defined as the total service time if no dissipation occurred, and the expected service time, ST, (4) CFLPSTi,TH = LPSTi,TH∕LPSTFe,TH. (4) 1 3 3 The International Journal of Life Cycle Assessment where the CFLPST for metal i is expressed in kg Fe-eq./kg. survey (BRGM) (Bru et al. 2015, p. 151–158). These are yttrium (Y), lanthanum (La), cerium (Ce), praseodymium (Pr), neodym- ium (Nd), samarium (Sm), europium (Eu), gadolinium (Gd), dysprosium (Dy), and terbium (Tb).f where the CFLPST for metal i is expressed in kg Fe-eq./kg. where CFPVLRi is measured in $US1998/kg.yr. 2.1.2 Endpoint characterization factors based on metal prices The price of niobium is calculated from data provided by Metalary (2021) considering a niobium content of 69.9% in niobium oxides (Nb2O5). Prices for ten of the rare earth elements (REEs) are compiled from data underlying a report of the French geological The ADR and LPST values are used for the calculation of endpoint CFs in order to keep the units in mass and mass. years (rather than their respective normalized midpoint CFs expressed in iron equivalents). The endpoint CFs for the ADR method represent a potential value loss rate (PVLR) due to average yearly dissipation rates of metals and are calculated as shown in Eq. (5): CFPVLRi = ADRi × pricei CFPVLRi = ADRi × pricei (5) where CFPVLRi is measured in $US1998/kg.yr. 1 3 3 The International Journal of Life Cycle Assessment Multiplying LPST with the price of metals indicates the lost potential value (LPV) due to the inaccessibility of met- als over time. It is assumed that the potential value of metals remains the same over time, thus no discounting is applied. Endpoint CFs for the LPV are calculated as shown in Eq. (6): of the ecoinvent database version 3.7.1 (Moreno Ruiz et al. 2020; Wernet et al. 2016), using allocation at point of sub- stitution (APOS). Market data sets represent the average consumption mixes for a given region and product (Wernet et al. 2016). We consider flows of metal resources included in the ecoinvent 3.7.1 database that are covered by the ADR and LPST methods. These include 45 flows, categorized as “met- als, in ground” (e.g., aluminum, in ground) as identified in Table S4 of the SI. In order to support the analysis of the contribution of different sections of economic activity to the inventory totals, the LCI data sets are subdivided by the International Standard Industrial Classification of All Eco- nomic Activities (ISIC) classification (United Nations 2008). Sections of economic activity regroup multiple sectors (e.g., section A includes the agriculture, forestry and fishing sec- tors). This facilitates the study’s intelligibility and indicates general trends expected from using the developed ADR and LPST methods. The number of data sets included in each economic section and the total mass of extracted metal flows for each section are detailed in section S4 of the SI. (6) CFLPVi = LPSTi × pricei CFLPVi = LPSTi × pricei (6) where CFLPVi is measured in $US1998/kg. 2.2.1 Selected LCIA methods We compare the midpoint and endpoint CFs for the LPST and ADR methods developed in this article with the latest ADP ultimate reserves method for elements based on the cumulative production in 2015 (van Oers et al. 2019, 2002) and the midpoint and endpoint CFs for the “mineral resource scarcity” category in the ReCiPe 2016 method (Huijbregts et al. 2017). The latter method includes 18 impact pathways covering three areas of protection. Its midpoint CFs for the mineral resource scarcity category originate from the SOP method (Vieira et al. 2017), and its endpoint CFs, from the SCP method (Vieira et al. 2016). 2.1.2 Endpoint characterization factors based on metal prices i The total PVLR (TPVLR) and the total LPV (TLPV) are calculated analogously to their corresponding midpoint cat- egory totals as defined by Charpentier Poncelet et al. (2021), as shown in Eqs. (7) and (8): (7) TPVLR = ∑n i=1mi × CFPVLRi (7) (8) TLPV = ∑n i=1mi × CFLPVi (8) where the TPVLR is measured in $US1998/yr and the TLPV is measured in $US1998. This approach is inspired by the study of Rørbech et al. (2014). However, unlike the latter study, we only charac- terize the impacts from metal resource flows and not other mineral forms of metallic elements nor energy minerals. It should be kept in mind that this study assesses the potential impacts of data sets whose functional units are not necessar- ily comparable (cf. “Sect. 3.2.3”). 2.1.3 Uncertainty A Monte Carlo simulation of 1000 iterations allowed computing 95% confidence intervals for the underlying dynamic MFA results (Charpentier Poncelet et al. 2022b). The uncertainty is also taken into account to compute the midpoint CFs derived from these results. It should be noted that the computed uncertainty reflects the uncertainty due to variation of input variables, and not the model’s ability to correctly reflect the real-life global system. Price uncer- tainties are not accounted for given that a potentially large and unquantifiable uncertainty can be associated with the assumption of the representativeness of the annual average market price information for the relative value of metals for humans over time. 2.2 Application of characterization factors to 6000 life cycle inventory data sets This objective of this application study is to investigate the general trends that could arise from using the ADR and LPST methods to evaluate the potential impacts due to the dissipation of mineral resources, and to compare these impact assessment results with those of widely used LCIA methods characterizing flows of metal resources in order to determine how they might differ. The assessment is realized by applying LCIA methods to multiple data sets rather than compare their CFs together. In order to do so, CFs from the selected LCIA methods are applied to the metal resource flows for all of the 5999 non-empty market LCI data sets The ADP ultimate reserves, SOP and SCP methods are briefly described in section S3 of the SI. Midpoint CFs are available for all methods, while no endpoint CFs are pro- posed in the ADP method. All of the CFs are normalized to kg Fe-eq./kg to facilitate the comparison of impact scores between metals and methods. The normalization is done by dividing all of the CFs by that of iron for the corresponding method. It should be noted that endpoint CFs for metals included in ReCiPe 2016 are equivalent to midpoint ones when normalized to iron equivalents because the former are 1 3 The International Journal of Life Cycle Assessment calculated from the latter using the same conversion factor calculated for copper (Berger et al. 2020). methods or not. Table S4 shows all of the CFs considered for the selected methods considered in the application study, including their value. pp g The ADP ultimate reserves method is currently recom- mended for use in the product environmental footprint (PEF) (Zampori and Pant 2019). The ADP ultimate reserves and SOP methods are recommended by the life cycle initiative to answer different questions linked with mineral resource use (Berger et al. 2020). 3.1.1 Midpoint and endpoint characterization factors Table 1 presents the ADR and STTOT for 61 metals as com- puted by Charpentier Poncelet et al. (2022b) and their cor- responding midpoint and endpoint CFs calculated for the ADR and LPST methods. The 95% confidence intervals for CFs are provided in Tables S1 and S2 of the SI. Higher CFADR and CFLPST indicate that a metal has a higher average dissipation rate and thus a shorter lifetime in the econ- omy (based on data for the recent past; see Charpentier Poncelet et al. 2022b). Given the coverage of 61 metals, we may not pro- vide extensive details for each of them. The underlying dynamic MFA model and results as well as its supplementary materials should be consulted for additional information on losses for each metal (Charpentier Poncelet et al. 2022b). We here discuss a few examples. It can be seen from the latter work that iron is second best preserved in the economy after gold. It has an expected lifetime of 154 years thanks to relatively long-lived applica- tions (e.g., infrastructure and mechanical equipment), a small percentage of dissipation in use, and a combined yield of about 80% for the collection and recycling processes (Charpentier Poncelet et al. 2022b). Its midpoint CFADR is 1.0 kg Fe-eq./kg. The corresponding CFADR of chromium (Cr), which is relatively well conserved in the economy (e.g., as chromium and stainless steels used in long-lived applications such as infrastructure and transport) although less than iron, is 5.5 kg Fe-eq./kg. In com- parison, gallium is quickly dissipated at the production phase (> 99%) for technical and economic reasons (Helbig et al. 2020; Løvik et al. 2016, 2015), resulting in relatively high midpoint CFADR, CFLPST25, and CFLPST100, with 1383 kg Fe-eq./kg, 6.0 kg Fe-eq./kg, and 3.2 kg Fe-eq./kg, respectively. At the same time, the CFADR of indium, which is also relatively rapidly lost after extraction due to overall low process yields and a lack of end-of- life recycling, is 105 kg Fe-eq./kg. The corresponding midpoint CFs for, e.g., the LPST100 method, are 2.3 kg Fe-eq./kg for chromium and 3.1 kg Fe-eq./kg for indium. While we here focus on two widely used LCIA meth- ods addressing mineral resource use, we refer readers to the study of Rørbech et al. (2014) and the critical review of the life cycle initiative’s task force on mineral resources (Berger et al. 2020; Sonderegger et al. 2020). Rørbech et al. 3.1.1 Midpoint and endpoint characterization factors (2014) applied the CFs from eleven LCIA methods to min- eral resource flows (including energy minerals, minerals, and metals) for all market data sets included in the ecoin- vent version 3.0 database. The life cycle initiative’s taskforce on mineral resources critically reviewed 27 LCIA methods assessing the impacts of mineral resource use (Sonderegger et al. 2020). 2.2 Application of characterization factors to 6000 life cycle inventory data sets The former is recommended to answer the question “How can I quantify the relative contribution of a product system to the depletion of mineral resources?.” The latter is interim recommended to answer the question “How can I quantify the relative consequences of the con- tribution of a product system to changing mineral resource quality?.” The midpoint ADR method could address the question “How can I quantify the relative contribution of a product system to the dissipation of mineral resources?” and the LPST method, “How can I quantify the relative contri- bution of a product system to the inaccessibility of mineral resources due to dissipation?.” Their endpoint versions could answer the question “How can I quantify the relative contri- bution of a product system to the potential mineral resource value lost due to dissipation?.” 2.2.2 Coverage of mineral resource flows by selected LCIA methods The ReCiPe 2016 method includes CFs for 75 resource flows of mineral resources, 26 of which are mineral compounds or ores. Parts of the latter 26 flows are no longer included in the ecoinvent database because many mineral compounds were converted to pure flows of metal content since ver- sion 3.6 (Moreno Ruiz et al. 2019). The latest update for the ADP ultimate reserves method for elements (van Oers et al. 2019) covers 76 elements. These include the 61 metallic ele- ments that are also covered by the ADR and LPST methods, in addition to calcium, cesium, potassium, sodium, and a few non-metals (e.g., halogens). Table S3 in the SI presents the 65 metallic elements included in the latest ADP method for elements (van Oers et al. 2019) and indicates whether they are included in the ReCiPe 2016 and the ADR/LPST Below, we discuss general trends per categories of met- als established by the UNEP (2011): ferrous, non-ferrous, precious, and specialty metals. Figures 2 and 3 show mid- point and endpoint CFs for the ADR and LPST100 methods, respectively. Figs. S2 and S3 provided in the SI depict CFs for the LPST25 and LPST500 methods. 1 3 3 The International Journal of Life Cycle Assessment 1 3 Table 1 Total expected service times (STTOT) corresponding to average lifetimes in the economy calculated by Charpentier Poncelet et al. (2022b), average dissipation rate (ADR), computed as loss rates by Charpentier Poncelet et al. (2022b), lost potential service time (LPST) with time horizons of 25, 100, and 500 years, and the associated midpoint and endpoint characterization fac- tors for the LPST and ADR methods STTOT (cf. ref. in caption) ADR (cf. ref. 2.2.2 Coverage of mineral resource flows by selected LCIA methods in caption) LPST25 LPST100 LPST500 Midpoint CFs Endpoint CFs CFADR CFLPST25 CFLPST100 CFLPST500 CFPVLR CFLPV25 CFLPV100 CFLPV500 Metal i kg.yr/kg kg/kg.yr kg.yr/kg kg.yr/kg kg.yr/kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg $US1998/ kg.yr $US1998/kg $US1998/kg $US1998/kg 03_Li 7.1E + 0 1.4E-1 1.9E + 1 9.3E + 1 4.9E + 2 2.2E + 1 4.5E + 0 3.0E + 0 1.4E + 0 4.2E-1 5.6E + 1 2.8E + 2 1.5E + 3 04_Be 2.6E + 1 3.8E-2 7.3E + 0 7.4E + 1 4.7E + 2 5.9E + 0 1.8E + 0 2.4E + 0 1.3E + 0 1.2E + 1 2.2E + 3 2.2E + 4 1.4E + 5 05_B 2.5E + 1 4.1E-2 1.1E + 1 7.5E + 1 4.8E + 2 6.3E + 0 2.6E + 0 2.4E + 0 1.3E + 0 2.4E-2 6.2E + 0 4.4E + 1 2.7E + 2 12_Mg 7.5E + 0 1.3E-1 2.1E + 1 9.2E + 1 4.9E + 2 2.0E + 1 5.0E + 0 2.9E + 0 1.4E + 0 4.8E-1 7.4E + 1 3.3E + 2 1.8E + 3 13_Al 7.6E + 1 1.3E-2 7.4E + 0 4.9E + 1 4.2E + 2 2.0E + 0 1.8E + 0 1.6E + 0 1.2E + 0 2.3E-2 1.3E + 1 8.4E + 1 7.3E + 2 14_Si 1.0E + 1 1.0E-1 1.6E + 1 9.0E + 1 4.9E + 2 1.5E + 1 3.9E + 0 2.9E + 0 1.4E + 0 1.7E-1 2.6E + 1 1.5E + 2 8.1E + 2 21_Sc 1.1E-2 8.7E + 1 2.5E + 1 1.0E + 2 5.0E + 2 1.3E + 4 6.0E + 0 3.2E + 0 1.4E + 0 4.6E + 5 1.3E + 5 5.3E + 5 2.6E + 6 22_Ti 1.4E + 1 7.4E-2 1.2E + 1 8.6E + 1 4.9E + 2 1.1E + 1 3.0E + 0 2.8E + 0 1.4E + 0 6.7E-1 1.1E + 2 7.8E + 2 4.4E + 3 23_V 8.0E + 0 1.3E-1 2.0E + 1 9.2E + 1 4.9E + 2 1.9E + 1 4.8E + 0 2.9E + 0 1.4E + 0 2.5E + 0 4.0E + 2 1.9E + 3 9.9E + 3 24_Cr 2.8E + 1 3.6E-2 1.1E + 1 7.3E + 1 4.7E + 2 5.5E + 0 2.8E + 0 2.3E + 0 1.3E + 0 6.2E-2 2.0E + 1 1.3E + 2 8.3E + 2 25_Mn 1.8E + 1 5.4E-2 1.6E + 1 8.2E + 1 4.8E + 2 8.3E + 0 3.8E + 0 2.6E + 0 1.4E + 0 5.7E-2 1.6E + 1 8.6E + 1 5.1E + 2 26_Fe 1.5E + 2 6.5E-3 4.1E + 0 3.1E + 1 3.5E + 2 1.0E + 0 1.0E + 0 1.0E + 0 1.0E + 0 3.7E-3 2.4E + 0 1.8E + 1 2.0E + 2 27_Co 4.1E + 0 2.5E-1 2.1E + 1 9.6E + 1 5.0E + 2 3.8E + 1 5.1E + 0 3.1E + 0 1.4E + 0 7.0E + 0 6.0E + 2 2.7E + 3 1.4E + 4 28_Ni 5.9E + 1 1.7E-2 8.5E + 0 5.7E + 1 4.4E + 2 2.6E + 0 2.0E + 0 1.8E + 0 1.2E + 0 2.6E-1 1.3E + 2 8.7E + 2 6.7E + 3 29_Cu 4.5E + 1 2.2E-2 6.5E + 0 5.9E + 1 4.5E + 2 3.4E + 0 1.6E + 0 1.9E + 0 1.3E + 0 1.2E-1 3.4E + 1 3.1E + 2 2.4E + 3 30_Zn 2.5E + 1 4.0E-2 1.1E + 1 7.6E + 1 4.8E + 2 6.2E + 0 2.7E + 0 2.4E + 0 1.3E + 0 7.1E-2 2.0E + 1 1.3E + 2 8.4E + 2 31_Ga 1.1E-1 9.0E + 0 2.5E + 1 1.0E + 2 5.0E + 2 1.4E + 3 6.0E + 0 3.2E + 0 1.4E + 0 3.3E + 3 9.2E + 3 3.7E + 4 1.8E + 5 32_Ge 4.7E-2 2.1E + 1 2.5E + 1 1.0E + 2 5.0E + 2 3.3E + 3 6.0E + 0 3.2E + 0 1.4E + 0 2.2E + 4 2.5E + 4 1.0E + 5 5.1E + 5 33_As 1.9E + 0 5.3E-1 2.4E + 1 9.8E + 1 5.0E + 2 8.2E + 1 5.7E + 0 3.1E + 0 1.4E + 0 2.5E-1 1.1E + 1 4.7E + 1 2.4E + 2 34_Se 5.3E-1 1.9E + 0 2.5E + 1 9.9E + 1 5.0E + 2 2.9E + 2 5.9E + 0 3.2E + 0 1.4E + 0 1.1E + 2 1.4E + 3 5.7E + 3 2.9E + 4 38_Sr 5.4E + 0 1.9E-1 2.0E + 1 9.5E + 1 4.9E + 2 2.9E + 1 4.9E + 0 3.0E + 0 1.4E + 0 1.3E-1 1.4E + 1 6.5E + 1 3.4E + 2 39_Y 5.7E + 0 1.7E-1 2.0E + 1 9.4E + 1 4.9E + 2 2.7E + 1 4.9E + 0 3.0E + 0 1.4E + 0 8.7E + 0 1.0E + 3 4.7E + 3 2.5E + 4 40_Zr 1.5E + 1 6.5E-2 1.3E + 1 8.5E + 1 4.8E + 2 1.0E + 1 3.2E + 0 2.7E + 0 1.4E + 0 5.9E-2 1.2E + 1 7.6E + 1 4.4E + 2 41_Nb 2.4E + 1 4.1E-2 1.1E + 1 7.6E + 1 4.8E + 2 6.4E + 0 2.7E + 0 2.4E + 0 1.3E + 0 1.7E + 0 4.5E + 2 3.1E + 3 2.0E + 4 42_Mo 2.4E + 1 4.1E-2 1.3E + 1 7.7E + 1 4.8E + 2 6.3E + 0 3.3E + 0 2.4E + 0 1.3E + 0 1.1E + 0 3.8E + 2 2.1E + 3 1.3E + 4 44_Ru 1.3E + 1 7.5E-2 1.5E + 1 8.7E + 1 4.9E + 2 1.2E + 1 3.5E + 0 2.8E + 0 1.4E + 0 2.4E + 2 4.7E + 4 2.8E + 5 1.6E + 6 45_Rh 2.4E + 1 4.2E-2 1.2E + 1 7.7E + 1 4.8E + 2 6.5E + 0 2.8E + 0 2.5E + 0 1.3E + 0 2.0E + 3 5.5E + 5 3.6E + 6 2.2E + 7 46_Pd 2.9E + 1 3.5E-2 1.0E + 1 7.3E + 1 4.7E + 2 5.4E + 0 2.5E + 0 2.3E + 0 1.3E + 0 3.7E + 2 1.1E + 5 7.8E + 5 5.1E + 6 47_Ag 4.4E + 1 2.3E-2 7.6E + 0 6.2E + 1 4.6E + 2 3.5E + 0 1.8E + 0 2.0E + 0 1.3E + 0 1.1E + 1 3.6E + 3 2.9E + 4 2.1E + 5 48_Cd 8.5E + 0 1.2E-1 1.7E + 1 9.2E + 1 4.9E + 2 1.8E + 1 4.2E + 0 2.9E + 0 1.4E + 0 3.0E-1 4.4E + 1 2.3E + 2 1.2E + 3 49_In 1.5E + 0 6.8E-1 2.4E + 1 9.9E + 1 5.0E + 2 1.1E + 2 5.7E + 0 3.1E + 0 1.4E + 0 3.0E + 2 1.0E + 4 4.3E + 4 2.2E + 5 50_Sn 1.2E + 1 8.4E-2 1.5E + 1 8.8E + 1 4.9E + 2 1.3E + 1 3.5E + 0 2.8E + 0 1.4E + 0 1.4E + 0 2.4E + 2 1.5E + 3 8.2E + 3 Table 1 Total expected service times (STTOT) corresponding to average lifetimes in the economy calculated by Charpentier Poncelet et al. 2.2.2 Coverage of mineral resource flows by selected LCIA methods (2022b), average dissipation rate (ADR), computed as loss rates by Charpentier Poncelet et al. (2022b), lost potential service time (LPST) with time horizons of 25, 100, and 500 years, and the associated midpoint and endpoint characterization fac- tors for the LPST and ADR methods STTOT (cf. ref. in caption) ADR (cf. ref. in caption) LPST25 LPST100 LPST500 Midpoint CFs Endpoint CFs CFADR CFLPST25 CFLPST100 CFLPST500 CFPVLR CFLPV25 CFLPV100 CFLPV500 Metal i kg.yr/kg kg/kg.yr kg.yr/kg kg.yr/kg kg.yr/kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg $US1998/ kg.yr $US1998/kg $US1998/kg $US1998/kg Total expected service times (STTOT) corresponding to average lifetimes in the economy calculated by Charpentier Poncelet et al. (2022b), average dissipation rate (ADR), computed as es by Charpentier Poncelet et al. (2022b), lost potential service time (LPST) with time horizons of 25, 100, and 500 years, and the associated midpoint and endpoint characterization fac- the LPST and ADR methods STTOT (cf. ref. in caption) ADR (cf. ref. in caption) LPST25 LPST100 LPST500 Midpoint CFs Endpoint CFs CFADR CFLPST25 CFLPST100 CFLPST500 CFPVLR CFLPV25 CFLPV100 CFLPV500 kg.yr/kg kg/kg.yr kg.yr/kg kg.yr/kg kg.yr/kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg $US1998/ kg.yr $US1998/kg $US1998/kg $US1998/kg 1 3 The International Journal of Life Cycle Assessment 1 3 ( ) STTOT (cf. ref. in caption) ADR (cf. ref. 2.2.2 Coverage of mineral resource flows by selected LCIA methods 2 Midpoint and endpoint characterization factors for the ADR method. CFs are shown in ascending order and log scale to facilitate comparison between methods. Black lines indicate the 95% confi- dence intervals. Average values for all CFs are provided in Table 1; values for the 95% confidence intervals are provided in Tables S1 and S2 of the SI 10 3 10 2 10 1 100 101 102 103 104 105 106 107 108 Fe Al B Ba Mn Zr Cr Pb Zn Cu Sr Si As Ni Cd Li Mg Sb Ti Mo Sn Nb Bi V Sm Ce La W Hg Co Er Ho Nd Yb Y Ag Pr Gd Be Ta Tm Dy Se Lu Au Tb Te Ru In Re Pd Eu Tl Pt Ir Rh Ga Hf Os Ge Sc Midpoint CFs - ADR (kg Fe-eq./kg) Endpoint CFs - ADR ($US1998/kg.yr) 10 1 100 101 102 103 104 105 106 Au Fe Al Ni Cu Ag Pt Pd Cr Pb Be Zn B Mo Nb Rh Mn Zr Er Ti Ho Ru Sb Sn Sm Re Si Ir Ta Cd Bi V Nd Mg Li Tl Dy Tm Yb Pr Gd Lu W Y Ce Sr Tb Hg La Co Os Ba Eu As In Te Se Ga Hf Ge Sc Ferrous metals Non-ferrous metals Specialty metals Precious metals 10 10 10 10 10 10 10 10 Endpoint CFs - ADR ($US1998/kg.yr) Ferrous metals Specialty metals Non-ferrous metals Fig. 2 Midpoint and endpoint characterization factors for the ADR method. CFs are shown in ascending order and log scale to facilitate comparison between methods. Black lines indicate the 95% confi- dence intervals. Average values for all CFs are provided in Table 1; values for the 95% confidence intervals are provided in Tables S1 and S2 of the SI CFs in the ADR method feature less prominently in the rank- ing of CFs. The six highest CFs are precious metals, i.e., rhodium (Rh), osmium (Os), platinum (Pt), iridium (Ir), gold (Au), and palladium (Pd). It shows that the price has a greater influence on the ranking of endpoint CFs for the LPST method than for the ADR method, because the mid- point CFLPST are less differentiated than the midpoint CFADR. The highest CFs for the midpoint ADR method are almost entirely specialty metals because they are typically dissi- pated the fastest. 2.2.2 Coverage of mineral resource flows by selected LCIA methods in caption) LPST25 LPST100 LPST500 Midpoint CFs Endpoint CFs CFADR CFLPST25 CFLPST100 CFLPST500 CFPVLR CFLPV25 CFLPV100 CFLPV500 Metal i kg.yr/kg kg/kg.yr kg.yr/kg kg.yr/kg kg.yr/kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg kg Fe-eq./kg $US1998/ kg.yr $US1998/kg $US1998/kg $US1998/kg 51_Sb 1.2E + 1 8.3E-2 1.5E + 1 8.8E + 1 4.9E + 2 1.3E + 1 3.6E + 0 2.8E + 0 1.4E + 0 5.1E-1 9.2E + 1 5.4E + 2 3.0E + 3 52_Te 6.3E-1 1.6E + 0 2.4E + 1 9.9E + 1 5.0E + 2 2.4E + 2 5.9E + 0 3.2E + 0 1.4E + 0 1.8E + 2 2.8E + 3 1.1E + 4 5.8E + 4 56_Ba 2.6E + 0 3.9E-1 2.2E + 1 9.7E + 1 5.0E + 2 6.0E + 1 5.4E + 0 3.1E + 0 1.4E + 0 4.8E-2 2.8E + 0 1.2E + 1 6.2E + 1 57_La 4.1E + 0 2.4E-1 2.1E + 1 9.6E + 1 5.0E + 2 3.8E + 1 5.1E + 0 3.1E + 0 1.4E + 0 4.5E + 0 3.9E + 2 1.7E + 3 9.0E + 3 58_Ce 5.6E + 0 1.8E-1 2.0E + 1 9.4E + 1 4.9E + 2 2.7E + 1 4.8E + 0 3.0E + 0 1.4E + 0 3.4E + 0 3.8E + 2 1.8E + 3 9.3E + 3 59_Pr 6.5E + 0 1.5E-1 1.9E + 1 9.4E + 1 4.9E + 2 2.4E + 1 4.6E + 0 3.0E + 0 1.4E + 0 1.1E + 1 1.3E + 3 6.5E + 3 3.4E + 4 60_Nd 7.8E + 0 1.3E-1 1.7E + 1 9.2E + 1 4.9E + 2 2.0E + 1 4.2E + 0 2.9E + 0 1.4E + 0 8.1E + 0 1.1E + 3 5.9E + 3 3.1E + 4 62_Sm 1.1E + 1 9.2E-2 1.5E + 1 8.9E + 1 4.9E + 2 1.4E + 1 3.5E + 0 2.8E + 0 1.4E + 0 3.0E + 0 4.8E + 2 2.9E + 3 1.6E + 4 63_Eu 2.6E + 0 3.9E-1 2.2E + 1 9.7E + 1 5.0E + 2 6.0E + 1 5.4E + 0 3.1E + 0 1.4E + 0 4.1E + 2 2.4E + 4 1.0E + 5 5.2E + 5 64_Gd 5.9E + 0 1.7E-1 1.9E + 1 9.4E + 1 4.9E + 2 2.6E + 1 4.6E + 0 3.0E + 0 1.4E + 0 1.1E + 1 1.2E + 3 6.0E + 3 3.1E + 4 65_Tb 5.1E + 0 2.0E-1 2.0E + 1 9.5E + 1 4.9E + 2 3.0E + 1 4.8E + 0 3.0E + 0 1.4E + 0 1.8E + 2 1.8E + 4 8.7E + 4 4.6E + 5 66_Dy 6.8E + 0 1.5E-1 1.8E + 1 9.3E + 1 4.9E + 2 2.3E + 1 4.4E + 0 3.0E + 0 1.4E + 0 6.4E + 1 8.0E + 3 4.1E + 4 2.1E + 5 67_Ho 1.4E + 1 7.4E-2 1.3E + 1 8.6E + 1 4.9E + 2 1.1E + 1 3.2E + 0 2.8E + 0 1.4E + 0 7.1E + 0 1.2E + 3 8.3E + 3 4.7E + 4 68_Er 1.4E + 1 7.0E-2 1.4E + 1 8.6E + 1 4.9E + 2 1.1E + 1 3.3E + 0 2.7E + 0 1.4E + 0 7.1E + 0 1.4E + 3 8.7E + 3 4.9E + 4 69_Tm 6.7E + 0 1.5E-1 1.9E + 1 9.3E + 1 4.9E + 2 2.3E + 1 4.5E + 0 3.0E + 0 1.4E + 0 3.2E + 1 4.0E + 3 2.0E + 4 1.1E + 5 70_Yb 6.6E + 0 1.5E-1 1.9E + 1 9.3E + 1 4.9E + 2 2.3E + 1 4.5E + 0 3.0E + 0 1.4E + 0 8.3E + 0 1.0E + 3 5.1E + 3 2.7E + 4 71_Lu 5.8E + 0 1.7E-1 2.0E + 1 9.4E + 1 4.9E + 2 2.6E + 1 4.8E + 0 3.0E + 0 1.4E + 0 1.4E + 2 1.6E + 4 7.6E + 4 4.0E + 5 72_Hf 7.2E-2 1.4E + 1 2.5E + 1 1.0E + 2 5.0E + 2 2.1E + 3 6.0E + 0 3.2E + 0 1.4E + 0 4.5E + 3 8.1E + 3 3.2E + 4 1.6E + 5 73_Ta 9.2E + 0 1.1E-1 1.7E + 1 9.1E + 1 4.9E + 2 1.7E + 1 4.0E + 0 2.9E + 0 1.4E + 0 1.6E + 1 2.4E + 3 1.3E + 4 7.1E + 4 74_W 5.7E + 0 1.7E-1 1.9E + 1 9.4E + 1 4.9E + 2 2.7E + 1 4.7E + 0 3.0E + 0 1.4E + 0 5.0E + 0 5.6E + 2 2.7E + 3 1.4E + 4 75_Re 1.0E + 1 9.8E-2 1.7E + 1 9.0E + 1 4.9E + 2 1.5E + 1 4.0E + 0 2.9E + 0 1.4E + 0 3.2E + 2 5.4E + 4 2.9E + 5 1.6E + 6 76_Os 3.6E + 0 2.8E-1 2.1E + 1 9.6E + 1 5.0E + 2 4.3E + 1 5.2E + 0 3.1E + 0 1.4E + 0 5.4E + 3 4.2E + 5 1.9E + 6 9.8E + 6 77_Ir 9.5E + 0 1.0E-1 1.7E + 1 9.0E + 1 4.9E + 2 1.6E + 1 4.1E + 0 2.9E + 0 1.4E + 0 1.3E + 3 2.1E + 5 1.1E + 6 6.0E + 6 78_Pt 4.1E + 1 2.4E-2 1.1E + 1 6.6E + 1 4.6E + 2 3.8E + 0 2.6E + 0 2.1E + 0 1.3E + 0 6.5E + 2 2.8E + 5 1.8E + 6 1.2E + 7 79_Au 1.9E + 2 5.2E-3 7.6E + 0 3.9E + 1 3.3E + 2 8.0E-1 1.8E + 0 1.2E + 0 9.4E-1 1.4E + 2 2.0E + 5 1.0E + 6 9.0E + 6 80_Hg 4.6E + 0 2.2E-1 2.1E + 1 9.5E + 1 5.0E + 2 3.3E + 1 5.0E + 0 3.0E + 0 1.4E + 0 5.8E + 0 5.5E + 2 2.5E + 3 1.3E + 4 81_Tl 7.0E + 0 1.4E-1 1.8E + 1 9.3E + 1 4.9E + 2 2.2E + 1 4.4E + 0 3.0E + 0 1.4E + 0 6.3E + 2 7.9E + 4 4.1E + 5 2.2E + 6 82_Pb 2.7E + 1 3.7E-2 9.7E + 0 7.4E + 1 4.7E + 2 5.6E + 0 2.3E + 0 2.3E + 0 1.3E + 0 6.3E-2 1.7E + 1 1.3E + 2 8.2E + 2 83_Bi 8.0E + 0 1.2E-1 1.7E + 1 9.2E + 1 4.9E + 2 1.9E + 1 4.1E + 0 2.9E + 0 1.4E + 0 2.0E + 0 2.7E + 2 1.5E + 3 7.8E + 3 1 The International Journal of Life Cycle Assessment 10 3 10 2 10 1 100 101 102 103 104 105 106 107 108 Fe Al B Ba Mn Zr Cr Pb Zn Cu Sr Si As Ni Cd Li Mg Sb Ti Mo Sn Nb Bi V Sm Ce La W Hg Co Er Ho Nd Yb Y Ag Pr Gd Be Ta Tm Dy Se Lu Au Tb Te Ru In Re Pd Eu Tl Pt Ir Rh Ga Hf Os Ge Sc Midpoint CFs - ADR (kg Fe-eq./kg) Endpoint CFs - ADR ($US1998/kg.yr) 10 1 100 101 102 103 104 105 106 Au Fe Al Ni Cu Ag Pt Pd Cr Pb Be Zn B Mo Nb Rh Mn Zr Er Ti Ho Ru Sb Sn Sm Re Si Ir Ta Cd Bi V Nd Mg Li Tl Dy Tm Yb Pr Gd Lu W Y Ce Sr Tb Hg La Co Os Ba Eu As In Te Se Ga Hf Ge Sc Ferrous metals Non-ferrous metals Specialty metals Precious metals Fig. 2.2.2 Coverage of mineral resource flows by selected LCIA methods Endpoint CFs are dramatically different for precious metals, whose price indexes are consistently the highest. The latter are among the highest ranked end- point CFs. Still, a few rapidly dissipating specialty metals with a high annual average market price, i.e., scandium (Sc), germanium (Ge), hafnium (Hf), and gallium (Ga), remain among the highest CFs, i.e., the first, second, fourth, and fifth CFs, respectively. Conversely, endpoint CFs for ferrous and non-ferrous metals remain in the bottom half of CFs for both the midpoint and the endpoint. 3.1.2 Limitations for the endpoint characterization factors Average values for all CFs are provided in Table 1; values for the 95% confidence intervals are provided in Tables S1 and S2 of the SI Midpoint CFs - LPST100 (kg Fe-eq./kg) Endpoint CFs - LPST100 ($US1998/kg) Ferrous metals Non-ferrous metals Specialty metals Precious metals 10 1 100 101 Fe Au Al Ni Cu Ag Pt Pd Cr Pb Be B Zn Nb Mo Rh Mn Zr Er Ti Ho Ru Sb Sn Sm Re Si Ir Ta Cd Bi V Nd Mg Li Tl Dy Tm Yb Pr Gd Lu W Y Ce Sr Tb Hg La Co Os Ba Eu As In Te Se Ga Hf Ge Sc 101 102 103 104 105 106 107 Ba Fe B As Sr Zr Al Mn Pb Cr Zn Si Cd Li Cu Mg Sb Ti Ni Bi Sn La Ce V Mo Hg W Co Sm Nb Y Yb Se Nd Gd Pr Ho Er Te Ta Tm Be Ag Hf Ga Dy In Lu Tb Ge Eu Ru Re Tl Sc Pd Au Ir Pt Os Rh Fig. 3 Midpoint and endpoint characterization factors for the LPST100 method. CFs are shown in ascending order and log scale to facilitate comparison between methods. Black lines indicate the 95% confidence intervals. Average values for all CFs are provided in Table 1; values for the 95% confidence intervals are provided in Tables S1 and S2 of the SI Midpoint CFs - LPST100 (kg Fe-eq./kg) 10 Ferrous metals Specialty metals 95% confidence intervals. Average values for all CFs are provided in Table 1; values for the 95% confidence intervals are provided in Tables S1 and S2 of the SI Fig. 3 Midpoint and endpoint characterization factors for the LPST100 method. CFs are shown in ascending order and log scale to facilitate comparison between methods. Black lines indicate the and Eggert 2021) that may cause a diversion from the assumed relationship between the price and the actual value of metals for humans over time. We decided to maintain this information in the price statistics because there is not much precise information available on the share of different production costs on the price of most metals (Huppertz et al. 2019), and because we assumed that the efforts put into pro- duction also partly reflect the utility of metals for humans. 3.1.2 Limitations for the endpoint characterization factors The limitations of the midpoint ADR and LPST methods were highlighted in previous work (Charpentier Poncelet et al. 2021). We here identify limitations linked with the use of price statistics for the computation of endpoint CFs. Firstly, the prices of metals include production costs influ- enced by several factors such as the price of energy (Watson As for the ADR method, the highest CFs for the midpoint LPST100 method are almost entirely those of specialty met- als, and endpoint CFs are much higher for precious metals. The most rapidly dissipating metals with the largest endpoint 1 3 The International Journal of Life Cycle Assessment Midpoint CFs - LPST100 (kg Fe-eq./kg) Endpoint CFs - LPST100 ($US1998/kg) Ferrous metals Non-ferrous metals Specialty metals Precious metals 10 1 100 101 Fe Au Al Ni Cu Ag Pt Pd Cr Pb Be B Zn Nb Mo Rh Mn Zr Er Ti Ho Ru Sb Sn Sm Re Si Ir Ta Cd Bi V Nd Mg Li Tl Dy Tm Yb Pr Gd Lu W Y Ce Sr Tb Hg La Co Os Ba Eu As In Te Se Ga Hf Ge Sc 101 102 103 104 105 106 107 Ba Fe B As Sr Zr Al Mn Pb Cr Zn Si Cd Li Cu Mg Sb Ti Ni Bi Sn La Ce V Mo Hg W Co Sm Nb Y Yb Se Nd Gd Pr Ho Er Te Ta Tm Be Ag Hf Ga Dy In Lu Tb Ge Eu Ru Re Tl Sc Pd Au Ir Pt Os Rh Fig. 3 Midpoint and endpoint characterization factors for the LPST100 method. CFs are shown in ascending order and log scale to facilitate comparison between methods. Black lines indicate the 95% confidence intervals. 3.1.2 Limitations for the endpoint characterization factors It should be noted that market prices tend to be more volatile than the value of metals over time that we intend to quantify using these data. Secondly, the large indirect investments and production costs associated with by-product metals may lead to both low production yields and market inefficien- cies, potentially keeping their prices up (Watson and Eggert 2021). Consequently, some by-product metals like gallium and scandium have relatively high midpoint CFs (production losses are here accounted for as dissipation, leading to highly dissipative profiles for these metals) and even greater end- point CFs (due to their relatively high prices). Thirdly, unquantified uncertainty may arise from using price data from different sources, for metals of possibly different quali- ties or purities, and in a few instances, over different time series. Fourthly, some metals that are often used in a com- pound form, such as the magnesium content of magnesia or the lanthanum content of lanthanum oxides, may not have the same potential value as their refined metal form. Nev- ertheless, most metals are almost exclusively used as pure metals or alloys (Graedel et al. 2022; UNEP 2011). Fifthly, the price of a few metals are determined from the price of compounds by considering the stoichiometric content of the metals in the compounds, which may not be consist- ent with how other price statistics are calculated. Finally, losses are here accounted for as dissipation, leading to highly dissipative profiles for these metals) and even greater end- point CFs (due to their relatively high prices). Thirdly, unquantified uncertainty may arise from using price data from different sources, for metals of possibly different quali- ties or purities, and in a few instances, over different time series. Fourthly, some metals that are often used in a com- pound form, such as the magnesium content of magnesia or the lanthanum content of lanthanum oxides, may not have the same potential value as their refined metal form. Nev- ertheless, most metals are almost exclusively used as pure metals or alloys (Graedel et al. 2022; UNEP 2011). Fifthly, the price of a few metals are determined from the price of compounds by considering the stoichiometric content of the metals in the compounds, which may not be consist- ent with how other price statistics are calculated. 3.2.1 Midpoint impact assessment per section of economic activity Figure 4 shows the relative contribution of metals to inven- tory results split in eight sections of economic activity and relative midpoint impacts for selected LCIA methods. Met- als contributing to over 10% of the total impacts for at least one method are shown individually on the figure for all columns of the corresponding economic section; others are grouped altogether. g p g Iron largely dominates resource extraction in the inven- tories among all sections, with 76% (section C: manufac- turing) to 98% (section B: mining and quarrying) of their respective inventory shares. Despite its CF being one of the two smallest for the ADR and LPST methods, iron consist- ently comes out as one of the main contributors for the ADR method, and is increasingly important for the LPST25, 100 and 500 methods. It should be noted that, given the overall great recyclability of iron (or steel) in most of its applica- tions (Pauliuk et al. 2017), its relative impacts would likely be much lower for LCI that account for the recycling in the end-of-life modeling beyond what is already accounted for in LCI databases (the same could be expected for other well- recycled metals). In this case, the extraction of metal would be allocated between different product systems (or between different life cycles) and therefore would be lower than a product system using only primary metal. As for other met- als, iron’s relative share of impacts is increasingly similar to its share of the inventory flows when considering longer time horizons for the assessment using the LPST method, because the midpoint CFLPST25 are more differentiated than the CFLPST100 and the CFLPST500 (i.e., they spread from 1.0 to 6.0, 1.0 to 3.2, and 1.0 to 1.4 kg Fe-eq./kg, respectively). Similarly to the ADR and LPST methods, widely extracted copper, iron, and nickel recurrently show up as important contributors to the impact assessment with the SOP method included in ReCiPe2016. For that method, iron contributes over 10% of the total impacts for all economic sections except section C (manufacturing), i.e., from 36% of the total impacts for section H (transportation and storage) to 87% for section B (mining and quarrying). 3.1.2 Limitations for the endpoint characterization factors Finally, 1 3 3 The International Journal of Life Cycle Assessment the volatility of market prices may have a significant effect on the average price of metals. For instance, the price of different REEs increased by a factor of 3 to 27 during the price peak of 2011 (Bru et al. 2015). For other metals, price variations typically ranged between ± 50 and 150% of the average price over the 2006–2015 period. Thus, using the average of yearly prices over a decade reduces the short-term variations associated with volatility and may provide more stable indications of the value of metals. Additional research is needed to address these challenging limitations. methods. Indeed, it is mostly used for gas and oil well drill- ing under its mineral form of barites, explaining both its extensive use in multiple economic sections and its highly dissipative profile that is best distinguished with the mid- point ADR method. Zinc also importantly contributes to the impact scores for the ADR and LPST methods for section C (manufacturing) and the other sections. As an indication, barium and zinc represent 0.7% and 4.5% of the inven- tory flows by weight for section C, while they respectively account for 18% and 12% of the total impacts for the mid- point ADR method. Nickel is widely used in section F (con- struction) with 3.1% of the inventory total, and comes up as the third highest contribution to the ADR and LPST assess- ments for that section with approximately 3 to 5% of their total impacts. Similarly, aluminum flows represent 4.7% of the inventory totals for section C (just over zinc), but its midpoint CFADR and CFLPST100 are 67% and 35% lower than the corresponding CFs for zinc, explaining its lower share of the impacts for that section (i.e., 4.0% for ADR and 5.8% for LPST100). For the same reason as aluminum and nickel, chromium and manganese do not contribute over 10% of the impacts for the midpoint ADR and LPST methods for any section of economic activity despite being widely extracted. 3.2.1 Midpoint impact assessment per section of economic activity For instance, barium is more often than not part of the main contributors for the total dissipation impacts as assessed with the ADR and LPST 1 3 The International Journal of Life Cycle Assessment Section B: Mining and quarrying Section A: Agriculture, forestry and fishing Iron Copper Gold Barium Zinc Tellurium Platinum Palladium Nickel Silver Magnesium Section C: Manufacturing Electricity, gas, steam & air conditioning supply Water supply; sewerage, waste management & remediation activities Section F: Construction Metals Others 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% Section D: Section E: Section B: Mining and quarrying Section A: Agriculture, forestry and fishing El & 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% 1 for sections A (18% of the impact total), B (15%), D (16%), E (12%), and F (16%), while its shares of inventory totals range from 0.94 to 1.5% across these economic sections. Aside from copper, only scarce metals with very low crustal concentrations recurrently come up as important contributors for the ADP ultimate reserves method. 3.2.1 Midpoint impact assessment per section of economic activity Meanwhile, nickel generally contributes to a larger share of the impacts for the SOP method than for the ADR and LPST methods, with, e.g., 45% of the relative impacts of section F (con- struction) and 11% of those of section H (transportation and storage). Finally, copper represents 12% of the total impacts for the SOP method for section A (agriculture, forestry and fishing), 14% for section D (electricity, gas, steam and air conditioning supply), 10% for section E (water supply; sew- erage, waste management and remediation activities), and 9.0% for section F (construction). In contrast to the other methods, the relative impacts of iron are negligible for ADP ultimate reserves because its CF is among the lowest for that method, while those of the scarcest metals are several orders of magnitude higher (van Oers et al. 2019). For example, the CF of gold in the latter method is 2 billion times higher than iron (cf. Table S4). Copper, which has a lower crustal concentration (28 ppm) than other widely extracted metals (e.g., 63 ppm for bar- ium, and 774 ppm for manganese: van Oers et al. 2019), is the only widely extracted metal with over 10% of the total impacts for the ADP method for at least one section of eco- nomic activity. It contributes most to the impact assessment Aside from iron, the only metals showing up among the highest relative contributions to the midpoint ADR and LPST impact assessments are other widely extracted metals typically representing around 1 to 5% of inventory shares for different economic sections. 3.2.1 Midpoint impact assessment per section of economic activity Metals contribut- ing over 10% of the total impacts for at least one impact method are shown individually for all columns of the corresponding section of economic activity; others are grouped altogether. The inventory col- umn presents the relative mass of resource flows in the compiled LCI data sets ◂ Fig. 4 Contribution of metals to inventory totals and impact scores for four midpoint LCIA methods. Graphs present results per section of economic activity established in the ISIC. ADP 2015: ADP ulti- mate reserves method for elements based on the cumulative produc- tion in 2015 (van Oers et al. 2019); ReCiPe 2016: results for the SOP method (Vieira et al. 2017) included in the latter. Metals contribut- ing over 10% of the total impacts for at least one impact method are shown individually for all columns of the corresponding section of economic activity; others are grouped altogether. The inventory col- umn presents the relative mass of resource flows in the compiled LCI data sets ◂ This convergence of results for the endpoint LPST100 and ReCiPe 2016 methods can be explained by the simi- larities between the surplus amount of ores assumed to be required to produce scarcer metals in the future, its assumed cost, and the higher economic value of these same metals. Indeed, production costs represent approximately 50–75% of the market price of metals (Huppertz et al. 2019). The price strongly influences the computation of endpoint CFs for the LPST methods, especially over longer time horizons (cf. Figure 3 for the LPST100 method, and Figs. S2 and S3 for the LPST25 and LPST500 methods, respectively). As shown in Fig. S5 of the SI, these endpoint methods are similarly sensitive to the relative shares of inventory totals for iron, copper, gold, and nickel. The dissipation of scarcer metals is costlier for society because they require relatively more ore to produce (as accounted for in the midpoint ReC- iPe 2016 method, i.e., SOP), and the endpoint ReCiPe 2016 method, i.e., SCP, assumes they are also more costly to pro- duce. The cost of production is also reflected in the price information underlying the endpoint CFs for the ADR and LPST methods. are gold, palladium, platinum, silver, and tellurium. 3.2.1 Midpoint impact assessment per section of economic activity Pre- cious metals are also revealed to be important contributors to the ReCiPe 2016’s impact assessments, but only for sec- tions with the highest shares of these metals in the inventory totals, i.e., gold, palladium, and platinum in sections C and H, and silver in other sections. Indeed, although the differ- ence is less marked than for ADP, the CFs for precious met- als are also among the highest in the SOP method because they require a lot of additional ores to produce. For instance, the CFs of gold and platinum are 5 and 6 orders of magni- tude greater than iron’s and rank thirtieth and thirty-third highest out of 33 CFs of the ReCiPe 2016 method consid- ered in this study, respectively (cf. Table S4). In contrast to the ADP and SOP methods, precious metals do not appear among highest contributors to total impact scores for the midpoint ADR and LPST methods because their midpoint CFs consistently rank among the smallest, as observable in Figs. 2 and 3. 3.2.3 Comparison of the impacts of data sets across methods Figure 6 shows the total impact scores for the selected LCIA methods applied to the forty-five studied metals across all market processes from the ecoinvent 3.1.7 database. Both X- and Y-axes are in log scale. Please note that the scale of the X-axes may vary between graphs. R2 values represent the correlation of the log–log regression of impact scores per data set. Lower R2 values indicate that the compared meth- ods present different impact hotspots among the covered metals. Three additional scatter plots are presented in the SI. 3.2.1 Midpoint impact assessment per section of economic activity These Iron Copper Gold Barium Zinc Tellurium Platinum Palladium Nickel Silver Magnesium Section C: Manufacturing Electricity, gas, steam & air conditioning supply Water supply; sewerage, waste management & remediation activities Section F: Construction Section H: Transportation and storage Other sections Inventory ADP 2015 ReCiPe 2016 LPST25 LPST100 LPST500 ADR Metals Others 0% 20% 40% 60% 80% 0% 20% 40% 60% 80% 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% Inventory ADP 2015 ReCiPe 2016 LPST25 LPST100 LPST500 ADR Section D: Section E: Barium Zinc Magnesium Section C: Manufacturing Electricity, gas, steam & air conditioning supply Water supply; sewerage waste Metals Others 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% Section D: Iron Copper Gold Tellurium Platinum Palladium Nickel Silver management & remediation activities Section F: Construction 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% Section E: Section H: Transportation and storage Other sections Inventory ADP 2015 ReCiPe 2016 LPST25 LPST100 LPST500 ADR 0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100% Inventory ADP 2015 ReCiPe 2016 LPST25 LPST100 LPST500 ADR Aside from copper, only scarce metals with very low crustal concentrations recurrently come up as important contributors for the ADP ultimate reserves method. These for sections A (18% of the impact total), B (15%), D (16%), E (12%), and F (16%), while its shares of inventory totals range from 0.94 to 1.5% across these economic sections. The International Journal of Life Cycle Assessment is more dramatic for the most expensive metals (e.g., gold) in the LPST method than the ADR method. Cheaper met- als such as barium and iron see their shares of the impacts diminish importantly in the endpoint assessment. Fig. 4 Contribution of metals to inventory totals and impact scores for four midpoint LCIA methods. Graphs present results per section of economic activity established in the ISIC. ADP 2015: ADP ulti- mate reserves method for elements based on the cumulative produc- tion in 2015 (van Oers et al. 2019); ReCiPe 2016: results for the SOP method (Vieira et al. 2017) included in the latter. 3.2.2 Midpoint versus endpoint impact assessment 6c, are rather well correlated (R2 = 0.991) given that they both rely on the service time 1 3 The International Journal of Life Cycle Assessment no endpoint CFs Total midpoint impacts 0% 20% 40% 60% 80% 100% Inventory ADP 2015 ReCiPe 2106 LPST25 LPST100 LPST500 ADR Total endpoint impacts Inventory ADP 2015 ReCiPe 2106 LPST25 LPST100 LPST500 ADR 0% 20% 40% 60% 80% 100% Tellurium Platinum Palladium Nickel Manganese Copper Chromium Aluminium Gallium Selenium Gold Barium Magnesium Metals Others Iron Fig. 5 Inventory contributions and total midpoint and endpoint impact scores for selected LCIA methods applied to 5999 market data sets from the ecoinvent 3.7.1 database. ADP 2015: ADP ultimate reserves method for elements based on the cumulative production in 2015 (van Oers et al. 2019); ReCiPe 2016 (midpoint): results for the SOP method (Vieira et al. 2017); ReCiPe 2016 (endpoint): results for the SCP method (Vieira et al. 2016). Metals representing the top five contributions to the total inventory flows and for each LCIA method are shown in their respective columns; others are grouped altogether Fig. 6 Comparison of impact assessment for six selected pairs of LCIA methods, covering 5,999 market data sets organized by section of economic activity defined in the ISIC. a Midpoint LPST100 vs. ADP 2015; b midpoint LPST100 vs. ReCiPe 2016; c midpoint ADR vs. LPST100; d midpoint ADR vs. ADP 2015; e endpoint ADR vs. LPST100; and f endpoint LPST100 vs. ReCiPe 2016. R2 values rep- resent the correlation of the log–log regression of impact scores per data set ◂ from iron flows because they are the largest resource flows, and from other widely extracted metals with higher CFs than others (e.g., barium and zinc). The impact scores of the LPST100 midpoint method are better correlated with ReCiPe 2016 (Fig. 6b; R2 = 0.985) than those of ADP 2015 (Fig. 6a; R2 = 0.937). Logically, ReCiPe 2016’s impact scores have a lower correlation with those of ADP 2015 (R2 = 0.967; cf. Fig. S6 of the SI) than with the other two methods. This can partially be explained by the fact that the CFs of the ReCiPe 2016 method present a stronger convergence with those of the LPST100 method than the ADP2015 method (cf. Table S4 in the SI). 3.2.2 Midpoint versus endpoint impact assessment Another reason for the stronger convergence between ReCiPe 2016 and the LPST100 results is that their CFs are much less differentiated than those of ADP 2015. Therefore, their impact scores are more deter- mined by the relative size of metal flows in the LCI in comparison to the ADP 2015 method. Contrastingly, the pronounced differentiation between CFs for the ADP 2015 method (cf. Table S4 in the SI) leads to acute hotspots for a few very scarce metals, despite their very small resource flows in the inventory. no endpoint CFs Inventory ADP 2015 ReCiPe 2106 LPST25 LPST100 LPST500 ADR 0% 20% 40% 60% 80% 100% Platinum Palladium Nickel Chromium Gold Iron Fig. 5 Inventory contributions and total midpoint and endpoint impact scores for selected LCIA methods applied to 5999 market data sets from the ecoinvent 3.7.1 database. ADP 2015: ADP ultimate reserves method for elements based on the cumulative production in 2015 (van Oers et al. 2019); ReCiPe 2016 (midpoint): results for the SOP method (Vieira et al. 2017); ReCiPe 2016 (endpoint): results for the SCP method (Vieira et al. 2016). Metals representing the top five contributions to the total inventory flows and for each LCIA method are shown in their respective columns; others are grouped altogether Fig. 5 Inventory contributions and total midpoint and endpoint impact scores for selected LCIA methods applied to 5999 market data sets from the ecoinvent 3.7.1 database. ADP 2015: ADP ultimate reserves method for elements based on the cumulative production in 2015 (van Oers et al. 2019); ReCiPe 2016 (midpoint): results for the SOP method (Vieira et al. 2017); ReCiPe 2016 (endpoint): results for the SCP method (Vieira et al. 2016). Metals representing the top five contributions to the total inventory flows and for each LCIA method are shown in their respective columns; others are grouped altogether l The correlation between endpoint ADR and LPST100 results is similar to their midpoint assessments, as shown by comparing Fig. 6c (R2 = 0.991) and Fig. 6e (R2 = 0.991). Even though the price of metals have a greater influence on the CFLPST100 than the CFADR, their endpoint CFs have similar rankings. The few metals with much larger CFs in the ADR method are metals with the highest dissipa- tion rates and high prices. 3.2.2 Midpoint versus endpoint impact assessment We here investigate differences between the midpoint and the endpoint impacts assessed with the selected LCIA meth- ods, as depicted in Fig. 5. Especially, we observe the effect of considering the price of metals in the endpoint assess- ment. For brevity, we here focus on general observations, and the comparison is elaborated with quantitative examples in section S5 of the SI. ReCiPe 2016’s relative impacts are the same for the midpoint and endpoint assessment for the reason exposed in “Sect. 2.2.1.” Contrastingly, there is an important shift in the relative shares of impacts between midpoint and endpoint for the ADR and LPST methods, because the price information allows further differentia- tion between metals. The impacts due to expensive metals increase dramatically, while those of cheaper ones diminish. Because of the way the CFADR are computed, the effect is most dramatic for metals that dissipate relatively quickly and that have relatively high annual average market prices (such as gallium) putting forward potentially substantial differ- ences between the endpoint impact assessment results for the ADR or the LPST methods. Since the prices have a stronger effect on the endpoint CFs for the LPST method, the increase The impact scores range over 18 orders of magnitude across the data sets. The relatively high correlations between impact scores can partially be explained by the differences in the relative size of functional units (i.e., the variations in their total resource extraction flows). Indeed, the sum of metal resource flows spreads over nine orders of magnitude across data sets. Normalizing or rescaling functional units could reduce such variations, going beyond what could be achieved in this article. For instance, the results of Rørbech et al. (2014) show that correlations between impact scores per data set could be expected to be much smaller after res- caling functional units to a more comparable scale. The impact scores for the midpoint ADR and LPST100 methods, as shown in Fig. 3.2.2 Midpoint versus endpoint impact assessment Relatively small resource flows of such metals in the inventories translate into relatively large shares of the total impacts, the most obvious exam- ple being gallium (as observable in Fig. 5). It should be noted that other similar metals with high CFs do not have resource flows in the ecoinvent 3.7.1 database (e.g., ger- manium and scandium). Thus, data sets are expected to present noticeably greater impact scores with the endpoint ADR method than the endpoint LPST100 method due to higher-than-average resource flows of highly dissipative, relatively expensive metals (e.g., gallium and germanium; cf. Figure 2). For instance, gallium contributes an aver- age of 0.016% to the inventory totals for the 100 data sets with the largest difference between the total impact scores for the endpoint ADR and LPST100 methods. In contrast, gallium’s average share of inventory totals is of (albeit with a different conceptual approach; cf. Equa- tions 1 and 2) and on the same underlying data. Thus, their CFs rank almost the same across metals. The comparison between impact scores for the midpoint ADR method and those of the ADP ultimate reserves method have the lowest correlation (Fig. 6d; R2 = 0.927), followed by those of mid- point LPST100 versus ADP methods (Fig. 6a; R2 = 0.937). Data sets responsible for most impacts for the ADP 2015 are expected to be quite different from those for the ADR and LPST methods. The former are influenced almost exclusively by data sets with larger resource flows of the scarcest elements in the crust (e.g., precious metals and tellurium), whereas the latter are influenced not only by the dissipation patterns of different metals but also by the mass of their resource flows (cf. “Sect. 3.2.1”). 3.2.2 Midpoint versus endpoint impact assessment The impact scores for the latter midpoint methods thus result mostly 1 3 3 3 3 ternational Journal of Life Cycle Assessment 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E+13 ADP 2015 (kg Fe-eq.) LPST 100 (midpoint, kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 ADR (midpoint, kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 LPST100 (midpoint, kg Fe-eq.) Section B: Mining and quarrying Section A: Agriculture, forestry and fishing Section C: Manufacturing Section D: Electricity, gas, steam & air conditionning supply Section F: Construction Section H: Transportation and storage Other sections Section E: Water supply; sewerage, waste management & remediation activities LPST 100 (midpoint, kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 ReCiPe 2016 (kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 b c ADR (midpoint, kg Fe-eq.) ADP 2015 (kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E+13 d a R = 0.937 2 R = 0.985 2 R = 0.927 2 R = 0.991 2 ADR (endpoint, kg Fe-eq.) LPST100 (endpoint, kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 e R = 0.991 2 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 ReCiPe 2016 (kg Fe-eq.) LPST 100 (endpoint, kg Fe-eq.) f R = 0.995 2 The International Journal of Life Cycle Assessment 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E+13 ADP 2015 (kg Fe-eq.) LPST 100 (midpoint, kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 ADR (midpoint, kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 LPST100 (midpoint, kg Fe-eq.) c a R = 0.937 2 R = 0.991 2 DR (endpoint, kg Fe-eq.) 1E+01 1E+04 1E+07 1E+10 e LPST 100 (midpoint, kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 ReCiPe 2016 (kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 b R = 0.985 2 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E+13 ADP 2015 (kg Fe-eq.) LPST 100 (midpoint, kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 a R = 0.937 2 a 0 5 ( g e eq.) ADR (midpoint, kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 LPST100 (midpoint, kg Fe-eq.) c R = 0.991 2 ADR (midpoint, kg Fe-eq.) ADP 2015 (kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E+13 d R = 0.927 2 d ADP 2015 (kg Fe-eq.) 1 1E 04 1E 07 1 ( p g q ) ADR (endpoint, kg Fe-eq.) LPST100 (endpoint, kg Fe-eq.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 e R = 0.991 2 ( g q.) 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 1E-08 1E-05 1E-02 1E+01 1E+04 1E+07 1E+10 ReCiPe 2016 (kg Fe-eq.) LPST 100 (endpoint, kg Fe-eq.) f R = 0.995 2 Section B: Mining and quarrying Section A: Agriculture, forestry and fishing Section C: Manufacturing Section D: Electricity, gas, steam & air conditionning supply Section F: Construction Section H: Transportation and storage Other sections Section E: Water supply; sewerage, waste management & remediation activities 1 3 The International Journal of Life Cycle Assessment Poncelet et al. 3.2.2 Midpoint versus endpoint impact assessment (2021) and complemented in this article. It could also be possible to adapt the model to assess the dissi- pation of plastics and that of biotic resources used in similar sectors as mineral resources (e.g., wood). 0.0054% across all of the data sets. Finally, the compari- son between endpoint category totals for the LPST100 method and those of the ReCiPe 2016 method (Fig. 6f; R2 = 0.995) shows an even higher convergence than for the respective midpoint assessments, as partly explained by the relation between the SCP and the price information used for the computation of endpoint CFs for the LPST method exposed in “Sect. 3.2.2.” Furthermore, additional research could aim to improve the evaluation of the economic and use values (i.e., the value of services provided by applications) of mineral resources, as defined by Charpentier Poncelet et al. (2022a). A differ- ent methodology could be developed to evaluate endpoint damage in a more reliable way than by using annual aver- age market prices. Other metrics could also be considered such as the global economic importance of resources com- puted for criticality studies (Graedel et al. 2012). Moreover, efforts could be spent on developing CFs considering dif- ferent assumptions on the future recovery of metals from, e.g., waste disposal facilities (see e.g. Dewulf et al. 2021). Such assumptions could build on different cultural perspec- tives, as discussed by Charpentier Poncelet et al. (2022a). Finally, CFs of the ADR and LPST methods could apply to dissipative flows (or “resource inaccessibility” flows) rather than extraction flows if these were identified in the LCI. For instance, dissipative flows due to a given product sys- tem could be estimated by subtracting functionally recycled metal flows from extracted metal flows. We discuss such perspectives in more details in section S8 of the SI, and investigate how different future recovery scenarios could be implemented in such an assessment under different cultural perspectives. 4 Conclusion The extended ADR and LPST methods presented in this article allow evaluating the impacts of dissipation of metals in LCA. Because of the lack of information on dissipative flows in the LCI, they so far apply to extraction flows, and thus no distinction is made between processes or product systems that allow for recycling. Potential workaround solu- tions addressing this issue are presented in section S8 of the SI, along with potential ways forward to account for the future recovery of metals in dissipation-oriented approaches. Further research is needed to compare LCIA results when applying the CFs from the ADR and LPST methods to dis- sipative flows in the inventory (e.g., using the process-based JRC approach; see Beylot et al. 2021) instead of extraction flows and evaluate how that would influence impact assess- ment results. 3.3 Evaluation of ADR and LPST methods We evaluated the extended ADR and LPST methods against five criteria adapted from the ILCD handbook (European Commission 2010). We hereby summarize our observations; the detailed evaluation can be found in section S7 of the SI. Completeness of scope: the methods provide a significant coverage of metal resources at the global scale, but do not include mineral compound flows (e.g., sand and talc), nor fossil fuels and their derived products like plastics. Relevance for the assessment of mineral resource use on the AoP natural resources: the methods rely on real-life statistics and provide a global evaluation of dissipation patterns for the studied met- als. The CFs apply to extraction flows, and hence do not allow distinguishing between different product systems that may dis- sipate metals differently. They address the safeguard subject for mineral resource use established by the Life Cycle Initia- tive, i.e., the “potential to make use of the value that mineral resources can hold for humans in the technosphere” (Berger et al. 2020). Scientific robustness and certainty: the methods rely on data and methods that are published or to be published as peer-reviewed articles. Uncertainty is addressed and where practical accounted for Documentation, Transparency, and Reproducibility: The underlying data and model are docu- mented transparently and are available in open-access (Helbig and Charpentier Poncelet 2022). The methods can be repli- cated, and additional CFs can be recomputed or newly devel- oped with the help of published articles (dynamic MFA meth- ods: Charpentier Poncelet et al. 2022b and LCIA methods: Charpentier Poncelet et al. 2021 and this article). Applicabil- ity: The large-scale application study presented in “Sect. 3.2” demonstrated that the ADR and LPST methods are readily usable with current LCI databases in LCA studies, albeit with some limitations to keep in mind (cf. “Sect. 3.1.2”). 3.4 Perspectives for the ADR and LPST methods Moreover, while price statistics were here considered to be the best applicable proxy to represent the value of metals for humans as defined by the Life Cycle Initiative’s taskforce on mineral resources (Berger et al. 2020; Sonderegger et al. 2020), other metrics might be considered in the future to improve the way the economic and use values of resources are taken into account. Future research could aim to improve upon price statistics or explore the use of other proxies to reflect the value of metals for humans. Therefore, we recommend the use of midpoint CFs to evaluate the physical dissipation rates of metals (ADR) or the related inaccessibility to potential users (LPST). The use of endpoint CFs provides additional infor- mation on the potentially lost value of metals, but should be done cautiously and keeping their limitations in mind. Supplementary Information The online version contains supplemen- tary material available at https://doi.org/10.1007/s11367-022-02093-2. The application study revealed that the midpoint impact assessments are expected to diverge between selected LCIA methods, as impact hotspots are expected to be due to differ- ent metals. The divergence is expected to be smaller between the ADR/LPST methods and ReCiPe 2016, because their CFs spread over less orders of magnitude, thus being more dependent on the size of resource flows than ADP 2015. Moreover, the scarcest metals (e.g., precious metals) consist- ently have higher CFs in the ADP 2015, midpoint and end- point ReCiPe 2016 methods (SOP and SCP, respectively), and endpoint ADR and LPST methods. However, their CFs are so high in the ADP 2015 method that they consistently dominate impact scores for that method, while they only show up in top impact scores for data sets with relatively high amounts of these scarce metals with the other methods. Rather, widely used metals with relatively large resource flows can generally be expected to come out as being respon- sible for the most impacts when assessed with the midpoint ADR, LPST, and midpoint/endpoint ReCiPe 2016 methods. This remains generally true for the endpoint ADR and LPST methods, although the metals that dissipate the fastest tend to be more represented in the endpoint assessment using the ADR method, and most expensive metals in the endpoint assessment using the LPST method. Acknowledgements We thank F. Ardente for providing a compila- tion of price statistics retrieved from the US Geological Survey data available on the latter’s website. 3.4 Perspectives for the ADR and LPST methods Some perspectives to improve the ADR and LPST methods can be identified. Additional CFs could be developed for other minerals such as construction aggregates and sand. It would be possible for third-party users to generate CFs for additional mineral resources by using the framework and methods for dynamic MFA of Charpentier Poncelet et al. (2022b), along with the methods presented by Charpentier The midpoint CFs for the ADR and LPST methods rely on replicable results from open-source data and model (Charpentier Poncelet et al. 2022b; Helbig and Charpentier Poncelet 2022). Endpoint CFs introduce the use of price statistics in the assessment, which is helpful to distinguish between the value of different metals but presents some 3 The International Journal of Life Cycle Assessment the impacts of mineral resource use in different ways depending on their beliefs. Therefore, the assessment of mineral resource use on the AoP natural resources in the LCA or life cycle sus- tainability assessment (LCSA) contexts could be realized by using the ADR or LPST methods along with other LCIA meth- ods with complementary pathways such as those covered by the ADP 2015 and ReCiPe 2016 methods (cf. Charpentier Poncelet et al. 2022a). Measuring the impacts of dissipation in terms of lost potential to use the economic and use values (in monetary units or in different units) could enable the comparison with the impacts for other biotic and abiotic resources as well as values obtained from ecosystems (cf. discussion in Charpentier Poncelet et al. 2022a). Nevertheless, as discussed by the authors, signifi- cant developments are needed to address impact pathways related to natural resource use consistently. Moreover, important method developments and harmonization between existing or new LCIA methods would be needed for an exhaustive impact assessment of different aspects of mineral resource use to become viable (Charpentier Poncelet et al. 2022a). limitations at this point (cf. “Sect. 3.1.2”). The identified limitations are expected to be most relevant for metals with very small resource flows in the LCI or for which no resource flows are currently reported in widespread data- bases like ecoinvent. For instance, it was observed that by- product metals like gallium are likely to be over-represented in the endpoint results for the ADR method. 3.4 Perspectives for the ADR and LPST methods We are grateful to the two anonymous reviewers whose comments greatly helped us improve the quality of this manuscript. Funding Open Access funding enabled and organized by Projekt DEAL. This article was mostly developed and written during the PhD thesis of A.C.P., which was co-financed by the French agency for the ecological transition (ADEME) and the French geological survey (BRGM). Data availability Data and results depicted in the article are provided in the Supporting information documents. The data and model used to compute midpoint characterization factors may be accessed online at https://doi.org/10.17605/OSF.IO/CWU3D (Helbig and Charpentier Poncelet 2022). Price data and references are detailed in the article and in the Supporting information. Data from the ecoinvent database are subject to copyright and may not be divulged. References https://doi.org/10.1016/j.resconrec.2020.104698 y p g j Bru K, Christmann P, Labbé J-F, Lefebvre G (2015) Panorama 2014 du marché des Terres Rares. Orléans, France Huijbregts MAJ, Steinmann ZJN, Elshout PMF, Stam G, Verones F, Vieira M, Zijp M, Hollander A, van Zelm R (2017) ReCiPe2016: a harmonised life cycle impact assessment method at midpoint and endpoint level. Int J Life Cycle Assess 22:138–147. https:// doi.org/10.1007/s11367-016-1246-y Charpentier Poncelet A, Beylot A, Loubet P, Laratte B, Muller S, Villeneuve J, Sonnemann G (2022a) Linkage of impact pathways to cultural per- spectives to account for multiple aspects of mineral resource use in life cycle assessment. Resour Conserv Recycl 176:105912. https://doi.org/ 10.1016/j.resconrec.2021.105912 Huppertz T, Weidema B, Standaert S, De Caevel B, van Overbeke E (2019) The social cost of sub-soil resource use. Resources 8:19. https://doi.org/10.3390/resources8010019 Charpentier Poncelet A, Helbig C, Loubet P, Beylot A, Muller S, Villeneuve J, Laratte B, Thorenz A, Tuma A, Sonnemann G (2022b) Losses and lifetimes of metals in the economy. Nat Sustain. https://doi.org/10. 1038/s41893-022-00895-8 Johnson CA, Piatak NM, Miller MM (2017) Barite (Barium), in: Criti- cal mineral resources of the United States—economic and envi- ronmental geology and prospects for future supply. U.S. Geologi- cal Survey, Reston, Virginia. Charpentier Poncelet A, Helbig C, Loubet P, Beylot A, Muller S, Villeneuve J, Laratte B, Thorenz A, Tuma A, Sonnemann G (2021) Life cycle impact assessment methods for estimating the impacts of dissipative flows of metals. J Ind Ecol jiec.13136. https://doi.org/10.1111/jiec.13136 Johnson Matthey (2021) Price charts [WWW Document]. http://www. platinum.matthey.com/prices/price-charts (Accessed 13 June 2021). Kelly T, Matos GR (2014) Historical statistics for mineral and material commodities in the United States (2016 version) [WWW Docu- ment]. US Geol Surv Data Ser. 140. https://www.usgs.gov/centers/ nmic/historical-statistics-mineral-and-material-commodities-united- states (Accessed 15 Aug 2021) Charpentier Poncelet A, Loubet P, Laratte B, Muller S, Villeneuve J, Sonnemann G (2019) A necessary step forward for proper non- energetic abiotic resource use consideration in life cycle assess- ment: the functional dissipation approach using dynamic material flow analysis data. Resour Conserv Recycl 151:104449. https:// doi.org/10.1016/j.resconrec.2019.104449i Labbé J-F, Dupuy J-J (2014) Panorama 2012 du marché des plati- noïdes, Brgm/Rp-63169-Fr. Orléans, France. Løvik AN, Restrepo E, Müller DB (2016) Byproduct metal availabil- ity constrained by dynamics of carrier metal cycle: the gallium- aluminum example. Environ Sci Technol 50:8453–8461. References European Commission (2010) Framework and requirements for life cycle impact assessment models and indicators, international reference life cycle data system (ILCD) handbook. https://doi.org/10.2788/38719 Berger M, Sonderegger T, Alvarenga R, Bach V, Cimprich A, Dewulf J, Frischknecht R, Guinée J, Helbig C, Huppertz T, Jolliet O, Motoshita M, Northey S, Peña CA, Rugani B, Sahnoune A, Schrijvers D, Schulze R, Sonnemann G, Valero A, Weidema BP, Young SB (2020) Mineral resources in life cycle impact assess- ment: part II – recommendations on application-dependent use of existing methods and on future method development needs. Int J Life Cycle Assess. https://doi.org/10.1007/s11367-020-01737-5 Graedel TE, Barr R, Chandler C, Chase T, Choi J, Christoffersen L, Friedlander E, Henly C, Jun C, Nassar NT, Schechner D, Warren S, Yang MY, Zhu C (2012) Methodology of metal criticality determination. Environ Sci Technol 46:1063–1070. https://doi.org/10.1021/es203534z Graedel TE, Harper EM, Nassar NT, Reck BK (2013) On the materi- als basis of modern society. Proc Natl Acad Sci 112:6295–6300. https://doi.org/10.1073/pnas.1312752110 Beylot A, Ardente F, Marques A, Mathieux F, Pant R, Sala S, Zampori L (2020a) Abiotic and biotic resources impact categories in LCA : development of new approaches. Publications Office of the Euro- pean Union, Luxembourg. https://doi.org/10.2760/232839 Graedel TE, Reck BK, Miatto A (2022) Alloy information helps pri- oritize material criticality lists. Nat Commun 13:1–8. https://doi. org/10.1038/s41467-021-27829-w Helbig C, Charpentier Poncelet A (2022) ODYM MaTrace dissipation - code and datasets. https://doi.org/10.17605/OSF.IO/CWU3D Beylot A, Ardente F, Sala S, Zampori L (2021) Mineral resource dis- sipation in life cycle inventories. Int J Life Cycle Assess. https:// doi.org/10.1007/s11367-021-01875-4 Helbig C, Kondo Y, Nakamura S (2021) Simultaneously tracing the fate of seven metals with MaTrace-multi (accepted manuscript). J Ind Ecol Beylot A, Ardente F, Sala S, Zampori L (2020b) Accounting for the dissipation of abiotic resources in LCA: status, key challenges and potential way forward. Resour Conserv Recycl 157:104748. https://doi.org/10.1016/j.resconrec.2020.104748 Helbig C, Thorenz A, Tuma A (2020) Quantitative assessment of dis- sipative losses of 18 metals. Resour Conserv Recycl. https://doi. org/10.1016/j.resconrec.2019.104537 Helbig C, Thorenz A, Tuma A (2020) Quantitative assessment of dis- sipative losses of 18 metals. Resour Conserv Recycl. https://doi. Henckens MLCM, van Ierland EC, Driessen PPJ, Worrell E (2016) Mineral resources: geological scarcity, market price trends, and future generations. Resour Policy 49:102–111. https://doi.org/10. 1016/j.resourpol.2016.04.012 Blomsma F, Tennant M (2020) Circular economy: preserving materials or products? Introducing the Resource States framework. Resour Conserv Recycl 156:104698. Competing interests The authors declare no competing interests. Competing interests The authors declare no competing interests. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Finally, we would like to conclude this article with a short discussion on assessing multiple aspects related to mineral resource use in the AoP natural resources. Charpentier Poncelet et al. (2022a) suggested that multiple LCIA methods should be used to assess the impacts of mineral resource use on the AoP natural resources exhaustively. Indeed, resources provide eco- nomic values and use values to different users, and practitioners with different cultural perspectives could attempt to characterize 1 3 The International Journal of Life Cycle Assessment European Commission (2010) Framework and requirements for life cycle impact assessment models and indicators, international reference life cycle data system (ILCD) handbook. https://doi.org/10.2788/38719 References https:// doi.org/10.1021/acs.est.6b02396 Dewulf J, Hellweg S, Pfister S, León MFG, Sonderegger T, de Matos CT, Blengini GA, Mathieux F (2021) Towards sustainable resource management: identification and quantification of human actions that compromise the accessibility of metal resources. Resour Conserv Recycl 167:105403. https://doi.org/10.1016/j.resconrec.2021.105403 Løvik AN, Restrepo E, Müller DB (2015) The global anthropogenic gallium system: determinants of demand, supply and efficiency improvements. Environ Sci Technol 49:5704–5712. https://doi. org/10.1021/acs.est.5b00320 Drielsma J, Russell-Vaccari AJ, Drnek T, Brady T, Weihed P, Mistry M, Simbor LP (2016) Mineral resources in life cycle impact assessment— defining the path forward. Int J Life Cycle Assess 21:85–105. https:// doi.org/10.1007/s11367-015-0991-7 Metalary (2021) Niobium price [WWW Document]. https://www.metalary. com/niobium-price/ Ecorys (2012) Mapping resource prices: the past and the future Ecorys (2012) Mapping resource prices: the past and the future Moreno Ruiz E, Valsasina L, FitzGerald D, Brunner F, Symeonidis A, Bourgault G, Wernet G (2019) Documentation of changes imple- mented in the ecoinvent database v3.6. Zürich, Switzerland European Commission (2020a) Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions. A new Circular Economy Action Plan For a cleaner and more competi- tive Europe. Brussels, Belgium. Moreno Ruiz E, Valsasina L, FitzGerald D, Symeonidis A, Turner D, Müller J, Minas N, Bourgault G, Vadenbo C, Ioannidou D, Wernet G (2020) Documentation of changes implemented in the ecoinvent database v3.7 & v3.7.1. Zürich, Switzerland. European Commission (2020b). Study on the EU’s list of Critical Raw Materials. Critical Raw Materials Factsheets. https://doi.org/10. 2873/92480 Nakamura S, Kondo Y, Kagawa S, Matsubae K, Nakajima K, Nagasaka T (2014) MaTrace: tracing the fate of materials over time and across 1 3 The International Journal of Life Cycle Assessment products in open-loop recycling. Environ Sci Technol 48:7207– 7214. https://doi.org/10.1021/es500820h van Oers L, de Koning A, Guinée JB, Huppes G (2002) Abiotic resource depletion in LCA: improving characterisation factors for abiotic resource depletion as recommended in the new Dutch LCA Handbook, Road and Hydraulic Engineering Institute. Pauliuk S, Kondo Y, Nakamura S, Nakajima K (2017) Regional dis- tribution and losses of end-of-life steel throughout multiple prod- uct life cycles—insights from the global multiregional MaTrace model. Resour Conserv Recycl 116:84–93. https://doi.org/10. 1016/j.resconrec.2016.09.029 van Oers L, Guinée JB, Heijungs R (2019) Abiotic resource depletion potentials (ADPs) for elements revisited—updating ultimate reserve estimates and introducing time series for production data. Int J Life Cycle Assess. https://doi.org/10.1007/s11367-019-01683-x j Reuter MA, van Schaik A, Gutzmer J, Bartie N, Abadías-Llamas A (2019) Challenges of the circular economy: a material, metallurgi- cal, and product design perspective. Annu Rev Mater Res 49:253– 274. https://doi.org/10.1146/annurev-matsci-070218-010057 Cycle Assess. https://doi.org/10.1007/s11367-019-01683-x van Oers L, Guinée JB, Heijungs R, Schulze R, Alvarenga RAF, Dewulf J, Drielsma J, Sanjuan-Delmás D, Kampmann TC, Bark G, Uriarte AG, Menger P, Lindblom M, Alcon L, Ramos MS, Torres JME (2020) Top-down characterization of resource use in LCA: from problem definition of resource use to operational characterization factors for dissipation of elements to the environment. Int J Life Cycle Assess. https://doi.org/10.1007/s11367-020-01819-4 Rørbech JT, Vadenbo C, Hellweg S, Astrup TF (2014) Impact assess- ment of abiotic resources in LCA: quantitative comparison of selected characterization models SUPP INFO. Environ Sci Tech- nol 48:11072–11081. g y doi.org/10.1007/s11367-020-01736-6 Wernet G, Bauer C, Steubing B, Reinhard J, Moreno-Ruiz E, Weidema B (2016) The ecoinvent database version 3 (part I): overview and methodology. Int J Life Cycle Assess 21:1218–1230. https://doi. org/10.1007/s11367-016-1087-8 Ecorys (2012) Mapping resource prices: the past and the future https://doi.org/10.1021/es5023976 Schulze R, Guinée J, van Oers L, Alvarenga R, Dewulf J, Drielsma J (2020) Abiotic resource use in life cycle impact assessment—Part II – Linking perspectives and modelling concepts. Resour Conserv Recycl. https://doi.org/10.1016/j.resconrec.2019.104595 Vieira M, Ponsioen T, Goedkoop M, Huijbregts M (2016) Surplus cost potential as a life cycle impact indicator for metal extraction. Resources 5:2. https://doi.org/10.3390/resources5010002 Vieira MDM, Ponsioen TC, Goedkoop MJ, Huijbregts MAJ (2017) Surplus ore potential as a scarcity indicator for resource extrac- tion. J Ind Ecol 21:381–390. https://doi.org/10.1016/j.jcrc.2017. 02.030 Sonderegger T, Berger M, Alvarenga R, Bach V, Cimprich A, Dewulf J, Frischknecht R, Guinée J, Helbig C, Huppertz T, Jolliet O, Motoshita M, Northey S, Rugani B, Schrijvers D, Schulze R, Sonnemann G, Valero A, Weidema BP, Young SB (2020) Mineral resources in life cycle impact assessment—part I: a critical review of existing methods. Int J Life Cycle Assess 25:784–797. https:// doi.org/10.1007/s11367-020-01736-6 Watson BJ, Eggert RG (2021) Understanding relative metal prices and availability: combining physical and economic perspectives. J Ind Ecol 25:890–899. https://doi.org/10.1111/jiec.13087 g Stormcrow (2014) Rare Earth Industry Report UNEP (2013) Metal recycling: opportunities, limits, infrastructure, A Report of the Working Group on the Global Metal Flows to the International Resource Panel. Zampori L, Pant R (2019) Suggestions for updating the product envi- ronmental footprint (PEF) method. Publications Office of the European Union, Luxembourg. https://doi.org/10.2760/424613 UNEP (2011) Recycling rates of metals: a status report, A Report of the Working Group on the Global Metal Flows to the Interna- tional Resource Panel. United Nations Environment Programme, Nairobi, Kenya.i Zampori L, Sala S (2017) Feasibility study to implement resource dis- sipation in LCA. Luxembourg. https://doi.org/10.2760/869503 United Nations (2008) International Standard Industrial Classification of All Economic Activities (ISIC), Rev. 4. United Nations Pub- lication, New York. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 1 3 1 3
https://openalex.org/W2292479725	https://www.atmos-meas-tech.net/9/5293/2016/amt-9-5293-2016.pdf	English	null	Continuous and high-precision atmospheric concentration measurements of COS, CO&lt;sub&gt;2&lt;/sub&gt;, CO and H&lt;sub&gt;2&lt;/sub&gt;O using a quantum cascade laser spectrometer (QCLS)	Atmospheric measurement techniques	2,016	cc-by	117,054	Correspondence to: Huilin Chen (huilin.chen@rug.nl) Received: 12 February 2016 – Published in Atmos. Meas. Tech. Discuss.: 29 February 2016 Revised: 10 September 2016 – Accepted: 27 September 2016 – Published: 1 November 2016 Received: 12 February 2016 – Published in Atmos. Meas. Tech. Discuss.: 29 February 2016 Revised: 10 September 2016 – Accepted: 27 September 2016 – Published: 1 November 2016 difference with the assigned cylinder value within 3.3 ppt COS, 0.05 ppm for CO2 and 1.7 ppb for CO over a period of 35 days. The different contributions to uncertainty in mea- surements of COS, CO2 and CO were summarized and the overall uncertainty was determined to be 7.5 ppt for COS, 0.23 ppm for CO2 and 3.3 ppb for CO for 1-minute data. A comparison of in situ QCLS measurements with those from concurrently ﬁlled ﬂasks that were subsequently measured by the QCLS showed a difference of −9.7 ± 4.6 ppt for COS. Comparison of the QCLS with a cavity ring-down spectrom- eter showed a difference of 0.12 ± 0.77 ppm for CO2 and −0.9 ± 3.8 ppb for CO. Abstract. Carbonyl sulﬁde (COS) has been suggested as a useful tracer for gross primary production as it is taken up by plants in a similar way as CO2. To explore and verify the application of this novel tracer, it is highly desired to de- velop the ability to perform continuous and high-precision in situ atmospheric measurements of COS and CO2. In this study we have tested a quantum cascade laser spectrome- ter (QCLS) for its suitability to obtain accurate and high- precision measurements of COS and CO2. The instrument is capable of simultaneously measuring COS, CO2, CO and H2O after including a weak CO absorption line in the ex- tended wavelength range. An optimal background and cali- bration strategy was developed based on laboratory tests to ensure accurate ﬁeld measurements. We have derived water vapor correction factors based on a set of laboratory exper- iments and found that for COS the interference associated with a water absorption line can dominate over the effect of dilution. This interference can be solved mathematically by ﬁtting the COS spectral line separately from the H2O spec- tral line. Furthermore, we improved the temperature stabil- ity of the QCLS by isolating it in an enclosed box and ac- tively cooling its electronics with the same thermoelectric chiller used to cool the laser. Atmos. Meas. Tech., 9, 5293–5314, 2016 www.atmos-meas-tech.net/9/5293/2016/ doi:10.5194/amt-9-5293-2016 © Author(s) 2016. CC Attribution 3.0 License. Atmos. Meas. Tech., 9, 5293–5314, 2016 www.atmos-meas-tech.net/9/5293/2016/ doi:10.5194/amt-9-5293-2016 © Author(s) 2016. CC Attribution 3.0 License. Correspondence to: Huilin Chen (huilin.chen@rug.nl) The QCLS was deployed at the Lutjewad atmospheric monitoring station (60 m; 6◦21′ E, 53◦24′ N; 1 m a.s.l.) in the Netherlands from July 2014 to April 2015. The QCLS measurements of independent work- ing standards while deployed in the ﬁeld showed a mean 1 Introduction Carbonyl sulﬁde (COS) has been suggested as a potential tracer for photosynthetic CO2 uptake (Sandoval-Soto et al., 2005; Montzka et al., 2007; Campbell et al., 2008; Berry et al., 2013; Asaf et al., 2013), as it follows the same uptake pathway into plants through stomata as CO2 but is not gener- ally re-emitted by plants (Protoschill-Krebs and Kesselmeier, 1992; Protoschill-Krebs et al., 1996; Stimler et al., 2010b). COS therefore provides a means to partition net ecosystem exchange into gross primary production (GPP) and respira- tion. As large uncertainties in the COS budget remain, ﬁeld measurements of COS and CO2 concentrations and ﬂuxes 1Centre for Isotope Research (CIO), University of Groningen, Groningen, the Netherlands 2Aerodyne Research Inc., MA, USA 1Centre for Isotope Research (CIO), University of Groningen, Groningen, the Netherlands 2Aerodyne Research Inc., MA, USA 1Centre for Isotope Research (CIO), University of Groningen, Groningen, the Netherlands 2Aerodyne Research Inc., MA, USA 3NOAA Earth System Research Laboratory, Boulder, CO, USA 4Cooperative Institute for Research in Environmental Sciences (CIRES), University of Colorado, Boulder, CO, USA 3NOAA Earth System Research Laboratory, Boulder, CO, USA 4Cooperative Institute for Research in Environmental Sciences (CIR NOAA Earth System Research Laboratory, Boulder, CO, USA Cooperative Institute for Research in Environmental Sciences (CIRES), University of Colorado, Bou Correspondence to: Huilin Chen (huilin.chen@rug.nl) Linda M. J. Kooijmans1, Nelly A. M. Uitslag1, Mark S. Zahniser2, David D. Nelson2, Stephen A. Montzka3, and Huilin Chen1,4 nda M. J. Kooijmans1, Nelly A. M. Uitslag1, Mark S. Zahniser2, David D. Nelson2, Stephen A uilin Chen1,4 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O The eddy-covariance technique requires high-frequency data (> 10 Hz), which typically adversely affect the precision of the measurements compared to 1 Hz data, and requires an averaging period of about 10 to 30 min. In contrast to the high frequency required for eddy-covariance measurements, lower-frequency measurements (1 Hz) provide useful results over extended measurement periods and enhance the preci- sion of any individual measurement. Furthermore, measure- ments for long-term monitoring do not require fast response, and thus it is not necessary to operate the instrument at high ﬂow rates. As a matter of fact, low ﬂow rates are preferred so that working standards can be used over a long period. This reduces the additional logistics needed for calibration gases, such as ﬁlling, calibration and transportation of the standards (Xiang et al., 2014). Besides in situ measurements, ﬂask or canister measurements can be a valuable tool for providing information about ambient concentrations of COS as well. For example, ﬂask measurements were used before when constructing an historical record from ﬁrn air (Montzka et al., 2004), during ﬁeld campaigns (White et al., 2010; Blon- quist et al., 2011) and for long-term monitoring (Montzka et al., 2007). In this research we developed a robust setup for high-precision and long-term monitoring of ambient concen- trations of COS, CO2, CO and H2O at different heights from the Lutjewad monitoring station in Groningen, the Nether- lands. To this end we employed a “QCL Mini Monitor” from Aerodyne Research Inc. (Billerica, MA, USA) that can op- erate autonomously and requires little operator attention. We designed an optimal strategy for “zero” air spectral correc- tion and calibration for accurate measurements and we as- sessed the correction for water vapor interference. In this pa- per we aim to evaluate and improve the performance of the instrument. We will show the precision and accuracy of the instrument with over half a year of ﬁeld data and measure- ments of working standards, and we compare the measure- ments with other instrumentation. Furthermore, we evaluate the total uncertainty of the measurements by combining the The required measurement precision (in this study we de- ﬁne precision as the standard deviation over a 2-minute pe- riod) for studies of exchange processes of COS and CO2 between biosphere and atmosphere depend on the concen- tration change that these gases undergo in any given experi- ment. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5294 sity of air ρ. To be able to capture COS ﬂuxes of, for ex- ample, 10 pmol m−2 s−1 over a height difference of 20 me- ters, the measurement precision of COS should be better than 0.5 ppt under high turbulent conditions (K = 10 m2 s−1) and 4.8 ppt under low turbulent conditions (K = 1 m2 s−1). If we were to infer the gross ﬂuxes from chamber measurements with 1CO2 (the difference between in- and outgoing cham- ber concentrations) measurable from 1 ppm, then, given the leaf-scale relative uptake (LRU) ratio of COS / CO2 1.5–4.0 (Stimler et al., 2010b; Seibt et al., 2010; Berkelhammer et al., 2014), our goal would be to have measurement precisions of COS better than 1.9–5.0 ppt for COS (calculated from LRU and scaling with 1CO2 and the ratio of COS/CO2 mole frac- tions gives, for example, 1.5 × 1 × (500/400) = 1.9 ppt at the ambient level of 500 ppt and 400 ppm CO2). from leaf to ecosystem and regional scale are required for the COS tracer method to be tested and validated (Wohlfahrt et al., 2012; Berkelhammer et al., 2014). Therefore, there is a need for high-frequency and high-precision measurements techniques of COS and CO2. q Several past studies on COS have relied on discrete (ﬂask) samples analyzed with gas chromatographic mass spectrom- etry (GC-MS; Montzka et al., 2007; Stimler et al., 2010b). For example, the global atmospheric ﬂask sampling network described by Montzka et al. (2007) has allowed a founda- tion for understanding COS concentrations over annual cy- cles on global scale. Although the GC-MS technique can be used for in situ measurements (Miller et al., 2008; Belviso et al., 2013), this technique does not typically allow for high- frequency measurements of 1 to 10 Hz. Recent developments of quantum cascade laser spectrometers (QCLSs) have en- abled in situ trace gas measurements including COS. These instruments have proven to be a valuable tool for continuous high-frequency measurements of COS and CO2 up to a fre- quency of 10 Hz (Stimler et al., 2010a, b; Asaf et al., 2013; Commane et al., 2013; Berkelhammer et al., 2014; Maseyk et al., 2014; Commane et al., 2015). Measurement instruments for long-term atmospheric trace gas concentration monitoring need to meet different require- ments than, for example, eddy-covariance measurements. Published by Copernicus Publications on behalf of the European Geosciences Union. Published by Copernicus Publications on behalf of the European Geosciences Union. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 2.1 Instrumentation The QCL Mini Monitor that we use is a tunable diode laser absorption spectrometer (TILDAS) using a single continuous-wave quantum cascade laser (Alpes Lasers), which is cooled with a Peltier element to −19.8 ◦C, and using a single photodiode infrared detector (Teledyne Judson Tech- nologies; McManus et al., 2010). The waste heat from both the laser and detector is removed with a recirculating mix- ture of water containing 25 % ethanol, which is temperature controlled with a thermoelectric chiller, ThermoCube 300 (Solid State Cooling Systems, USA). The instrument was ini- tially set to simultaneously measure COS, CO2 and H2O at wavenumbers 2050.397, 2050.566 and 2050.638 cm−1, re- spectively. We extended the range of the laser current to include measurements of CO at 2050.854 cm−1. Figure 1 shows the simulated transmission spectrum of ambient con- centrations of COS, CO2, CO and H2O as obtained through the HITRAN 2012 database (Rothman et al., 2013). The pre- cision and accuracy of the measurements will be discussed in Sect. 3.1. and the HITRAN 2012 database cross sections as a function of wavelength. The spectral ﬁt for CO is separated from the ﬁt for COS, CO2 and H2O as there is slight interaction of the CO peak with a second absorption line of COS. The COS ﬁt close to the CO peak is linked to the COS peak at lower wavenumbers to improve the ﬁtting for CO. This is achieved by ﬁtting the spectra in two steps: ﬁrst the mole fractions are determined for both COS peaks independently; second the CO concentration is recalculated with the ﬁxed COS concen- tration derived from the separated COS peak in the ﬁrst step. p p p The TDLWINTEL software has the option to store raw spectra. These spectra can later be used for re-analysis using the so-called “playback” mode of the software. The spectral parameters (line shape and position) for the ﬁts are taken from the HITRAN database (Rothman et al., 2013). The sample spectra are normalized with a “zero” air spectrum to remove background spectral structures and to remove ab- sorbance external to the multi-pass cell (Stimler et al., 2010a; Santoni et al., 2012). The “zero” air spectrum is periodically determined when the cell is ﬂushed with high-purity nitrogen (99.99999 %), which we will now refer to as “background” measurement. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5295 2050.3 2050.4 2050.5 2050.6 2050.7 2050.8 2050.9 0.994 0.995 0.996 0.997 0.998 0.999 1.000 Wavenumber [cm−1] Transmittance COS CO2 H2O CO COS CO2 H2O CO T = 298 K, P = 40 Torr, L = 76 m Figure 1. Simulated transmission spectrum of ambient concen- trations of COS (500 ppt), CO2 (500 ppm), H2O (1.5 %) and CO (200 ppb) with sample cell conditions: temperature 298 K, pressure 53.3 hPa (40 Torr) and the absorption path length 76 m. A small wa- ter band at 2050.5 cm−1 can interfere with COS at 2050.4 cm−1 and can affect the COS correction for water vapor without a split ﬁt at 2050.45 cm−1 (Sect. 2.3). uncertainties of scale transfer, water vapor corrections and the measurement precision. Based on the precision and accu- racy that we derive from these experiments, we discuss the suitability of COS measurements on this instrument for dif- ferent purposes, that is, for interpreting proﬁle measurements and comparing concentrations across sites. In addition to the experimental setup for continuous in situ measurements we developed a setup to analyze ﬂasks, which we used to make a comparison with GC-MS measurements of ﬂasks and to assess the laboratory-derived correction for water vapor in- terference. 2050.3 2050.4 2050.5 2050.6 2050.7 2050.8 2050.9 0.994 0.995 0.996 0.997 0.998 0.999 1.000 Wavenumber [cm−1] Transmittance COS CO2 H2O CO COS CO2 H2O CO T = 298 K, P = 40 Torr, L = 76 m 2 Experimental setup Before the actual deployment of the instrument in the ﬁeld, we performed laboratory tests to assess the accuracy and traceability of the QCLS measurements and to develop pro- cedures for applying corrections as needed. Here we describe the laboratory tests and we give detailed information about the instrumentation and ﬁeld setup. Figure 1. Simulated transmission spectrum of ambient concen- trations of COS (500 ppt), CO2 (500 ppm), H2O (1.5 %) and CO (200 ppb) with sample cell conditions: temperature 298 K, pressure 53.3 hPa (40 Torr) and the absorption path length 76 m. A small wa- ter band at 2050.5 cm−1 can interfere with COS at 2050.4 cm−1 and can affect the COS correction for water vapor without a split ﬁt at 2050.45 cm−1 (Sect. 2.3). www.atmos-meas-tech.net/9/5293/2016/ L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O On the regional scale, COS shows seasonal variations typically between ∼100 and 150 ppt at continental sites in the Northern Hemisphere (NH) and between 40 and 70 ppt in the Southern Hemisphere and at marine sites (Montzka et al., 2007). CO2 seasonal variations typically reach up to 15 ppm in the NH and are as low as 2 ppm at the South Pole (Zhao and Zeng, 2014). For the leaf scale, COS and CO2 concentration changes can be substantially larger; for example, Berkelhammer et al. (2014) showed that during branch bag measurements COS generally decreased by 180 to 240 ppt during active photosynthesis and CO2 concentra- tions can easily change by 200 ppm, depending on the setup. Besides the difference in requirements for precision between different experimental setups, the type of analyses intended for a dataset also determines the requirements for precision and accuracy of the measurements. If the intention is to com- pare atmospheric concentrations across sites, then accuracy is important because data from different sites must be on con- sistent scales. In contrast, short-term precision is more im- portant than accuracy when differences between heights are to be interpreted (e.g., as in estimation of ﬂuxes from pro- ﬁle measurements). Following the K-parameterization for- mulation of the ﬂux-gradient method (e.g., Meredith et al., 2014), F = −K1C/1zρ, the precision required to capture the concentration differences between heights (1C) mostly depends on the size of the ﬂuxes F, the height difference 1z, the turbulence conditions, which are represented by the eddy diffusivity K, and to a lesser extent on the molar den- Atmos. Meas. Tech., 9, 5293–5314, 2016 www.atmos-meas-tech.net/9/5293/2016/ L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5296 GGG GGG GG G GG GG GGGGGGGGG G G GG G G GGG G GGG GGGG GG G GG GG GGG G G GGGGGGGGGG GGGGGGGG G GG GGGGG GG GGG G G G G GG GGG GG G GGG G GGG G GG GGGG GGGG G GG GGGG GG GG G GGGG GG G GG GG GG GG GG GGG G G GG G GGGGGGG GGGG G GG G GG GGGGGGG GG GGGGGGGGGGGGG G GGG G G G GGG G G GGG GGG GGGGGGGGGGG GG GGGGGG GG GG GGG GGGGGG GGG GGGGG G GGG GGGGG GG GG GGG GGGG G GGG GGGGGGGG GGGGGGGGG GGG GGGGGG GGG G G GG G GG G G G G G GGG GGGGGG GGGG G GGG GGG G GG GG GG GGG G GGGGGG G GGGG GGG G G GGGG G GG G GGG GG G GGG GG GG G GGGGGGG GGGG GG GG GGG GGG GGGG GGGGG GGGG GGG G GG G G GG GGGGG GG GG G G GG GGG G GG G GG GG GG GGG G GG GGG GG GGGGGG GGG GGG G G GGGGGGG GG G GGG GGGGGGGGGGGGGGG GGGG GG GGG G G GGG GGG G GG GGGGGG GGG GG GG G G G G G G G GG G G GGG G GG GG GGG GGGGG GGGGGGGG G GG G GG GG GGG GGGGGG GGGGGGG GG GGG GG G GG GG GGGGG GG G GGG GGG G G G GGG G G GGG GGG GG GGGGGGGG GGG G GG G G GG GG GG GG GG G G GG GG G GG G GGGGGG GGG GGG GG GG GG GG GGG G GG G G GG GG GG G G G GG GGGGG GG GG G GG GG G G G G G GG GG G GGGGGGGG GGGG GGG G GG GGG G GG GGGGG GGG GGG G GG GGGGGG GG GG G G GGGG G GGGG GGG G GGGGGGGG GG GG GG GGG GG GG GG GGG GGG GG GGG GGG GG GGG GG G G GGGGGGG G GGGG GGG GG GGG G GG GG GGGGGGG GGG G G GGG GGGGG GG GGG GGG G GGGG G G GGG G GGG G GGG GG GGGGG GG GGGGGG GG G GG G GGGG GGG GGGG GG GGGGGG GGGG GGGG GG G G GG GGG GGG G GGG G GGGG GG GGG GGG GGG GGG GG GGG GG G GG GGG GGG GG G G G G GG GGG GG GGG G GG G GGG GG GG G GG G GGGGGG GG G GG G G GG GGG G GGG GG G GG GG GGG GG GGGGG GG GG GGGGGGGG G GGG GG GG GG GGG GGGG GGGGG GGGGG GG GGG G GG GG GG GG GG GGG GG GGGG GGG GGGG GGGG G GGG GGG GG GGG GGG GG GG G GGGGGG G G G GGGGG GGG G GGGG GG G GGGG GGG GG G G GGGG GG GG G GGGGGG G GG GG G GGGGGGGG GG GG G GG G GG GGGG G G GG G G GG GG GG G GGG GG G GGGGG G GGGGG GGGGG GG G GG GG GGGG GG GG GGG GGG GGG GGGGG GGGG G GG GG GG GG GG GGG GG GG GGGGGGGGGGGG G GG GGG GG G G G GG G G G G G GG GGGGGGG GGGGG GGGGGG GGGG GGGGGGG G G G GG GGGGG GG GG G G G GGG GGG GG GG GGGGG GGG GG G GG GG G G G GG GG GG G G GGG G GG GGG GGGG GGGGG GG GG G GG GGGG GGGGGG GG G GG GG G GG G GGG GGG G G GG GGGGGG GG G GG GG GG GGGG GGGGGGGGGGG GGGGGG G G G G GG GGG GGG GG G GG G G GG G GG GGGGGG GGGG GG G GGG G GG G GG GG GG GGG GGG G GGGG GGGGG GG GG GGG GG GG GGG GG GG GGG G GG G GGG GG GGGGG GGG GG GGG G GG GGG GGG GGGG GGG G GGG GG GG GG GGGG GGG GGG GG G G GGGG GG GG GGGGGGG GGGG GGGGG GG GG GGG G GGG G G GG GG GGG G GGGG G GGG GGGG G GG GGG G GGGGGGGG G GGGG GG GG GG GG GG GG GGG G GGG G G GG G GG G G G GG GGGG G GG GGG G GGGG GGG GG G GGGGG GG GGG GGGGG GGGG G GG GG G GG G GGG G G GGG GG G GG G GGG GGG G GGGG GGGG GG GG G GGGGGG 100 150 200 250 300 350 400 100 150 200 250 300 350 400 Measured CO [ppb] Calibrated CO [ppb] GGG GGG GG G GG GG GGGGGGGGG G G GG G G GGG G GGG GGGG GG G GG GG GGG G G GGGGGGGGGG GGGGGGGG G GG GGGGG GG GGG G G G G GG GGG GG G GGG G GGG G GG GGGG GGGG G GG GGGG GG GG G GGGG GG G GG GG GG GG GG GGG G G GG G GGGGGGG GGGG G GG G GG GGGGGGG GG GGGGGGGGGGGGG G GGG G G G GGG G G GGG GGG GGGGGGGGGGG GG GGGGGG GG GG GGG GGGGGG GGG GGGGG G GGG GGGGG GG GG GGG GGGG G GGG GGGGGGGG GGGGGGGGG GGG GGGGGG GGG G G GG G GG G G G G G GGG GGGGGG GGGG G GGG GGG G GG GG GG GGG G GGGGGG G GGGG GGG G G GGGG G GG G GGG GG G GGG GG GG G GGGGGGG GGGG GG GG GGG GGG GGGG GGGGG GGGG GGG G GG G G GG GGGGG GG GG G G GG GGG G GG G GG GG GG GGG G GG GGG GG GGGGGG GGG GGG G G GGGGGGG GG G GGG GGGGGGGGGGGGGGG GGGG GG GGG G G GGG GGG G GG GGGGGG GGG GG GG G G G G G G G GG G G GGG G GG GG GGG GGGGG GGGGGGGG G GG G GG GG GGG GGGGGG GGGGGGG GG GGG GG G GG GG GGGGG GG G GGG GGG G G G GGG G G GGG GGG GG GGGGGGGG GGG G GG G G GG GG GG GG GG G G GG GG G GG G GGGGGG GGG GGG GG GG GG GG GGG G GG G G GG GG GG G G G GG GGGGG GG GG G GG GG G G G G G GG GG G GGGGGGGG GGGG GGG G GG GGG G GG GGGGG GGG GGG G GG GGGGGG GG GG G G GGGG G GGGG GGG G GGGGGGGG GG GG GG GGG GG GG GG GGG GGG GG GGG GGG GG GGG GG G G GGGGGGG G GGGG GGG GG GGG G GG GG GGGGGGG GGG G G GGG GGGGG GG GGG GGG G GGGG G G GGG G GGG G GGG GG GGGGG GG GGGGGG GG G GG G GGGG GGG GGGG GG GGGGGG GGGG GGGG GG G G GG GGG GGG G GGG G GGGG GG GGG GGG GGG GGG GG GGG GG G GG GGG GGG GG G G G G GG GGG GG GGG G GG G GGG GG GG G GG G GGGGGG GG G GG G G GG GGG G GGG GG G GG GG GGG GG GGGGG GG GG GGGGGGGG G GGG GG GG GG GGG GGGG GGGGG GGGGG GG GGG G GG GG GG GG GG GGG GG GGGG GGG GGGG GGGG G GGG GGG GG GGG GGG GG GG G GGGGGG G G G GGGGG GGG G GGGG GG G GGGG GGG GG G G GGGG GG GG G GGGGGG G GG GG G GGGGGGGG GG GG G GG G GG GGGG G G GG G G GG GG GG G GGG GG G GGGGG G GGGGG GGGGG GG G GG GG GGGG GG GG GGG GGG GGG GGGGG GGGG G GG GG GG GG GG GGG GG GG GGGGGGGGGGGG G GG GGG GG G G G GG G G G G G GG GGGGGGG GGGGG GGGGGG GGGG GGGGGGG G G G GG GGGGG GG GG G G G GGG GGG GG GG GGGGG GGG GG G GG GG G G G GG GG GG G G GGG G GG GGG GGGG GGGGG GG GG G GG GGGG GGGGGG GG G GG GG G GG G GGG GGG G G GG GGGGGG GG G GG GG GG GGGG GGGGGGGGGGG GGGGGG G G G G GG GGG GGG GG G GG G G GG G GG GGGGGG GGGG GG G GGG G GG G GG GG GG GGG GGG G GGGG GGGGG GG GG GGG GG GG GGG GG GG GGG G GG G GGG GG GGGGG GGG GG GGG G GG GGG GGG GGGG GGG G GGG GG GG GG GGGG GGG GGG GG G G GGGG GG GG GGGGGGG GGGG GGGGG GG GG GGG G GGG G G GG GG GGG G GGGG G GGG GGGG G GG GGG G GGGGGGGG G GGGG GG GG GG GG GG GG GGG G GGG G G GG G GG G G G GG GGGG G GG GGG G GGGG GGG GG G GGGGG GG GGG GGGGG GGGG G GG GG G GG G GGG G G GGG GG G GG G GGG GGG G GGGG GGGG GG GG G GGGGGG 100 150 200 250 300 350 400 100 150 200 250 300 350 400 Measured CO [ppb] Calibrated CO [ppb] G Cylinder meas. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1:1 y = 0.975x + 2.173 G G G G 100 200 300 −1.0 0.0 1.0 Calibrated CO [ppb] Residuals [ppb] GGGGGGGGGGGGGGG GGGGG G G GG GGGG GGGG GG GGGGGGGGGG GGG GGG GG GG GGG GG GGG GGG GGGGGG GG GG GG GGG GGGGGGG GGG GG GG GG G GGG GGGGGGGGG GGGGGGGGG GGGGGGGG G GGG GG G G GGGG GG GGGGGGG GG GG GG GGGGG GGG GG GGG GG GGGG GGG G GG GG GGGGGGGGGG GG GGGGG GG GGG GGGG G GGGGGG GGG GGGG GGG GG GGG GGGGGGGG GGG GGGG GGGGG GG GGG GGG GGG G GG GGGGGG GG GGG GG GG GG GGG GG GGGGGGG G GGGGG GGGGG GGGG GGGG GGG GG GGG GGGGGGGGGGGG G GG G GGG GGGGGGG GGGG GGGGGGG GGGG GGGGGGG GG GGGGGGGGGGGG GGGGG GGGGGGGGG GGGG GGG GGGGGGGGGG GG G GGGG GGGGGG GGG GGGGGGG GGGGG GGGG GGGGG G GG GGGGG GGGG GGG G GG G GGGGGGGGGGGGGGGG GGG GGG GGGGGG GGG G GG GG GGGG GGGGGGGG GGGG GGG GG GG GGG GGGGGG GGGGG GGG GGGGGGG GGG G GGGGGG GGG GGGGGG GGGGGGG GGGGG GGGG GGGG GG GGG GGG G G GG GGG GGGG GG GGGG GG GGG GGGG GGGG GGGGGGGGGG GG GG GG GGG GGGGG GGGGGGGG GG GG GG GGGGGGGGG GGGG GG GGGGGGGG GG GG GG GGGGGGGG GG GG GGGG GGGGGG GGGGG GG GGGGGGGGG GG GGG GGGGGG GGGG GGG G GGGGG GGGGG GG GGGGG GGGGGG GGG GGGG GGGGGG GGG GGG GGGG GG GGG GG GGGGGGGGGG GGGGG GGGGGG GGGGG G GGGGG GGG GGGGGG GGGGG GG G G G GG GGGG GGGGG GGGGGG GGGGG GGG GGG GGGG GG GGGG G GGGGG G GG GGGGGG GGGGG GGG G GGGGGGGGG GGGGGGGGGGGG GG GG G GG GGG G GGGGGGGGGGG G GGG GGGG GGGGGGGGG G G GGGGGGGG G GGGGG GG GGGGGG GG GGGGGG GGGGG GGGG GGGGGGG GGG GGGGGGGG GG GGGG G GGGGGGGGG GGGGGGGG GGGGGGGG GGG G GG GGGG GGGG GGG GGGGGGG GGGGGG GGGGGGG GG GG GGGGGG G GGGGGGGG GG GGGG GGGG GGGGGGGG GGG GGG G GGG GG GGG G GGGGGG GGGG G GGG GG GGGGG GG G GGGG GGGGGGG GGG GG GG GGGG GGGGG G GG GGGGG GGGGGG GGGGGG GG GGGGGGGGGGGGGGGGGGGG GGGG GGG G GG GGGGG GGGGG GGG GGG GGGGGGGGGGGGGGGG GGG GGGGGGG GG GG GGGGGG GG GGG GGGG GGGGG GGG G GGGGGGG GGG GGG GGG GGGGGGG GGGGGGGGGGGGGGGGGG GGGGGG GGGG GGG GGG GGGG GGGGG GGG GG GGGGGGGGGGGGGGGGGG GGGG GGGG GGGGGG GGGGGGG GGG GG GG GGG GGGGGGG GGGG GG GGGGG G GGGG GG GGG GGG GGGGGGG GGG GGG GG GGGGG GGGGGG GGGGGGGGGGGGGGG GGG GGG GGGG GGG GG GGG GG GGGG GGGGGGGGGG GG GGGGG GG GG GGGG GGGGG GGG GG GG GGGG GGGG GG GG GGGG GGGGGGGG GGGGG GGGGGGG GGGGG GG GGGG GGGGG GGGGGGG GGGG GGGG GGG GG G GG GG GG G GGG GG GGGG GGG GG GGGGG GG G GG GG GGGG GGG GG GGGGG GGGGGGG GGGG G GGG GGG G GG GGGG GGG GG G GGG GGGGG GG GGGGGGGGGGGG GG G GGG GG GGGGGGGG GG GG GGGGGGG G GGGGGGGGGGGGG GGGG GGG GGGGGG GGG GG GGGG G GG GG G GGG GGGGGGG GG GG G GGGGGG GG G GG G GGGGGG GG GGG GG GGGGGGG GGG G GGGG GG GGGGGGGGGGGGGGGGGGGGGG GGG G GGG G GGG GGGG GG GGGGG GG GG GGG GGGG GGGGGGG G GGG GGGGGG GGGG GGGGGGG GGGGG GG GGGGGG GGGGGG G GGG GGG G GG GGGG GG GGGG GGG GGGGGG GGGG GGGG G GG GG GGGGGGGGGG GG GG GG GG GGGGG GG GGGG GGGGG GGG GG GG G GGG GGGGG GGGGGGG GGGGG GGGG G GGG GG GGGGGGGGG GG GGG GGG GGG GGGGG GGGGGGG GGGGGG GG GGGGGG GGG GGGGG GGGGG GGGGG GGG G GGG GGGGGGGGG GG GG GGG GGGG GGG GGGGGGGGGGG GGGGGGG G GGG GGGGGGG G G 340 360 380 400 420 340 360 380 400 420 Measured CO2 [ppm] Calibrated CO2 [ppm] GGGGGGGGGGGGGGG GGGGG G G GG GGGG GGGG GG GGGGGGGGGG GGG GGG GG GG GGG GG GGG GGG GGGGGG GG GG GG GGG GGGGGGG GGG GG GG GG G GGG GGGGGGGGG GGGGGGGGG GGGGGGGG G GGG GG G G GGGG GG GGGGGGG GG GG GG GGGGG GGG GG GGG GG GGGG GGG G GG GG GGGGGGGGGG GG GGGGG GG GGG GGGG G GGGGGG GGG GGGG GGG GG GGG GGGGGGGG GGG GGGG GGGGG GG GGG GGG GGG G GG GGGGGG GG GGG GG GG GG GGG GG GGGGGGG G GGGGG GGGGG GGGG GGGG GGG GG GGG GGGGGGGGGGGG G GG G GGG GGGGGGG GGGG GGGGGGG GGGG GGGGGGG GG GGGGGGGGGGGG GGGGG GGGGGGGGG GGGG GGG GGGGGGGGGG GG G GGGG GGGGGG GGG GGGGGGG GGGGG GGGG GGGGG G GG GGGGG GGGG GGG G GG G GGGGGGGGGGGGGGGG GGG GGG GGGGGG GGG G GG GG GGGG GGGGGGGG GGGG GGG GG GG GGG GGGGGG GGGGG GGG GGGGGGG GGG G GGGGGG GGG GGGGGG GGGGGGG GGGGG GGGG GGGG GG GGG GGG G G GG GGG GGGG GG GGGG GG GGG GGGG GGGG GGGGGGGGGG GG GG GG GGG GGGGG GGGGGGGG GG GG GG GGGGGGGGG GGGG GG GGGGGGGG GG GG GG GGGGGGGG GG GG GGGG GGGGGG GGGGG GG GGGGGGGGG GG GGG GGGGGG GGGG GGG G GGGGG GGGGG GG GGGGG GGGGGG GGG GGGG GGGGGG GGG GGG GGGG GG GGG GG GGGGGGGGGG GGGGG GGGGGG GGGGG G GGGGG GGG GGGGGG GGGGG GG G G G GG GGGG GGGGG GGGGGG GGGGG GGG GGG GGGG GG GGGG G GGGGG G GG GGGGGG GGGGG GGG G GGGGGGGGG GGGGGGGGGGGG GG GG G GG GGG G GGGGGGGGGGG G GGG GGGG GGGGGGGGG G G GGGGGGGG G GGGGG GG GGGGGG GG GGGGGG GGGGG GGGG GGGGGGG GGG GGGGGGGG GG GGGG G GGGGGGGGG GGGGGGGG GGGGGGGG GGG G GG GGGG GGGG GGG GGGGGGG GGGGGG GGGGGGG GG GG GGGGGG G GGGGGGGG GG GGGG GGGG GGGGGGGG GGG GGG G GGG GG GGG G GGGGGG GGGG G GGG GG GGGGG GG G GGGG GGGGGGG GGG GG GG GGGG GGGGG G GG GGGGG GGGGGG GGGGGG GG GGGGGGGGGGGGGGGGGGGG GGGG GGG G GG GGGGG GGGGG GGG GGG GGGGGGGGGGGGGGGG GGG GGGGGGG GG GG GGGGGG GG GGG GGGG GGGGG GGG G GGGGGGG GGG GGG GGG GGGGGGG GGGGGGGGGGGGGGGGGG GGGGGG GGGG GGG GGG GGGG GGGGG GGG GG GGGGGGGGGGGGGGGGGG GGGG GGGG GGGGGG GGGGGGG GGG GG GG GGG GGGGGGG GGGG GG GGGGG G GGGG GG GGG GGG GGGGGGG GGG GGG GG GGGGG GGGGGG GGGGGGGGGGGGGGG GGG GGG GGGG GGG GG GGG GG GGGG GGGGGGGGGG GG GGGGG GG GG GGGG GGGGG GGG GG GG GGGG GGGG GG GG GGGG GGGGGGGG GGGGG GGGGGGG GGGGG GG GGGG GGGGG GGGGGGG GGGG GGGG GGG GG G GG GG GG G GGG GG GGGG GGG GG GGGGG GG G GG GG GGGG GGG GG GGGGG GGGGGGG GGGG G GGG GGG G GG GGGG GGG GG G GGG GGGGG GG GGGGGGGGGGGG GG G GGG GG GGGGGGGG GG GG GGGGGGG G GGGGGGGGGGGGG GGGG GGG GGGGGG GGG GG GGGG G GG GG G GGG GGGGGGG GG GG G GGGGGG GG G GG G GGGGGG GG GGG GG GGGGGGG GGG G GGGG GG GGGGGGGGGGGGGGGGGGGGGG GGG G GGG G GGG GGGG GG GGGGG GG GG GGG GGGG GGGGGGG G GGG GGGGGG GGGG GGGGGGG GGGGG GG GGGGGG GGGGGG G GGG GGG G GG GGGG GG GGGG GGG GGGGGG GGGG GGGG G GG GG GGGGGGGGGG GG GG GG GG GGGGG GG GGGG GGGGG GGG GG GG G GGG GGGGG GGGGGGG GGGGG GGGG G GGG GG GGGGGGGGG GG GGG GGG GGG GGGGG GGGGGGG GGGGGG GG GGGGGG GGG GGGGG GGGGG GGGGG GGG G GGG GGGGGGGGG GG GG GGG GGGG GGG GGGGGGGGGGG GGGGGGG G GGG GGGGGGG G G 340 360 380 400 420 340 360 380 400 420 Measured CO2 [ppm] Calibrated CO2 [ppm] G Cylinder meas. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1:1 y = 1.038x − 0.839 G G G G G 360 380 400 420 −0.15 −0.05 0.05 0.15 Calibrated CO2 [ppm] Residuals [ppm] G G G G GG GG G G G G GG GG G GG GG G G G G GG GGG GG G G G GGGG G GG G G G G G G G G G G G G G G G G G GG GG G GG G G G GG G G G GGG G GG G G G G G G G GG G G G G G G GG G GGG G G G GGG G G GG G GG G G G G G G G G G GGG GG G G G GGG G G G G G G G G GG G G GG G GG GG G G G G G G G G G G G GG G GGGG G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G GG G GG G G G G G G G G G G G G G G G G G G G G G G G GG GGG G G G G G G GG G G GGGG G GG G GG G G G GGG G G GG G G G G G G G G GGG GG G G G G G G G G G G GG GGG G G G G G G G G G G G G G G G GGG G G G GG G G GG G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G GG GG G GGG G G G G G G G G G GG G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G GG G G GG G G G G G G GG G G G GGG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G 400 420 440 460 480 500 520 400 420 440 460 480 500 520 Measured COS [ppt] Calibrated COS [ppt] G G G G GG GG G G G G GG GG G GG GG G G G G GG GGG GG G G G GGGG G GG G G G G G G G G G G G G G G G G G GG GG G GG G G G GG G G G GGG G GG G G G G G G G GG G G G G G G GG G GGG G G G GGG G G GG G GG G G G G G G G G G GGG GG G G G GGG G G G G G G G G GG G G GG G GG GG G G G G G G G G G G G GG G GGGG G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G GG G GG G G G G G G G G G G G G G G G G G G G G G G G GG GGG G G G G G G GG G G GGGG G GG G GG G G G GGG G G GG G G G G G G G G GGG GG G G G G G G G G G G GG GGG G G G G G G G G G G G G G G G GGG G G G GG G G GG G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G GG GG G GGG G G G G G G G G G GG G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G GG G G GG G G G G G G GG G G G GGG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G 400 420 440 460 480 500 520 400 420 440 460 480 500 520 Measured COS [ppt] Calibrated COS [ppt] G Cylinder meas. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1:1 y = 0.975x + 2.173 G G G G 100 200 300 −1.0 0.0 1.0 Calibrated CO [ppb] Residuals [ppb] Figure 2. Response curves determined with NOAA/ESRL calibration standards, including the residuals of the ﬁt. The NOAA/ESRL calibra- tion standards are calibrated on NOAA-2004 COS scale, WMO-X2007 CO2 scale (Zhao et al., 2006) and WMO-2004 CO scale (Novelli et al., 2003). Each standard was measured for 5 min. The measured mole fractions shown on the x axes are mole fractions calculated through the TDLWINTEL software but are corrected for drift using a reference standard. The drift was on the order of 8 ppt for COS, 3 ppm for CO2 and 5 ppb for CO and would have given biased results if this were not corrected for. response is linear for COS as well. The NOAA/ESRL cal- ibrated standards cover a wide range of CO2 and CO con- centrations and response curves that were determined for these gases were consistent when the experiment was re- peated. The factor 1.038 to transfer the QCLS CO2 measure- ments to the WMO scale is slightly different from the factor 1.05 found by Commane et al. (2013). For COS the concen- tration range of the NOAA/ESRL calibration standards was smaller, between 447.8 and 486.6 ppt, and combined with the lower precision of these measurements, the response curve has large uncertainty. Therefore, repeating the experiment resulted in varying response curves, with an average slope equal to 0.99 and the minimum and maximum slope equal to 0.87 and 1.04 respectively. The response curve shown in Fig. 2 is the curve where the two NOAA/ESRL calibration standards measurements showed the lowest standard devia- tions: 2.7 and 3.0 ppt, with the average standard deviation over different experiments equal to 4.4 ppt. In the next sec- tion we will discuss different calibration methods to deal with the higher uncertainty of the COS response curve. Further- more, we tested the stability of the response of COS, CO2 and CO with hourly measurements of three working stan- dards over a period of 35 days during ﬁeld measurements. The results of this will be shown in Sect. 3.1. ﬁlled with 8 hPa (6 Torr) COS and 27 hPa (20 Torr) CO. Ini- tially, we could use the peak position of COS in the reference cell to determine the frequency of the laser. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1:1 y = 1.038x − 0.839 G G G G G 360 380 400 420 −0.15 −0.05 0.05 0.15 Calibrated CO2 [ppm] Residuals [ppm] GGG GGG GG G GG GG GGGGGGGGG G G GG G G GGG G GGG GGGG GG G GG GG GGG G G GGGGGGGGGG GGGGGGGG G GG GGGGG GG GGG G G G G GG GGG GG G GGG G GGG G GG GGGG GGGG G GG GGGG GG GG G GGGG GG G GG GG GG GG GG GGG G G GG G GGGGGGG GGGG G GG G GG GGGGGGG GG GGGGGGGGGGGGG G GGG G G G GGG G G GGG GGG GGGGGGGGGGG GG GGGGGG GG GG GGG GGGGGG GGG GGGGG G GGG GGGGG GG GG GGG GGGG G GGG GGGGGGGG GGGGGGGGG GGG GGGGGG GGG G G GG G GG G G G G G GGG GGGGGG GGGG G GGG GGG G GG GG GG GGG G GGGGGG G GGGG GGG G G GGGG G GG G GGG GG G GGG GG GG G GGGGGGG GGGG GG GG GGG GGG GGGG GGGGG GGGG GGG G GG G G GG GGGGG GG GG G G GG GGG G GG G GG GG GG GGG G GG GGG GG GGGGGG GGG GGG G G GGGGGGG GG G GGG GGGGGGGGGGGGGGG GGGG GG GGG G G GGG GGG G GG GGGGGG GGG GG GG G G G G G G G GG G G GGG G GG GG GGG GGGGG GGGGGGGG G GG G GG GG GGG GGGGGG GGGGGGG GG GGG GG G GG GG GGGGG GG G GGG GGG G G G GGG G G GGG GGG GG GGGGGGGG GGG G GG G G GG GG GG GG GG G G GG GG G GG G GGGGGG GGG GGG GG GG GG GG GGG G GG G G GG GG GG G G G GG GGGGG GG GG G GG GG G G G G G GG GG G GGGGGGGG GGGG GGG G GG GGG G GG GGGGG GGG GGG G GG GGGGGG GG GG G G GGGG G GGGG GGG G GGGGGGGG GG GG GG GGG GG GG GG GGG GGG GG GGG GGG GG GGG GG G G GGGGGGG G GGGG GGG GG GGG G GG GG GGGGGGG GGG G G GGG GGGGG GG GGG GGG G GGGG G G GGG G GGG G GGG GG GGGGG GG GGGGGG GG G GG G GGGG GGG GGGG GG GGGGGG GGGG GGGG GG G G GG GGG GGG G GGG G GGGG GG GGG GGG GGG GGG GG GGG GG G GG GGG GGG GG G G G G GG GGG GG GGG G GG G GGG GG GG G GG G GGGGGG GG G GG G G GG GGG G GGG GG G GG GG GGG GG GGGGG GG GG GGGGGGGG G GGG GG GG GG GGG GGGG GGGGG GGGGG GG GGG G GG GG GG GG GG GGG GG GGGG GGG GGGG GGGG G GGG GGG GG GGG GGG GG GG G GGGGGG G G G GGGGG GGG G GGGG GG G GGGG GGG GG G G GGGG GG GG G GGGGGG G GG GG G GGGGGGGG GG GG G GG G GG GGGG G G GG G G GG GG GG G GGG GG G GGGGG G GGGGG GGGGG GG G GG GG GGGG GG GG GGG GGG GGG GGGGG GGGG G GG GG GG GG GG GGG GG GG GGGGGGGGGGGG G GG GGG GG G G G GG G G G G G GG GGGGGGG GGGGG GGGGGG GGGG GGGGGGG G G G GG GGGGG GG GG G G G GGG GGG GG GG GGGGG GGG GG G GG GG G G G GG GG GG G G GGG G GG GGG GGGG GGGGG GG GG G GG GGGG GGGGGG GG G GG GG G GG G GGG GGG G G GG GGGGGG GG G GG GG GG GGGG GGGGGGGGGGG GGGGGG G G G G GG GGG GGG GG G GG G G GG G GG GGGGGG GGGG GG G GGG G GG G GG GG GG GGG GGG G GGGG GGGGG GG GG GGG GG GG GGG GG GG GGG G GG G GGG GG GGGGG GGG GG GGG G GG GGG GGG GGGG GGG G GGG GG GG GG GGGG GGG GGG GG G G GGGG GG GG GGGGGGG GGGG GGGGG GG GG GGG G GGG G G GG GG GGG G GGGG G GGG GGGG G GG GGG G GGGGGGGG G GGGG GG GG GG GG GG GG GGG G GGG G G GG G GG G G G GG GGGG G GG GGG G GGGG GGG GG G GGGGG GG GGG GGGGG GGGG G GG GG G GG G GGG G G GGG GG G GG G GGG GGG G GGGG GGGG GG GG G GGGGGG 100 150 200 250 300 350 400 100 150 200 250 300 350 400 Measured CO [ppb] Calibrated CO [ppb] GGG GGG GG G GG GG GGGGGGGGG G G GG G G GGG G GGG GGGG GG G GG GG GGG G G GGGGGGGGGG GGGGGGGG G GG GGGGG GG GGG G G G G GG GGG GG G GGG G GGG G GG GGGG GGGG G GG GGGG GG GG G GGGG GG G GG GG GG GG GG GGG G G GG G GGGGGGG GGGG G GG G GG GGGGGGG GG GGGGGGGGGGGGG G GGG G G G GGG G G GGG GGG GGGGGGGGGGG GG GGGGGG GG GG GGG GGGGGG GGG GGGGG G GGG GGGGG GG GG GGG GGGG G GGG GGGGGGGG GGGGGGGGG GGG GGGGGG GGG G G GG G GG G G G G G GGG GGGGGG GGGG G GGG GGG G GG GG GG GGG G GGGGGG G GGGG GGG G G GGGG G GG G GGG GG G GGG GG GG G GGGGGGG GGGG GG GG GGG GGG GGGG GGGGG GGGG GGG G GG G G GG GGGGG GG GG G G GG GGG G GG G GG GG GG GGG G GG GGG GG GGGGGG GGG GGG G G GGGGGGG GG G GGG GGGGGGGGGGGGGGG GGGG GG GGG G G GGG GGG G GG GGGGGG GGG GG GG G G G G G G G GG G G GGG G GG GG GGG GGGGG GGGGGGGG G GG G GG GG GGG GGGGGG GGGGGGG GG GGG GG G GG GG GGGGG GG G GGG GGG G G G GGG G G GGG GGG GG GGGGGGGG GGG G GG G G GG GG GG GG GG G G GG GG G GG G GGGGGG GGG GGG GG GG GG GG GGG G GG G G GG GG GG G G G GG GGGGG GG GG G GG GG G G G G G GG GG G GGGGGGGG GGGG GGG G GG GGG G GG GGGGG GGG GGG G GG GGGGGG GG GG G G GGGG G GGGG GGG G GGGGGGGG GG GG GG GGG GG GG GG GGG GGG GG GGG GGG GG GGG GG G G GGGGGGG G GGGG GGG GG GGG G GG GG GGGGGGG GGG G G GGG GGGGG GG GGG GGG G GGGG G G GGG G GGG G GGG GG GGGGG GG GGGGGG GG G GG G GGGG GGG GGGG GG GGGGGG GGGG GGGG GG G G GG GGG GGG G GGG G GGGG GG GGG GGG GGG GGG GG GGG GG G GG GGG GGG GG G G G G GG GGG GG GGG G GG G GGG GG GG G GG G GGGGGG GG G GG G G GG GGG G GGG GG G GG GG GGG GG GGGGG GG GG GGGGGGGG G GGG GG GG GG GGG GGGG GGGGG GGGGG GG GGG G GG GG GG GG GG GGG GG GGGG GGG GGGG GGGG G GGG GGG GG GGG GGG GG GG G GGGGGG G G G GGGGG GGG G GGGG GG G GGGG GGG GG G G GGGG GG GG G GGGGGG G GG GG G GGGGGGGG GG GG G GG G GG GGGG G G GG G G GG GG GG G GGG GG G GGGGG G GGGGG GGGGG GG G GG GG GGGG GG GG GGG GGG GGG GGGGG GGGG G GG GG GG GG GG GGG GG GG GGGGGGGGGGGG G GG GGG GG G G G GG G G G G G GG GGGGGGG GGGGG GGGGGG GGGG GGGGGGG G G G GG GGGGG GG GG G G G GGG GGG GG GG GGGGG GGG GG G GG GG G G G GG GG GG G G GGG G GG GGG GGGG GGGGG GG GG G GG GGGG GGGGGG GG G GG GG G GG G GGG GGG G G GG GGGGGG GG G GG GG GG GGGG GGGGGGGGGGG GGGGGG G G G G GG GGG GGG GG G GG G G GG G GG GGGGGG GGGG GG G GGG G GG G GG GG GG GGG GGG G GGGG GGGGG GG GG GGG GG GG GGG GG GG GGG G GG G GGG GG GGGGG GGG GG GGG G GG GGG GGG GGGG GGG G GGG GG GG GG GGGG GGG GGG GG G G GGGG GG GG GGGGGGG GGGG GGGGG GG GG GGG G GGG G G GG GG GGG G GGGG G GGG GGGG G GG GGG G GGGGGGGG G GGGG GG GG GG GG GG GG GGG G GGG G G GG G GG G G G GG GGGG G GG GGG G GGGG GGG GG G GGGGG GG GGG GGGGG GGGG G GG GG G GG G GGG G G GGG GG G GG G GGG GGG G GGGG GGGG GG GG G GGGGGG 100 150 200 250 300 350 400 100 150 200 250 300 350 400 Measured CO [ppb] Calibrated CO [ppb] G Cylinder meas. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1:1 y = 1.012x Figure 2. Response curves determined with NOAA/ESRL calibration standards, including the residuals of the ﬁt. The NOAA/ESRL calibra- tion standards are calibrated on NOAA-2004 COS scale, WMO-X2007 CO2 scale (Zhao et al., 2006) and WMO-2004 CO scale (Novelli et al., 2003). Each standard was measured for 5 min. The measured mole fractions shown on the x axes are mole fractions calculated through the TDLWINTEL software but are corrected for drift using a reference standard. The drift was on the order of 8 ppt for COS, 3 ppm for CO2 and 5 ppb for CO and would have given biased results if this were not corrected for. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O Calibrated CO [ppm] Calibrated CO [ppm] G G G G GG GG G G G G GG GG G GG GG G G G G GG GGG GG G G G GGGG G GG G G G G G G G G G G G G G G G G G GG GG G GG G G G GG G G G GGG G GG G G G G G G G GG G G G G G G GG G GGG G G G GGG G G GG G GG G G G G G G G G G GGG GG G G G GGG G G G G G G G G GG G G GG G GG GG G G G G G G G G G G G GG G GGGG G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G GG G GG G G G G G G G G G G G G G G G G G G G G G G G GG GGG G G G G G G GG G G GGGG G GG G GG G G G GGG G G GG G G G G G G G G GGG GG G G G G G G G G G G GG GGG G G G G G G G G G G G G G G G GGG G G G GG G G GG G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G GG GG G GGG G G G G G G G G G GG G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G GG G G GG G G G G G G GG G G G GGG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G 400 420 440 460 480 500 520 400 420 440 460 480 500 520 Measured COS [ppt] Calibrated COS [ppt] G G G G GG GG G G G G GG GG G GG GG G G G G GG GGG GG G G G GGGG G GG G G G G G G G G G G G G G G G G G GG GG G GG G G G GG G G G GGG G GG G G G G G G G GG G G G G G G GG G GGG G G G GGG G G GG G GG G G G G G G G G G GGG GG G G G GGG G G G G G G G G GG G G GG G GG GG G G G G G G G G G G G GG G GGGG G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G GG G GG G G G G G G G G G G G G G G G G G G G G G G G GG GGG G G G G G G GG G G GGGG G GG G GG G G G GGG G G GG G G G G G G G G GGG GG G G G G G G G G G G GG GGG G G G G G G G G G G G G G G G GGG G G G GG G G GG G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G GG GG G GGG G G G G G G G G G GG G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G GG G G GG G G G G G G GG G G G GGG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G 400 420 440 460 480 500 520 400 420 440 460 480 500 520 Measured COS [ppt] Calibrated COS [ppt] G Cylinder meas. 2.2.1 Instrument response To characterize the response of the instrument for COS, CO2 and CO mole fractions we used six aluminum cylinders that were calibrated at the NOAA Boulder laboratories on NOAA-2004 COS scale, WMO-X2007 CO2 scale (Zhao et al., 2006) and WMO-X2004 CO scale (Novelli et al., 2003). Five of the six standards were calibrated for CO2, four for CO and two for COS. The internal surfaces of the COS calibrated cylinders were Aculife treated. Figure 2 shows that the re- sponse curves are linear for CO2 (between 354 and 426 ppm) and CO (between 94 and 384 ppb). Furthermore, the resid- uals of these curves do not show signs of nonlinearity. For COS, conclusions about the linearity of the response are not possible given that we had only two calibrated standards. As the response is linear for CO2 and CO, we assume that the 2.2 Calibration strategy To allow comparison of QCLS measurements with other in- strumentation and across different sites requires traceabil- ity to a primary scale. Laboratory tests were conducted to characterize the response of the instrument against ambient air standards from NOAA/ESRL, which were subsequently used to transfer the calibration scale to working standards. Moreover, we performed tests to understand the frequency required for background and reference measurements to en- sure reliable and accurate results. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O However, COS did not last longer than a few months in the reference cell so thereafter the laser frequency was locked only based on the peak position of CO, which did not impact the results. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1:1 y = 1.012x Calibrated CO [ppb] Calibrated CO [ppb] GGGGGGGGGGGGGGG GGGGG G G GG GGGG GGGG GG GGGGGGGGGG GGG GGG GG GG GGG GG GGG GGG GGGGGG GG GG GG GGG GGGGGGG GGG GG GG GG G GGG GGGGGGGGG GGGGGGGGG GGGGGGGG G GGG GG G G GGGG GG GGGGGGG GG GG GG GGGGG GGG GG GGG GG GGGG GGG G GG GG GGGGGGGGGG GG GGGGG GG GGG GGGG G GGGGGG GGG GGGG GGG GG GGG GGGGGGGG GGG GGGG GGGGG GG GGG GGG GGG G GG GGGGGG GG GGG GG GG GG GGG GG GGGGGGG G GGGGG GGGGG GGGG GGGG GGG GG GGG GGGGGGGGGGGG G GG G GGG GGGGGGG GGGG GGGGGGG GGGG GGGGGGG GG GGGGGGGGGGGG GGGGG GGGGGGGGG GGGG GGG GGGGGGGGGG GG G GGGG GGGGGG GGG GGGGGGG GGGGG GGGG GGGGG G GG GGGGG GGGG GGG G GG G GGGGGGGGGGGGGGGG GGG GGG GGGGGG GGG G GG GG GGGG GGGGGGGG GGGG GGG GG GG GGG GGGGGG GGGGG GGG GGGGGGG GGG G GGGGGG GGG GGGGGG GGGGGGG GGGGG GGGG GGGG GG GGG GGG G G GG GGG GGGG GG GGGG GG GGG GGGG GGGG GGGGGGGGGG GG GG GG GGG GGGGG GGGGGGGG GG GG GG GGGGGGGGG GGGG GG GGGGGGGG GG GG GG GGGGGGGG GG GG GGGG GGGGGG GGGGG GG GGGGGGGGG GG GGG GGGGGG GGGG GGG G GGGGG GGGGG GG GGGGG GGGGGG GGG GGGG GGGGGG GGG GGG GGGG GG GGG GG GGGGGGGGGG GGGGG GGGGGG GGGGG G GGGGG GGG GGGGGG GGGGG GG G G G GG GGGG GGGGG GGGGGG GGGGG GGG GGG GGGG GG GGGG G GGGGG G GG GGGGGG GGGGG GGG G GGGGGGGGG GGGGGGGGGGGG GG GG G GG GGG G GGGGGGGGGGG G GGG GGGG GGGGGGGGG G G GGGGGGGG G GGGGG GG GGGGGG GG GGGGGG GGGGG GGGG GGGGGGG GGG GGGGGGGG GG GGGG G GGGGGGGGG GGGGGGGG GGGGGGGG GGG G GG GGGG GGGG GGG GGGGGGG GGGGGG GGGGGGG GG GG GGGGGG G GGGGGGGG GG GGGG GGGG GGGGGGGG GGG GGG G GGG GG GGG G GGGGGG GGGG G GGG GG GGGGG GG G GGGG GGGGGGG GGG GG GG GGGG GGGGG G GG GGGGG GGGGGG GGGGGG GG GGGGGGGGGGGGGGGGGGGG GGGG GGG G GG GGGGG GGGGG GGG GGG GGGGGGGGGGGGGGGG GGG GGGGGGG GG GG GGGGGG GG GGG GGGG GGGGG GGG G GGGGGGG GGG GGG GGG GGGGGGG GGGGGGGGGGGGGGGGGG GGGGGG GGGG GGG GGG GGGG GGGGG GGG GG GGGGGGGGGGGGGGGGGG GGGG GGGG GGGGGG GGGGGGG GGG GG GG GGG GGGGGGG GGGG GG GGGGG G GGGG GG GGG GGG GGGGGGG GGG GGG GG GGGGG GGGGGG GGGGGGGGGGGGGGG GGG GGG GGGG GGG GG GGG GG GGGG GGGGGGGGGG GG GGGGG GG GG GGGG GGGGG GGG GG GG GGGG GGGG GG GG GGGG GGGGGGGG GGGGG GGGGGGG GGGGG GG GGGG GGGGG GGGGGGG GGGG GGGG GGG GG G GG GG GG G GGG GG GGGG GGG GG GGGGG GG G GG GG GGGG GGG GG GGGGG GGGGGGG GGGG G GGG GGG G GG GGGG GGG GG G GGG GGGGG GG GGGGGGGGGGGG GG G GGG GG GGGGGGGG GG GG GGGGGGG G GGGGGGGGGGGGG GGGG GGG GGGGGG GGG GG GGGG G GG GG G GGG GGGGGGG GG GG G GGGGGG GG G GG G GGGGGG GG GGG GG GGGGGGG GGG G GGGG GG GGGGGGGGGGGGGGGGGGGGGG GGG G GGG G GGG GGGG GG GGGGG GG GG GGG GGGG GGGGGGG G GGG GGGGGG GGGG GGGGGGG GGGGG GG GGGGGG GGGGGG G GGG GGG G GG GGGG GG GGGG GGG GGGGGG GGGG GGGG G GG GG GGGGGGGGGG GG GG GG GG GGGGG GG GGGG GGGGG GGG GG GG G GGG GGGGG GGGGGGG GGGGG GGGG G GGG GG GGGGGGGGG GG GGG GGG GGG GGGGG GGGGGGG GGGGGG GG GGGGGG GGG GGGGG GGGGG GGGGG GGG G GGG GGGGGGGGG GG GG GGG GGGG GGG GGGGGGGGGGG GGGGGGG G GGG GGGGGGG G G 340 360 380 400 420 340 360 380 400 420 Measured CO2 [ppm] Calibrated CO2 [ppm] GGGGGGGGGGGGGGG GGGGG G G GG GGGG GGGG GG GGGGGGGGGG GGG GGG GG GG GGG GG GGG GGG GGGGGG GG GG GG GGG GGGGGGG GGG GG GG GG G GGG GGGGGGGGG GGGGGGGGG GGGGGGGG G GGG GG G G GGGG GG GGGGGGG GG GG GG GGGGG GGG GG GGG GG GGGG GGG G GG GG GGGGGGGGGG GG GGGGG GG GGG GGGG G GGGGGG GGG GGGG GGG GG GGG GGGGGGGG GGG GGGG GGGGG GG GGG GGG GGG G GG GGGGGG GG GGG GG GG GG GGG GG GGGGGGG G GGGGG GGGGG GGGG GGGG GGG GG GGG GGGGGGGGGGGG G GG G GGG GGGGGGG GGGG GGGGGGG GGGG GGGGGGG GG GGGGGGGGGGGG GGGGG GGGGGGGGG GGGG GGG GGGGGGGGGG GG G GGGG GGGGGG GGG GGGGGGG GGGGG GGGG GGGGG G GG GGGGG GGGG GGG G GG G GGGGGGGGGGGGGGGG GGG GGG GGGGGG GGG G GG GG GGGG GGGGGGGG GGGG GGG GG GG GGG GGGGGG GGGGG GGG GGGGGGG GGG G GGGGGG GGG GGGGGG GGGGGGG GGGGG GGGG GGGG GG GGG GGG G G GG GGG GGGG GG GGGG GG GGG GGGG GGGG GGGGGGGGGG GG GG GG GGG GGGGG GGGGGGGG GG GG GG GGGGGGGGG GGGG GG GGGGGGGG GG GG GG GGGGGGGG GG GG GGGG GGGGGG GGGGG GG GGGGGGGGG GG GGG GGGGGG GGGG GGG G GGGGG GGGGG GG GGGGG GGGGGG GGG GGGG GGGGGG GGG GGG GGGG GG GGG GG GGGGGGGGGG GGGGG GGGGGG GGGGG G GGGGG GGG GGGGGG GGGGG GG G G G GG GGGG GGGGG GGGGGG GGGGG GGG GGG GGGG GG GGGG G GGGGG G GG GGGGGG GGGGG GGG G GGGGGGGGG GGGGGGGGGGGG GG GG G GG GGG G GGGGGGGGGGG G GGG GGGG GGGGGGGGG G G GGGGGGGG G GGGGG GG GGGGGG GG GGGGGG GGGGG GGGG GGGGGGG GGG GGGGGGGG GG GGGG G GGGGGGGGG GGGGGGGG GGGGGGGG GGG G GG GGGG GGGG GGG GGGGGGG GGGGGG GGGGGGG GG GG GGGGGG G GGGGGGGG GG GGGG GGGG GGGGGGGG GGG GGG G GGG GG GGG G GGGGGG GGGG G GGG GG GGGGG GG G GGGG GGGGGGG GGG GG GG GGGG GGGGG G GG GGGGG GGGGGG GGGGGG GG GGGGGGGGGGGGGGGGGGGG GGGG GGG G GG GGGGG GGGGG GGG GGG GGGGGGGGGGGGGGGG GGG GGGGGGG GG GG GGGGGG GG GGG GGGG GGGGG GGG G GGGGGGG GGG GGG GGG GGGGGGG GGGGGGGGGGGGGGGGGG GGGGGG GGGG GGG GGG GGGG GGGGG GGG GG GGGGGGGGGGGGGGGGGG GGGG GGGG GGGGGG GGGGGGG GGG GG GG GGG GGGGGGG GGGG GG GGGGG G GGGG GG GGG GGG GGGGGGG GGG GGG GG GGGGG GGGGGG GGGGGGGGGGGGGGG GGG GGG GGGG GGG GG GGG GG GGGG GGGGGGGGGG GG GGGGG GG GG GGGG GGGGG GGG GG GG GGGG GGGG GG GG GGGG GGGGGGGG GGGGG GGGGGGG GGGGG GG GGGG GGGGG GGGGGGG GGGG GGGG GGG GG G GG GG GG G GGG GG GGGG GGG GG GGGGG GG G GG GG GGGG GGG GG GGGGG GGGGGGG GGGG G GGG GGG G GG GGGG GGG GG G GGG GGGGG GG GGGGGGGGGGGG GG G GGG GG GGGGGGGG GG GG GGGGGGG G GGGGGGGGGGGGG GGGG GGG GGGGGG GGG GG GGGG G GG GG G GGG GGGGGGG GG GG G GGGGGG GG G GG G GGGGGG GG GGG GG GGGGGGG GGG G GGGG GG GGGGGGGGGGGGGGGGGGGGGG GGG G GGG G GGG GGGG GG GGGGG GG GG GGG GGGG GGGGGGG G GGG GGGGGG GGGG GGGGGGG GGGGG GG GGGGGG GGGGGG G GGG GGG G GG GGGG GG GGGG GGG GGGGGG GGGG GGGG G GG GG GGGGGGGGGG GG GG GG GG GGGGG GG GGGG GGGGG GGG GG GG G GGG GGGGG GGGGGGG GGGGG GGGG G GGG GG GGGGGGGGG GG GGG GGG GGG GGGGG GGGGGGG GGGGGG GG GGGGGG GGG GGGGG GGGGG GGGGG GGG G GGG GGGGGGGGG GG GG GGG GGGG GGG GGGGGGGGGGG GGGGGGG G GGG GGGGGGG G G 340 360 380 400 420 340 360 380 400 420 Measured CO2 [ppm] Calibrated CO2 [ppm] G Cylinder meas. 2.1 Instrumentation The nitrogen is ﬁrst passed over a gas puri- ﬁer (Gatekeeper, CE-500K-I-4R) to remove CO that is often found in such nitrogen cylinders. The frequency of the laser is locked based on the spectrum measurement of the high strength CO2 line at 2050.566 as shown in Fig. 1. For au- tomatic start-up, a gas sealed in an aluminum reference cell can be ﬂipped into the optical beam. The reference cell was The instrument consists of a 0.5 L astigmatic Herriott style multi-pass absorption cell (McManus et al., 2010) with an effective path length of 76 m. The cell has a temperature be- tween 20 and 24 ◦C, depending on the room temperature and the temperature setting of the thermoelectric chiller. The cell is kept at a constant pressure of 53.3 hPa (40 Torr) with an inlet valve that is controlled by the TDLWINTEL program (Aerodyne Research Inc., Billerica, MA, USA) based on the measured cell pressure. The same software manages the data acquisition and spectral analysis (Nelson et al., 2004) and calculates dry air mole fractions in real time (1 Hz) through nonlinear least square spectral ﬁts combined with the mea- sured cell temperature and pressure, a constant path length www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO a L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O Results of calibrations of three working standards with the QCLS are shown in Table 1 for CO2 and CO. We calibrated the same working standards with a cavity ring-down spectrometer (CRDS; Picarro Inc. model G2401- m) using the same standards linked to the WMO scale. Cali- brations with the QCLS for CO2 and CO agree with calibra- tions from the CRDS within their uncertainties, which gives conﬁdence in our calibration method. For COS we could not compare our calibrations with that from other instrumenta- tion. may be limited in labs, especially for COS, as this gas is usu- ally not one of the standard measured species. The question is therefore what method is suitable to transfer the scale to the working standards for COS. To test this, we re-analyzed cal- ibration measurements in different ways: (1) with response from the two NOAA/ESRL calibration standards and the curve forced through a zero point, (2) with two NOAA/ESRL calibration standards and the curve not forced through zero and (3) with a single bias correction using a NOAA/ESRL calibration standard that has the concentration closest to the working standard. g Table 2 shows the assigned values on the working stan- dards considering the different approaches. The calibration measurement was repeated three times and results are here shown as the average over the three measurement. The fact that the cylinder values need to be extrapolated causes larger deviations compared to when the curve is forced through a zero point (see, e.g., cylinder “A” in Table 2). There- fore, when working standards have concentrations outside the range covered by the NOAA/ESRL calibration standards, method 1 is preferred because it avoids extrapolation of the calibration curve by using the zero point. However, as we saw in Sect. 2.2.1 it is difﬁcult to accurately determine a calibration curve due to the lower precision of the mea- surements and the narrow COS concentration range of the NOAA/ESRL calibration standards, which holds for both methods 1 and 2. In this study we therefore calibrate cylin- ders and ﬁeld measurements under the assumption that the data follow a calibration curve with a slope equal to 1 (in Sect. 2.2.1 we saw that the average slope of different cali- bration experiments was equal to 0.99) and we apply a single bias correction, which is applied in Table 2 as method 3. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O Table 1. CO2 and CO calibration values together with 1σ uncertainties of the calibration curves as obtained with the QCLS and a cavity ring-down spectrometer (CRDS) for comparison of calibration results. COS calibration values could not be compared with that from other instrumentation. CO2 (ppm) CO (ppb) QCLS CRDS QCLS CRDS Cyl. no. 1 412.33 ± 0.12 412.43 ± 0.08 97.8 ± 1.7 97.6 ± 3.0 Cyl. no. 2 398.30 ± 0.12 398.13 ± 0.08 119.1 ± 1.7 119.1 ± 3.0 Table 2. COS calibration values for our real-air working standards as obtained from three different calibration approaches: (1) with response from the two NOAA/ESRL calibration standards and a zero point, (2) with the two NOAA/ESRL calibration standards and the curve not forced through a zero point and (3) using a single bias correction. The NOAA/ESRL calibration standards have COS concentrations 447.8 and 486.6 ppt. The calibration measurement was repeated three times; results are shown as the average over the three measurements and uncertainties indicate the standard deviation over the three measurements. Cyl. A Cyl. B Cyl. C Cyl. D 1. Through NOAA/ESRL standards and 0 393.0 ± 2.2 448.8 ± 4.7 473.6 ± 1.5 504.1 ± 0.7 2. Through NOAA/ESRL standards and not 0 379.2 ± 8.3 445.5 ± 4.4 474.7 ± 1.9 510.8 ± 4.7 3. Single bias correction 390.9 ± 2.6 445.8 ± 3.9 476.6 ± 2.4 506.6 ± 2.1 as target standards). The working standards used in this study are aluminum gas cylinders (Luxfer, max. 200 bar) ﬁlled to high pressure with ambient air at the Center for Isotope Re- search of the University of Groningen using an oil-free air compressor (RIX SA-3) and are used in combination with two-stage regulators (Scott Specialty Gases, model 14). The working standards differ from the NOAA/ESRL calibration standards in the sense that they are uncoated. To trace the working standards that we used in the ﬁeld we refer to these with numbers 1, 2 and 3 throughout the paper. Other work- ing standards that were used in this paper but not in the ﬁeld are referred to as “A”,“B”, “C” and “D”. Using the linear re- gression curve that we found in Fig. 2, we determined mole fractions in these working standards by considering response curves derived from measurements of the NOAA/ESRL cal- ibration standards. 2.2.2 Working standards In this study we use working standards to represent high- pressure cylinders that are calibrated against NOAA or WMO standards in our laboratory. The working standards are used for two purposes: (1) to correct for instrument drift during ﬁeld measurements and link the measurements to the NOAA or WMO scales (in this case we refer to these stan- dards as reference standards) and (2) to assess the accuracy of the measurements (in this case we refer to these standards www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 5297 www.atmos-meas-tech.net/9/5293/2016/ L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5298 Table 3. Standard deviation (SD) over minute-averaged data of COS, CO2 and CO over the 19 h measurement period for different correctio frequencies using reference measurements. Uncorrected 1 h corr. 30 min corr. 15 min corr. COS SD (ppt) 11.8 9.6 5.3 4.6 CO2 SD (ppm) 0.26 0.12 0.09 0.08 CO SD (ppb) 0.8 0.7 0.3 0.3 D) over minute-averaged data of COS, CO2 and CO over the 19 h measurement period for different correction asurements. reality, but here we observe that the difference of method 3 against methods 1 and 2 is small (on average 1.3 ppt) and the result is often in between those of methods 1 and 2. We therefore consider the single bias correction as a good com- promise when a calibration curve cannot accurately be deter- mined. of 100 ppt (COS), 2 ppm (CO2) and 10 ppb (CO) per degree change of the electronics temperature in a few hours when the instrument was placed in an enclosed box and when the electronics section was actively cooled (see Sect. 2.5). To test the frequency of reference measurements that is needed to re- move drift due to temperature changes, we measured cylin- der air over a period of 19 h, alternating with a reference gas every 15 min. During this measurement period the elec- tronics temperature changed by 0.1 ◦C due to changes of the room temperature, where the most rapid change was 0.06 ◦C in 30 min. Without correcting the data, the standard devia- tion of the minute-averaged data is 11.8 ppt for COS, with drift up to 40 ppt. Correcting the data with reference mea- surements from every 15 min lowers the standard deviation to 4.6 ppt, while correcting only every 30 and 60 min gives a standard deviation of 5.3 and 9.6 ppt respectively. See Table 3 for an overview of the standard deviations of COS, CO2 and CO. Based on these results we can say that reference mea- surements every 30 min are sufﬁcient to remove drift within 5.3 ppt with temperature changes up to at least 0.06 ◦C per 30 min. Field measurements have shown that with correc- tions from 30 min reference measurements, drift is still suf- ﬁciently removed with temperature changes up to 0.2 ◦C per 30 min. Furthermore, improving the temperature stability of the instrument could reduce the drift such that reference mea- surements are needed less frequently. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O The effect of improved temperature stability was tested during ﬁeld operation of the QCLS at the Lutjewad station (see Sect. 2.5). 2.2.3 Background and reference strategy Background measurements with dry nitrogen are required to reduce the effect of curvature of the baseline spectra and are typically done every 2, 5 or 30 min by other users of simi- lar QCLS analyzers (Stimler et al., 2010a; Commane et al., 2013). When we tested the required frequency of background measurements we found that COS concentrations can shift to another level, either up or down, uncorrelated with in- strument parameters such as temperature, and even when the background measurements were done shortly after each other (every 10 min). In a test where we measured cylinder air over a period of 5 h, alternating with background measurements every 10 min, we found that on average the COS concentra- tion shifts by 5 ppt after every background measurement, but in 32 % of the cases the shift is larger than 10 ppt, with peaks up to 37 ppt. We could not correlate these shifts to changes in instrument temperature or inlet pressure or to the length of the background measurement itself. As we could not ﬁnd correlations with any parameter, the reason for these shifting concentrations after background measurements is unknown. To be able to correct for the concentration shifts for accu- rate concentration measurements requires measurements of reference gases at least once within every background cy- cle. Therefore, increasing the background frequency auto- matically leads to greater use of a reference cylinder gas. As we also need to consider cylinder logistics (ﬁlling, calibrat- ing and transportation), we decided to do less frequent back- ground measurements (once every 6 h) with reference cylin- der measurements in between. Lowering the background fre- quency to 6 h did not negatively affect the measurement pre- cision, but the accuracy of the measurements is affected due to instrument drift over the period in between the background measurements. The frequency of reference cylinder measure- ments that is needed to correct for the instrument drift de- pends on the rate with which the instrument temperature changes. Hourly measurements of reference standards in the ﬁeld (during a later ﬁeld campaign in Hyytiälä, Finland, not shown here) showed that concentrations drift on the order L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O The- oretically, methods 1 and 2 would give the results closest to As we saw in Sect. 2.2.1, the response curve of COS is dif- ﬁcult to determine due to the lower instrument precision and the narrow COS concentration range of the NOAA/ESRL calibration standards. These factors introduce uncertainty in assigning values to the working standards that will be used on-site, especially for those having COS mole fractions out- side of this range. Besides that, the number of available cali- brated standards used to transfer the scales to other cylinders www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 Atmos. Meas. Tech., 9, 5293–5314, 2016 2.3 Water vapor interference correction The concentration measurements of gases determined through light absorption spectrometry can be affected by water vapor in the sample air in two ways: (1) by spectro- scopic effects (enhanced pressure broadening or direct spec- tral interference), which will directly modify the absorption spectrum, and (2) through dilution of the sample air, which linearly depends on water vapor concentrations. The wa- ter vapor interference can be prevented by drying the air before measurement. However, this requires adding a drier to the sampling inlet lines and, depending on the drier, it can require additional maintenance, which is not favorable for unattended autonomous measurements. As the QCLS in- cludes measurements of H2O, water vapor interferences can be accurately accounted for in the calculation of dry air mole fractions of the other gases. We determined the water va- www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 . J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5299 Figure 3. Left: water vapor experiment where cylinder air was humidiﬁed with wet silica gel and no water vapor correction was applied to the data. Right: wet over dry ratio of COS, CO2 and CO vs. H2O as obtained from multiple water vapor experiments with humidiﬁed cylinder air. The data show the linear dependence with H2O when the TDLWINTEL water vapor correction was turned off with the standard ﬁt (blue) and with the split ﬁt (green). Figure 3. Left: water vapor experiment where cylinder air was humidiﬁed with wet silica gel and no water vapor correction was applied to the data. Right: wet over dry ratio of COS, CO2 and CO vs. H2O as obtained from multiple water vapor experiments with humidiﬁed cylinder air. The data show the linear dependence with H2O when the TDLWINTEL water vapor correction was turned off with the standard ﬁt (blue) and with the split ﬁt (green). late to H2O when the TDLWINTEL water correction was turned off (blue). For this ﬁgure we combined three water vapor tests, of which one of the three was done with the TDLWINTEL water correction turned off and the other two had the TDLWINTEL water correction turned on. 2.3 Water vapor interference correction We ran the playback mode of the TDLWINTEL software to get the data of these water vapor experiments for the case that the correc- tion was turned off, when the real time data were obtained with the correction on. For CO only two of the three experi- ments are used in the analysis because the third showed larger scatter (colored in gray), indicating instability of the QCLS. We will further discuss the precision of the measurements and its variation over time in Sect. 3.1. The linear regres- sion is derived from averages over every 0.1 % H2O range such that the regression is not inﬂuenced by the fact that there are more data points towards lower H2O concentra- tions. Figure 3 shows that the COS, CO2 and CO wet / dry ratios are all linearly dependent on H2O. When wet air mole fractions are measured (that is, the TDLWINTEL water va- por correction is turned off), these curves could act as a water vapor correction factor to obtain dry air mole fractions. The effect of H2O on the species is +3.0 % for COS, −1.46 % for CO2 and −0.8 % for CO per % H2O. The standard de- por dependence of mole fractions reported by the instrument (TDLWINTEL software) for COS, CO2 and CO from labo- ratory experiments. To do this we measured dry and humid- iﬁed cylinder air alternately. We humidiﬁed the cylinder air with wet silica gel in a ﬁlter, giving a water vapor proﬁle from 2.1 % down to 0.2 % H2O in the sample air when the line was ﬂushed (Rella et al., 2013). The interaction of sil- ica gel with COS, CO2 and CO was tested by drying the air with magnesium perchlorate or a cryogenic system after the air had passed the wet silica gel. No interaction with COS, CO2 and CO was found for silica gel, magnesium perchlorate and the cryogenic system: the average difference between the unaffected and humidiﬁed/dried silica gel air was max- imum 1.0 ± 1.9 ppt for COS, 0.11 ± 0.13 ppm for CO2 and 0.3 ± 0.4 ppb for CO. Furthermore, we found that a 0.3 nm molecular sieve, which is commonly used to remove water vapor, removes all COS from cylinder air. The dry air cylin- der measurements were used as a reference during the exper- iment to account for instrumental drift. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O This means that vary- ing the broadening coefﬁcient of COS from 1.0 to 1.5 only changes the COS concentration by 2.9 ppt (at a concentration of 450 ppt COS). The large uncertainties of the COS and CO broadening coefﬁcients therefore indicate that these coefﬁ- cients have relatively little effect on the concentrations. Different water vapor correction strategies can now be considered and are summarized in Table 4. An appropriate direct water correction for COS is not possible with the com- bination of the standard ﬁt and the TDLWINTEL correction on. However, a correction curve can still be applied to these data with a curve that is determined with the same broad- ening coefﬁcients as are used for the original data to be cor- rected. Here we found that for a broadening coefﬁcient of 1.5 with the standard ﬁt the slope of the curve for COS is equal to 0.031. We continued to test the performance of the water vapor correction based on the standard ﬁt and the TDLWIN- TEL correction off. We applied the correction curves to ﬁeld measurements over a period of 35 days in March and April 2015. In Sect. 3.2 we will compare the dry air mole fractions with measurements of a collocated CRDS for CO2, CO and H2O and with dry air ﬂask sample measurements for COS. Using the playback mode we tried to ﬁnd the most optimal water broadening coefﬁcients to sufﬁciently correct for water vapor using TDLWINTEL. We did this for the three differ- ent water vapor tests and for both the standard ﬁt and the split ﬁt. For the standard ﬁt we could not ﬁnd optimized broaden- ing coefﬁcients for COS because turning the TDLWINTEL correction on caused an opposite correction and resulted in larger deviations from the assigned cylinder value due to the effect of the H2O peak on the baseline. For the standard ﬁt, the optimized broadening coefﬁcients varied between 2.1 and 2.2 for CO2 and between 1.0 and 2.0 for CO for the different experiments. For the split ﬁt we did ﬁnd optimized coefﬁ- cients for COS between 1.0 and 1.4 for the different experi- ments; for CO2 and CO the same results were found as for the standard ﬁt. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O Different water vapor correction strategies based on the software ﬁtting parameters (standard or split ﬁt) and with the TDL- WINTEL correction turned on or off. When the TDLWINTEL cor- rection is turned on the values indicate the broadening coefﬁcient used for the different species, and when it is turned off the values indicate the slope of the correction curve as determined in Fig. 3 with y = slope × H2O + 1 with H2O in percent and y the wet / dry ratio of the gas. TDLWINTEL Broadening coefﬁcient or slope correction on/off COS CO2 CO Standard ﬁt on –* 2.15 1.0 off 0.030 −0.0146 −0.008 Split ﬁt on 1.0 2.15 1.0 off −0.008 −0.0149 −0.008 * No broadening coefﬁcient could be derived; however, we found that for a broadening coefﬁcient of 1.5 with the standard ﬁt the slope of the curve for COS is equal to 0.031, which can be applied as an extra correction on top of the TDLWINTEL correction. TDLWINTEL Broadening coefﬁcient or slope correction on/off COS CO2 CO Standard ﬁt on –* 2.15 1.0 off 0.030 −0.0146 −0.008 Split ﬁt on 1.0 2.15 1.0 off −0.008 −0.0149 −0.008 * No broadening coefﬁcient could be derived; however, we found that for a broadening coefﬁcient of 1.5 with the standard ﬁt the slope of the curve for COS is equal to 0.031, which can be applied as an extra correction on top of the TDLWINTEL correction. * No broadening coefﬁcient could be derived; however, we found that for a broadening coefﬁcient of 1.5 with the standard ﬁt the slope of the curve for COS is equal to 0.031, which can be applied as an extra correction on top of the TDLWINTEL correction. * No broadening coefﬁcient could be derived; however, we found that for a broadening coefﬁcient of 1.5 with the standard ﬁt the slope of the curve for COS is equal to 0.031, which can be applied as an extra correction on top of the TDLWINTEL correction. ﬁt, and the value for COS is only suitable for the split ﬁt. As the optimization of the broadening coefﬁcients depends on the curves in Fig. 3 (most optimal is when the slopes are equal to zero), the uncertainty of the water correction also entails an uncertainty in the broadening coefﬁcients. We ﬁnd that the uncertainties of the broadening coefﬁcients are equal to 0.5 (COS), 0.03 (CO2) and 0.7 (CO). L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O When the different experiments were combined, the optimized broadening coefﬁcients are equal to 1.0, 2.15 and 1.0 for COS, CO2 and CO respectively, where the values for CO2 and CO can be used for both the standard and split L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O viation of the residuals of the ﬁt translates to an uncertainty in concentrations of 2.9 ppt for COS (at a concentration of 450 ppt), 0.10 ppm for CO2 (at 400 ppm) and 1.1 ppb for CO (at 150 ppb). These uncertainties are dependent on the COS, CO2 and CO concentrations. The fact that CO2 and CO show an inverse correlation with H2O indicates that these species are primarily affected by the dilution effect of water vapor. To the contrary, COS shows a positive relation with H2O. The reason for this positive correlation is that the baseline shape of the spectra is distorted by a small H2O peak on the left of the larger H2O peak to the right of the large CO2 peak (see Fig. 1). A potential solution to this would be to split the ﬁt between the COS peak and the small H2O peak, which we will now refer to as “split ﬁt” (in contrast to the “standard ﬁt”). In Fig. 3 (right) the results with the split ﬁt are shown in green. For CO2 and CO the green curve does not differ signiﬁcantly from the blue curve, as would be expected. For COS the correlation with H2O is now negative, with an ef- fect of −0.8 % per % H2O. A slope of −1 % would be ex- pected due only to pure dilution; however, pressure broaden- ing by water vapor affects the mole fractions as well. Within TDLWINTEL the width of the peaks is ﬁxed based on the line shape information from the HITRAN database, which is only for air broadening with N2 and O2. Water broadening can be greater than air broadening. To correct for the pres- sure broadening effect the software modiﬁes the width of the absorption line through the so-called air “broadening coefﬁ- cients”, which are speciﬁc for every spectral line of a certain species. The later version of TDLWINTEL can also use the water broadening effect by increasing the air broadening co- efﬁcients from HITRAN. The residual error of the ﬁt, which is caused by the broadened absorption line but not properly adjusted line width, affects the mole fraction and thereby the slope of the curves in Fig. 3. This can explain why the slopes differ for the different species. Besides that, the measured H2O concentration has an uncertainty, which contributes to the deviation of the slope from −1 %. Table 4. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5300 viation of the residuals of the ﬁt translates to an uncertainty in concentrations of 2.9 ppt for COS (at a concentration of 450 ppt), 0.10 ppm for CO2 (at 400 ppm) and 1.1 ppb for CO (at 150 ppb). These uncertainties are dependent on the COS, CO2 and CO concentrations. The fact that CO2 and CO show an inverse correlation with H2O indicates that these species are primarily affected by the dilution effect of water vapor. To the contrary, COS shows a positive relation with H2O. The reason for this positive correlation is that the baseline shape of the spectra is distorted by a small H2O peak on the left of the larger H2O peak to the right of the large CO2 peak (see Fig. 1). A potential solution to this would be to split the ﬁt between the COS peak and the small H2O peak, which we will now refer to as “split ﬁt” (in contrast to the “standard ﬁt”). In Fig. 3 (right) the results with the split ﬁt are shown in green. For CO2 and CO the green curve does not differ signiﬁcantly from the blue curve, as would be expected. For COS the correlation with H2O is now negative, with an ef- fect of −0.8 % per % H2O. A slope of −1 % would be ex- pected due only to pure dilution; however, pressure broaden- ing by water vapor affects the mole fractions as well. Within TDLWINTEL the width of the peaks is ﬁxed based on the line shape information from the HITRAN database, which is only for air broadening with N2 and O2. Water broadening can be greater than air broadening. To correct for the pres- sure broadening effect the software modiﬁes the width of the absorption line through the so-called air “broadening coefﬁ- cients”, which are speciﬁc for every spectral line of a certain species. The later version of TDLWINTEL can also use the water broadening effect by increasing the air broadening co- efﬁcients from HITRAN. The residual error of the ﬁt, which is caused by the broadened absorption line but not properly adjusted line width, affects the mole fraction and thereby the slope of the curves in Fig. 3. This can explain why the slopes differ for the different species. Besides that, the measured H2O concentration has an uncertainty, which contributes to the deviation of the slope from −1 %. 2.3 Water vapor interference correction Figure 3 (left) shows the mole fractions of the humidiﬁed air measurements dur- ing the experiment with the TDLWINTEL water vapor cor- rection turned off. The right plots in Fig. 3 show how the wet / dry ratios of the COS, CO2 and CO concentrations re- www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 2.5 Temperature stability tween the inlet at the tower and the laboratory where the an- alyzer is positioned (up to 60 m length in the tower and 30 m from the tower to the lab building) with a ﬂow of 2 L min−1. Just in front of the pumps we split the line with a tee junction to get a subsample from the tower lines. These subsamples are pulled through the sample cell with an oil-free dry scroll pump (Varian TriScroll) downstream of the cell of the QCLS. Half-inch Synﬂex (Decabon) tubing is used in the tower and in front of the KNF pumps. We have tested the interaction of Teﬂon and Synﬂex with COS: no signiﬁcant differences in COS mole fractions were found when the air ﬂow was alter- nately passed over stainless steel and Teﬂon (0.9 ± 1.9 ppt) or Synﬂex (0.7 ± 2.6 ppt). We did observe COS production from the KNF N-86 pumps, but this did not cause problems for the measurement sampling as these pumps were placed on the bypass lines. We use a 5.0 µm Teﬂon ﬁlter at the in- let of the tower sample lines to prevent dust, sand and salt coming into the lines. In front of the analyzer we use another 0.5 µm stainless steel ﬁlter (Swagelok) to prevent pollution of the sample cell. A check valve is placed between the an- alyzer and the vacuum pump to prevent unﬁltered room air to enter the cell through the pump in case the vacuum pump suddenly stops. A needle valve was placed between the cell and the vacuum pump to control the ﬂow of the system at 0.16 L min−1, resulting in a 90 % response time of ∼15 s. We use a multi-position Valco valve (VICI; Valco Instru- ments Co. Inc.) to switch between sample lines from differ- ent heights in the measurement tower and to working stan- dards. The Valco valve is controlled by the TDLWINTEL software in which one can set the sequence and duration of the valve ports to be measured. An interval is set for auto- matic hourly repetition of the sequence. Until January 2015 we used a solenoid valve (Parker) to switch between dry ni- trogen and cylinder or ambient air. However, we found that this solenoid valve was leaking and was thereby diluting the cylinder and ambient air measurements with dry nitrogen. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5301 2.5 Temperature stability On 7 January 2015 we changed the setup such that the Valco valve also controlled switching to the dry nitrogen. The Valco valve, KNF pumps, air puriﬁer and 0.5 µm ﬁlter are built into a 19 in. rack with 1/8 in. stainless steel tubing. We noticed that under lab conditions, when the temperature is controlled within ∼0.2 ◦C, the precision of the instrument was typically better than with the highly changing tempera- tures up to 2 ◦C at the measurement station during the course of the day. These temperature dependencies were also ob- served with other QCLS analyzers in Xiang et al. (2014) and Berkelhammer et al. (2014). Before we made any modiﬁca- tions, the electronics temperature (Telectr) varied with 0.91 ◦C with every degree of changing room temperature (Troom) and the cell temperature (Tcell) varied with 0.11 ◦C ◦C−1. Xiang et al. (2014) showed the potential to improve the tempera- ture stability with an active temperature control using an Oa- sis 3 chiller (Solid State Cooling Systems, USA) of which the set point can be controlled with the TDLWINTEL soft- ware. They improved the Tcell variability of 0.03 ◦C ◦C−1 without active temperature control to 0.0005 ◦C ◦C−1 with active control. We improved the temperature stability of the instrument with only the ThermoCube chiller, for which we extended the cooling loop to a heat exchanger attached to the fan of the electronics section. This way, the air going into the electronics section for ventilation is actively cooled. Moreover, we have put the analyzer in an enclosed box to add an extra layer of temperature isolation. With these mod- iﬁcations the temperature variability of Telectr improved to 0.21 ◦C ◦C−1 and to 0.07 ◦C ◦C−1 for Tcell. 2.4 Flow schematics for measurements at the Lutjewad station In July 2014 we deployed the QCLS in the ﬁeld for measure- ments at the Lutjewad monitoring station in the Netherlands (60 m; 6◦21′ E, 53◦24′ N; 1 m a.s.l). We use three diaphragm pumps (KNF N-86) to keep the inlet lines well ﬂushed be- Atmos. Meas. Tech., 9, 5293–5314, 2016 www.atmos-meas-tech.net/9/5293/2016/ www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 . Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5302 2XW 7 m 40 m 60 m $LULQOHW &DOLEUDWLRQF\OLQGHUQR 12 position Valco valve 4XDQWXPFDVFDGH ODVHUVSHFWURPHWHU &DOLEUDWLRQF\OLQGHUQR Optical table Electronics 1LWURJHQF\OLQGHU &DOLEUDWLRQF\OLQGHUQR Swagelok filter 0.5 μm ,QOHW YDOYH &KHFN YDOYH Varian TriScroll Pump Thermo- cube chiller KNF N-86 pump QHHGOHYDOYH within 19'' rack Teflon filter 5.0 μm 3UHVVXUH VHQVRU p Pump with shut-off valve )ORZ PHWHU Flask analysis only $LU SXULILHU Figure 4. Schematic overview of the instrument setup for tower proﬁle measurements at the Lutjewad monitoring station. The pressure sensor and pump with shut-off valve were added for ﬂask measurements only (see Sect. 2.6). $LULQOHW 60 m 40 m Swagelok filter 0.5 μm ,QOHW YDOYH )ORZ PHWHU Figure 4. Schematic overview of the instrument setup for tower proﬁle measurements at the Lutjewad monitoring station. The pressure sensor and pump with shut-off valve were added for ﬂask measurements only (see Sect. 2.6). inlet pressure below ambient. The measurement results will be shown and compared with the in situ QCLS measurements in Sect. 3.2. surement setup as it is used for ﬂask measurements only. Air from reference and target tanks was introduced before mea- suring the actual ﬂask sample to calibrate the measurements to the NOAA scale. We did not use measurements of a ref- erence gas after the ﬂask measurement as the stability of the measurements is affected by the larger pressure difference between the ﬂask and the reference cylinder. Two measure- ments of 2 min each were done on each ﬂask and the results were averaged to derive the ﬁnal value. The results of these ﬂask measurements will be shown in Sect. 3.2. surement setup as it is used for ﬂask measurements only. Air from reference and target tanks was introduced before mea- suring the actual ﬂask sample to calibrate the measurements to the NOAA scale. We did not use measurements of a ref- erence gas after the ﬂask measurement as the stability of the measurements is affected by the larger pressure difference between the ﬂask and the reference cylinder. Two measure- ments of 2 min each were done on each ﬂask and the results were averaged to derive the ﬁnal value. The results of these ﬂask measurements will be shown in Sect. 3.2. 2.6 Measurements of COS from ﬂasks Besides testing the QCLS for continuous in situ measure- ments, we explore the suitability of the QCLS for measur- ing air from ﬂasks. We developed a means to analyze ﬂasks; this allowed a comparison between the QCLS and GC-MS measurements from the NOAA Boulder laboratory (Montzka et al., 2007). For this analysis we ﬁlled ﬂasks from cylin- ders, such that, besides the instrument comparison, we were able to assess the accuracy of the ﬂask measurements and check the methodology for assigning values to our working standards based on the NOAA/ESRL calibration standards. Four pairs of glass ﬂasks with a volume of 2.5 L were ﬁlled to 2.5 bar with dry air from two of our working standards and the two NOAA/ESRL calibration standards. The work- ing standards were calibrated for COS by comparing their response on the QCLS to the NOAA/ESRL calibration stan- dards as described in Sect. 2.2.2. The four cylinders con- tained COS concentrations between 447.8 and 486.6 ppt. The ﬂask measurement setup using the QCLS is similar to that of routine tower measurements but with a few modiﬁcations. A pressure sensor was added to monitor the pressure in the ﬂasks. Furthermore, a diaphragm pump with a shut-off valve was used to remove residual air from previous ﬂask measure- ments in the connecting tube and to test the lines for leaks. In Fig. 4 the placement of the pressure sensor and diaphragm pump is indicated; they are separated from the ﬁeld mea- For our setup, every hour starts with a measurement from a reference and target standard (3 min each). Subsequently, the system alternates between three measurement heights where every height is measured for 8 min, meaning that every height is measured twice an hour. The reference standard was mea- sured every half hour to remove instrument drift. In March 2015 we measured an extra reference gas once every hour, which was used to test the stability of the instrument re- sponse over a period of 35 days (the results will be shown in Sect. 3.1). Background measurements were done every 6 h with dry nitrogen over 60 s. Before the actual background measurement is done, the cell is ﬁrst ﬂushed for 2.5 min to make sure that water vapor is removed from the cell by 99 %. www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 www.atmos-meas-tech.net/9/5293/2016/ 3.1 Precision and accuracy 1: 495.7 ppt COS, 412.33 ppm CO , 2 97.8 ppb CO no. 2: 468.5 ppt COS, 398.30 ppm CO , 2 119.1 ppb CO no. 3: 454.5 ppt COS, 421.50 ppm CO , 2 237.2 ppb CO GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GG GGGGGGGGGGGGGGGG GGG G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGG GGGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGG GGG GGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGG GGGGGG GGGGGGGGG GGGGGGGGGG GGG GGG GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GG GGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGG GG G G GGGGGGGGGGGGGGGG GGGGGGGGGGGGGG G GGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGG GG GGGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGGG G GGGGGGG GG G G G GGG GGGGGGG GGGG GGGGGG GGGG GG G GGGGG G G G GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG Telectr. [K] 292 296 300 304 Aug'14 Sep'14 Oct'14 Nov'14 Dec'14 Jan'15 Feb'15 Mar'15 Apr'15 May'15 Figure 5. Mean concentrations of hourly measurements of working standards conducted while the instrument was in the ﬁeld from August 2014 until April 2015 together with electronics temperature. The data shown here are uncorrected data and are not calibrated with a response curve. The concentrations are therefore not necessarily close to the assigned cylinder values. The solid vertical line at 7 January 2015 indicates the moment when we changed the setup with a solenoid valve to a Valco valve for switching to nitrogen, as the solenoid valve was leaking. The dashed vertical line at 25 March 2015 indicates the moment when we improved the temperature stability by actively cooling the electronics section and putting the analyzer in an enclosed box, which resulted in substantially smaller temperature ﬂuctuations than before. From this moment onwards an extra working standard was also measured every hour to ascertain if the instrument response was stable over a period of 35 days (orange). The gap in the data record in December 2014 and February–March 2015 is because of tests with the QCLS in the laboratory. periods before and after improvement of the temperature sta- bility, with the black/gray colors corresponding to that in the left ﬁgure. 3.1 Precision and accuracy We assessed the measurement uncertainty and accuracy with the hourly measurements of the reference and the target gases over the period from August 2014 until April 2015. As men- tioned previously, each reference and target gas was mea- sured for 2 min. The mean value of the hourly instrument- reported and uncorrected 2-minute measurements are shown in Fig. 5. In Fig. 6 the standard deviation of these measure- ments is shown. Figure 5 also includes the electronics tem- perature of the QCLS. The cylinder measurements show that concentrations can drift substantially; i.e., COS concentra- tions easily vary by 50 to 100 ppt. However, on the long term, concentration changes are not correlated with tem- perature, which changed by 13 K throughout the year. The concentration shift on 7 January 2015 (especially visible in CO2) happened after eliminating the leaking solenoid valve that caused mixing of nitrogen into the tubing that deliv- We also measured dry air ﬂask samples to test if the wa- ter vapor correction that we determined in Sect. 2.3 sufﬁ- ciently removes the effect of water vapor on calculated mole fractions. To that aim, ﬂasks were ﬁlled to ambient pressure as part of a standard ﬂask sampling routine at the Lutje- wad station (Neubert et al., 2004). The air samples are dried with a cryogenic system prior to collecting. The ﬂasks were stored for maximum 1.5 months before being measured. The same measurement strategy as for the NOAA/ESRL com- parison was used, and for these ﬂasks two measurements of ∼1.5 min could be done before an inlet pressure of 0.3 bar was reached. We did not observe any dependence of mea- sured dry air mole fractions in air from the ﬂasks with the Atmos. Meas. Tech., 9, 5293–5314, 2016 www.atmos-meas-tech.net/9/5293/2016/ L. M. J. 3.1 Precision and accuracy Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5303 G GGGGG G G GGGG G G G GG G G GGG GG G GGGG G G G G GG G G G G G G G G GG G G G G GGG GG G G GG G G G G G G GG G GGGG G G G GG G G G G GG G G G G GG GG G G G G G G G GG GG G G GG G G G G G G G G GG G G GGG G G GG G G G G G G G G G G G GG G G G G G GGG G GG G G G G G G GG G G GG G G G G G G GG G G GG GG G G G G G G GG G G GG G GG G G G G GG G G G G G GG G G G G G G GG G GGG GG G GGG G G G G G GGG G G GGG GG GGGGG G G G G GG G GG GGG G G GGGG G GGG GG G GGGGG G G GGG G G GGG G G G G GGGG G G G G GGG G G G GG GG GG G G G GGG GG G GGGGG G GGG GG G GGGGG G G GGGG G G GG GG GG GG G G G G G GGG G G G G G G G GG GGGG GG GG G G G GG GG G G GG GG G GGGG G G GG G G G G G GG G G G GG G GG G G GGGG G GG GG G G GGGG G G G G GG G G G GGGGG G G GGG G G GGGG G GG GG G G G GG GGG G G G G G GG G G G G G G GG G GG G GG G G G G GGGG G G G G GGGG G G GG GGG G G G G G G G G G G GGG GGGG G G GGGG G G G GGG G G G G GGG G GG G GG G GG GGG G G GGGG G GG GGG G GGG GG G GG G G GG GGGGG G G GG GG G GGG G G G GG G G G G GGGG G G GG G GG G G G GGG G COS [ppt] G G GGG G G G G G G G G GGG GG G G G G G GG G G G G GG G GG G GG G G G G GGG G G G GG G G G GGGG G G G G G G G GG G GGG G G GG GGG G G G GG GG G G G G GG GGGG G G GG G G G G G GGG G G G G GG GGG G G GG G G GG G G G GG G GG G G G G GG G GG G G G G G G G G G G G G G G GG G GG GG G G GG GG GG G G G G GG GG G GG G GG GGG G GG G G G G G G GG GG G G G G G G G GG G GG G G G G G G G G GG G G GG G G GGGG GG GG GGG G GG G GGGGGGG G G GG GG G G GGGGG G GG G G G G G G G GGGGGGG GG G GG GG G GG G GG GGGG G G G GG G G G G G G G G G G G G GG G G G G G GGG G G G G G G G G G G G G GG G G G G G G GGG G G G G G G G G G G G G G G G G GG GG G G G G GG G G G GGGG G G G G G GGG G G GG G GG GGG G GG G G GG GG G G GGG GG GGG G GG GG G G G GGGGG G G G G G G G G G G G GG GGGGG GG GG G G GGGGG GG G G GG G G G G G G G G GGG G G G G GG GG G GGGG G G G G GG GG G GGG G GG GG GG GG GGG GG GGG GGGG GGG GGGG GGG G GGGG G GG GG GGGGGG GG G G G GGGG GGG GG G GGG GGG G G GG G GG GG GG G G GGGG G G GGG G GGG GGG G G G G G G G GG G G G G G G G GGG GG GG GG GGGG G G G G G G GG GG G GGG G GG G GG GG G GGGG G GG GG G G GG GGGGGGGG GG GG G G GG GGGGG G G GG G GG G G G GG G GGG GGGG G G GG G G G G GG GG G G G GGGGG GGGGG GGGG GGGGGG GG GGGGG GGG GGGGG G GG G GG GGGG G G GG GGGGGG GGG G G G GG GGG G GG GG G GGGG GGGG G G GG G GGGG G G G G GG GG G G G G G G GGGG G G G GGGGGG GG G G G G GGGG G GG GGG G GG G G G GGG G GG G G GGG G GG G GGG G G G G G G G G GGG G G G G G G G G G G GGG G GGG GG G G GGG G GG G G GG G GG GG G GGGG G GG G G G G G GGG G G G G G GGG GG GG G G GGGG GG G G GG G GG G G G GG GG G G G G G G G G GGG GGG GG GG GGG GGGGG GGG GGG G GG G G G GG G GG G G G G GG G G GGGGG G GG G G GGGGG G G GG GG G GGG G G GG G GGG G GGGGGG GG GGGGGG GG GG G G G GGG GG G GGGGG GGGG G G G G G G GGGG GGGGG G GG GG G G GG G GGGGGGG G GGG G G GG G G GGG GGG G GG GGGGGGG G GGGGGGGGGG G G GG G GG GG G G GGGGG G GG GG G G G GG G G G G GGGG G G G GGG GGG G G G G G G G GGG GG GGG GG GG G GGG G GG GG G GGGG GGGGGGGGGGGGGG G GGGGG G GGGG GGGG G GG GGG G G GGG G GGG G G GG GGGG G GGGG G G G G GG G GG G G GG G GG GG GGGG G G GG G GG G G GG GGG G GG G G G GG G G G G G G G GG G G G G G G GG G G G GGG G G G G GG G G G GGGG G GGGG G G G GG GG G G G GG GG G GG G G G G G G G G G GG G G G G G GGG G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G G G GGGG G G G G G G G G GG G G G G G G G GG G G G G G GG G G GG G GGG GG G GGG G G G GG G G GGG G G G G GG GGG G GG G G G G G G G GG G G GG G GG G G G G G G GG G G G G G G GG GGG G G GGG G G G G G G G G G GGG G G G G G G G G G G G G G G G GG G GG G G GG G G G GG G GG G G G GG GG G G G G G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G GGGG GG G G G GG G G G GG G GG G GG G G G G GGGG G G G G G G G G GGG G G G G G G G G GG G G G G G G G G GG GGG GGG G G G G GG GGG G GG G G G G G G G G GG GG G G G G G GGG G G GG GG G GG GG G GGG GG GGGG G GGG GG G GG G G G GG G G GG GGG G G GG G G GG G G G GGG G G G GG GG GGG GG GG GG GG GG GG G GG GG GGG GGGG G GGG G G G G G G GGG G G G G GGGG GGGGG GGG G GG G GG G GG GG G G GG G GGG G GG G G G GGGG GG G G GG GG G G G G G G G GG GG G G G G G G GG G G GG G GGG G G GG GG G GG GG G G GG G G G G GG G G G G G G G GG G G G G GGG G G G G G G G G GG G G G G GG G G GG G GG G G G GG G G GG GGG G G G G G G G G GG G G G G G GG G G G G GG G GG G GGG G GG G G G GG GG GGG G G GGG GGG GGG G G GG G G GG G GGGG GGGG GG GGGG G GGG GG GG G GGGGGG G G G G G G GG GG GGGGG GG G GG G G G GGG G G G GG GGG G G G GGG G G G GG GGG GG G GG GG G G GG GG GGGG G GGGG G GG G G G GG GG G GG G G G GG G GG G G G GG G G GGG GGG GG GG GG G GG GGG G G G GG G G G G G G GGG G GGG G GGG GGGGG GGG GGGG G GG GG G G GGGG GGG G G G G G GG G G G GGG GGGG GG GG GG G GG G G G GGG G G G GG GGGG G G G GGGGGG GG GGGGG G GG G G G GG G G G G G GG G GG G G GG GGG G G G GG G G GGG GG G G G G G G G G G G GGGG GGG G G G GGG GG GGGG GG GGG GG G G G G G GG GG GG GG GG G G GG G G G GGG G G G GG G G G G G G GG G G G GGG GG GG G G GG GG GGGGG GG GGGG G GG G G GGGGG G G G G GGG GGGG G G G G G G G G G GG GG G G G GGG GG GG G GG G GGGG GG GG G GG GG G GGG GG GG G GG G GG GG GGGGGG GG GG G GG GGG G G G G G GG G GG G GG G G G G G G GG G GGGG G GG GGGG GGG G GG GGGGGG GG G GGGG G G GGG GG GGGG G G G G GGG G GGGGG G G GGGG GG G GG G GGGGGGGGG GGG GG G G G G G GGG GG G G G G G G GG GG GGG G G G G GG G GG G GG G GGG GG G GG GG GG GG G G GGG G G GGGG GG GG G GGGG G G G GG G G G G G GGG GG GG G G G G G GG GG G G GGGGG G GG GGGG G GGG G GGG G G G G GG GGG G G G G G G G GGG GG G G GG GG GGG GG GGG GG G GG GGGG G G G G GGGGG G G GG GG G G G GGG G GGG GG G GG GG G G GG G G G G G G G G GG G G G GG G GGG G G G G GG GG G GG G GG G GGG GG G G G G GGG G G G GG G G GG G G G GG G GG G GG G G G G GG G G G G G G GG GG G G GGG G GG G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G GG G G G G G G G G G GGG G GG GGG G G G G G G G G G G GG G G GG GG G GGG G G G G GG G G G G G G G G G GGGG G G GGG G G G G GGGG G G GG G GG GGGGG G GG G GG G G G GGG G G GGGG G GGG G G G GGGGG G G G GG G G G GG GG G G G GGG G G G G G GG G G G GG G G G G G G G GGG GG G GGGG G G GGGG G G GGGGG G G GGG G G G G GG G GG GG G G G G GGG G G G G G G G G G GGGG GG GG G G G GG G G G G GG G G G GGGGG G G G G G G G G GG G G G G G G G G G G GGG GG G GG G G G G GG G G G G G GG G G G G G GG G GG GGG G G GGGG G GGGGG G GG G G G G G GG G G G G G G G G G G GG G G G G GGGG G G GGGG G G G G G G GG G G G G GG G G G G G G G G G G G G GG GGGG G G GGGG G G G GG G G G G G GGG G GG GGG G GGGGG G GGGGG G GGG GG G G GGGG G GG G GGG GGGGG G G G G GG G GGGGG G GG G G G G GGGGG G GG G G G G G G GGG G COS [ppt] G G G G G G G G G G GGGG G GG G G G G G G G G G G GG G G GG G GG G G G G G G G G G G GG G G G GG GG G G G G G GGG G G G GG G G G G GGG G G G G G G GGG GG G G GGGGGG GG G G G G GG G GG G G GGG GGGG G G G G G GG G GGGG G GG G G GGG G G G G G G G GG GG G GGGG G G GG G GGG G G GG G G G G G G G G GG G GG G G GG G G GGG G GGGG G G G GG G GG G G G G GG G GG G GG G GG GG G G GGG GG GGGG G GGGG G GG GGG G GG G GGG G G GG G G GG GG G G GG GG G GGG G G G G GG G GGG G G GG G GG GG GG GG G GG G GGG GGG G GG GG G GG G G G G G G G G GG GG G G G G G G G G G GG G G G G G G G GG G G G G G GG G G GG G G GG GGGG G G G G GG G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G GG G GGG GGG GG GG GGG G G GG GGG GG G GG GGGG G G G G G GGGG GGGG GG G GG GGG G G GG G GG G GGGGG G G G G G GG G G G GGG G GGGG G G G G GG G GG G G G G GG GG G GGG GGG G GGG G GGGGGG GGGGG GGG G G GGGG GGGG GGGG G G GG GG G GGG GGG G GG G GGGGG GG GGG G G G GG G G GG GGGG GG GGGGGG GG G GGGG G G G G G GG G G G G G GG G GG GG GG GG G GGG G G G G G GGG GG GGGGG GGG G G GG G GG GG G GGG G G G GGG G GGGG GG GG GG G G GG G GGGG G G G GG GG G G G GGG GGG G GGG G G GG G G GG GG GG G G G GGGGG GGGGG GG GGGGGG G G GGGGGGGGGG GG GG GGG G G GGGGGGG G G G GGGGGG GGG G G G G GGGG G GG GG GGGGGGGG G G G GG G G GGGG G G G GGG G G G G G G G GGGG G G G G G GG G G G G G G G G G G GG G GG GGG G G G G G G GG G G GG G G GGG G GG G G G G G G G G G G G G GG G G G G G G G G GG G G GG G GGG GG G G GGG G GG G G G G G GG GG G GGGG G G G GG G G G GGG G G G G G GGG GG G G G G G G GG GG G G GGG GG G G G G G GG G G GG G G G G GG G GG G G G GG GGG GGGGG GGGG G GG GG G G G G G G GG G G G GGGG G GG GGG G GG GGGGGGG G G GG GG G GGG G G GG G GGG G G GGGGG G GGGGGG G GG GG G G G GG G GG GGGGGG GGGG G G G G G G G GGG GGGGG G GG GG G G GG GGGG GGG G G G GG G G GG G G GGG GGG G G G GGGGGGG G GG GGGGGGGGGGG G G GG GG G G GGG GG G G G GG G G G GG GG G G G GG G G G G GGG GGG G G G G G G G GG G GG G GG GG G G G GGG G GG G G G GGG G GGGGGGGGGGGG GG G GGG G G G G GGG GGGG G GG GGG G G GGG GGGG G G G G GGGGG GGGG G G G G GG G G GG G G G G G G GGG GG GG G GG G GGGG G G GGGGGG G G GGG G GGGG G G G G GG G G G G G G GG G GG G G G G G G G G G GG G G G G G GG G GG G GG G G G G G GG GG G G G G G G G G GG G GGG G G G G GGGG GG G G G G G G G G G GGG GG G G G G G G G G GGG G G GG G GG GG G G G G G G G G G GG G G G G G G G G GG G G G G G G G G G GG G G GGGG G G G G G GG GG G G G G G G GGG GG G G GG G G G G G G G GGGG G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G GG GG G G GG G G G GG G GGG GG G G G GG G G G G G G G G G G G G G G G G G G G G GG G GGGGG G G G GG G G G G G G G G G G GGG G G G G G G G GG G G G G G GG G GGG GG G G G G G G G G G G G G GG G G G G G G G G GGG G GG G G G GG G G G G G G G GG G G GG GG G G GG G GG GG GGG G G GG GGGG G G GG GG G G G G G G G GG G G G G GG G GG G G G G G GG GGG G GG G G G G G G G G GGG G G G GG GGGG G GG GG G GG G G GGGGG G GG GG G GGG G G G G G G GG GG G G G G GGG G G GG G G GG G G GG GG GG G G G GG G GG G GG GG G GGG G G G G GGGGGGG GGGGGGGG GG G G G GGG G G G G G GGGG GG GGG GG G G GG G GG G GG GG G G GG G GGG G GG G G G GGGGG G G G GG G G G G G G G G G GG GGGGG GG G GG G G GG G GGG GG G GGG G GG GG G G G G G G GG GGG GG G G G G G G G GG G GG G GG G G G G G G GG G G G G G G G G GG G G G G G G GG GGG G G GGG G G G G GG G GG G GG GG GG G G G G G G G GG G GGG G G G G GG G G G G G G G G G GG G GG GG GG G GG G GG GG G GGGGGG GGGG GG G GG GGG G GG G G G G GGG G G G G G G G GG GG G GGG G G G GGGG G G GGG G GG G G GGGG G G G G G G G G GGG GGGG G GG GG GGGG G G GGGG G GG G G G GGG GG G G GGG G GG G G GG G GG G GG G G G G GGG GG GG GGG GGG GG G GG G G G G G G G GG G G GGG G G G GG GGG GGGGG GGG GGGG G GGGG GG GG G GGGG G G G G GGG GG GG G GGGGG GG GG GGGGGG GGGGG G G GG G GG GG G GG GGGGGG GG GGGGG G GG GG G GG G G G G G GG GGG G G GG GGG G G GG G GG G GG G GG G G G GGG GG G GGGG G GG G G G GGG GGGGGGG G G GGG GG GG G GG GG G G GG GG GG G G GG G G G G G G G G G GG G G G G G G GG G GGGGGGGGGGG GGGG GG G G G GGGG GGG GG G G G GG GG GGG G GGGG GG G GG G GG G GG GG GG G G G G GGGGGG G G GG GGG G G G G G GG GG G G G G G G GG GGG G G GGG GG GGGGGG GG GGG G G G G GG G G GGGG GG GG GGG GG GG G GG G GGGGG GG GGGGGGG G G G G G GG GG GG GG G GG G G G GG GG GGG GGGG G GG G G G GGG G G G GGGG GG G GG GGGGG GGGGG GGG GG G GGG G G GG G G G G G GG G GG GG G GG G GGG GG G GG G GG G GGG GG G G G GG GG GG G G GGG G GG GGGG G GG G GGGG G G G G G GGG G G G GG GG G G G G G G G GG GG G G GGGG G G GGGGGG GGGGGGGG GG G G G G G G GG G G G G G GG GGGG GG G G GG G GG GG GGGGG G GG GGGG G G G GGGGGGGGGG G GG GG GG GGGG G GG GG GGGGGG G GG GG GG GGGG G GGGGGGGGGG G GG GG GG GG GG G GG GG GG GG GG G GGGGGG GGGG G GGGGGG GG GG G GG GG GG GGGG G GGG G GG GGGG G GGGG GGGGGG G GG GGGG GG G G G GG GGGG GGGG G GG GGGG GG GG G GGGG G G GG GG G GG G G GG GG G GG GG GG GG GGG GG GG GG GG GG G GG GG GGGGGGG GG GG GGGG GG G GG GGGG GG GG G GG GG GG GG GGG G G GG GG GG GG G GG GG GGGGGG G G G G G GG GG GG G GG GG GG GG GG G G G GGGG GG GGG GG GGG G GG GG G GG GG GG GG GG G GG GG GG GG GG G GG GG GGGG G G GGG GG G GGG GG G GG GG GGGG GG G GG GG GG GG GG G GG GGGGGGGG G G G G G G GGG GGG GG GG GG GGGG G GG GGGG GGGG G GG GGGGGGGGG GG GGGG GG GG G GG GG GG GGGG G GG GG GGG GGGG GG GG GG GGGG G GG GGGGG GGG G GG GG GG GG GGG GG GG GGGGGG G GG GG GG GGGGG GG GG GG GG GG GGG GG GG GG GG G GG GG GG G G GG GGG GG GG GG GG G GG GG GGGGGG G GG GGGGGGGG G GG GG GGGGGGG GG GG GGGGGGG GG GG GG GG COS [ppt] 350 450 550 650 GGGGGGG G GGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGG G GGGGG GGGGG GGGG G GGGG G GGGG GGGGGGGGGG G GGGGG GGGGGGGGGG GGGG G GGGG GGGGGGGGGGG GGGGG GGGG G GGGGGGGGGG GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGGG GGGGG GGGG GGGGGG GGGGGGGGG G GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG GGGGG GGGGGGGGGGGGGGGG GGGG GGGGG GGGGGG GGGGG GGGG G GGGG G GGGG G GGGG GGGGG G GGGGG G GGGG GGGGG GGGGGG G G GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GGGGG GGGGGGGGGG GGGGGGGGGGG GGGG GGGGGG GGGG G GGGG GGGGGG G GGGG G GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGG GGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGG GGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GG G G GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG CO2 [ppm] GGGGG G G GGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG G GG GGGGGGGGGGG G GGGGG GGGG G GGGG G GGGG G GGGG G GGGGG GGGG G GGGG G GGGGGGGGG G GGGG G GGGG G GGGG GGGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGGG GGGG G GGGGG GGGGG GGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGG GGGGGGGGGG G GGGGGG G GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG G GGGG G GGGGGGGGGG GGGG G GGGG G GGGG G GGGGG GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGGG G GGGG GGGGG GGGGGG G G GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GGGG G GGGGG GGGG GGGGGG GGGG G GGGG G GGGGG GGGG G GGGG GGGGGG G GGGG G GGGG GGGGG G GGGG G GGGGG GGGGG GGGGG GGGG GGGGGG G GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGG GG GGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG CO2 [ppm] GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGG G GGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG CO2 [ppm] 370 390 410 GGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGG G GG G GGGGGGGGGGGGG G GG G GGG GG GGGGGGGG GGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGG G GGGGGGGGGGGGG GGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GGGGGGGGG G GGGGGGG GGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGG GGGGG GGGGG GGGGG GGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGGGGGGG G G G GG GGG G G GGGG G GG G G GG GGG GG GGG GGGGGGGG GGGGG GG GGGGGGGG GGGGG GG GGGG GGGGGGGGGGGGGGG GGGGGG GGG GG G GGGG GGGGG GG G GGGG GGGGGGG GGG G G GGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGG GGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGG GGGGGGGGGG GGGGGGGGG G GGGGGGGGGGGGGGG GGGGG GGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GG G GGGGGGGGG GGG GG GGGGGGGGGG GGGGGGGGGGGGGG GGGGGG GGGGGGG G GGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGG GG GGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGG GG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGG GGGGGGGGGGGGG GGGGG GG GGGGG G GGGGGGGG GG G G GGGGG GGGG G GGGGG G GGGGGGGGGGGGGG GGGG GGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGG G GGGGGGGGGGGGGGGGGGGGGG G G GGGGGGGGGGGGGG GGGGGGGGGGGGG GGGGGG GGGGGGG GGGGG GG GGG GGGGGGGGG GG GGGG GGGGGGGGGGGGG GGGGG G GGGGG GGGGGGGGGGGGGGGGGG GGG GGG GGGGGGGGGGGGG GG G GGGGGGGG GGGGGGGGGGGGGGGGGG GGGGGG GGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGG GGGGG GGGGGG GGGGGG GGGGGG GGG GGGGG GGGGGGGGGGGG GGGGGG GGGGGGGGGG GGGG GGGGGG GG GGGGGGGGGGGGG G GGGGG G GGG G GGGGG GG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGG GGG GGGGGG GGGGGGGGGGGGG GGGGGGGG GGGGGGGGGGG G GGGGGGGGGGGGGGG GGG GGGGGGGGGGGG GGGGGGGGGGGG G G G GGGGGGGGGGGG GGGGGG GGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGG GG GGGGG GGGGG GGGG GGGGGGGG GGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GG GGGGGG GGGG GGGGGGGG GGGGGG G GGGGG G CO [ppb] GGGGG G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGG GGGGGGGGGG G GGGG GGGGGGGG GG G GGGG GGGGG GGGGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGG GGGGGGGGGGGGGGGGG G GG GGGGG GGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGG GGGG G G GGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GGGGGGGGGG GGGGGGG GGGGGGGGGGGGGGGGGGGGG GG GGGGGG GGGGGGGGG GGGGGGGGGG GGGGG GGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGG G G G GGGGG G G GG GG GGGG G GG G GG GGGGG G GGGGG G G GGGGG GG G GG GGGGG GGGG G GGGGGG GGGG GGGGGGG GGG G GGGGGG GGGGG GGGGG G GGG GGG G GGGGG GG G GG G GGG G G GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGG GGG GGGGGGGGGG GGGGGGGGG G GGGGGGGGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG G GGGGGGGG GGGGGGGG GGGGGGGGGGGG G GGGGGGGGGGGG GGG GGGGG GGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGG GGGG G G GGGGGGGGGG GGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGG GGGGGGG GGGGGGGGGGGGG GGG G G GGGGG GGGGG GGGGG G GGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGG G GGGGGGGGGG GGGGGGGGGGGG GG GGGGGGGG GGGGGGGGG GG GGGGGGGGGGG GGGGGGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGG GGGG GGGGGGGGGGG GGGGGGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGGG GGGGGG GGG GGGG GGGGGGG GG GGGGGGGGGGG GG GGGGGGGGGGGGGGGGGG GGGG GG GGG G G GG GGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGG GGGG GG GGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGG GGG GGGGGGGGGG G GG GGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGG GGGGGGGGGGG GGGGGGGGGGG GG GGGGGGGGG GGGG GGGGGGGGGGGGGGGGGG GG GG GGGGGGGGGGG GGGGGGGGGGGGGGGGG G GGGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG GGGGG GG GGGG CO [ppb] G GGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGG GG G GGGG GGGGG GG GGGGGGGGGGGG GGGGGGGGGGGGGG GGGGGGGGG G GGGGGG GGGGGG G GGGGGGGGGGG GGG GGG GGGGGGGGGGGGGGGG GGGG GGGG GGGGGGGGGGGGG GGGGG GGGGGG GGGG GGG GGGG GGGGGGGGGG GGGGGGG GGGGGGGGGGGG GG GGGGGGGGGGGGG G GGGG GGGGGG GGGGGG GGGGGG G GG GGGGGGGGGGG GGGGGG GGGGGG GGGGGGGGGG GGGG GGGGGG GGG G GGGGGGGGGGG GGGGGG G GGG G GGGGGGG G GGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGG GG GGG GGGGGG GGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGG G GGG GGG GGGGGGGGGGGG GGGGGGGGGGGG G G GGGGGGGGGGGG GGGGGG GGGGGG GGG GGGGGGGGGGG GGGGGGGGGGGG G GGGGGG GG GGG G GGGGGGGGGGG GGGGGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGGGG GG GGGG GGGG GGGGGGGG GGGGGG G GGGGG G CO [ppb] 100 150 200 250 G G G Assigned cylinder values: no. 3.1 Precision and accuracy 1: 495.7 ppt COS, 412.33 ppm CO , 2 97.8 ppb CO no. 2: 468.5 ppt COS, 398.30 ppm CO , 2 119.1 ppb CO no. 3: 454.5 ppt COS, 421.50 ppm CO , 2 237.2 ppb CO GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GG GGGGGGGGGGGGGGGG GGG G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGG GGGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGG GGG GGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGG GGGGGG GGGGGGGGG GGGGGGGGGG GGG GGG GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GG GGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGG GG G G GGGGGGGGGGGGGGGG GGGGGGGGGGGGGG G GGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGG GG GGGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGGG G GGGGGGG GG G G G GGG GGGGGGG GGGG GGGGGG GGGG GG G GGGGG G G G GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG Telectr. [K] 292 296 300 304 Aug'14 Sep'14 Oct'14 Nov'14 Dec'14 Jan'15 Feb'15 Mar'15 Apr'15 May'15 Figure 5. Mean concentrations of hourly measurements of working standards conducted while the instrument was in the ﬁeld from August 2014 until April 2015 together with electronics temperature. The data shown here are uncorrected data and are not calibrated with a response curve. The concentrations are therefore not necessarily close to the assigned cylinder values. The solid vertical line at 7 January 2015 indicates the moment when we changed the setup with a solenoid valve to a Valco valve for switching to nitrogen, as the solenoid valve was leaking. The dashed vertical line at 25 March 2015 indicates the moment when we improved the temperature stability by actively cooling the electronics section and putting the analyzer in an enclosed box, which resulted in substantially smaller temperature ﬂuctuations than before. From this moment onwards an extra working standard was also measured every hour to ascertain if the instrument response was stable over a period of 35 days (orange). The gap in the data record in December 2014 and February–March 2015 is because of tests with the QCLS in the laboratory. 3.1 Precision and accuracy G GGGGG G G GGGG G G G GG G G GGG GG G GGGG G G G G GG G G G G G G G G GG G G G G GGG GG G G GG G G G G G G GG G GGGG G G G GG G G G G GG G G G G GG GG G G G G G G G GG GG G G GG G G G G G G G G GG G G GGG G G GG G G G G G G G G G G G GG G G G G G GGG G GG G G G G G G GG G G GG G G G G G G GG G G GG GG G G G G G G GG G G GG G GG G G G G GG G G G G G GG G G G G G G GG G GGG GG G GGG G G G G G GGG G G GGG GG GGGGG G G G G GG G GG GGG G G GGGG G GGG GG G GGGGG G G GGG G G GGG G G G G GGGG G G G G GGG G G G GG GG GG G G G GGG GG G GGGGG G GGG GG G GGGGG G G GGGG G G GG GG GG GG G G G G G GGG G G G G G G G GG GGGG GG GG G G G GG GG G G GG GG G GGGG G G GG G G G G G GG G G G GG G GG G G GGGG G GG GG G G GGGG G G G G GG G G G GGGGG G G GGG G G GGGG G GG GG G G G GG GGG G G G G G GG G G G G G G GG G GG G GG G G G G GGGG G G G G GGGG G G GG GGG G G G G G G G G G G GGG GGGG G G GGGG G G G GGG G G G G GGG G GG G GG G GG GGG G G GGGG G GG GGG G GGG GG G GG G G GG GGGGG G G GG GG G GGG G G G GG G G G G GGGG G G GG G GG G G G GGG G COS [ppt] G G GGG G G G G G G G G GGG GG G G G G G GG G G G G GG G GG G GG G G G G GGG G G G GG G G G GGGG G G G G G G G GG G GGG G G GG GGG G G G GG GG G G G G GG GGGG G G GG G G G G G GGG G G G G GG GGG G G GG G G GG G G G GG G GG G G G G GG G GG G G G G G G G G G G G G G G GG G GG GG G G GG GG GG G G G G GG GG G GG G GG GGG G GG G G G G G G GG GG G G G G G G G GG G GG G G G G G G G G GG G G GG G G GGGG GG GG GGG G GG G GGGGGGG G G GG GG G G GGGGG G GG G G G G G G G GGGGGGG GG G GG GG G GG G GG GGGG G G G GG G G G G G G G G G G G G GG G G G G G GGG G G G G G G G G G G G G GG G G G G G G GGG G G G G G G G G G G G G G G G G GG GG G G G G GG G G G GGGG G G G G G GGG G G GG G GG GGG G GG G G GG GG G G GGG GG GGG G GG GG G G G GGGGG G G G G G G G G G G G GG GGGGG GG GG G G GGGGG GG G G GG G G G G G G G G GGG G G G G GG GG G GGGG G G G G GG GG G GGG G GG GG GG GG GGG GG GGG GGGG GGG GGGG GGG G GGGG G GG GG GGGGGG GG G G G GGGG GGG GG G GGG GGG G G GG G GG GG GG G G GGGG G G GGG G GGG GGG G G G G G G G GG G G G G G G G GGG GG GG GG GGGG G G G G G G GG GG G GGG G GG G GG GG G GGGG G GG GG G G GG GGGGGGGG GG GG G G GG GGGGG G G GG G GG G G G GG G GGG GGGG G G GG G G G G GG GG G G G GGGGG GGGGG GGGG GGGGGG GG GGGGG GGG GGGGG G GG G GG GGGG G G GG GGGGGG GGG G G G GG GGG G GG GG G GGGG GGGG G G GG G GGGG G G G G GG GG G G G G G G GGGG G G G GGGGGG GG G G G G GGGG G GG GGG G GG G G G GGG G GG G G GGG G GG G GGG G G G G G G G G GGG G G G G G G G G G G GGG G GGG GG G G GGG G GG G G GG G GG GG G GGGG G GG G G G G G GGG G G G G G GGG GG GG G G GGGG GG G G GG G GG G G G GG GG G G G G G G G G GGG GGG GG GG GGG GGGGG GGG GGG G GG G G G GG G GG G G G G GG G G GGGGG G GG G G GGGGG G G GG GG G GGG G G GG G GGG G GGGGGG GG GGGGGG GG GG G G G GGG GG G GGGGG GGGG G G G G G G GGGG GGGGG G GG GG G G GG G GGGGGGG G GGG G G GG G G GGG GGG G GG GGGGGGG G GGGGGGGGGG G G GG G GG GG G G GGGGG G GG GG G G G GG G G G G GGGG G G G GGG GGG G G G G G G G GGG GG GGG GG GG G GGG G GG GG G GGGG GGGGGGGGGGGGGG G GGGGG G GGGG GGGG G GG GGG G G GGG G GGG G G GG GGGG G GGGG G G G G GG G GG G G GG G GG GG GGGG G G GG G GG G G GG GGG G GG G G G GG G G G G G G G GG G G G G G G GG G G G GGG G G G G GG G G G GGGG G GGGG G G G GG GG G G G GG GG G GG G G G G G G G G G GG G G G G G GGG G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G G G GGGG G G G G G G G G GG G G G G G G G GG G G G G G GG G G GG G GGG GG G GGG G G G GG G G GGG G G G G GG GGG G GG G G G G G G G GG G G GG G GG G G G G G G GG G G G G G G GG GGG G G GGG G G G G G G G G G GGG G G G G G G G G G G G G G G G GG G GG G G GG G G G GG G GG G G G GG GG G G G G G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G GGGG GG G G G GG G G G GG G GG G GG G G G G GGGG G G G G G G G G GGG G G G G G G G G GG G G G G G G G G GG GGG GGG G G G G GG GGG G GG G G G G G G G G GG GG G G G G G GGG G G GG GG G GG GG G GGG GG GGGG G GGG GG G GG G G G GG G G GG GGG G G GG G G GG G G G GGG G G G GG GG GGG GG GG GG GG GG GG G GG GG GGG GGGG G GGG G G G G G G GGG G G G G GGGG GGGGG GGG G GG G GG G GG GG G G GG G GGG G GG G G G GGGG GG G G GG GG G G G G G G G GG GG G G G G G G GG G G GG G GGG G G GG GG G GG GG G G GG G G G G GG G G G G G G G GG G G G G GGG G G G G G G G G GG G G G G GG G G GG G GG G G G GG G G GG GGG G G G G G G G G GG G G G G G GG G G G G GG G GG G GGG G GG G G G GG GG GGG G G GGG GGG GGG G G GG G G GG G GGGG GGGG GG GGGG G GGG GG GG G GGGGGG G G G G G G GG GG GGGGG GG G GG G G G GGG G G G GG GGG G G G GGG G G G GG GGG GG G GG GG G G GG GG GGGG G GGGG G GG G G G GG GG G GG G G G GG G GG G G G GG G G GGG GGG GG GG GG G GG GGG G G G GG G G G G G G GGG G GGG G GGG GGGGG GGG GGGG G GG GG G G GGGG GGG G G G G G GG G G G GGG GGGG GG GG GG G GG G G G GGG G G G GG GGGG G G G GGGGGG GG GGGGG G GG G G G GG G G G G G GG G GG G G GG GGG G G G GG G G GGG GG G G G G G G G G G G GGGG GGG G G G GGG GG GGGG GG GGG GG G G G G G GG GG GG GG GG G G GG G G G GGG G G G GG G G G G G G GG G G G GGG GG GG G G GG GG GGGGG GG GGGG G GG G G GGGGG G G G G GGG GGGG G G G G G G G G G GG GG G G G GGG GG GG G GG G GGGG GG GG G GG GG G GGG GG GG G GG G GG GG GGGGGG GG GG G GG GGG G G G G G GG G GG G GG G G G G G G GG G GGGG G GG GGGG GGG G GG GGGGGG GG G GGGG G G GGG GG GGGG G G G G GGG G GGGGG G G GGGG GG G GG G GGGGGGGGG GGG GG G G G G G GGG GG G G G G G G GG GG GGG G G G G GG G GG G GG G GGG GG G GG GG GG GG G G GGG G G GGGG GG GG G GGGG G G G GG G G G G G GGG GG GG G G G G G GG GG G G GGGGG G GG GGGG G GGG G GGG G G G G GG GGG G G G G G G G GGG GG G G GG GG GGG GG GGG GG G GG GGGG G G G G GGGGG G G GG GG G G G GGG G GGG GG G GG GG G G GG G G G G G G G G GG G G G GG G GGG G G G G GG GG G GG G GG G GGG GG G G G G GGG G G G GG G G GG G G G GG G GG G GG G G G G GG G G G G G G GG GG G G GGG G GG G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G GG G G G G G G G G G GGG G GG GGG G G G G G G G G G G GG G G GG GG G GGG G G G G GG G G G G G G G G G GGGG G G GGG G G G G GGGG G G GG G GG GGGGG G GG G GG G G G GGG G G GGGG G GGG G G G GGGGG G G G GG G G G GG GG G G G GGG G G G G G GG G G G GG G G G G G G G GGG GG G GGGG G G GGGG G G GGGGG G G GGG G G G G GG G GG GG G G G G GGG G G G G G G G G G GGGG GG GG G G G GG G G G G GG G G G GGGGG G G G G G G G G GG G G G G G G G G G G GGG GG G GG G G G G GG G G G G G GG G G G G G GG G GG GGG G G GGGG G GGGGG G GG G G G G G GG G G G G G G G G G G GG G G G G GGGG G G GGGG G G G G G G GG G G G G GG G G G G G G G G G G G G GG GGGG G G GGGG G G G GG G G G G G GGG G GG GGG G GGGGG G GGGGG G GGG GG G G GGGG G GG G GGG GGGGG G G G G GG G GGGGG G GG G G G G GGGGG G GG G G G G G G GGG G COS [ppt] G G G G G G G G G G GGGG G GG G G G G G G G G G G GG G G GG G GG G G G G G G G G G G GG G G G GG GG G G G G G GGG G G G GG G G G G GGG G G G G G G GGG GG G G GGGGGG GG G G G G GG G GG G G GGG GGGG G G G G G GG G GGGG G GG G G GGG G G G G G G G GG GG G GGGG G G GG G GGG G G GG G G G G G G G G GG G GG G G GG G G GGG G GGGG G G G GG G GG G G G G GG G GG G GG G GG GG G G GGG GG GGGG G GGGG G GG GGG G GG G GGG G G GG G G GG GG G G GG GG G GGG G G G G GG G GGG G G GG G GG GG GG GG G GG G GGG GGG G GG GG G GG G G G G G G G G GG GG G G G G G G G G G GG G G G G G G G GG G G G G G GG G G GG G G GG GGGG G G G G GG G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G GG G GGG GGG GG GG GGG G G GG GGG GG G GG GGGG G G G G G GGGG GGGG GG G GG GGG G G GG G GG G GGGGG G G G G G GG G G G GGG G GGGG G G G G GG G GG G G G G GG GG G GGG GGG G GGG G GGGGGG GGGGG GGG G G GGGG GGGG GGGG G G GG GG G GGG GGG G GG G GGGGG GG GGG G G G GG G G GG GGGG GG GGGGGG GG G GGGG G G G G G GG G G G G G GG G GG GG GG GG G GGG G G G G G GGG GG GGGGG GGG G G GG G GG GG G GGG G G G GGG G GGGG GG GG GG G G GG G GGGG G G G GG GG G G G GGG GGG G GGG G G GG G G GG GG GG G G G GGGGG GGGGG GG GGGGGG G G GGGGGGGGGG GG GG GGG G G GGGGGGG G G G GGGGGG GGG G G G G GGGG G GG GG GGGGGGGG G G G GG G G GGGG G G G GGG G G G G G G G GGGG G G G G G GG G G G G G G G G G G GG G GG GGG G G G G G G GG G G GG G G GGG G GG G G G G G G G G G G G G GG G G G G G G G G GG G G GG G GGG GG G G GGG G GG G G G G G GG GG G GGGG G G G GG G G G GGG G G G G G GGG GG G G G G G G GG GG G G GGG GG G G G G G GG G G GG G G G G GG G GG G G G GG GGG GGGGG GGGG G GG GG G G G G G G GG G G G GGGG G GG GGG G GG GGGGGGG G G GG GG G GGG G G GG G GGG G G GGGGG G GGGGGG G GG GG G G G GG G GG GGGGGG GGGG G G G G G G G GGG GGGGG G GG GG G G GG GGGG GGG G G G GG G G GG G G GGG GGG G G G GGGGGGG G GG GGGGGGGGGGG G G GG GG G G GGG GG G G G GG G G G GG GG G G G GG G G G G GGG GGG G G G G G G G GG G GG G GG GG G G G GGG G GG G G G GGG G GGGGGGGGGGGG GG G GGG G G G G GGG GGGG G GG GGG G G GGG GGGG G G G G GGGGG GGGG G G G G GG G G GG G G G G G G GGG GG GG G GG G GGGG G G GGGGGG G G GGG G GGGG G G G G GG G G G G G G GG G GG G G G G G G G G G GG G G G G G GG G GG G GG G G G G G GG GG G G G G G G G G GG G GGG G G G G GGGG GG G G G G G G G G G GGG GG G G G G G G G G GGG G G GG G GG GG G G G G G G G G G GG G G G G G G G G GG G G G G G G G G G GG G G GGGG G G G G G GG GG G G G G G G GGG GG G G GG G G G G G G G GGGG G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G GG GG G G GG G G G GG G GGG GG G G G GG G G G G G G G G G G G G G G G G G G G G GG G GGGGG G G G GG G G G G G G G G G G GGG G G G G G G G GG G G G G G GG G GGG GG G G G G G G G G G G G G GG G G G G G G G G GGG G GG G G G GG G G G G G G G GG G G GG GG G G GG G GG GG GGG G G GG GGGG G G GG GG G G G G G G G GG G G G G GG G GG G G G G G GG GGG G GG G G G G G G G G GGG G G G GG GGGG G GG GG G GG G G GGGGG G GG GG G GGG G G G G G G GG GG G G G G GGG G G GG G G GG G G GG GG GG G G G GG G GG G GG GG G GGG G G G G GGGGGGG GGGGGGGG GG G G G GGG G G G G G GGGG GG GGG GG G G GG G GG G GG GG G G GG G GGG G GG G G G GGGGG G G G GG G G G G G G G G G GG GGGGG GG G GG G G GG G GGG GG G GGG G GG GG G G G G G G GG GGG GG G G G G G G G GG G GG G GG G G G G G G GG G G G G G G G G GG G G G G G G GG GGG G G GGG G G G G GG G GG G GG GG GG G G G G G G G GG G GGG G G G G GG G G G G G G G G G GG G GG GG GG G GG G GG GG G GGGGGG GGGG GG G GG GGG G GG G G G G GGG G G G G G G G GG GG G GGG G G G GGGG G G GGG G GG G G GGGG G G G G G G G G GGG GGGG G GG GG GGGG G G GGGG G GG G G G GGG GG G G GGG G GG G G GG G GG G GG G G G G GGG GG GG GGG GGG GG G GG G G G G G G G GG G G GGG G G G GG GGG GGGGG GGG GGGG G GGGG GG GG G GGGG G G G G GGG GG GG G GGGGG GG GG GGGGGG GGGGG G G GG G GG GG G GG GGGGGG GG GGGGG G GG GG G GG G G G G G GG GGG G G GG GGG G G GG G GG G GG G GG G G G GGG GG G GGGG G GG G G G GGG GGGGGGG G G GGG GG GG G GG GG G G GG GG GG G G GG G G G G G G G G G GG G G G G G G GG G GGGGGGGGGGG GGGG GG G G G GGGG GGG GG G G G GG GG GGG G GGGG GG G GG G GG G GG GG GG G G G G GGGGGG G G GG GGG G G G G G GG GG G G G G G G GG GGG G G GGG GG GGGGGG GG GGG G G G G GG G G GGGG GG GG GGG GG GG G GG G GGGGG GG GGGGGGG G G G G G GG GG GG GG G GG G G G GG GG GGG GGGG G GG G G G GGG G G G GGGG GG G GG GGGGG GGGGG GGG GG G GGG G G GG G G G G G GG G GG GG G GG G GGG GG G GG G GG G GGG GG G G G GG GG GG G G GGG G GG GGGG G GG G GGGG G G G G G GGG G G G GG GG G G G G G G G GG GG G G GGGG G G GGGGGG GGGGGGGG GG G G G G G G GG G G G G G GG GGGG GG G G GG G GG GG GGGGG G GG GGGG G G G GGGGGGGGGG G GG GG GG GGGG G GG GG GGGGGG G GG GG GG GGGG G GGGGGGGGGG G GG GG GG GG GG G GG GG GG GG GG G GGGGGG GGGG G GGGGGG GG GG G GG GG GG GGGG G GGG G GG GGGG G GGGG GGGGGG G GG GGGG GG G G G GG GGGG GGGG G GG GGGG GG GG G GGGG G G GG GG G GG G G GG GG G GG GG GG GG GGG GG GG GG GG GG G GG GG GGGGGGG GG GG GGGG GG G GG GGGG GG GG G GG GG GG GG GGG G G GG GG GG GG G GG GG GGGGGG G G G G G GG GG GG G GG GG GG GG GG G G G GGGG GG GGG GG GGG G GG GG G GG GG GG GG GG G GG GG GG GG GG G GG GG GGGG G G GGG GG G GGG GG G GG GG GGGG GG G GG GG GG GG GG G GG GGGGGGGG G G G G G G GGG GGG GG GG GG GGGG G GG GGGG GGGG G GG GGGGGGGGG GG GGGG GG GG G GG GG GG GGGG G GG GG GGG GGGG GG GG GG GGGG G GG GGGGG GGG G GG GG GG GG GGG GG GG GGGGGG G GG GG GG GGGGG GG GG GG GG GG GGG GG GG GG GG G GG GG GG G G GG GGG GG GG GG GG G GG GG GGGGGG G GG GGGGGGGG G GG GG GGGGGGG GG GG GGGGGGG GG GG GG GG COS [ppt] 350 450 550 650 GGGGGGG G GGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGG G GGGGG GGGGG GGGG G GGGG G GGGG GGGGGGGGGG G GGGGG GGGGGGGGGG GGGG G GGGG GGGGGGGGGGG GGGGG GGGG G GGGGGGGGGG GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGGG GGGGG GGGG GGGGGG GGGGGGGGG G GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG GGGGG GGGGGGGGGGGGGGGG GGGG GGGGG GGGGGG GGGGG GGGG G GGGG G GGGG G GGGG GGGGG G GGGGG G GGGG GGGGG GGGGGG G G GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GGGGG GGGGGGGGGG GGGGGGGGGGG GGGG GGGGGG GGGG G GGGG GGGGGG G GGGG G GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGG GGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGG GGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GG G G GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG CO2 [ppm] GGGGG G G GGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG G GG GGGGGGGGGGG G GGGGG GGGG G GGGG G GGGG G GGGG G GGGGG GGGG G GGGG G GGGGGGGGG G GGGG G GGGG G GGGG GGGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGGG GGGG G GGGGG GGGGG GGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGG GGGGGGGGGG G GGGGGG G GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG G GGGG G GGGGGGGGGG GGGG G GGGG G GGGG G GGGGG GGGG G GGGG G GGGG G GGGG G GGGG G GGGG G GGGGG G GGGG GGGGG GGGGGG G G GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GGGG G GGGGG GGGG GGGGGG GGGG G GGGG G GGGGG GGGG G GGGG GGGGGG G GGGG G GGGG GGGGG G GGGG G GGGGG GGGGG GGGGG GGGG GGGGGG G GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGG GG GGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG CO2 [ppm] GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGG G GGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG CO2 [ppm] 370 390 410 GGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGG G GG G GGGGGGGGGGGGG G GG G GGG GG GGGGGGGG GGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGG G GGGGGGGGGGGGG GGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GGGGGGGGG G GGGGGGG GGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGG GGGGG GGGGG GGGGG GGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGGGGGGG G G G GG GGG G G GGGG G GG G G GG GGG GG GGG GGGGGGGG GGGGG GG GGGGGGGG GGGGG GG GGGG GGGGGGGGGGGGGGG GGGGGG GGG GG G GGGG GGGGG GG G GGGG GGGGGGG GGG G G GGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGG GGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGG GGGGGGGGGG GGGGGGGGG G GGGGGGGGGGGGGGG GGGGG GGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GG G GGGGGGGGG GGG GG GGGGGGGGGG GGGGGGGGGGGGGG GGGGGG GGGGGGG G GGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGG GG GGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGG GG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGG GGGGGGGGGGGGG GGGGG GG GGGGG G GGGGGGGG GG G G GGGGG GGGG G GGGGG G GGGGGGGGGGGGGG GGGG GGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGG G GGGGGGGGGGGGGGGGGGGGGG G G GGGGGGGGGGGGGG GGGGGGGGGGGGG GGGGGG GGGGGGG GGGGG GG GGG GGGGGGGGG GG GGGG GGGGGGGGGGGGG GGGGG G GGGGG GGGGGGGGGGGGGGGGGG GGG GGG GGGGGGGGGGGGG GG G GGGGGGGG GGGGGGGGGGGGGGGGGG GGGGGG GGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGG GGGGG GGGGGG GGGGGG GGGGGG GGG GGGGG GGGGGGGGGGGG GGGGGG GGGGGGGGGG GGGG GGGGGG GG GGGGGGGGGGGGG G GGGGG G GGG G GGGGG GG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGG GGG GGGGGG GGGGGGGGGGGGG GGGGGGGG GGGGGGGGGGG G GGGGGGGGGGGGGGG GGG GGGGGGGGGGGG GGGGGGGGGGGG G G G GGGGGGGGGGGG GGGGGG GGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGG GG GGGGG GGGGG GGGG GGGGGGGG GGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GG GGGGGG GGGG GGGGGGGG GGGGGG G GGGGG G CO [ppb] GGGGG G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGG GGGGGGGGGG G GGGG GGGGGGGG GG G GGGG GGGGG GGGGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGG GGGGGGGGGGGGGGGGG G GG GGGGG GGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGG GGGG G G GGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GGGGGGGGGG GGGGGGG GGGGGGGGGGGGGGGGGGGGG GG GGGGGG GGGGGGGGG GGGGGGGGGG GGGGG GGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGG G G G GGGGG G G GG GG GGGG G GG G GG GGGGG G GGGGG G G GGGGG GG G GG GGGGG GGGG G GGGGGG GGGG GGGGGGG GGG G GGGGGG GGGGG GGGGG G GGG GGG G GGGGG GG G GG G GGG G G GGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGGG GGG GGGGGGGGGG GGGGGGGGG G GGGGGGGGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG G GGGGGGGG GGGGGGGG GGGGGGGGGGGG G GGGGGGGGGGGG GGG GGGGG GGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGG GGGG G G GGGGGGGGGG GGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGG GGGGGGG GGGGGGGGGGGGG GGG G G GGGGG GGGGG GGGGG G GGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGG G GGGGGGGGGG GGGGGGGGGGGG GG GGGGGGGG GGGGGGGGG GG GGGGGGGGGGG GGGGGGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGG GGGG GGGGGGGGGGG GGGGGGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGG GGGGGGG GGGGGG GGG GGGG GGGGGGG GG GGGGGGGGGGG GG GGGGGGGGGGGGGGGGGG GGGG GG GGG G G GG GGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGG GGGG GG GGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGG GGG GGGGGGGGGG G GG GGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGG GGGG GGGGGGGGGGG GGGGGGGGGGG GG GGGGGGGGG GGGG GGGGGGGGGGGGGGGGGG GG GG GGGGGGGGGGG GGGGGGGGGGGGGGGGG G GGGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG GGGGG GG GGGG CO [ppb] G GGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGG GG G GGGG GGGGG GG GGGGGGGGGGGG GGGGGGGGGGGGGG GGGGGGGGG G GGGGGG GGGGGG G GGGGGGGGGGG GGG GGG GGGGGGGGGGGGGGGG GGGG GGGG GGGGGGGGGGGGG GGGGG GGGGGG GGGG GGG GGGG GGGGGGGGGG GGGGGGG GGGGGGGGGGGG GG GGGGGGGGGGGGG G GGGG GGGGGG GGGGGG GGGGGG G GG GGGGGGGGGGG GGGGGG GGGGGG GGGGGGGGGG GGGG GGGGGG GGG G GGGGGGGGGGG GGGGGG G GGG G GGGGGGG G GGGGGGGGGGGGG GGG GGGGGGGGGGGGGGGG GG GGG GGGGGG GGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGG G GGG GGG GGGGGGGGGGGG GGGGGGGGGGGG G G GGGGGGGGGGGG GGGGGG GGGGGG GGG GGGGGGGGGGG GGGGGGGGGGGG G GGGGGG GG GGG G GGGGGGGGGGG GGGGGGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGGGG GG GGGG GGGG GGGGGGGG GGGGGG G GGGGG G CO [ppb] 100 150 200 250 G G G Assigned cylinder values: no. Atmos. Meas. Tech., 9, 5293–5314, 2016 www.atmos-meas-tech.net/9/5293/2016/ 3.1 Precision and accuracy The data show that for COS and CO2 the period after improving the temperature stability has an improved mean standard deviation compared to the period before the temperature improvement (from 6.6 to 4.8 ppt for COS and from 0.14 to 0.06 ppm for CO2). However, as there were also moments with better precision in August and Septem- ber 2014, there is no consistent relation between temperature stability and instrument precision. What we do see is that the instrument precision changed every time that the mirror alignment was changed, indicated by blue vertical lines. Dur- ing periods that the mirror alignment was not changed (Jan- uary and April 2015), the instrument precision was stable. These results show that the instrument precision can largely be affected by the mirror alignment. ers the reference and target gases. In October 2014 the span between the two cylinder measurements changed, which is again mostly visible in CO2. The reason for this change is that the regulator pressure of one of the two cylinders was slightly changed, which affected the amount of dilution of nitrogen into the sample line. Although it is known that COS mole fractions can drift in cylinders over time, we did not ﬁnd indications that the mole fractions drifted within the mea- surement period. Later in this section we discuss the cylinder drift in more detail. Figure 6 shows the 1-second standard deviations of the hourly reference gas measurements between August 2014 and April 2015. It is clear from Fig. 6 that the instrument precision cannot be captured with one single value due to its variation. For the period marked in red the room temperature was characterized by rapid changes caused by an air condi- tioner. This was a period in which the instrument precision was adversely affected, indicating that temperature stability inﬂuences the instrument precision. In the right plot of Fig. 6 the histograms of the standard deviations are shown for the Allan deviation plots are an effective way to show how far the random noise level can be reduced by averaging and at what timescale the drift effect starts (Allan, 1987). How- ever, we found that Allan deviation plots of the measure- www.atmos-meas-tech.net/9/5293/2016/ L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. [ppb] 0 1 2 3 Aug'14 Sep'14 Oct'14 Nov'14 Dec'14 Jan'15 Feb'15 Mar'15 Apr'15 May'15 Figure 6. Left: standard deviation over 2 min of hourly measurements of one of the working standards as measured from August 2014 until April 2015. The gray (black) colored data indicate the period after (before) modiﬁcations were made to improve the temperature stability (such that Telectr varied with 0.21 ◦C with every degree of changing Troom, against 0.91 ◦C ◦C−1 before modiﬁcations; see Sect. 2.5). For the period that is marked with a red background the room temperature was characterized by rapid changes caused by an air conditioner. Blue vertical (dashed) lines indicate moments where the mirror alignment was changed. Right: histogram of standard deviations as shown in the left ﬁgure with the black/gray colors corresponding to the colors in the left ﬁgure. Note that the dark histogram is transparent. Also included are the mean standard deviations as well as an overview of percentiles because the data do not show a Gaussian distribution. individual uncertainties. The overall uncertainty that is pre- sented in Table 5 characterizes the typical measurement un- certainty but may vary depending on the conditions. For ex- ample, the measurement precision varies, as we discussed in the previous paragraph, and the water vapor correction uncer- tainty scales with the COS, CO2 and CO concentrations. In Table 5 we present 1-minute precisions that depend on the frequency of reference measurements (see Sect. 2.2.3 and Table 3). We adopted the 1-minute precision that is based on reference measurements made every 30 min, which rep- resents the calibration strategy of our ﬁeld measurements. Furthermore, precision depends on the length of the averag- ing period, where it can either decrease for Gaussian noise or increase for instrument drift. In Fig. 6 we can see that the 1-second precision of working standard measurements in March–April is 4.3 ppt for COS (50 % percentile of the mea- surements). The precision of 2-minute averaged target mea- surements in Fig. 8 (discussed later in this section) varies ments from the instrument is not consistent over time. As an example, we show two Allan deviation plots in Fig. 7. The data used in the two Allan deviation plots were retrieved from cylinder air measurements during two laboratory ex- periments under similar conditions on the same day. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5304 j g p , , COS std. dev. [ppt] Relative frequency 0 5 10 15 0.00 0.02 0.04 0.06 0.08 0.10 Percentiles: 5 20 50 80 95 Mean: Aug−Feb 3.1 3.8 5.2 8.1 13.2 6.6 ± 4.7 Mar−Apr 2.9 3.4 4.3 5.9 8.6 4.8 ± 4.7 CO2 std. dev. [ppm] Relative frequency 0.00 0.05 0.10 0.15 0.20 0.00 0.02 0.04 0.06 0.08 0.10 Percentiles: 5 20 50 80 95 Mean: Aug−Feb 0.03 0.04 0.06 0.14 0.24 0.14 ± 2.43 Mar−Apr 0.03 0.03 0.04 0.06 0.08 0.06 ± 0.13 CO std. dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppb] Relative frequency 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.00 0.02 0.04 0.06 0.08 0.10 Percentiles: 5 20 50 80 95 Mean: Aug−Feb 0.2 0.3 0.5 0.9 1.3 0.7 ± 2.2 Mar−Apr 0.3 0.5 0.9 1.3 1.7 0.9 ± 0.4 G G G G G G G G G G G G GGG G G GGG G GG GGG G G G G G G G G GG G G G G G GGGG G G G G G G G GGG G G G G G G GG G GG G G GG G G G G G G G G G GGGG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G GG GG G G G G G GG G GG G G G G G G G G G G G G G G G G G G G G G GG G GGG G G G G G G G GG G G G G G G G GG G GG G G G G G G GG G G G G G G G G G GG G GG G G G G GG G G G GG G GG G G G G G GG GG G GG G G G G GG G G G G G G G G G G G G G GGGG G G GG GGGG G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GGG G GG G GG G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G GGG G GG G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G GG G GG GGGGG GGG G G G G G G G G G G G G G G G G G GG G G G GG GG G G G G GGG G G GG G G G G GGG G G GG GG G G G G GG G G G G GG G G G G G GGGG G G G G G G G G G G GG G G G G G G GG G GG G G G GG G G G G G G G G GGG G G GG G GG G GGG G GG GG G G G G G G G G G G G G G G G G GGGG G G G G G G GG GGGG GGGG GGGGGGG GG GG G GG G GG G GG GG GGG G GGGG G G G G G GG G G G G G GGG G GG G G G G G GG G G G GGGGG G G G GG G GG G G G G G GGGG GGG GGG GG G G GG GG G GGGG G G GGGGG GG G G G G GG G GG G G G G GGGGG G GG GGG G GGG GGG GG G G GG GG G G G G G GGGGG G GG G G G G GGGGGG GGGGGGGG G GGG G G GGG GG GG G G GGGG G G G G GG G G G G G G GG G G GG G G GGGG G G G G G G G G G G G GG G G G G G GG G G G G G G GG G GG GG GG G G GGG G G G G GG GG G G G GGGGGGG G GG G G G G G G G GG GG G GGGGG G GG G G G G G GG GG GGG GGGG G G G GG GGGG G G G G GGG GG G G G G GG G G G G GG G G G G G G GGG G G G GG G GGG GG G G GG GGG G GGG G G G GG G GG G GGGG G G G G G GG G GGGG GGG G GG G G G GG G G G G G G G G G G G G G G G G G G GGGG G G G GG G G GG GGGG G GGG GG GGGG GGG G G GGG GG G G G G G G GGG G G G GG G GG GGGG G G G G G G GGG GGGG G G G G G G G G G G GG G G G G GGGG G G G G G G GGGG GG GGG G G G G G G GG G GG G G G G G G G G GGGG G G G G G G G G G GG G G G G G GG GG G G G G G G G GG GGG G GG G GG G G G G G GGG GG G GG GG G GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G GGGG G G G GGG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GG G G G G G G G G G G G G GG GG G G G G G G GG G G G G G G G G G GG G GG GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G GG G G G G G G G G GG G GG G G G G G G G G G G GGG G G G G G GG G G G G G GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G GG G GG G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G GG G G GGG G G G G G G GG G G G G GG G G G G GG G G GG G G G G G G G G G G G G G G G G G G G GGG G G G GG G G G G G G G G G G G GG G G G GG G G G G G G G GGG GG G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G G G GG G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G GG GG G G G G G GGG G GG GGG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G GG G G G GGG GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GGG G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G GG GG G G G G G GG G G G G G G G G G G G G G G G GGGG G G G G GG G G GG G GG G G G G G G G G G G G G G G G GG G G G GG G GG G G G G G G G G G G G G G G G GG G G G GG G G G G G GGGG G G GGG G G G GG G G GG G GGG GG G G G G GG G G G G G G G GG G G G G G G GG G G GG G G G G G G G G GGG G G G G G G G G G G G GG GG G G G G G G G G GG GG G GG G G G G G G G G G G GGG G G G G G G G GG G G G G G G GG GGG G G G G GGG G G GG G GG G G G G G G G G G GG G G G G GG G GG GG G G G G G GG G G G G G G G G G G G G G GG G GG G G G G G G G G G G G GG G G G GG G G G G G GGG G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G GG G G G G G G G GG GG G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G GGG G G G G GGG G G G G G G G G G G GG G G G G G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G GG G G GG G GG GGGG G GG GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G G GGG G G G G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G COS std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppt] G G G G G G G G G G G G GGG G G GGG G GG GGG G G G G G G G G GG G G G G G GGGG G G G G G G G GGG G G G G G G GG G GG G G GG G G G G G G G G G GGGG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G GG GG G G G G G GG G GG G G G G G G G G G G G G G G G G G G G G G GG G GGG G G G G G G G GG G G G G G G G GG G GG G G G G G G GG G G G G G G G G G GG G GG G G G G GG G G G GG G GG G G G G G GG GG G GG G G G G GG G G G G G G G G G G G G G GGGG G G GG GGGG G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GGG G GG G GG G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G GGG G GG G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G GG G GG GGGGG GGG G G G G G G G G G G G G G G G G G GG G G G GG GG G G G G GGG G G GG G G G G GGG G G GG GG G G G G GG G G G G GG G G G G G GGGG G G G G G G G G G G GG G G G G G G GG G GG G G G GG G G G G G G G G GGG G G GG G GG G GGG G GG GG G G G G G G G G G G G G G G G G GGGG G G G G G G GG GGGG GGGG GGGGGGG GG GG G GG G GG G GG GG GGG G GGGG G G G G G GG G G G G G GGG G GG G G G G G GG G G G GGGGG G G G GG G GG G G G G G GGGG GGG GGG GG G G GG GG G GGGG G G GGGGG GG G G G G GG G GG G G G G GGGGG G GG GGG G GGG GGG GG G G GG GG G G G G G GGGGG G GG G G G G GGGGGG GGGGGGGG G GGG G G GGG GG GG G G GGGG G G G G GG G G G G G G GG G G GG G G GGGG G G G G G G G G G G G GG G G G G G GG G G G G G G GG G GG GG GG G G GGG G G G G GG GG G G G GGGGGGG G GG G G G G G G G GG GG G GGGGG G GG G G G G G GG GG GGG GGGG G G G GG GGGG G G G G GGG GG G G G G GG G G G G GG G G G G G G GGG G G G GG G GGG GG G G GG GGG G GGG G G G GG G GG G GGGG G G G G G GG G GGGG GGG G GG G G G GG G G G G G G G G G G G G G G G G G G GGGG G G G GG G G GG GGGG G GGG GG GGGG GGG G G GGG GG G G G G G G GGG G G G GG G GG GGGG G G G G G G GGG GGGG G G G G G G G G G G GG G G G G GGGG G G G G G G GGGG GG GGG G G G G G G GG G GG G G G G G G G G GGGG G G G G G G G G G GG G G G G G GG GG G G G G G G G GG GGG G GG G GG G G G G G GGG GG G GG GG G GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G GGGG G G G GGG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GG G G G G G G G G G G G G GG GG G G G G G G GG G G G G G G G G G GG G GG GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G GG G G G G G G G G GG G GG G G G G G G G G G G GGG G G G G G GG G G G G G GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G GG G GG G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G GG G G GGG G G G G G G GG G G G G GG G G G G GG G G GG G G G G G G G G G G G G G G G G G G G GGG G G G GG G G G G G G G G G G G GG G G G GG G G G G G G G GGG GG G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G G G GG G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G GG GG G G G G G GGG G GG GGG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G GG G G G GGG GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GGG G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G GG GG G G G G G GG G G G G G G G G G G G G G G G GGGG G G G G GG G G GG G GG G G G G G G G G G G G G G G G GG G G G GG G GG G G G G G G G G G G G G G G G GG G G G GG G G G G G GGGG G G GGG G G G GG G G GG G GGG GG G G G G GG G G G G G G G GG G G G G G G GG G G GG G G G G G G G G GGG G G G G G G G G G G G GG GG G G G G G G G G GG GG G GG G G G G G G G G G G GGG G G G G G G G GG G G G G G G GG GGG G G G G GGG G G GG G GG G G G G G G G G G GG G G G G GG G GG GG G G G G G GG G G G G G G G G G G G G G GG G GG G G G G G G G G G G G GG G G G GG G G G G G GGG G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G GG G G G G G G G GG GG G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G GGG G G G G GGG G G G G G G G G G G GG G G G G G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G GG G G GG G GG GGGG G GG GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G G GGG G G G G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G COS std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppt] G G G GG G G G GGG G G G G G G GG G G G GGG G G GG GG GG G GGG G G GG G G G G G G GG GG GG G G GG G G G G GG G GG G G G G G G G G GG G G G G G GGG G G G G G GGG G G G GG G G G G G GG GGG G G G G G G G GG GG G G GGGGG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G G G GG G G GG G G G G G G G G G G G G G G GGG G G G GGG G GG G G G G G G G G G G G G G G GG GG G G G G G GG GG GG G G G G G G G G G G G GG G GG G GG G G G G G G G GG GG G GGG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G GGG G G GG GG G G G G G G G G G G G GG G G GG GGGG G G G G G G G G G G GG G G G G G G G G GG G G GG G G G G G G G G GG G GG G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GG G GG G G G G G G G G G G G G G GGG GGGG G G G GG G GGGG G G G G G G G G G G G G G G G G GG G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G GGGGG GG G G G G G G GGG G G G G G G GGGG G G G G GG G G GG G G G G G G G G G G G G GG G GG G G GGG G G GG G G G GGGGG COS std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppt] 0 5 10 15 20 25 30 G G GGG G G G G GG G G G GG G G GG GG GG GG G G G G G G G G G G GG G G GG G G G G G G G GG G G G G GGGG G G G G G G G GGGG G G GG G G G G G G G G GGG G G GG GG G G G G GG G GG GGG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G GG G G G G G G G G G G G GG G GGG G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G GG G G G G G G G G GG GG G G G G G G G G GG G G G GGG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G GG G GG G G G GG G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GGG GGG G G G G G G G G G G GG G G G G G G G G G GG G GG G G GG G G G G G GGG G G G G G GGG G GGG G GG G G G GGGG G G G G G G GG G G G GGG G GG G G G G GGG GGG GGGG G G G G GGGGG G G G G G G G GGG G G GGG G G GGG G G GG G GG G GGG G G G GG GGG G G G G G G G GG G G GG G G G GG GGG G G G G G G GG GG G GGG GGG G G G G G G G G G G G G G G G G G GG G G G GG G G GG G G G G G G G G GGG G G G G GGG G G G GG G G G GG G G GGG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G GG G G GG G G G GG G G GG G GG G G G G G GG G G G G G G GGG G G G G G G GG G G G G G G G G G G G G GG G G GG GG G G GG G G GGG GGG G GG G G GG G G G G GGG G G G GG G G GG G G G G G G GG G G G G G G G G G GG G G G G GG G G GG G GG G GG G G G GGGGG G G G GG G G G G G G GGG G G G G G GG GGG G GG G GG GGGG GG G G GG G G G G G G G G G G G G G GG G G G G G GG G G GG G G G GGGGG G G GGGG G GG G G G G G G G G G G G G G GGG G GGGGG GG GG G GG G G G G GGG G GG GG GG GGGG G GGG GGG G GGG GGGG G G G G G G GGGGGG GG G G GG G GGG GG G GG GG G GGGG G G G G G G G G G G G G G G G G G GG GGGG GG G G G GG G GG G G G G G G GG GG GG GGGGGGGGGG G G G GG GGGG G G GG G G G GG G G G GG G GGGGGGG GGG GGGGG G GGG G GGGGG GGG G G G G G GG G G G G G G GG G GGG G GG GG G G GG G G G G GG G GG GGG G G G GGGG G G G G G GG G G G GGG G G G GG GG G G GGGG G G GG G G G G G GGGGGG G G G G G GG G G G G G G GG GG GG G G GG G G G G G G G G G G GG GG G G G G G G G G GGG G G GG GG G G G GGGGGG G GGG G G G G G G G G G GGG G G G G G G G GG G G G G G GGG G G G G GGG G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GGG G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G GG GG G G G G G G G G G GG G GG G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G GG G GG G G G G GG G G G GG G G GG G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G GG GG GG G G G G GG G G GG GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G G GG G G G G G GGG G G G G G GG G G G G G G G G GG G G GG G G G GG GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G GG GGG G G G G G G GG G G G GG G G G G G G G G G GG G GG G G GG G G G G G G G G G G G GG G G G G GG G G G G G G G G G G GG G G GG G G G G G G G G GG G G GGG GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G GG G GGG G G G G G G G G G GG G G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G G G GG G G G G G GG G G GG G G G G G G GG G GG G G G G G G GG G G G G GGG G G G G GG G G G G GG GG G G G G GG G G G G G G G G G G GG G G G G G G GG G G GG G G G G G G G G G G G GG G G G G G G GG G G G G GG G G GG GG GG GG G G G G GG GGG G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GG GG G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GGG G G G G G GG G G G GG G G GGG G G G G G G G G G G G G G G G G GG G GGGG G G GG G G G G G G G G G G G G GGG G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G GG G G GGG GGG G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G GG G G GG G G G GG G G G GG G G G G GG G G G G G GG G GG G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G GG G GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G GG G GG GG G G G G G G G G G G G G GG G G GG G G GGGGGG G G G G G G G G GG G G G G G G GG G G G G G G G G G GG G G G G G G G G G G GG G GG G G GG G GG G G G G G G G G G GG G G G G CO2 std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppm] G G GGG G G G G GG G G G GG G G GG GG GG GG G G G G G G G G G G GG G G GG G G G G G G G GG G G G G GGGG G G G G G G G GGGG G G GG G G G G G G G G GGG G G GG GG G G G G GG G GG GGG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G GG G G G G G G G G G G G GG G GGG G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G GG G G G G G G G G GG GG G G G G G G G G GG G G G GGG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G GG G GG G G G GG G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GGG GGG G G G G G G G G G G GG G G G G G G G G G GG G GG G G GG G G G G G GGG G G G G G GGG G GGG G GG G G G GGGG G G G G G G GG G G G GGG G GG G G G G GGG GGG GGGG G G G G GGGGG G G G G G G G GGG G G GGG G G GGG G G GG G GG G GGG G G G GG GGG G G G G G G G GG G G GG G G G GG GGG G G G G G G GG GG G GGG GGG G G G G G G G G G G G G G G G G G GG G G G GG G G GG G G G G G G G G GGG G G G G GGG G G G GG G G G GG G G GGG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G GG G G GG G G G GG G G GG G GG G G G G G GG G G G G G G GGG G G G G G G GG G G G G G G G G G G G G GG G G GG GG G G GG G G GGG GGG G GG G G GG G G G G GGG G G G GG G G GG G G G G G G GG G G G G G G G G G GG G G G G GG G G GG G GG G GG G G G GGGGG G G G GG G G G G G G GGG G G G G G GG GGG G GG G GG GGGG GG G G GG G G G G G G G G G G G G G GG G G G G G GG G G GG G G G GGGGG G G GGGG G GG G G G G G G G G G G G G G GGG G GGGGG GG GG G GG G G G G GGG G GG GG GG GGGG G GGG GGG G GGG GGGG G G G G G G GGGGGG GG G G GG G GGG GG G GG GG G GGGG G G G G G G G G G G G G G G G G G GG GGGG GG G G G GG G GG G G G G G G GG GG GG GGGGGGGGGG G G G GG GGGG G G GG G G G GG G G G GG G GGGGGGG GGG GGGGG G GGG G GGGGG GGG G G G G G GG G G G G G G GG G GGG G GG GG G G GG G G G G GG G GG GGG G G G GGGG G G G G G GG G G G GGG G G G GG GG G G GGGG G G GG G G G G G GGGGGG G G G G G GG G G G G G G GG GG GG G G GG G G G G G G G G G G GG GG G G G G G G G G GGG G G GG GG G G G GGGGGG G GGG G G G G G G G G G GGG G G G G G G G GG G G G G G GGG G G G G GGG G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GGG G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G GG GG G G G G G G G G G GG G GG G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G GG G GG G G G G GG G G G GG G G GG G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G GG GG GG G G G G GG G G GG GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G G GG G G G G G GGG G G G G G GG G G G G G G G G GG G G GG G G G GG GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G GG GGG G G G G G G GG G G G GG G G G G G G G G G GG G GG G G GG G G G G G G G G G G G GG G G G G GG G G G G G G G G G G GG G G GG G G G G G G G G GG G G GGG GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G GG G GGG G G G G G G G G G GG G G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G G G GG G G G G G GG G G GG G G G G G G GG G GG G G G G G G GG G G G G GGG G G G G GG G G G G GG GG G G G G GG G G G G G G G G G G GG G G G G G G GG G G GG G G G G G G G G G G G GG G G G G G G GG G G G G GG G G GG GG GG GG G G G G GG GGG G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GG GG G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GGG G G G G G GG G G G GG G G GGG G G G G G G G G G G G G G G G G GG G GGGG G G GG G G G G G G G G G G G G GGG G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G GG G G GGG GGG G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G GG G G GG G G G GG G G G GG G G G G GG G G G G G GG G GG G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G GG G GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G GG G GG GG G G G G G G G G G G G G GG G G GG G G GGGGGG G G G G G G G G GG G G G G G G GG G G G G G G G G G GG G G G G G G G G G G GG G GG G G GG G GG G G G G G G G G G GG G G G G CO2 std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppm] GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GGGG G GG G G G GG GG G GG G G G G G G G G G GG G G G G G G G G G G G GG GG G GGGG G G G GGG GG G G G G G G G GG G GG G G G G GG GG GG G G GG G G G G G GG G GG G G GG G G G GG GGGG G G GG G G G G G G G GG GG G G G GG G G G G GG G G G GG G G G G G G G G G G GG GG G GG G GGG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG GG G G GG GG G G GGGG G G G G G G G G G G G G G G G G GG GGG GG GG G GGG G GG G GG G G GG G G GGGG G G G G GG G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG GG G G G G G G G G G G G GG GG G G G G GG GG GG G G G G G G G G G G G GG G G G G G G G GG GG G G G G G G G G G GG GG G GG G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G GG G G G G G G GG G G GG G G G G G G GG G G G G GG G G G G G G G G G G G G G G GG G G G G GG G G GG G G G G G G GG G G G G G G G GG G G G GG G GGG GG G G G G G G GG G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG CO2 std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppm] 0.0 0.1 0.2 0.3 G G G G G G G G G G G G G G G G G GG G G G GGG G G GG G GG G G G G G G GG G G G G G G G G G G G G G G G G G G GGG G G G G G G G G G GG G G G G G G GG G GG G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G G GG GG GG G GG G G G GG G G G G G G G G G G GG G G G GG G G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G GG GG G G GG G G G GG G G G GGGG G G G G G G G GGGGGGGG G G G G GG GG G GG GG G G G GG G GGG GG G G G G G G G GG G GGG G G G G GGGGGGGGG G GG G G G G GGG G G G GG G G GGG GG G G G GG G G G G G G G G G G G GG G GGG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G GG G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G GG G G G G GGGG G G G G GGG G G G G G GG G G GG GG G G G G G G G G G G GGGG G G G G GG G G GGG G G G GG G GG G G G G G G G G G GG G GG G GG G G G GG G G G G G G G G G G G G G GG G G GG GG GG G G G G G G G G G G GGG GG G G GGG G G G G GG G G G G G G G G G GG GG GG G G GG GGGGG G GGG GG G G GGGGG GG GG GG G G G G G G G GGG GG G G G G GGG G GG G G GGG G G GG G GG GG G G G G GG G GG G G G G G G G GG GGGGGG G G GGGG G GGG GG GGGG GG GGG GG G G G G G G G G G G GGG G GGGGG GG G GG G G GG GGG G G G G G G GG G GGG G G G G GG G G G G GG GGGG GG G G G G GG G G G G G G G G G G G G G G G G GGGG GGGGG G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G GGGG G G G G G G G G G G G G GG G G G G GG G G GG GG G G G G GG G GG G G G G GG G G G G G GG G G G G G G GGG G G G G G GG GG GG G G G G G G G G G G G G GG G G G GG G G GG G G G G G GG G G G G GG G G G G G G G G G G G G GG G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G GG GGG G G GG G G G G G G G G G G GGGG G GG G GG G G G G G G GG G G G G GG G G G G G G G G G G GG G G G G G GG G G G G G G G G G G G G GG G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G GG G G G G G GG G G G G G G G GG G G G G G G GG G G G G G G GG G G GGG G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G GG G G GG G G G G G G G G G G G GG GG G G G G G G G G GGGGG GGGG GGGGG G G G G G G G G GG G G G G GG G GGG G GG G G G G G G GG G G G GGG G GG G G GG G G GG G G G G GGGG G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G GG G GG G G GGG G G G G GG G G GGG G G G GGG G G G G G GG GGG GG G G G G GG G GG G GG G G GG G GGGGG G G G G GGG GG G G GG GGGG G G G G G G GG G G G G GG G G G GG GG G G GG GG G G G G G G G G G G G GGG G G G G G G GG G GG G G G G G G GG G GGG G G GG G G G G G G G G G G G GG G GG G G G G G G G G G G GG G G G GG G G G G G G G G G G GGG G G GGG G G GG G G G G G G G G G G G G G G G G G GGGG GG G G G G G GG G G G G G GG G G G G GGG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G GG G G GGGGG G G G G GG G G G G G GGGG G G G G G G G G GGG G G G G G G G G G G G GGG G G G GG G G GG G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG GG G G G G G G G G G G G G GG G G G G GGG G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G GG G G G G G GGG G G G G G GG G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G CO std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppb] G G G G G G G G G G G G G G G G G GG G G G GGG G G GG G GG G G G G G G GG G G G G G G G G G G G G G G G G G G GGG G G G G G G G G G GG G G G G G G GG G GG G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G G GG GG GG G GG G G G GG G G G G G G G G G G GG G G G GG G G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G GG GG G G GG G G G GG G G G GGGG G G G G G G G GGGGGGGG G G G G GG GG G GG GG G G G GG G GGG GG G G G G G G G GG G GGG G G G G GGGGGGGGG G GG G G G G GGG G G G GG G G GGG GG G G G GG G G G G G G G G G G G GG G GGG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G GG G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G GG G G G G GGGG G G G G GGG G G G G G GG G G GG GG G G G G G G G G G G GGGG G G G G GG G G GGG G G G GG G GG G G G G G G G G G GG G GG G GG G G G GG G G G G G G G G G G G G G GG G G GG GG GG G G G G G G G G G G GGG GG G G GGG G G G G GG G G G G G G G G G GG GG GG G G GG GGGGG G GGG GG G G GGGGG GG GG GG G G G G G G G GGG GG G G G G GGG G GG G G GGG G G GG G GG GG G G G G GG G GG G G G G G G G GG GGGGGG G G GGGG G GGG GG GGGG GG GGG GG G G G G G G G G G G GGG G GGGGG GG G GG G G GG GGG G G G G G G GG G GGG G G G G GG G G G G GG GGGG GG G G G G GG G G G G G G G G G G G G G G G G GGGG GGGGG G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G GGGG G G G G G G G G G G G G GG G G G G GG G G GG GG G G G G GG G GG G G G G GG G G G G G GG G G G G G G GGG G G G G G GG GG GG G G G G G G G G G G G G GG G G G GG G G GG G G G G G GG G G G G GG G G G G G G G G G G G G GG G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G GG GGG G G GG G G G G G G G G G G GGGG G GG G GG G G G G G G GG G G G G GG G G G G G G G G G G GG G G G G G GG G G G G G G G G G G G G GG G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G GG G G G G G GG G G G G G G G GG G G G G G G GG G G G G G G GG G G GGG G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G GG G G GG G G G G G G G G G G G GG GG G G G G G G G G GGGGG GGGG GGGGG G G G G G G G G GG G G G G GG G GGG G GG G G G G G G GG G G G GGG G GG G G GG G G GG G G G G GGGG G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G GG G GG G G GGG G G G G GG G G GGG G G G GGG G G G G G GG GGG GG G G G G GG G GG G GG G G GG G GGGGG G G G G GGG GG G G GG GGGG G G G G G G GG G G G G GG G G G GG GG G G GG GG G G G G G G G G G G G GGG G G G G G G GG G GG G G G G G G GG G GGG G G GG G G G G G G G G G G G GG G GG G G G G G G G G G G GG G G G GG G G G G G G G G G G GGG G G GGG G G GG G G G G G G G G G G G G G G G G G GGGG GG G G G G G GG G G G G G GG G G G G GGG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G GG G G GGGGG G G G G GG G G G G G GGGG G G G G G G G G GGG G G G G G G G G G G G GGG G G G GG G G GG G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG GG G G G G G G G G G G G G GG G G G G GGG G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G GG G G G G G GGG G G G G G GG G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G CO std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppb] G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G G G G GG GG G G G G G G G G G GGG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G GG G G G G G G G G G G G G G G GG G G G GG G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G CO std. www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. [ppb] 0 1 2 3 Aug'14 Sep'14 Oct'14 Nov'14 Dec'14 Jan'15 Feb'15 Mar'15 Apr'15 May'15 Figure 6. Left: standard deviation over 2 min of hourly measurements of one of the working standards as measured from August 2014 until April 2015. The gray (black) colored data indicate the period after (before) modiﬁcations were made to improve the temperature stability (such that Telectr varied with 0.21 ◦C with every degree of changing Troom, against 0.91 ◦C ◦C−1 before modiﬁcations; see Sect. 2.5). For the period that is marked with a red background the room temperature was characterized by rapid changes caused by an air conditioner. Blue vertical (dashed) lines indicate moments where the mirror alignment was changed. Right: histogram of standard deviations as shown in the left ﬁgure with the black/gray colors corresponding to the colors in the left ﬁgure. Note that the dark histogram is transparent. Also included are the mean standard deviations as well as an overview of percentiles because the data do not show a Gaussian distribution. COS std. dev. [ppt] Relative frequency 0 5 10 15 0.00 0.02 0.04 0.06 0.08 0.10 Percentiles: 5 20 50 80 95 Mean: Aug−Feb 3.1 3.8 5.2 8.1 13.2 6.6 ± 4.7 Mar−Apr 2.9 3.4 4.3 5.9 8.6 4.8 ± 4.7 CO2 std. dev. [ppm] Relative frequency 0.00 0.05 0.10 0.15 0.20 0.00 0.02 0.04 0.06 0.08 0.10 Percentiles: 5 20 50 80 95 Mean: Aug−Feb 0.03 0.04 0.06 0.14 0.24 0.14 ± 2.43 Mar−Apr 0.03 0.03 0.04 0.06 0.08 0.06 ± 0.13 CO std. dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppb] Relative frequency 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.00 0.02 0.04 0.06 0.08 0.10 Percentiles: 5 20 50 80 95 Mean: Aug−Feb 0.2 0.3 0.5 0.9 1.3 0.7 ± 2.2 Mar−Apr 0.3 0.5 0.9 1.3 1.7 0.9 ± 0.4 G G G G G G G G G G G G GGG G G GGG G GG GGG G G G G G G G G GG G G G G G GGGG G G G G G G G GGG G G G G G G GG G GG G G GG G G G G G G G G G GGGG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G GG GG G G G G G GG G GG G G G G G G G G G G G G G G G G G G G G G GG G GGG G G G G G G G GG G G G G G G G GG G GG G G G G G G GG G G G G G G G G G GG G GG G G G G GG G G G GG G GG G G G G G GG GG G GG G G G G GG G G G G G G G G G G G G G GGGG G G GG GGGG G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GGG G GG G GG G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G GGG G GG G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G GG G GG GGGGG GGG G G G G G G G G G G G G G G G G G GG G G G GG GG G G G G GGG G G GG G G G G GGG G G GG GG G G G G GG G G G G GG G G G G G GGGG G G G G G G G G G G GG G G G G G G GG G GG G G G GG G G G G G G G G GGG G G GG G GG G GGG G GG GG G G G G G G G G G G G G G G G G GGGG G G G G G G GG GGGG GGGG GGGGGGG GG GG G GG G GG G GG GG GGG G GGGG G G G G G GG G G G G G GGG G GG G G G G G GG G G G GGGGG G G G GG G GG G G G G G GGGG GGG GGG GG G G GG GG G GGGG G G GGGGG GG G G G G GG G GG G G G G GGGGG G GG GGG G GGG GGG GG G G GG GG G G G G G GGGGG G GG G G G G GGGGGG GGGGGGGG G GGG G G GGG GG GG G G GGGG G G G G GG G G G G G G GG G G GG G G GGGG G G G G G G G G G G G GG G G G G G GG G G G G G G GG G GG GG GG G G GGG G G G G GG GG G G G GGGGGGG G GG G G G G G G G GG GG G GGGGG G GG G G G G G GG GG GGG GGGG G G G GG GGGG G G G G GGG GG G G G G GG G G G G GG G G G G G G GGG G G G GG G GGG GG G G GG GGG G GGG G G G GG G GG G GGGG G G G G G GG G GGGG GGG G GG G G G GG G G G G G G G G G G G G G G G G G G GGGG G G G GG G G GG GGGG G GGG GG GGGG GGG G G GGG GG G G G G G G GGG G G G GG G GG GGGG G G G G G G GGG GGGG G G G G G G G G G G GG G G G G GGGG G G G G G G GGGG GG GGG G G G G G G GG G GG G G G G G G G G GGGG G G G G G G G G G GG G G G G G GG GG G G G G G G G GG GGG G GG G GG G G G G G GGG GG G GG GG G GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G GGGG G G G GGG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GG G G G G G G G G G G G G GG GG G G G G G G GG G G G G G G G G G GG G GG GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G GG G G G G G G G G GG G GG G G G G G G G G G G GGG G G G G G GG G G G G G GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G GG G GG G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G GG G G GGG G G G G G G GG G G G G GG G G G G GG G G GG G G G G G G G G G G G G G G G G G G G GGG G G G GG G G G G G G G G G G G GG G G G GG G G G G G G G GGG GG G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G G G GG G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G GG GG G G G G G GGG G GG GGG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G GG G G G GGG GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GGG G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G GG GG G G G G G GG G G G G G G G G G G G G G G G GGGG G G G G GG G G GG G GG G G G G G G G G G G G G G G G GG G G G GG G GG G G G G G G G G G G G G G G G GG G G G GG G G G G G GGGG G G GGG G G G GG G G GG G GGG GG G G G G GG G G G G G G G GG G G G G G G GG G G GG G G G G G G G G GGG G G G G G G G G G G G GG GG G G G G G G G G GG GG G GG G G G G G G G G G G GGG G G G G G G G GG G G G G G G GG GGG G G G G GGG G G GG G GG G G G G G G G G G GG G G G G GG G GG GG G G G G G GG G G G G G G G G G G G G G GG G GG G G G G G G G G G G G GG G G G GG G G G G G GGG G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G GG G G G G G G G GG GG G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G GGG G G G G GGG G G G G G G G G G G GG G G G G G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G GG G G GG G GG GGGG G GG GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G G GGG G G G G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G COS std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppt] G G G G G G G G G G G G GGG G G GGG G GG GGG G G G G G G G G GG G G G G G GGGG G G G G G G G GGG G G G G G G GG G GG G G GG G G G G G G G G G GGGG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G GG GG G G G G G GG G GG G G G G G G G G G G G G G G G G G G G G G GG G GGG G G G G G G G GG G G G G G G G GG G GG G G G G G G GG G G G G G G G G G GG G GG G G G G GG G G G GG G GG G G G G G GG GG G GG G G G G GG G G G G G G G G G G G G G GGGG G G GG GGGG G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GGG G GG G GG G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G GGG G GG G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G GG G GG GGGGG GGG G G G G G G G G G G G G G G G G G GG G G G GG GG G G G G GGG G G GG G G G G GGG G G GG GG G G G G GG G G G G GG G G G G G GGGG G G G G G G G G G G GG G G G G G G GG G GG G G G GG G G G G G G G G GGG G G GG G GG G GGG G GG GG G G G G G G G G G G G G G G G G GGGG G G G G G G GG GGGG GGGG GGGGGGG GG GG G GG G GG G GG GG GGG G GGGG G G G G G GG G G G G G GGG G GG G G G G G GG G G G GGGGG G G G GG G GG G G G G G GGGG GGG GGG GG G G GG GG G GGGG G G GGGGG GG G G G G GG G GG G G G G GGGGG G GG GGG G GGG GGG GG G G GG GG G G G G G GGGGG G GG G G G G GGGGGG GGGGGGGG G GGG G G GGG GG GG G G GGGG G G G G GG G G G G G G GG G G GG G G GGGG G G G G G G G G G G G GG G G G G G GG G G G G G G GG G GG GG GG G G GGG G G G G GG GG G G G GGGGGGG G GG G G G G G G G GG GG G GGGGG G GG G G G G G GG GG GGG GGGG G G G GG GGGG G G G G GGG GG G G G G GG G G G G GG G G G G G G GGG G G G GG G GGG GG G G GG GGG G GGG G G G GG G GG G GGGG G G G G G GG G GGGG GGG G GG G G G GG G G G G G G G G G G G G G G G G G G GGGG G G G GG G G GG GGGG G GGG GG GGGG GGG G G GGG GG G G G G G G GGG G G G GG G GG GGGG G G G G G G GGG GGGG G G G G G G G G G G GG G G G G GGGG G G G G G G GGGG GG GGG G G G G G G GG G GG G G G G G G G G GGGG G G G G G G G G G GG G G G G G GG GG G G G G G G G GG GGG G GG G GG G G G G G GGG GG G GG GG G GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G GGGG G G G GGG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GG G G G G G G G G G G G G GG GG G G G G G G GG G G G G G G G G G GG G GG GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G GG G G G G G G G G GG G GG G G G G G G G G G G GGG G G G G G GG G G G G G GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G GG G GG G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G GG G G GGG G G G G G G GG G G G G GG G G G G GG G G GG G G G G G G G G G G G G G G G G G G G GGG G G G GG G G G G G G G G G G G GG G G G GG G G G G G G G GGG GG G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G G G GG G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G GG GG G G G G G GGG G GG GGG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G GG G G G GGG GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GGG G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G GG GG G G G G G GG G G G G G G G G G G G G G G G GGGG G G G G GG G G GG G GG G G G G G G G G G G G G G G G GG G G G GG G GG G G G G G G G G G G G G G G G GG G G G GG G G G G G GGGG G G GGG G G G GG G G GG G GGG GG G G G G GG G G G G G G G GG G G G G G G GG G G GG G G G G G G G G GGG G G G G G G G G G G G GG GG G G G G G G G G GG GG G GG G G G G G G G G G G GGG G G G G G G G GG G G G G G G GG GGG G G G G GGG G G GG G GG G G G G G G G G G GG G G G G GG G GG GG G G G G G GG G G G G G G G G G G G G G GG G GG G G G G G G G G G G G GG G G G GG G G G G G GGG G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G GG G G G G G G G GG GG G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G GGG G G G G GGG G G G G G G G G G G GG G G G G G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G GG G G GG G GG GGGG G GG GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G G GGG G G G G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G COS std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppt] G G G GG G G G GGG G G G G G G GG G G G GGG G G GG GG GG G GGG G G GG G G G G G G GG GG GG G G GG G G G G GG G GG G G G G G G G G GG G G G G G GGG G G G G G GGG G G G GG G G G G G GG GGG G G G G G G G GG GG G G GGGGG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G G G GG G G GG G G G G G G G G G G G G G G GGG G G G GGG G GG G G G G G G G G G G G G G G GG GG G G G G G GG GG GG G G G G G G G G G G G GG G GG G GG G G G G G G G GG GG G GGG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G GGG G G GG GG G G G G G G G G G G G GG G G GG GGGG G G G G G G G G G G GG G G G G G G G G GG G G GG G G G G G G G G GG G GG G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GG G GG G G G G G G G G G G G G G GGG GGGG G G G GG G GGGG G G G G G G G G G G G G G G G G GG G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G GGGGG GG G G G G G G GGG G G G G G G GGGG G G G G GG G G GG G G G G G G G G G G G G GG G GG G G GGG G G GG G G G GGGGG COS std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppt] 0 5 10 15 20 25 30 G G GGG G G G G GG G G G GG G G GG GG GG GG G G G G G G G G G G GG G G GG G G G G G G G GG G G G G GGGG G G G G G G G GGGG G G GG G G G G G G G G GGG G G GG GG G G G G GG G GG GGG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G GG G G G G G G G G G G G GG G GGG G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G GG G G G G G G G G GG GG G G G G G G G G GG G G G GGG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G GG G GG G G G GG G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GGG GGG G G G G G G G G G G GG G G G G G G G G G GG G GG G G GG G G G G G GGG G G G G G GGG G GGG G GG G G G GGGG G G G G G G GG G G G GGG G GG G G G G GGG GGG GGGG G G G G GGGGG G G G G G G G GGG G G GGG G G GGG G G GG G GG G GGG G G G GG GGG G G G G G G G GG G G GG G G G GG GGG G G G G G G GG GG G GGG GGG G G G G G G G G G G G G G G G G G GG G G G GG G G GG G G G G G G G G GGG G G G G GGG G G G GG G G G GG G G GGG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G GG G G GG G G G GG G G GG G GG G G G G G GG G G G G G G GGG G G G G G G GG G G G G G G G G G G G G GG G G GG GG G G GG G G GGG GGG G GG G G GG G G G G GGG G G G GG G G GG G G G G G G GG G G G G G G G G G GG G G G G GG G G GG G GG G GG G G G GGGGG G G G GG G G G G G G GGG G G G G G GG GGG G GG G GG GGGG GG G G GG G G G G G G G G G G G G G GG G G G G G GG G G GG G G G GGGGG G G GGGG G GG G G G G G G G G G G G G G GGG G GGGGG GG GG G GG G G G G GGG G GG GG GG GGGG G GGG GGG G GGG GGGG G G G G G G GGGGGG GG G G GG G GGG GG G GG GG G GGGG G G G G G G G G G G G G G G G G G GG GGGG GG G G G GG G GG G G G G G G GG GG GG GGGGGGGGGG G G G GG GGGG G G GG G G G GG G G G GG G GGGGGGG GGG GGGGG G GGG G GGGGG GGG G G G G G GG G G G G G G GG G GGG G GG GG G G GG G G G G GG G GG GGG G G G GGGG G G G G G GG G G G GGG G G G GG GG G G GGGG G G GG G G G G G GGGGGG G G G G G GG G G G G G G GG GG GG G G GG G G G G G G G G G G GG GG G G G G G G G G GGG G G GG GG G G G GGGGGG G GGG G G G G G G G G G GGG G G G G G G G GG G G G G G GGG G G G G GGG G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GGG G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G GG GG G G G G G G G G G GG G GG G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G GG G GG G G G G GG G G G GG G G GG G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G GG GG GG G G G G GG G G GG GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G G GG G G G G G GGG G G G G G GG G G G G G G G G GG G G GG G G G GG GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G GG GGG G G G G G G GG G G G GG G G G G G G G G G GG G GG G G GG G G G G G G G G G G G GG G G G G GG G G G G G G G G G G GG G G GG G G G G G G G G GG G G GGG GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G GG G GGG G G G G G G G G G GG G G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G G G GG G G G G G GG G G GG G G G G G G GG G GG G G G G G G GG G G G G GGG G G G G GG G G G G GG GG G G G G GG G G G G G G G G G G GG G G G G G G GG G G GG G G G G G G G G G G G GG G G G G G G GG G G G G GG G G GG GG GG GG G G G G GG GGG G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GG GG G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GGG G G G G G GG G G G GG G G GGG G G G G G G G G G G G G G G G G GG G GGGG G G GG G G G G G G G G G G G G GGG G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G GG G G GGG GGG G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G GG G G GG G G G GG G G G GG G G G G GG G G G G G GG G GG G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G GG G GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G GG G GG GG G G G G G G G G G G G G GG G G GG G G GGGGGG G G G G G G G G GG G G G G G G GG G G G G G G G G G GG G G G G G G G G G G GG G GG G G GG G GG G G G G G G G G G GG G G G G CO2 std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppm] G G GGG G G G G GG G G G GG G G GG GG GG GG G G G G G G G G G G GG G G GG G G G G G G G GG G G G G GGGG G G G G G G G GGGG G G GG G G G G G G G G GGG G G GG GG G G G G GG G GG GGG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G GG G G G G G G G G G G G GG G GGG G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G GG G G G G G G G G GG GG G G G G G G G G GG G G G GGG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G GG G GG G G G GG G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G GGG GGG G G G G G G G G G G GG G G G G G G G G G GG G GG G G GG G G G G G GGG G G G G G GGG G GGG G GG G G G GGGG G G G G G G GG G G G GGG G GG G G G G GGG GGG GGGG G G G G GGGGG G G G G G G G GGG G G GGG G G GGG G G GG G GG G GGG G G G GG GGG G G G G G G G GG G G GG G G G GG GGG G G G G G G GG GG G GGG GGG G G G G G G G G G G G G G G G G G GG G G G GG G G GG G G G G G G G G GGG G G G G GGG G G G GG G G G GG G G GGG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G GG G G GG G G G GG G G GG G GG G G G G G GG G G G G G G GGG G G G G G G GG G G G G G G G G G G G G GG G G GG GG G G GG G G GGG GGG G GG G G GG G G G G GGG G G G GG G G GG G G G G G G GG G G G G G G G G G GG G G G G GG G G GG G GG G GG G G G GGGGG G G G GG G G G G G G GGG G G G G G GG GGG G GG G GG GGGG GG G G GG G G G G G G G G G G G G G GG G G G G G GG G G GG G G G GGGGG G G GGGG G GG G G G G G G G G G G G G G GGG G GGGGG GG GG G GG G G G G GGG G GG GG GG GGGG G GGG GGG G GGG GGGG G G G G G G GGGGGG GG G G GG G GGG GG G GG GG G GGGG G G G G G G G G G G G G G G G G G GG GGGG GG G G G GG G GG G G G G G G GG GG GG GGGGGGGGGG G G G GG GGGG G G GG G G G GG G G G GG G GGGGGGG GGG GGGGG G GGG G GGGGG GGG G G G G G GG G G G G G G GG G GGG G GG GG G G GG G G G G GG G GG GGG G G G GGGG G G G G G GG G G G GGG G G G GG GG G G GGGG G G GG G G G G G GGGGGG G G G G G GG G G G G G G GG GG GG G G GG G G G G G G G G G G GG GG G G G G G G G G GGG G G GG GG G G G GGGGGG G GGG G G G G G G G G G GGG G G G G G G G GG G G G G G GGG G G G G GGG G G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GGG G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G GG GG G G G G G G G G G GG G GG G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G GG G GG G G G G GG G G G GG G G GG G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G GG GG GG G G G G GG G G GG GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G G GG G G G G G GGG G G G G G GG G G G G G G G G GG G G GG G G G GG GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G GG GGG G G G G G G GG G G G GG G G G G G G G G G GG G GG G G GG G G G G G G G G G G G GG G G G G GG G G G G G G G G G G GG G G GG G G G G G G G G GG G G GGG GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G GG G GGG G G G G G G G G G GG G G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G G G GG G G G G G GG G G GG G G G G G G GG G GG G G G G G G GG G G G G GGG G G G G GG G G G G GG GG G G G G GG G G G G G G G G G G GG G G G G G G GG G G GG G G G G G G G G G G G GG G G G G G G GG G G G G GG G G GG GG GG GG G G G G GG GGG G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GG GG G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GGG G G G G G GG G G G GG G G GGG G G G G G G G G G G G G G G G G GG G GGGG G G GG G G G G G G G G G G G G GGG G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G GG G G GGG GGG G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G GG G G GG G G G GG G G G GG G G G G GG G G G G G GG G GG G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G GG G GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G GG G GG GG G G G G G G G G G G G G GG G G GG G G GGGGGG G G G G G G G G GG G G G G G G GG G G G G G G G G G GG G G G G G G G G G G GG G GG G G GG G GG G G G G G G G G G GG G G G G CO2 std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppm] GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GGGG G GG G G G GG GG G GG G G G G G G G G G GG G G G G G G G G G G G GG GG G GGGG G G G GGG GG G G G G G G G GG G GG G G G G GG GG GG G G GG G G G G G GG G GG G G GG G G G GG GGGG G G GG G G G G G G G GG GG G G G GG G G G G GG G G G GG G G G G G G G G G G GG GG G GG G GGG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG GG G G GG GG G G GGGG G G G G G G G G G G G G G G G G GG GGG GG GG G GGG G GG G GG G G GG G G GGGG G G G G GG G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG GG G G G G G G G G G G G GG GG G G G G GG GG GG G G G G G G G G G G G GG G G G G G G G GG GG G G G G G G G G G GG GG G GG G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G GG G G G G G G GG G G GG G G G G G G GG G G G G GG G G G G G G G G G G G G G G GG G G G G GG G G GG G G G G G G GG G G G G G G G GG G G G GG G GGG GG G G G G G G GG G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG CO2 std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppm] 0.0 0.1 0.2 0.3 G G G G G G G G G G G G G G G G G GG G G G GGG G G GG G GG G G G G G G GG G G G G G G G G G G G G G G G G G G GGG G G G G G G G G G GG G G G G G G GG G GG G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G G GG GG GG G GG G G G GG G G G G G G G G G G GG G G G GG G G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G GG GG G G GG G G G GG G G G GGGG G G G G G G G GGGGGGGG G G G G GG GG G GG GG G G G GG G GGG GG G G G G G G G GG G GGG G G G G GGGGGGGGG G GG G G G G GGG G G G GG G G GGG GG G G G GG G G G G G G G G G G G GG G GGG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G GG G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G GG G G G G GGGG G G G G GGG G G G G G GG G G GG GG G G G G G G G G G G GGGG G G G G GG G G GGG G G G GG G GG G G G G G G G G G GG G GG G GG G G G GG G G G G G G G G G G G G G GG G G GG GG GG G G G G G G G G G G GGG GG G G GGG G G G G GG G G G G G G G G G GG GG GG G G GG GGGGG G GGG GG G G GGGGG GG GG GG G G G G G G G GGG GG G G G G GGG G GG G G GGG G G GG G GG GG G G G G GG G GG G G G G G G G GG GGGGGG G G GGGG G GGG GG GGGG GG GGG GG G G G G G G G G G G GGG G GGGGG GG G GG G G GG GGG G G G G G G GG G GGG G G G G GG G G G G GG GGGG GG G G G G GG G G G G G G G G G G G G G G G G GGGG GGGGG G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G GGGG G G G G G G G G G G G G GG G G G G GG G G GG GG G G G G GG G GG G G G G GG G G G G G GG G G G G G G GGG G G G G G GG GG GG G G G G G G G G G G G G GG G G G GG G G GG G G G G G GG G G G G GG G G G G G G G G G G G G GG G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G GG GGG G G GG G G G G G G G G G G GGGG G GG G GG G G G G G G GG G G G G GG G G G G G G G G G G GG G G G G G GG G G G G G G G G G G G G GG G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G GG G G G G G GG G G G G G G G GG G G G G G G GG G G G G G G GG G G GGG G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G GG G G GG G G G G G G G G G G G GG GG G G G G G G G G GGGGG GGGG GGGGG G G G G G G G G GG G G G G GG G GGG G GG G G G G G G GG G G G GGG G GG G G GG G G GG G G G G GGGG G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G GG G GG G G GGG G G G G GG G G GGG G G G GGG G G G G G GG GGG GG G G G G GG G GG G GG G G GG G GGGGG G G G G GGG GG G G GG GGGG G G G G G G GG G G G G GG G G G GG GG G G GG GG G G G G G G G G G G G GGG G G G G G G GG G GG G G G G G G GG G GGG G G GG G G G G G G G G G G G GG G GG G G G G G G G G G G GG G G G GG G G G G G G G G G G GGG G G GGG G G GG G G G G G G G G G G G G G G G G G GGGG GG G G G G G GG G G G G G GG G G G G GGG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G GG G G GGGGG G G G G GG G G G G G GGGG G G G G G G G G GGG G G G G G G G G G G G GGG G G G GG G G GG G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG GG G G G G G G G G G G G G GG G G G G GGG G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G GG G G G G G GGG G G G G G GG G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G CO std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppb] G G G G G G G G G G G G G G G G G GG G G G GGG G G GG G GG G G G G G G GG G G G G G G G G G G G G G G G G G G GGG G G G G G G G G G GG G G G G G G GG G GG G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G G GG GG GG G GG G G G GG G G G G G G G G G G GG G G G GG G G G G G G G G GG G G G G G G G G GG G G G G G G G G G G G G G GG GG G G GG G G G GG G G G GGGG G G G G G G G GGGGGGGG G G G G GG GG G GG GG G G G GG G GGG GG G G G G G G G GG G GGG G G G G GGGGGGGGG G GG G G G G GGG G G G GG G G GGG GG G G G GG G G G G G G G G G G G GG G GGG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G GG G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G GG G G G G GGGG G G G G GGG G G G G G GG G G GG GG G G G G G G G G G G GGGG G G G G GG G G GGG G G G GG G GG G G G G G G G G G GG G GG G GG G G G GG G G G G G G G G G G G G G GG G G GG GG GG G G G G G G G G G G GGG GG G G GGG G G G G GG G G G G G G G G G GG GG GG G G GG GGGGG G GGG GG G G GGGGG GG GG GG G G G G G G G GGG GG G G G G GGG G GG G G GGG G G GG G GG GG G G G G GG G GG G G G G G G G GG GGGGGG G G GGGG G GGG GG GGGG GG GGG GG G G G G G G G G G G GGG G GGGGG GG G GG G G GG GGG G G G G G G GG G GGG G G G G GG G G G G GG GGGG GG G G G G GG G G G G G G G G G G G G G G G G GGGG GGGGG G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G G G G GGGG G G G G G G G G G G G G GG G G G G GG G G GG GG G G G G GG G GG G G G G GG G G G G G GG G G G G G G GGG G G G G G GG GG GG G G G G G G G G G G G G GG G G G GG G G GG G G G G G GG G G G G GG G G G G G G G G G G G G GG G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G GG GGG G G GG G G G G G G G G G G GGGG G GG G GG G G G G G G GG G G G G GG G G G G G G G G G G GG G G G G G GG G G G G G G G G G G G G GG G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G GG G G G G G GG G G G G G G G GG G G G G G G GG G G G G G G GG G G GGG G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G GG G G GG G G G G G G G G G G G GG GG G G G G G G G G GGGGG GGGG GGGGG G G G G G G G G GG G G G G GG G GGG G GG G G G G G G GG G G G GGG G GG G G GG G G GG G G G G GGGG G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G GG G GG G G GGG G G G G GG G G GGG G G G GGG G G G G G GG GGG GG G G G G GG G GG G GG G G GG G GGGGG G G G G GGG GG G G GG GGGG G G G G G G GG G G G G GG G G G GG GG G G GG GG G G G G G G G G G G G GGG G G G G G G GG G GG G G G G G G GG G GGG G G GG G G G G G G G G G G G GG G GG G G G G G G G G G G GG G G G GG G G G G G G G G G G GGG G G GGG G G GG G G G G G G G G G G G G G G G G G GGGG GG G G G G G GG G G G G G GG G G G G GGG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G GG G G GGGGG G G G G GG G G G G G GGGG G G G G G G G G GGG G G G G G G G G G G G GGG G G G GG G G GG G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G GG GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G GG G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG GG G G G G G G G G G G G G GG G G G G GGG G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G GG G G G G G GGG G G G G G GG G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G CO std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O dev. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O [ppb] G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G G G G GG GG G G G G G G G G G GGG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G GG G G G G G G G G G G G G G G GG G G G GG G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G CO std. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O We provide the numbers for COS here, but the same holds for CO2 and CO. 2 We have now presented the total uncertainty for long-term concentration monitoring at an atmospheric station, but not all uncertainties are relevant for every type of analyses. If the data are to be compared across different sites, then the data should be on the same scale and the accuracy of the mea- surements is important. However, if data from the same site and the same instrumentation are compared, for example to do ﬂux-gradient analysis, then there is a less stringent need for accuracy and the short-term precision is more important. If the uncertainties related to the transfer of the scales are not taken into account, then the total uncertainty would be 6.0 ppt for COS, 0.13 ppm for CO2 and 1.1 ppb for CO. This in- cludes both the uncertainty of the water vapor correction and the 1-minute measurement precision. The uncertainty of the water vapor correction cannot simply be ignored as gradients in H2O can exist as well. We observed that COS concentra- tions between 40 and 7 m differ on average by 1.7 ± 5.3 ppt during daytime and 11.4 ± 33.0 ppt during nighttime. Given that the gradients should be larger than the uncertainties men- tioned above, the daytime gradients are too small to be able to use the ﬂux-gradient method. The suitability of the ﬂux- gradient method is further dependent on the choice of the measurement height (gradients are larger closer to the sur- face), the size of the ﬂuxes at a given site and the turbu- lence conditions. Even though the use of the ﬂux-gradient method is limited by the measurement uncertainty that we found here, COS proﬁle measurements from this instrument can still be useful to derive storage ﬂuxes, as nighttime gra- dients are typically larger. Figure 7. Two Allan deviation plots of COS retrieved from cylinder air measurements during two laboratory experiments under similar conditions on the same day. The Allan deviation plots show that the lowest random noise level of drift is not constant over time. Additional to the uncertainties presented in Table 5, we have observed that COS can decrease over time in uncoated aluminum cylinders. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5305 Table 5. Uncertainty contributions and the overall uncertainty for measurements of COS (447.8–486.6 ppt), CO2 (354–426 ppm) and CO (94–384 ppb), for the range of H2O (0–2.1 %). ons and the overall uncertainty for measurements of COS (447.8–486.6 ppt), CO2 (354–426 ppm) and CO H2O (0–2.1 %). Table 5. Uncertainty contributions and the overall uncertainty for measurements of COS (447.8–486.6 ppt), CO2 (354–426 ppm) and CO (94–384 ppb), for the range of H2O (0–2.1 %). Uncertainty contributions COS [ppt] CO2 [ppm] CO [ppb] Repeatability of the NOAA or WMO scalea 2.1 0.07 2.0 Transfer scale to working standards (1σ)b 2.8 0.12 1.7 Measurement calibrationc 2.8 0.12 1.7 Water vapor correction (1σ) 2.9 0.10 1.1 Measurement precision (1 min)d 5.3 0.09 0.3 Overall uncertainty 7.5 0.23 3.3 a For COS: deﬁned as the standard deviation of the measurements associated with the cylinder calibration. For CO2 and CO: certiﬁed by the WMO central calibration laboratory (NOAA/ESRL). b Average uncertainty over four cylinders in Table 2 (method 3). c Using the single bias correction (see Sect. 2.2.2) it is the same as transferring the scale to the working standards. d The standard deviation over minute-averaged cylinder measurements after drift correction with reference measurements every 30 min (Table 3). L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O G GG GGG GGGG G G GG GG GG G G G GGG GGG G G GG G G GGGG GG G G GGG G G GG GGGG GG GGGGGG GGG GGG G G G G GG G GG GG GG G G G G G GGG G G G GG GG G GGG G GGG GG G GGGGG G G G G G G GGG GG G G G G G G G G G G GG G GGG GG G GG GG G GGG G GG GG GG G GG G G GGGG G G G G GG G G G G G G G G G GGGGG GG GG G GGGG G G GG G G GG G G GG G G GG G G GG G GG G G G G GG G G G GG G GG G GGGGG GG GG G G G GGGG G GGGG G G G G GG GGG G GGGG G G G G GG G GGG G G GG GG GGG G G G G GGGGGG G G GG G G G G G G G G GG G GG GG G G G GG G G GGGG GGG G G GG G GGG GGG G GG G GGG G G G G G GG G G G G G G GGG G GGGGG G G GG G G GG GG GG GGGG G G G G G GG G G G GG G G GG G GG G G G GG GG GG GGG G G G G G G G G GGG G G G G G G GGG G G G G GGG GG GG G G G GG G GG G GG G G GG GG G G GG G G G G GG GG GG GGG G G G G G GG G G G G G G G GGGG G GG G G G GG G G GG G G G G G G GG G G GG GG G GG G G GGGG GGG GG G G GGG G GGG G G G G GGG G G GG GG G G G GG G G G G GGG G GGG G G G G G GG GG G GG G GG G GG G G GG G G G G GG G G G G GG G G G GG G GGG G G G GG G G GG GG GG GGG G G GG GG G G GG G G G G GG GG G G GGG G G G GG G GGGGG G G G GG GG G GG G GG G GG G G G GGG G G GGG G G G G G G G G G G GG G G G GGG GGG GG GG G G GGG G GG G G GGG G GG GGG GGGG G GGGG G G G G G G GG G G GG G G GG GG GGG GG GG GG GG G GG G G GGG GG G G G GG GGG G G GG GGG G G GG GG G GG G G G G G G G G G G GG GG G G G G G GG G G G G G G GG G G G G GG GGGG GGG G G G GG G GG GGGG GG G G G G GG G G G GG G G GG G GG GGG GG G G G G G G GG GG G G G GGG G G G G G G GG G GG GGG G G G GGG G GGG G G GGGG G GG G GGGGG G GGGGG G GG G G G G G G GG G G G GG G GG GG G G G G G G G G G G G G G GGGGGGG G GG GG G GGGGGG G GG G G G GGGGGG G GGG G G G G G G GG G G G GG G GG GGG GG G G GG GG GG G G G G G GG G G G GG GG G GG GG G G G G G G G GG G G G GG G G G G GGG G G G G G G GGG GG G GGGG G GG G G GG G GG G GG GG GG G G G G GGG GG G G G G GGGGG G GGG G GG GG G GG GG GG G GG GGG G GG GGG G GG G GG GG G GG GGGG G GGG G G G GGGG G GGGGGG G G G G GG GGGG GG GG G GG G GG G G G G G G GG G GG G G GG G G GGGGG GGGG GGGG G G GG GG G G GGG G G G G GG G G G G G G G G G G G G G G G G G GGG G G GG G GG G G G GGG GG G GG G GG G G GG G GG G G G GG G GGG G G G GG GG G G GG GGG G G G GG G GGG G G G GGG G G GGGG G GGG G G G G G GG G G G G G G G GGG GGG G GG GGG G GG G G G GGGG G G GG GG GG GGG G G G G G G G G GGGG G G GG G GG G G G G G GGGG G G G G G GGG GGG GGG G G GGGG GG GGG GGG G GG G G G GG GG GG GGG GGG G GG G G GGG G G G GGG G G GG GG G G G G G G G G G G GGG G GG G G G G G G G G G GG GGG G G G G GG GG GG G G G G G GG GG G G G GG G GGG G G G G G G G GG G G GG G G GG G G G G G G G G GGG G GG G GG G G G GGG GGG GG GG G GGG G G G GG GGGGGGGG G G GGGGG G GGGGG G G G G G GG G G G GG GG GG G G G G GGGG G G G G GG G GG GGGGG G G G GG GGGG G GGG G G GG G G GGG G G G GGG G G GG G G G G G G G GGG G G G G G G G G G GG GG G G G GG GG G G G G G G G G G GGG G G G G G G G G G G GGGG GG G G G G GGG GG G GG G G GG GGG G G G G GGG GGG G G G GG G GGGG GG G G G G G G G GGG GG G GG G GG G GG G G G G GG GG GG G G G G G G GGG GG G G G G GG G GGG GG GG G G GG GGGG G G GG G GG G GGGG GG G G GGGG GGG G GG G G G GGG G GG G GGGGGG G GG G GG G GG G G G GGGG G G GGG G G GGGGG G G GG G GGGGG G GG GGG GGGGGG G GG G GG GGGGG GGG G G G G G GGGGGGG G G GGG G G G G G G G GG GGG GG G G G GGG G G G G G G G GGGG G G G GG GG G G G G GGGG GGG GG GG G G G G GGG G GG GG GGGG G G G G G G G G G GG G G G G G G G G G G G G G GGGGG G GG GG G G G G G GG G G GG G GGG G GGGG G GG G GG G G G G G GG GG G G G G G G G G G GG GG G G G GGG GG G G GGGGG G G GG G GG G G G G G G G GGG G G G G GG GG GGG G GG GG G GG GG G G G G G G GG G G GG G G GG G GGGG G G G GGGG GG GG G GGG G G G GGG G G G G GG G G GG G G G G GG G G G GG GG GG GG G G GG G GGG G G G G G G G G G G GG G GG GG G G G GGG G G G G G GG GG GG G GG GG G GGGG G G G G G G G G G GGG GG G GG G G G GG G GGG G G G GG GG GG GGG G G G G G G G G G GG GG G G GG GG G GG GG G GG GGGGGG GG GG G G GG GG GG GG G GG G G G G G G G G GG GG G G G G G G G GG G GGGGG GGGGGG G GG G G G GG G G G G G G G G G G G GGG GG GG GG G G G G G G G G G G G GG G G G G G G G G GG G G GG G GG G GG G G G G G GGGG GG GG GG G GGG GG GGGG G G GGGG GG GGGGG GGG G G G GG G GGG GGG G G GGG G G G GG G G GG G G G G G G G G GG G G G G G G G G G GGG G G GG G G GGGG GGG GG G G G GGG GG GG GGG G G GG G GGG GG G GG G G G GG GG GG G G GGG G GG GGG G G G GGGGGGG GG GG GGGGG G G G GGG G G G GG G G G G GG GG GGGG G G GG G G G G GGGG G GGG GG G G GGG G G GG GG GG GGG GGG G G G GGGG G GGG GGG GG GGGGGGG G GG GG G GGGG G GG GGG G GG GGG GGGG GG G G G G G G G G GGG GG G G G G GG GG GG GG GG G G G GG G G GG G G GG G G G GGGG G G GGG G G GG G GGG G G G GGG G GGG G GG G G G G GGGGG G G G G GG G G GG GGGGG GGGG G G G G GG G GGGGGG GG G GG GGG G G G GG GGG G G G G G G G GG G GG G G G G GGGG GGGG G G G G GGGG GG G G G GG G G G GG G G GGG G G G GGG G G G G G G G GG G G G G GG GGG G G G GG G G GG G G GGG GG G G G G G GG G GGGGG G GG GG G G G G G GG G GG GG G G G G GGGG GG G G G G G G GG G G G G G G GGGG GG GGG GG GG G G G G G G G G GG GG GG GGG GGG G GGG GG GG G G GG G GG G G G G G G G G G GG G GG G G G G GG G G GG G G G G G G G G G GGG GG G GG GG G G G GG GG G G G G G G G G G G G G G G G G G G G G GGGG G GGG G GG G G G GG G GGGG G G G G G G G GG G G GG G G G G GG G G G G G G G G G G GG G G G GGGG G GGG G G G G G G G GG G G G G GG G GG G G GG GG G GGG G G GG G GG GG GG G GG GG G GG G G G GGG G GG G GG GG G G GG GG G GG G GG G G G GG GGGGG G G G GGGG GGGGGG G GGG G GGG G G G GG G G G GG GGGGG G G GG G GG G GG GG G GG G G G G G GGG G G G G GG G G G G G G GG G G G GG GGG G G G G G G G GGG GG G GG G G GGGG G G G G G G GG GG GG G GG GG G G G GG GG G GGG G G GG G G GG G GG G G G G G G G GGGG GGGGG GG GGG GG G GG GG GG GG G G G G GGG GG G GGG G GG GGGGG G G G GG GGG G G G G GG GGG G GGG GG G G G G GG G G G G GG G G GGG G GGGGG G G GG G G G GG G G G G GG GG G G G GGG GG G G G GGGG G GGG G G G G G G GG GG G G G G G G G GGGGG G G G G G G G GG GGGG G G GG GG GGG G G GG G GG GGGGG GGG GGGGGG GGG G GG G G G G G G GG G G G G G GGG GG GG GGGGG G G GGG G GG G GGG G G GGG GG GGG GGG G G G G G G GG GG G G GG G G GG G G G G G G G G G GGG G G G GG GGGG G G G G G GG G G G G GG G G GGG GGGG G GGGG GG GGGGG G G G G G GGG GG G G G G G GGGG G G GG G G GGG GG G G GG G GG G G G G G G G G GG G GGG GG G G G G G G GG G GGGG G G G GG G G G GG G G GGGG GGG G G GGG GG G GGG G GG GG G G G GG GG G G G G G G G GG GGGG GG G GGG G GGGG GG GGG G GG GGG G GG G G G G G GG GGG GG G G G GGG G G GG G GG G G GG G G G G G G G GG GG G GGG GG G G G GGG G G GGGGGGG G G GG GG G GG G GG GGGG G GG GGGGGG G G G G G G G GG G G G G GGG G G GG GGGGGG G G G G G G GG G G GG G G G G GG GGGGG G GGG G GG GGG G G G G GGG G G GG G G GGG G G G GGG G G G GGG G G GGGG G G GG G G G GG GG G G G GG GGG G G G GGG G GGG G G G G G G GGGG G G GG GG G GGGGG GGGG GGG GGGGG G G G G G GG G GG G G G GGG G G G G COS [ppt] 12:10 12:20 12:30 12:40 12:50 13:00 13:10 13:20 310 330 350 G G G G G G GGG GG GG G G G G G GG G G GG G G G G GGGGGG G GG GGGGGGG GG G GG G GG G GG G G G G G GG GG G GGGGG G GG G G GG G G G G G G GGG GGGG G G GG G G G G GG G G G GGGG G G G G G G GG G G GG G G G G G G G G GG GG G G G GGG GG GG G G GGG G G GG GG G G G GG G G G GG G G G G GG GGGG GG G GG G G G G G G G G G GGGG G GG GGG G GG GG GG GGG G G GGGG GGGG GG G G G GG G G G GG G G G GG G G GG G G G G GGG G G G G G G GG G GG G GG G G G G G G G G G G G G G G GGG GGG GG GG G GG GG GGGG GGG GG GGGGGGGGGG G GGG G G G G GG G G G G G GGGGGG GGG GG G GG G G G G GG GG G GGG G GG G G G GG GG G G GGG GG GG G G G G G GGGG GGG G G GG G G G GGG G G GG G G GGG GGG GG G G G G GGG GGGG G GG G G GG G GG GGGG G G G G G G G GG GG G G GGG G GG G GG G GGG GGGG G G G G G G GG G G G GG G G G G G GG GG G GG G GG G GGG G G G G G GGG G G G G G GG GG G GG G GG GGG G G GGG G GGGG GG G G G G G G GG G G GG G GG G G G G G GGG G G GG G G G G GG GGG G G G GGGG G G GGG G G GG GGG GG G G G GG GGG GGG GG G G GGG G GGGGG GG G G G GG G GGGGGGGG G GG G G G G G G G GG G G G G G G GG G G G G G G G GGG GG G G G G GGG GG G G G G G G GG GG G G G G G G G GG GG G G G GGGGG G G G GG G G G GGGG G G GG GG G G GGGGG GG G GG GGG GGG G GG GG G G G GG G GGG G G G GG G GG G GG G G G G G G GG G G GG G G G G G G GG GG GGG G GG GG G G G GG G G G G G G G G G GGG GG GG G G G GG GGGG G GG GG GG G GGGGGG GG G G GGGG G G G G G GG G G GG GG G G G GGG GGG G G GGG GGGG GG GGG G G GG G G G G G GGG G G GG GG GG G GG G GG G G G G G GG G G G G G G G G G GGG G G G G G G G G GGGG G GG G G G GG G G G GG GGGGG GG G G GG G G G GG G G GGG G G G G G G GG G G GGGGGGGGG G G G G G GGGGG G GGG GG GG GG G G GG GG GG G G G G G G G G GGG G G G G G G G G G G GG GG G G GGG G GGG G G GGGG G G GG G G GG G GGG GG G G G G G GGGG GGGGG G G G G G G GG GGG GGG G GGG G GG G G GG G G G G G G G G G G GGG GG GG G GGG G G G GG GG G GG G G G G G GG GG G G G G GG G G G G GG GGGG G G G GGG GG GG G G GGG G G G G GG G GGG G G G GG G GG GG G G G G G G G G G G G GGG G G G G GGGGGGGG G G G G G G G G G GGG G G G GG G GG G GG G GGGG G G G G G G G G GGGGG GG G G GG G GG G G G GGGGG GGG GG G G GGG G G G GG G G G GG G G GG G GG G G G GG G GG GGGG G G GG G GGG G G GG GG G G G GGGGG GGG G GG G G G G G G GG GGG G GG G G G GGG G GG G GG GG G G G G GG G G G G G GGGG G G G G GG G GGGGGGG G G GGG G GG GG GG G G G G GG G G G G GG G G GG GG G GGG G GG GGGG G GGGGGG G G G GGG G G GG G G G G G G GG G G GGGG G G G GG GGG GGG GG GG GG G GG G G G GG GGG G G G G GGGGGG G G GG G G G GGGGGGG G GG G G GG G G G GGGGGGGG GG GGG G GGG G GG GG G G GGGG GG G GGGG G G G G GG GGGG G GG G G G G GG GGG G G G G GGGGGGG G G G G GG G G G GGGGG GG G G G G G G G GGG G G G G GG G G G G GG G G G G G G G G GG G G G G GG G GGG GGG G G G G GG G G GG G G G G GGGG GG GGGG G G G GG G G G GGG G G G G GG GG GG GG G GGGG GG GG G G GG GG G GGG G GG G G GGG GG GGGG G G G GGG GG GGG G G G G G GG G G GG G G GG G G G G G G GG G G G G G G G G G G G GG G G GG GG G GG GG G G G GG G GG G GGG G GG G GGGGG G G G G G GG G GGG G G GGGGGG G GG G G G G G G G G G G GG G G G GG G G G G GGGGG G G G G G GG G GG G GG G G GG G GG G G G G G G G G G G GG G GG G GGGGGGG G G G G G G G GGGG G GG G G GG GG G G G G G GGGG G G G G GGG G GG GG G GGGG GGG G G G GG GG G G G G GG G G GG GG G GGG G GG GG G G G G GGGG GG G GG G G G G G GG GGGG GG G G GG G G G G G G G G G G G GG G GG G G G G G G GG G G GG G GGG G GG G G G G G GGG GG G GG G G GG GG GG G G G GG G GGGG G G GG G G G G GG G G GGGG G GG G GG G G G G GGG GG GG GGGG GG GG G G GGGGGGG GG G GGGGG GG G G GGG G G GG G GGG GG G GG G GG GG GG G G GGG G G GG G GG G G GG GG G G G GG GGG GGG G G G G GG G G G G GG GGGGG GG GG G G G GG GG GGG G G G GG G GG GG GGGG G G GGGGGG G G GGG G G G GG G G G GG GG GG G GGG GG G GG G GGG G G G G G G GG GG GGG G G G G GG GGG G G GGG G GGGG G G G G GG G G GGG GG G G G G G G G GGGGGG G G G GG GG GG G GG GGGG G G GG GG GG GG G GG G GG G GGGG G G G GG G G G G G G GG G G GGG G G G G G GGG GG GG GGG GG G G G G G GGG G G GG GGG G GG GG G G G G GG G G G G G G G G G G GGG G G G GG G GG GG GG G G GG G G GG G G G G G G GG G GGG G G G GGG G G G GGG GGG G GG GGG G G G GGG G G G G G GG G G G G GG G G G GGG G G G G G G GG G GGG G G G G G G G GG G GG GGG G G GG G G G G G G G GG G GGG GG G G G G G G G G GGG G G G GG G G G G G G GG G G G GGGG G GGG G GG G GG G G GGG G GG GG G G G G GG GGG G G G GG G G G GGG GGGGG G G G G G G G GG G G G G G GGG G G GGG G GGG GGG G G GG G GGGGG G G G G GGG GGGG G G G G G G G GG G G G G G GG G G G G G G GG GG G GG G G G G G GG G G GG GG GG G GG G G G G G GG GG GG G GGGG G GGGGGGG GG GG G G G G GG G GGG G G GG G G G G GGG G GGG G G GG GG GGG G G GG GG GGG GG GG G GGGGG G G GG G G G GGGG GG G G GGG GG GGG GGGG G G G G G G GGG GG G G G GG G GGG G G GG GG GG G G GG G G G GG GG G GG G G G G GGG G GG GG G GGGG G G GG GGG GG G G G G G G GG G GG GG G G G GG G G GG G G GG GGGG GG G G G G G G GGGGGG GGG G G G G G G G GG GGG G GGG GG G GG G GG G G G GG G G G G G G G G GGGG G G GGGGG GGGGG G G GGG G G GGGG G GG G GGG G G G G G GG GG G GG G G GG G G G G GGGGGG GG G G G GGG G G G G G G GG GG GG G G GGGG G GGG G G G GG G G GGG GG GG G GG GGGGGG GG G G G G G G G GGG G G G G GGG G G G GG G G GG G G G G G GG G G G GGG GGGG GG G G GGGG GGGG G GG GG GGG G GG GGG GG G GG G GGGGGG G GGGGGG GGGGG G G GG G G GGGG G GGGG GG G GG GGGG GGG GGGG G GG G G G G G G G GG G GG G GG GGG G G G G GG G G GG GG G G G G G GG G G G G GG G G GG GG GG G GGGGGG G G G G GG G G G GG GG G G G G GG GG G G GG G G GGGG G GG GG G G GG G GGG GGG GGGG GG G G GG GGGG G G GG G GG G G GG GG G G GGG GG G G G GGG GG GGG G G GG G G GGGGGG G GG GG G G G GG G GGG G G G GG G G GG G GGG G G G G G G GG G GG G GGGGG G G G G G G GGG GG G G G G G GGGGGG G G G GG G G G G G G GG GG G G G G G G G G G G GGG G GG GG G GG G GG G GG G G GGGGGGGG G G G G G G G GGGGG G GG G G G G GGGG G G G GGG G GGG GG G G GGGG G G G G G G GGGG GGG G G GGG G G GG G G G G G GGG GGGGGG GG GG G G GG G G G G GG G G G G G GG GG GG G G GGG G G G G GG G G G G GG G G G G G G G GGGGG G G GGG G G GG GGG GG GGGG G G G G G GGG G G G G GG GGG GGG GGG GGG G G G G G G G G G GGGG GGG G GG G G GG GGG G G G GG G G G GGG GGGG G G G GG G G GGGG G GG G GGG G G GG G G G GG GG G GG G GGGG G G G GGG GGG GG GG G G G GGG G G G GGG GG G G G GGG G G G G G G G G G G GG G G GG G GG G G GGG G GG G GG G G G G GG G G GGG G G GG G G G G GG GGG G G G GGG G GG G GG G G GGGG G GG GG G G G G G G G G GG G G GG GGG GG GGG G GG GG G GG G G GG G G GGG G G G G G GGGG G GG GG G G G GGG G G G G G GG GGGGGG GG GGG G G GG G GGGGG G G G G GG G G GGGG G GG G G G GG GGGG G GGG G G GG G G GGG G GG G G GGG G G G G G G G GG G GGG GGG GG GGG GG G G GG GGGG GG GG GGGG G GG G G GG G G G G G GGG G G G G G G GGGGGGGG G G G G G G GGG GGGG G GG G G GG G GGGGG G G G G G G G G G G GG GGG GGG G G G GGG GGGG GGG GG G G G G G G G GGG GG G GG GGG GG GGG GG G GG G G G GG G G GG G G G GG COS [ppt] 21:20 21:30 21:40 21:50 22:00 22:10 22:20 22:30 370 390 410 Allan deviation ( )σ 1 10 100 1000 0.01 0.10 1.00 10.00 Integration time [s] 1.0 s: 10.0 s: 100.0 s: 2.2 ppt 2.2 ppt 3.3 ppt 2.2 ppt 2.2 ppt 3.3 ppt 2.2 ppt 2.2 ppt 3.3 ppt 2.2 ppt 2.2 ppt 3.3 ppt 2.2 ppt 2.2 ppt 3.3 ppt 2.2 ppt 2.2 ppt 3.3 ppt 2.1 ppt 1.4 ppt 0.8 ppt 2.1 ppt 1.4 ppt 0.8 ppt 2.1 ppt 1.4 ppt 0.8 ppt 2.1 ppt 1.4 ppt 0.8 ppt 2.1 ppt 1.4 ppt 0.8 ppt 2.1 ppt 1.4 ppt 0.8 ppt Figure 7. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O G GG GGG GGGG G G GG GG GG G G G GGG GGG G G GG G G GGGG GG G G GGG G G GG GGGG GG GGGGGG GGG GGG G G G G GG G GG GG GG G G G G G GGG G G G GG GG G GGG G GGG GG G GGGGG G G G G G G GGG GG G G G G G G G G G G GG G GGG GG G GG GG G GGG G GG GG GG G GG G G GGGG G G G G GG G G G G G G G G G GGGGG GG GG G GGGG G G GG G G GG G G GG G G GG G G GG G GG G G G G GG G G G GG G GG G GGGGG GG GG G G G GGGG G GGGG G G G G GG GGG G GGGG G G G G GG G GGG G G GG GG GGG G G G G GGGGGG G G GG G G G G G G G G GG G GG GG G G G GG G G GGGG GGG G G GG G GGG GGG G GG G GGG G G G G G GG G G G G G G GGG G GGGGG G G GG G G GG GG GG GGGG G G G G G GG G G G GG G G GG G GG G G G GG GG GG GGG G G G G G G G G GGG G G G G G G GGG G G G G GGG GG GG G G G GG G GG G GG G G GG GG G G GG G G G G GG GG GG GGG G G G G G GG G G G G G G G GGGG G GG G G G GG G G GG G G G G G G GG G G GG GG G GG G G GGGG GGG GG G G GGG G GGG G G G G GGG G G GG GG G G G GG G G G G GGG G GGG G G G G G GG GG G GG G GG G GG G G GG G G G G GG G G G G GG G G G GG G GGG G G G GG G G GG GG GG GGG G G GG GG G G GG G G G G GG GG G G GGG G G G GG G GGGGG G G G GG GG G GG G GG G GG G G G GGG G G GGG G G G G G G G G G G GG G G G GGG GGG GG GG G G GGG G GG G G GGG G GG GGG GGGG G GGGG G G G G G G GG G G GG G G GG GG GGG GG GG GG GG G GG G G GGG GG G G G GG GGG G G GG GGG G G GG GG G GG G G G G G G G G G G GG GG G G G G G GG G G G G G G GG G G G G GG GGGG GGG G G G GG G GG GGGG GG G G G G GG G G G GG G G GG G GG GGG GG G G G G G G GG GG G G G GGG G G G G G G GG G GG GGG G G G GGG G GGG G G GGGG G GG G GGGGG G GGGGG G GG G G G G G G GG G G G GG G GG GG G G G G G G G G G G G G G GGGGGGG G GG GG G GGGGGG G GG G G G GGGGGG G GGG G G G G G G GG G G G GG G GG GGG GG G G GG GG GG G G G G G GG G G G GG GG G GG GG G G G G G G G GG G G G GG G G G G GGG G G G G G G GGG GG G GGGG G GG G G GG G GG G GG GG GG G G G G GGG GG G G G G GGGGG G GGG G GG GG G GG GG GG G GG GGG G GG GGG G GG G GG GG G GG GGGG G GGG G G G GGGG G GGGGGG G G G G GG GGGG GG GG G GG G GG G G G G G G GG G GG G G GG G G GGGGG GGGG GGGG G G GG GG G G GGG G G G G GG G G G G G G G G G G G G G G G G G GGG G G GG G GG G G G GGG GG G GG G GG G G GG G GG G G G GG G GGG G G G GG GG G G GG GGG G G G GG G GGG G G G GGG G G GGGG G GGG G G G G G GG G G G G G G G GGG GGG G GG GGG G GG G G G GGGG G G GG GG GG GGG G G G G G G G G GGGG G G GG G GG G G G G G GGGG G G G G G GGG GGG GGG G G GGGG GG GGG GGG G GG G G G GG GG GG GGG GGG G GG G G GGG G G G GGG G G GG GG G G G G G G G G G G GGG G GG G G G G G G G G G GG GGG G G G G GG GG GG G G G G G GG GG G G G GG G GGG G G G G G G G GG G G GG G G GG G G G G G G G G GGG G GG G GG G G G GGG GGG GG GG G GGG G G G GG GGGGGGGG G G GGGGG G GGGGG G G G G G GG G G G GG GG GG G G G G GGGG G G G G GG G GG GGGGG G G G GG GGGG G GGG G G GG G G GGG G G G GGG G G GG G G G G G G G GGG G G G G G G G G G GG GG G G G GG GG G G G G G G G G G GGG G G G G G G G G G G GGGG GG G G G G GGG GG G GG G G GG GGG G G G G GGG GGG G G G GG G GGGG GG G G G G G G G GGG GG G GG G GG G GG G G G G GG GG GG G G G G G G GGG GG G G G G GG G GGG GG GG G G GG GGGG G G GG G GG G GGGG GG G G GGGG GGG G GG G G G GGG G GG G GGGGGG G GG G GG G GG G G G GGGG G G GGG G G GGGGG G G GG G GGGGG G GG GGG GGGGGG G GG G GG GGGGG GGG G G G G G GGGGGGG G G GGG G G G G G G G GG GGG GG G G G GGG G G G G G G G GGGG G G G GG GG G G G G GGGG GGG GG GG G G G G GGG G GG GG GGGG G G G G G G G G G GG G G G G G G G G G G G G G GGGGG G GG GG G G G G G GG G G GG G GGG G GGGG G GG G GG G G G G G GG GG G G G G G G G G G GG GG G G G GGG GG G G GGGGG G G GG G GG G G G G G G G GGG G G G G GG GG GGG G GG GG G GG GG G G G G G G GG G G GG G G GG G GGGG G G G GGGG GG GG G GGG G G G GGG G G G G GG G G GG G G G G GG G G G GG GG GG GG G G GG G GGG G G G G G G G G G G GG G GG GG G G G GGG G G G G G GG GG GG G GG GG G GGGG G G G G G G G G G GGG GG G GG G G G GG G GGG G G G GG GG GG GGG G G G G G G G G G GG GG G G GG GG G GG GG G GG GGGGGG GG GG G G GG GG GG GG G GG G G G G G G G G GG GG G G G G G G G GG G GGGGG GGGGGG G GG G G G GG G G G G G G G G G G G GGG GG GG GG G G G G G G G G G G G GG G G G G G G G G GG G G GG G GG G GG G G G G G GGGG GG GG GG G GGG GG GGGG G G GGGG GG GGGGG GGG G G G GG G GGG GGG G G GGG G G G GG G G GG G G G G G G G G GG G G G G G G G G G GGG G G GG G G GGGG GGG GG G G G GGG GG GG GGG G G GG G GGG GG G GG G G G GG GG GG G G GGG G GG GGG G G G GGGGGGG GG GG GGGGG G G G GGG G G G GG G G G G GG GG GGGG G G GG G G G G GGGG G GGG GG G G GGG G G GG GG GG GGG GGG G G G GGGG G GGG GGG GG GGGGGGG G GG GG G GGGG G GG GGG G GG GGG GGGG GG G G G G G G G G GGG GG G G G G GG GG GG GG GG G G G GG G G GG G G GG G G G GGGG G G GGG G G GG G GGG G G G GGG G GGG G GG G G G G GGGGG G G G G GG G G GG GGGGG GGGG G G G G GG G GGGGGG GG G GG GGG G G G GG GGG G G G G G G G GG G GG G G G G GGGG GGGG G G G G GGGG GG G G G GG G G G GG G G GGG G G G GGG G G G G G G G GG G G G G GG GGG G G G GG G G GG G G GGG GG G G G G G GG G GGGGG G GG GG G G G G G GG G GG GG G G G G GGGG GG G G G G G G GG G G G G G G GGGG GG GGG GG GG G G G G G G G G GG GG GG GGG GGG G GGG GG GG G G GG G GG G G G G G G G G G GG G GG G G G G GG G G GG G G G G G G G G G GGG GG G GG GG G G G GG GG G G G G G G G G G G G G G G G G G G G G GGGG G GGG G GG G G G GG G GGGG G G G G G G G GG G G GG G G G G GG G G G G G G G G G G GG G G G GGGG G GGG G G G G G G G GG G G G G GG G GG G G GG GG G GGG G G GG G GG GG GG G GG GG G GG G G G GGG G GG G GG GG G G GG GG G GG G GG G G G GG GGGGG G G G GGGG GGGGGG G GGG G GGG G G G GG G G G GG GGGGG G G GG G GG G GG GG G GG G G G G G GGG G G G G GG G G G G G G GG G G G GG GGG G G G G G G G GGG GG G GG G G GGGG G G G G G G GG GG GG G GG GG G G G GG GG G GGG G G GG G G GG G GG G G G G G G G GGGG GGGGG GG GGG GG G GG GG GG GG G G G G GGG GG G GGG G GG GGGGG G G G GG GGG G G G G GG GGG G GGG GG G G G G GG G G G G GG G G GGG G GGGGG G G GG G G G GG G G G G GG GG G G G GGG GG G G G GGGG G GGG G G G G G G GG GG G G G G G G G GGGGG G G G G G G G GG GGGG G G GG GG GGG G G GG G GG GGGGG GGG GGGGGG GGG G GG G G G G G G GG G G G G G GGG GG GG GGGGG G G GGG G GG G GGG G G GGG GG GGG GGG G G G G G G GG GG G G GG G G GG G G G G G G G G G GGG G G G GG GGGG G G G G G GG G G G G GG G G GGG GGGG G GGGG GG GGGGG G G G G G GGG GG G G G G G GGGG G G GG G G GGG GG G G GG G GG G G G G G G G G GG G GGG GG G G G G G G GG G GGGG G G G GG G G G GG G G GGGG GGG G G GGG GG G GGG G GG GG G G G GG GG G G G G G G G GG GGGG GG G GGG G GGGG GG GGG G GG GGG G GG G G G G G GG GGG GG G G G GGG G G GG G GG G G GG G G G G G G G GG GG G GGG GG G G G GGG G G GGGGGGG G G GG GG G GG G GG GGGG G GG GGGGGG G G G G G G G GG G G G G GGG G G GG GGGGGG G G G G G G GG G G GG G G G G GG GGGGG G GGG G GG GGG G G G G GGG G G GG G G GGG G G G GGG G G G GGG G G GGGG G G GG G G G GG GG G G G GG GGG G G G GGG G GGG G G G G G G GGGG G G GG GG G GGGGG GGGG GGG GGGGG G G G G G GG G GG G G G GGG G G G G COS [ppt] 12:10 12:20 12:30 12:40 12:50 13:00 13:10 13:20 310 330 350 G G G G G G GGG GG GG G G G G G GG G G GG G G G G GGGGGG G GG GGGGGGG GG G GG G GG G GG G G G G G GG GG G GGGGG G GG G G GG G G G G G G GGG GGGG G G GG G G G G GG G G G GGGG G G G G G G GG G G GG G G G G G G G G GG GG G G G GGG GG GG G G GGG G G GG GG G G G GG G G G GG G G G G GG GGGG GG G GG G G G G G G G G G GGGG G GG GGG G GG GG GG GGG G G GGGG GGGG GG G G G GG G G G GG G G G GG G G GG G G G G GGG G G G G G G GG G GG G GG G G G G G G G G G G G G G G GGG GGG GG GG G GG GG GGGG GGG GG GGGGGGGGGG G GGG G G G G GG G G G G G GGGGGG GGG GG G GG G G G G GG GG G GGG G GG G G G GG GG G G GGG GG GG G G G G G GGGG GGG G G GG G G G GGG G G GG G G GGG GGG GG G G G G GGG GGGG G GG G G GG G GG GGGG G G G G G G G GG GG G G GGG G GG G GG G GGG GGGG G G G G G G GG G G G GG G G G G G GG GG G GG G GG G GGG G G G G G GGG G G G G G GG GG G GG G GG GGG G G GGG G GGGG GG G G G G G G GG G G GG G GG G G G G G GGG G G GG G G G G GG GGG G G G GGGG G G GGG G G GG GGG GG G G G GG GGG GGG GG G G GGG G GGGGG GG G G G GG G GGGGGGGG G GG G G G G G G G GG G G G G G G GG G G G G G G G GGG GG G G G G GGG GG G G G G G G GG GG G G G G G G G GG GG G G G GGGGG G G G GG G G G GGGG G G GG GG G G GGGGG GG G GG GGG GGG G GG GG G G G GG G GGG G G G GG G GG G GG G G G G G G GG G G GG G G G G G G GG GG GGG G GG GG G G G GG G G G G G G G G G GGG GG GG G G G GG GGGG G GG GG GG G GGGGGG GG G G GGGG G G G G G GG G G GG GG G G G GGG GGG G G GGG GGGG GG GGG G G GG G G G G G GGG G G GG GG GG G GG G GG G G G G G GG G G G G G G G G G GGG G G G G G G G G GGGG G GG G G G GG G G G GG GGGGG GG G G GG G G G GG G G GGG G G G G G G GG G G GGGGGGGGG G G G G G GGGGG G GGG GG GG GG G G GG GG GG G G G G G G G G GGG G G G G G G G G G G GG GG G G GGG G GGG G G GGGG G G GG G G GG G GGG GG G G G G G GGGG GGGGG G G G G G G GG GGG GGG G GGG G GG G G GG G G G G G G G G G G GGG GG GG G GGG G G G GG GG G GG G G G G G GG GG G G G G GG G G G G GG GGGG G G G GGG GG GG G G GGG G G G G GG G GGG G G G GG G GG GG G G G G G G G G G G G GGG G G G G GGGGGGGG G G G G G G G G G GGG G G G GG G GG G GG G GGGG G G G G G G G G GGGGG GG G G GG G GG G G G GGGGG GGG GG G G GGG G G G GG G G G GG G G GG G GG G G G GG G GG GGGG G G GG G GGG G G GG GG G G G GGGGG GGG G GG G G G G G G GG GGG G GG G G G GGG G GG G GG GG G G G G GG G G G G G GGGG G G G G GG G GGGGGGG G G GGG G GG GG GG G G G G GG G G G G GG G G GG GG G GGG G GG GGGG G GGGGGG G G G GGG G G GG G G G G G G GG G G GGGG G G G GG GGG GGG GG GG GG G GG G G G GG GGG G G G G GGGGGG G G GG G G G GGGGGGG G GG G G GG G G G GGGGGGGG GG GGG G GGG G GG GG G G GGGG GG G GGGG G G G G GG GGGG G GG G G G G GG GGG G G G G GGGGGGG G G G G GG G G G GGGGG GG G G G G G G G GGG G G G G GG G G G G GG G G G G G G G G GG G G G G GG G GGG GGG G G G G GG G G GG G G G G GGGG GG GGGG G G G GG G G G GGG G G G G GG GG GG GG G GGGG GG GG G G GG GG G GGG G GG G G GGG GG GGGG G G G GGG GG GGG G G G G G GG G G GG G G GG G G G G G G GG G G G G G G G G G G G GG G G GG GG G GG GG G G G GG G GG G GGG G GG G GGGGG G G G G G GG G GGG G G GGGGGG G GG G G G G G G G G G G GG G G G GG G G G G GGGGG G G G G G GG G GG G GG G G GG G GG G G G G G G G G G G GG G GG G GGGGGGG G G G G G G G GGGG G GG G G GG GG G G G G G GGGG G G G G GGG G GG GG G GGGG GGG G G G GG GG G G G G GG G G GG GG G GGG G GG GG G G G G GGGG GG G GG G G G G G GG GGGG GG G G GG G G G G G G G G G G G GG G GG G G G G G G GG G G GG G GGG G GG G G G G G GGG GG G GG G G GG GG GG G G G GG G GGGG G G GG G G G G GG G G GGGG G GG G GG G G G G GGG GG GG GGGG GG GG G G GGGGGGG GG G GGGGG GG G G GGG G G GG G GGG GG G GG G GG GG GG G G GGG G G GG G GG G G GG GG G G G GG GGG GGG G G G G GG G G G G GG GGGGG GG GG G G G GG GG GGG G G G GG G GG GG GGGG G G GGGGGG G G GGG G G G GG G G G GG GG GG G GGG GG G GG G GGG G G G G G G GG GG GGG G G G G GG GGG G G GGG G GGGG G G G G GG G G GGG GG G G G G G G G GGGGGG G G G GG GG GG G GG GGGG G G GG GG GG GG G GG G GG G GGGG G G G GG G G G G G G GG G G GGG G G G G G GGG GG GG GGG GG G G G G G GGG G G GG GGG G GG GG G G G G GG G G G G G G G G G G GGG G G G GG G GG GG GG G G GG G G GG G G G G G G GG G GGG G G G GGG G G G GGG GGG G GG GGG G G G GGG G G G G G GG G G G G GG G G G GGG G G G G G G GG G GGG G G G G G G G GG G GG GGG G G GG G G G G G G G GG G GGG GG G G G G G G G G GGG G G G GG G G G G G G GG G G G GGGG G GGG G GG G GG G G GGG G GG GG G G G G GG GGG G G G GG G G G GGG GGGGG G G G G G G G GG G G G G G GGG G G GGG G GGG GGG G G GG G GGGGG G G G G GGG GGGG G G G G G G G GG G G G G G GG G G G G G G GG GG G GG G G G G G GG G G GG GG GG G GG G G G G G GG GG GG G GGGG G GGGGGGG GG GG G G G G GG G GGG G G GG G G G G GGG G GGG G G GG GG GGG G G GG GG GGG GG GG G GGGGG G G GG G G G GGGG GG G G GGG GG GGG GGGG G G G G G G GGG GG G G G GG G GGG G G GG GG GG G G GG G G G GG GG G GG G G G G GGG G GG GG G GGGG G G GG GGG GG G G G G G G GG G GG GG G G G GG G G GG G G GG GGGG GG G G G G G G GGGGGG GGG G G G G G G G GG GGG G GGG GG G GG G GG G G G GG G G G G G G G G GGGG G G GGGGG GGGGG G G GGG G G GGGG G GG G GGG G G G G G GG GG G GG G G GG G G G G GGGGGG GG G G G GGG G G G G G G GG GG GG G G GGGG G GGG G G G GG G G GGG GG GG G GG GGGGGG GG G G G G G G G GGG G G G G GGG G G G GG G G GG G G G G G GG G G G GGG GGGG GG G G GGGG GGGG G GG GG GGG G GG GGG GG G GG G GGGGGG G GGGGGG GGGGG G G GG G G GGGG G GGGG GG G GG GGGG GGG GGGG G GG G G G G G G G GG G GG G GG GGG G G G G GG G G GG GG G G G G G GG G G G G GG G G GG GG GG G GGGGGG G G G G GG G G G GG GG G G G G GG GG G G GG G G GGGG G GG GG G G GG G GGG GGG GGGG GG G G GG GGGG G G GG G GG G G GG GG G G GGG GG G G G GGG GG GGG G G GG G G GGGGGG G GG GG G G G GG G GGG G G G GG G G GG G GGG G G G G G G GG G GG G GGGGG G G G G G G GGG GG G G G G G GGGGGG G G G GG G G G G G G GG GG G G G G G G G G G G GGG G GG GG G GG G GG G GG G G GGGGGGGG G G G G G G G GGGGG G GG G G G G GGGG G G G GGG G GGG GG G G GGGG G G G G G G GGGG GGG G G GGG G G GG G G G G G GGG GGGGGG GG GG G G GG G G G G GG G G G G G GG GG GG G G GGG G G G G GG G G G G GG G G G G G G G GGGGG G G GGG G G GG GGG GG GGGG G G G G G GGG G G G G GG GGG GGG GGG GGG G G G G G G G G G GGGG GGG G GG G G GG GGG G G G GG G G G GGG GGGG G G G GG G G GGGG G GG G GGG G G GG G G G GG GG G GG G GGGG G G G GGG GGG GG GG G G G GGG G G G GGG GG G G G GGG G G G G G G G G G G GG G G GG G GG G G GGG G GG G GG G G G G GG G G GGG G G GG G G G G GG GGG G G G GGG G GG G GG G G GGGG G GG GG G G G G G G G G GG G G GG GGG GG GGG G GG GG G GG G G GG G G GGG G G G G G GGGG G GG GG G G G GGG G G G G G GG GGGGGG GG GGG G G GG G GGGGG G G G G GG G G GGGG G GG G G G GG GGGG G GGG G G GG G G GGG G GG G G GGG G G G G G G G GG G GGG GGG GG GGG GG G G GG GGGG GG GG GGGG G GG G G GG G G G G G GGG G G G G G G GGGGGGGG G G G G G G GGG GGGG G GG G G GG G GGGGG G G G G G G G G G G GG GGG GGG G G G GGG GGGG GGG GG G G G G G G G GGG GG G GG GGG GG GGG GG G GG G G G GG G G GG G G G GG COS [ppt] 21:20 21:30 21:40 21:50 22:00 22:10 22:20 22:30 370 390 410 Allan deviation ( )σ 1 10 100 1000 0.01 0.10 1.00 10.00 Integration time [s] 1.0 s: 10.0 s: 100.0 s: 2.2 ppt 2.2 ppt 3.3 ppt 2.2 ppt 2.2 ppt 3.3 ppt 2.2 ppt 2.2 ppt 3.3 ppt 2.2 ppt 2.2 ppt 3.3 ppt 2.2 ppt 2.2 ppt 3.3 ppt 2.2 ppt 2.2 ppt 3.3 ppt 2.1 ppt 1.4 ppt 0.8 ppt 2.1 ppt 1.4 ppt 0.8 ppt 2.1 ppt 1.4 ppt 0.8 ppt 2.1 ppt 1.4 ppt 0.8 ppt 2.1 ppt 1.4 ppt 0.8 ppt 2.1 ppt 1.4 ppt 0.8 ppt to instrument drift. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O Two Allan deviation plots of COS retrieved from cylinder air measurements during two laboratory experiments under similar conditions on the same day. The Allan deviation plots show that the lowest random noise level of drift is not constant over time. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O We ﬁnd that the lowest random noise level of drift is not constant over time. The left Allan deviation plot in Fig. 7 shows that the noise level goes down to 0.6 ppt at 100 s, whereas in the right plot averaging does not lower the precision but rather increases it to 3.5 ppt at 100 s. The averaging time at which drift dominates Gaussian noise is around 2 s. The overall uncertainty of the measurements consists of uncertainties associated with the scale transfer, water vapor corrections, drift correction and the measurement precision. Table 5 summarizes the different uncertainty contributions as well as the overall uncertainty for measurements of COS, CO2 and CO. The overall uncertainty is calculated as the quadrature sum (square root of the sum of squares) of the www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O First, we did not ﬁnd indications that our working standards drifted during ﬁeld measurements at the Lutjewad station; calibrations in July 2014 and March 2015 showed a decrease of only 2.2 and 1.1 ppt for cylinder nos. 1 and 2 over this 8-month period, which is well within between 4.9 and 6.6 ppt. These numbers reﬂect the typically higher values for precision on longer timescales than 1 s due L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 2 mean = −0.2 , SD = 0.9 , SE = 0.04 GG G G G G GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GGGG G G G GG G G GG G G G G G G G G G GG G GG G G GG G G GG GG G G G G G G G G G G G G G GG G GG G ● G G G G G G G G G G GG G G G G GG ● G G GG GG G GGGGG G G G G G G GG G GG G G G G G G G GG GG G G G G GG G GG G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G GGGG G G G G G G G GG G G G G G GG G G G G G G G GG G G G GG G G GG G G G G G GG G G G G G G G G G GG G G G G GG G G G G GG G G G GGG G G G GG G GGG G GG G G G G G G G G G G G G G GG G G G G GGGG G GG G G G G G G G G G G G G G G G ● G G G G G G G GG G G G G G G G G G G G G GGGGG G G G G G G G G G G G G G GG GG G G G G GG G G G G G G G G G G G G G G G G GGGG G G G G GG GGG G G G G G G G G G GG G G G GGG G G GGG G G G G G G G G G G G G G G GG G G G G G G G GG G G G G time COS [ppt] GGG G G G G G G G G G G G G G G GGG G G G G G G G G G G G G G G G GGG G GG G G G G G G G GG G G G G G G GGG G G G G G G G GG G G G G GG G GG G G G G G G GGG G G G G G G GGG G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G GGG G G GG G G G GG G G GG G G G G GG G G GG GGG G G G G G G G GG G G G G G GG G GG G G G G G GG G GG G G G G GG GGG G G G GG G G G G GGGG GGGG G G GG G G G G GGG GG G G G G G G G G G G G G G G G G GGG G G G G G G GG G GGGG G G G G G G G G GG G G GG G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GG GG GGG G G G GG G G GG G G G G G G G GGGGG G G G G G G G G G G GG G G G G G G GGGG G G G GG G G G G GG GG G G GG G GG GG G G G G G G GG G G GG G GG G G G G GG G G G G G G G G G GG G G G G G GGG G G G GG G G GG GG G GG G G G G G G GGGG G G G G GG G GG G G G G time COS [ppt] −30 −15 Mean +15 +30 Fixed response curve / Hourly response curve Single bias correction No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1 mean = −1.7, SD = 6.6, SE = 0.3 No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 2 mean = −3.3, SD = 4.9, SE = 0.2 GGG G G G G G G G G G G GG G G GGG G G G G G G G G GG G G G G G GGG G GG G G G G G G G GGG G G G G G GG G G G G G G G G GGG G G G G G G GG G G G G G G GGG GG G G G G GGG G G GGG G G G G G G G G G G G G G GG GG GG G G G G G G G G G GGG G G G G G G G GG G G G G G G G G G G G G GG G GG G G G G G G G GG G G G G G GG G G G G GG G G GG G GG G G G G GG GGG G G GGG G G GG GGGG GG GG G G GG G G G G GGG GG G G G G G G G G GG G G G G G G GGG G GG G G GGG G GGGG G G G G G G G G G G G G GG G G G G GG G G G G GGG G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G GG GGG G G G GG G G G G G G G G G GGG G GGG G G G GG G G G G GG G G G G G GGGGG G G G GG G G GG GG G G G G GG G GG GG G G G G G G GG G G GG G GG G G G G GG G G G G GG G G G GG G G G G G GGG G G G GG G G GGGG GGG G G G G G G GGGG G G G G GG G GG G G G G time −30 −15 Mean +15 +30 No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 2 mean = −1.7, SD = 5.3, SE = 0.2 G G G G G G G G GG G G G GG G G G GG G G G G G G G G G GGGGG G GG GG G G G GG G GGG G G GGGG G GGG G GG G G GG G G G G G GGGGG G GGGGG GG G G GGG GG G G G G G GGG G GGGGGGGG GG G G G GGGGG G G GG GGG GG G G G GG G G G G GGGG G G G GG G GGGGGGGGG GG GG GG G G G GGG GG G G GGG G G G G G G GGGG G G G G GGG G G G G GGGGG GG G GGGG GG G GG G G G G G GGG G G G G G GG GG G G GGG GG G G GG GGGGG G G G GG G G G G G G G GG GGG G G G G GGGGGG G G G G G GG G G G GG GG G GGG G G GG G G GG G G G G G GG G G GG G G G G G GGG G GG GGG G GGG GG G G G G G G G G G G GG GGG G G G GG G G G G G GG G GGG G GG G GGG G G G G G GG G G GGG GG G G G G G G G G GG G GGG G GG G G G G G GG G GG G G G G G G GG G G GGG G G G G G G GG G G G G G G GGG G G G G G GG G G G G G G GGGG G GG GG G G G G G G GG G G G GGG G G G GG GG G G GG G GG G GG G G G GG G G G G G G G G G GG G G GG G CO2 [ppm] G GG G G G G G G G G G G G G GG G GG G G G G G G G G G G GG G GG GG G G GG G G G G GG G G GG G GGGG G G G GGGGGG G GG G G G GGG G GGGGG GGGGG GG GG G G GG G G GG GG GGGGGGGGG G G GG GG G G G G GGGGG GG GG G GG G GG G GG GGG GG G G G G G G GGG G G G G G G G GG G G GG GGGG G G G GGGG GGGG GGG G GGGG G G G G G G G G G G GG G G GGGGGG G G G GGGG GG GG G G G GGG G GG GG G GGG G G GG G G G G G G G GGG G G G G G G G GGG G G GG G GG G GGGG G G G G G G G G G GGG G G GG G G G G G G GGGG G G G G G G GG G GGG G GG GG G G GG G G G GG GG GGG G G GG G G G G G G G GGG G G G GG G G GG G G G GG GGGG G G G G G GG G G G G G G G G G G G G GG GG G G G G GG G G G G GGG G G G G G G G G G G G GG G GG G G G G G G G GGG G G G G G G G GG G GG G GG G G G GG G G G G G G GG GG GG G GG G G G G G G G G G G GG GG G G G GG G G G G G G G G GG G G G G G GG G GG G G G G G G G GG GG GG G G G G CO2 [ppm] −0.4 −0.2 Mean +0.2 +0.4 No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1 mean = −0.05, SD = 0.08, SE = 0.003 No. 3 mean = −0.04, SD = 0.08, SE = 0.004 G G G G G G GGGGG G G GG G GG G G G G G GG G G G G GGGG G GGGGG G GGGG G GGG G G GGG G G G GG G G G G GG G G G G G G G GGGG G GGGG G GG G G G G GGGG G G GG GGGGG G G G GGG G G GG G GG G G G G G GG GG GGG G GGGG GGG G GGG G G G G G G G GG GGGGGG G G GGG GG G G G GGG G G GG GGG G GG G G G GGG G G G G G G G GGG G G G G G G G G G G G G GG GGGGG G G GG G GG G G G G G G GGG G G GG G G G GG GGG G G G GG G G G G G G G GG G G G G GG GG G G G G G G GGGG G G GG G G G G G G G G G G G GGG G G GG G GG G G G G G G G G G G G GG GG GGG GGG G G GG G G GGGGG G G G G G G G GG G GGG G G G GGG G G G G GG G G G GGG G G G G G GG G G G G G GG GGGGG GG GG G G G G G G GGG GGG G GG G G G G G GGG GG G G G G G GG G G G G G G G G G G GG G G G G G G G G GGG G G G G G G G G GG GG G G G G G G GG G G GG G G G GG GG G GG GGG G G G G G G G G G G GG GGG G G G GGGGGGG GGG G G G G G GG G G −0.4 −0.2 Mean +0.2 +0.4 No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1 mean = −0.1, SD = 0.06, SE = 0.003 G G G G G G G G GGG G G GG G G GGG G G GGG G GGG G G G G G G GG G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G GGG GGG G G G G G GG G G G G GG G G G GG G G GG G G G G G G G GGG G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G GGG G GG G GGGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG GG G GGG G G G G G G G GG G GG G G G G GGG GG G G GGGG GG G G G G G GGG G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G GG G GG G G G G G G G GG G GGG G G G G GGGG G G G G GGGG G GGG G GGG GG GG G G G GG G G G GG G GG G G G G GG G G G GGG G G G GG G G G G G GGG G G G GGG GGG G G G G G G GG G G G G GG G G G GG GG GG G G G G GG G GGG G GG GG G GG G G G G G G G G G G G G G G G G G G G G G G G GGG G GG G G G G G G GGG GGG G G GGG G GG GG G G G G GG G GG G G G GG G GGG GG GGG G G GG G G G G G G G G G G GG G GG G G GG G G GG G G G G G G G G G G time CO [ppb] GG G G GG G G G G G G G GG G GG G G G G G G GG GG G G G G G G ● G G GG G G GG GG GG G G G G G G G G G G G GG G ● G G G G G G G GG GG G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG GG G G GG G G G G ● G G GG ● G GGGGG G G GG G G ● ● G G G G G GG ● G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G GG GG G G GG GG G GG G GG G GG G GG G G G GG GGGGG G G GGG G G G G G GG GGG G G G G G G G G G G G G G G GGG G GG G G G G G G G GG GG G GG G G G G G G G GGG ● G G GG G G G G GGGG G G G G GG GG GG G G GG G GGGG GG G G G G G G G G GG G G GG ● G G G G G G G G GG G G G G G G G GG G G G G GGGG GG GG G ● G G G GG G G G G G GG GG G G G G G GG G G GGG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G GG G G G G G G G G GG G G G G GGG G G GG G G GG G G G G G GGG G G G G G GGG G G G G G GGG G G GG G G G GG G G G G G G G GGG G G G G G G GG G G G G G G time CO [ppb] −6 −3 Mean +3 +6 25 Mar 06 Apr 18 Apr 29 Apr No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 2 mean = −0.6 , SD = 1.0 , SE = 0.04 No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 3 mean = −1.7 , SD = 1.2 , SE = 0.1 GG G G G G G G G GG G G GG G G GGG G G GG G G GG G G GG G G G GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GGGG GGG G G G G G G G G G G G GG G G G GG G GGG G G G G G G G GG G G G G G G G GG G G G G G GG G G G G G G G G GG G G G G G G G G G G G GG GG G G GGG GG G GG GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G GGG G G G G G G G GG G GG G G G G G GG GG G G GGGG GG G G G G G GGGG G G G G G G G GG G G G G G G G GG G G G GG G G GG G G GG G GG G G G G G G G GGG G GG G G G G GGGG G G GGGGGG G GGG G GGG GG GG G G G G G G G G G G G G G G G G G GGG G G GGG G G G GG G G G G G GGG G G G GG G GG G G G G G GG G G G G GG GG G G G G G GG GG G G G G GG G GGG G GG G G G GG G G G G G G G G G G G G G G G GG G GG G G GG G G G GG G G G G G G GGG GG G G G GGG G GG GG G G GG GGG GG G GG GG G GGG GGGG G G G GG G G G G G G G G G G G GG GG G G GGG G GG G G GG G G G G G G time −6 −3 Mean +3 +6 25 Mar 06 Apr 18 Apr 29 Apr No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 2 mean = −0.2 , SD = 0.9 , SE = 0.04 Figure 8. Mole fraction offsets in target standards after application of (left) corrections with a ﬁxed response curve for CO2 and CO that was determined in the laboratory (see Sect. 2.2.1) and with a single bias correction for COS, and (right) corrections with changing response curves determined from the hourly measurements of working standards. For the ﬁxed response curve measurements of two target standards are shown, for the hourly response curve only one is shown, as two of the three working standards were needed to determine the response curve. The mean deviation from the assigned cylinder value, standard deviation (SD) and standard error (SE) are given for each target standard. The cylinder numbers (nos. 1, 2, 3) correspond with the cylinder numbers and assigned cylinders values given in Fig. 5. Figure 8. Mole fraction offsets in target standards after application of (left) corrections with a ﬁxed response curve for CO2 and CO that was determined in the laboratory (see Sect. 2.2.1) and with a single bias correction for COS, and (right) corrections with changing response curves determined from the hourly measurements of working standards. For the ﬁxed response curve measurements of two target standards are shown, for the hourly response curve only one is shown, as two of the three working standards were needed to determine the response curve. The mean deviation from the assigned cylinder value, standard deviation (SD) and standard error (SE) are given for each target standard. The cylinder numbers (nos. 1, 2, 3) correspond with the cylinder numbers and assigned cylinders values given in Fig. 5. sired COS concentrations (LaFranchi et al., 2015), provided the ppb-level standard is less susceptible to drifting COS con- centrations. Applying this method could improve the accu- racy of the calibration if the COS concentrations can be ac- curately and precisely provided over a broad range, thereby allowing for a more accurate determination of instrument re- sponse and a calibration curve. Besides that, this method will aid in assessing the stability of working standards. the measurement uncertainty. However, a re-calibration in November 2015 showed a decrease of 18.2 and 24.1 ppt in these cylinders, a decrease with a rate of 2.3 and 3 ppt per month, while cylinder no. 3 only changed by 1.9 ppt. Cylin- der no. 3 is not different from nos. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O GG G G G G GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GGGG G G G GG G G GG G G G G G G G G G GG G GG G G GG G G GG GG G G G G G G G G G G G G G GG G GG G ● G G G G G G G G G G GG G G G G GG ● G G GG GG G GGGGG G G G G G G GG G GG G G G G G G G GG GG G G G G GG G GG G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G GGGG G G G G G G G GG G G G G G GG G G G G G G G GG G G G GG G G GG G G G G G GG G G G G G G G G G GG G G G G GG G G G G GG G G G GGG G G G GG G GGG G GG G G G G G G G G G G G G G GG G G G G GGGG G GG G G G G G G G G G G G G G G G ● G G G G G G G GG G G G G G G G G G G G G GGGGG G G G G G G G G G G G G G GG GG G G G G GG G G G G G G G G G G G G G G G G GGGG G G G G GG GGG G G G G G G G G G GG G G G GGG G G GGG G G G G G G G G G G G G G G GG G G G G G G G GG G G G G time COS [ppt] GGG G G G G G G G G G G G G G G GGG G G G G G G G G G G G G G G G GGG G GG G G G G G G G GG G G G G G G GGG G G G G G G G GG G G G G GG G GG G G G G G G GGG G G G G G G GGG G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G GGG G G GG G G G GG G G GG G G G G GG G G GG GGG G G G G G G G GG G G G G G GG G GG G G G G G GG G GG G G G G GG GGG G G G GG G G G G GGGG GGGG G G GG G G G G GGG GG G G G G G G G G G G G G G G G G GGG G G G G G G GG G GGGG G G G G G G G G GG G G GG G G G G G G G G G G GG G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G GG GG GGG G G G GG G G GG G G G G G G G GGGGG G G G G G G G G G G GG G G G G G G GGGG G G G GG G G G G GG GG G G GG G GG GG G G G G G G GG G G GG G GG G G G G GG G G G G G G G G G GG G G G G G GGG G G G GG G G GG GG G GG G G G G G G GGGG G G G G GG G GG G G G G time COS [ppt] −30 −15 Mean +15 +30 Fixed response curve / Hourly response curve Single bias correction No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1 mean = −1.7, SD = 6.6, SE = 0.3 No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 2 mean = −3.3, SD = 4.9, SE = 0.2 GGG G G G G G G G G G G GG G G GGG G G G G G G G G GG G G G G G GGG G GG G G G G G G G GGG G G G G G GG G G G G G G G G GGG G G G G G G GG G G G G G G GGG GG G G G G GGG G G GGG G G G G G G G G G G G G G GG GG GG G G G G G G G G G GGG G G G G G G G GG G G G G G G G G G G G G GG G GG G G G G G G G GG G G G G G GG G G G G GG G G GG G GG G G G G GG GGG G G GGG G G GG GGGG GG GG G G GG G G G G GGG GG G G G G G G G G GG G G G G G G GGG G GG G G GGG G GGGG G G G G G G G G G G G G GG G G G G GG G G G G GGG G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG G G G GG GGG G G G GG G G G G G G G G G GGG G GGG G G G GG G G G G GG G G G G G GGGGG G G G GG G G GG GG G G G G GG G GG GG G G G G G G GG G G GG G GG G G G G GG G G G G GG G G G GG G G G G G GGG G G G GG G G GGGG GGG G G G G G G GGGG G G G G GG G GG G G G G time −30 −15 Mean +15 +30 No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 2 mean = −1.7, SD = 5.3, SE = 0.2 G G G G G G G G GG G G G GG G G G GG G G G G G G G G G GGGGG G GG GG G G G GG G GGG G G GGGG G GGG G GG G G GG G G G G G GGGGG G GGGGG GG G G GGG GG G G G G G GGG G GGGGGGGG GG G G G GGGGG G G GG GGG GG G G G GG G G G G GGGG G G G GG G GGGGGGGGG GG GG GG G G G GGG GG G G GGG G G G G G G GGGG G G G G GGG G G G G GGGGG GG G GGGG GG G GG G G G G G GGG G G G G G GG GG G G GGG GG G G GG GGGGG G G G GG G G G G G G G GG GGG G G G G GGGGGG G G G G G GG G G G GG GG G GGG G G GG G G GG G G G G G GG G G GG G G G G G GGG G GG GGG G GGG GG G G G G G G G G G G GG GGG G G G GG G G G G G GG G GGG G GG G GGG G G G G G GG G G GGG GG G G G G G G G G GG G GGG G GG G G G G G GG G GG G G G G G G GG G G GGG G G G G G G GG G G G G G G GGG G G G G G GG G G G G G G GGGG G GG GG G G G G G G GG G G G GGG G G G GG GG G G GG G GG G GG G G G GG G G G G G G G G G GG G G GG G CO2 [ppm] G GG G G G G G G G G G G G G GG G GG G G G G G G G G G G GG G GG GG G G GG G G G G GG G G GG G GGGG G G G GGGGGG G GG G G G GGG G GGGGG GGGGG GG GG G G GG G G GG GG GGGGGGGGG G G GG GG G G G G GGGGG GG GG G GG G GG G GG GGG GG G G G G G G GGG G G G G G G G GG G G GG GGGG G G G GGGG GGGG GGG G GGGG G G G G G G G G G G GG G G GGGGGG G G G GGGG GG GG G G G GGG G GG GG G GGG G G GG G G G G G G G GGG G G G G G G G GGG G G GG G GG G GGGG G G G G G G G G G GGG G G GG G G G G G G GGGG G G G G G G GG G GGG G GG GG G G GG G G G GG GG GGG G G GG G G G G G G G GGG G G G GG G G GG G G G GG GGGG G G G G G GG G G G G G G G G G G G G GG GG G G G G GG G G G G GGG G G G G G G G G G G G GG G GG G G G G G G G GGG G G G G G G G GG G GG G GG G G G GG G G G G G G GG GG GG G GG G G G G G G G G G G GG GG G G G GG G G G G G G G G GG G G G G G GG G GG G G G G G G G GG GG GG G G G G CO2 [ppm] −0.4 −0.2 Mean +0.2 +0.4 No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1 mean = −0.05, SD = 0.08, SE = 0.003 No. 3 mean = −0.04, SD = 0.08, SE = 0.004 G G G G G G GGGGG G G GG G GG G G G G G GG G G G G GGGG G GGGGG G GGGG G GGG G G GGG G G G GG G G G G GG G G G G G G G GGGG G GGGG G GG G G G G GGGG G G GG GGGGG G G G GGG G G GG G GG G G G G G GG GG GGG G GGGG GGG G GGG G G G G G G G GG GGGGGG G G GGG GG G G G GGG G G GG GGG G GG G G G GGG G G G G G G G GGG G G G G G G G G G G G G GG GGGGG G G GG G GG G G G G G G GGG G G GG G G G GG GGG G G G GG G G G G G G G GG G G G G GG GG G G G G G G GGGG G G GG G G G G G G G G G G G GGG G G GG G GG G G G G G G G G G G G GG GG GGG GGG G G GG G G GGGGG G G G G G G G GG G GGG G G G GGG G G G G GG G G G GGG G G G G G GG G G G G G GG GGGGG GG GG G G G G G G GGG GGG G GG G G G G G GGG GG G G G G G GG G G G G G G G G G G GG G G G G G G G G GGG G G G G G G G G GG GG G G G G G G GG G G GG G G G GG GG G GG GGG G G G G G G G G G G GG GGG G G G GGGGGGG GGG G G G G G GG G G −0.4 −0.2 Mean +0.2 +0.4 No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1 mean = −0.1, SD = 0.06, SE = 0.003 G G G G G G G G GGG G G GG G G GGG G G GGG G GGG G G G G G G GG G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G GGG GGG G G G G G GG G G G G GG G G G GG G G GG G G G G G G G GGG G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G GG G GGG G GG G GGGG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG GG G GGG G G G G G G G GG G GG G G G G GGG GG G G GGGG GG G G G G G GGG G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G GG G GG G G G G G G G GG G GGG G G G G GGGG G G G G GGGG G GGG G GGG GG GG G G G GG G G G GG G GG G G G G GG G G G GGG G G G GG G G G G G GGG G G G GGG GGG G G G G G G GG G G G G GG G G G GG GG GG G G G G GG G GGG G GG GG G GG G G G G G G G G G G G G G G G G G G G G G G G GGG G GG G G G G G G GGG GGG G G GGG G GG GG G G G G GG G GG G G G GG G GGG GG GGG G G GG G G G G G G G G G G GG G GG G G GG G G GG G G G G G G G G G G time CO [ppb] GG G G GG G G G G G G G GG G GG G G G G G G GG GG G G G G G G ● G G GG G G GG GG GG G G G G G G G G G G G GG G ● G G G G G G G GG GG G G G GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GG GG G G GG G G G G ● G G GG ● G GGGGG G G GG G G ● ● G G G G G GG ● G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G G G G G G GG GG G G GG GG G GG G GG G GG G GG G G G GG GGGGG G G GGG G G G G G GG GGG G G G G G G G G G G G G G G GGG G GG G G G G G G G GG GG G GG G G G G G G G GGG ● G G GG G G G G GGGG G G G G GG GG GG G G GG G GGGG GG G G G G G G G G GG G G GG ● G G G G G G G G GG G G G G G G G GG G G G G GGGG GG GG G ● G G G GG G G G G G GG GG G G G G G GG G G GGG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G GG G G G G G G G G GG G G G G GGG G G GG G G GG G G G G G GGG G G G G G GGG G G G G G GGG G G GG G G G GG G G G G G G G GGG G G G G G G GG G G G G G G time CO [ppb] −6 −3 Mean +3 +6 25 Mar 06 Apr 18 Apr 29 Apr No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 2 mean = −0.6 , SD = 1.0 , SE = 0.04 No. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 3 mean = −1.7 , SD = 1.2 , SE = 0.1 GG G G G G G G G GG G G GG G G GGG G G GG G G GG G G GG G G G GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G GGGG GGG G G G G G G G G G G G GG G G G GG G GGG G G G G G G G GG G G G G G G G GG G G G G G GG G G G G G G G G GG G G G G G G G G G G G GG GG G G GGG GG G GG GG G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G GG G G G GGG G G G G G G G GG G GG G G G G G GG GG G G GGGG GG G G G G G GGGG G G G G G G G GG G G G G G G G GG G G G GG G G GG G G GG G GG G G G G G G G GGG G GG G G G G GGGG G G GGGGGG G GGG G GGG GG GG G G G G G G G G G G G G G G G G G GGG G G GGG G G G GG G G G G G GGG G G G GG G GG G G G G G GG G G G G GG GG G G G G G GG GG G G G G GG G GGG G GG G G G GG G G G G G G G G G G G G G G G GG G GG G G GG G G G GG G G G G G G GGG GG G G G GGG G GG GG G G GG GGG GG G GG GG G GGG GGGG G G G GG G G G G G G G G G G G GG GG G G GGG G GG G G GG G G G G G G time −6 −3 Mean +3 +6 25 Mar 06 Apr 18 Apr 29 Apr No. www.atmos-meas-tech.net/9/5293/2016/ www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 5306 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O We did notice that the response curve was changing only in the period after the in- strument was transported. A potential reason for the change L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5307 tween March and April. Except for the fact that the hourly response curve corrects for the changing response curve for CO2, the changing response curve does not signiﬁcantly re- move scatter compared to the ﬁxed response curve for CO2 and CO; that is, the standard deviation of target measure- ments for the ﬁxed response curve is not substantially lower when the hourly response curve was applied. Also, the tar- get measurements are not consistently closer to the assigned values when the hourly response curve was applied. Further- more, the fact that the use of hourly response curves does not give lower standard deviations for target measurements of COS compared to when the single bias correction is used indicates again that the response curves cannot accurately be determined for this species, which we discussed in Sect. 2.2.1 and 2.2.2 as well. As we have not seen indications that the response curve for CO2 changes outside the period between March and April, we do not see the need to frequently deter- mine the response curves with multiple standards. Moreover, if the response would change outside of the period in March and April as well, then the effect only reaches up to 0.2 ppm for CO2. Taking into account logistical reasons (use of cylin- der gases), we suggest correcting data with a single bias cor- rection for COS and with a ﬁxed response curve for CO2 and CO as determined once with NOAA/ESRL calibration stan- dards (see Sect. 2.2.1), together with a single bias correction from a reference standard. application of two corrections: (1) a correction factor as ob- tained from response curves of CO2 and CO to transfer the data to the WMO scale and (2) a bias correction using a ref- erence standard to remove instrument drift and to calibrate the data with the NOAA scale for COS. For measurements in March and April 2015 we determined hourly response curves from measurements of two working standards. For every species we took the standards with the outer concen- tration values such that the response curves span a wide con- centration range. To analyze the need to determine hourly re- sponse curves, we corrected the data in two ways: (1) with a single bias correction for COS and a ﬁxed response curve for CO2 and CO, i.e., the one that we determined in the laboratory (see Fig. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 2), shown in the left plots of Fig. 8; and (2) with changing response curves determined from the hourly working standard measurements, shown in the right plots of Fig. 8. For the ﬁxed response curve, measurements of two target standards are shown; for the hourly response curve only one is shown, as two of the three working standards were needed to determine the response curve. After the cor- rections are applied, the mean offset of the measurements is within 3.3 ± 0.2 ppt for COS, 0.05 ± 0.003 ppm for CO2 and 1.7 ± 0.1 ppb for CO over the period of 35 days. The standard errors indicate that the mean offset is signiﬁcantly different from 0. Still, the offsets are within the expected uncertainty based on the relevant uncertainties listed in Table 5, namely the transfer to working standards and measurement calibra- tion of which the quadrature sum is 4.0 ppt COS, 0.17 ppm CO2 and 2.4 ppb CO. The instrument drift is signiﬁcant, which is why the standard deviation of the series of 2-minute target measurements shown in Fig. 8 is higher than the stan- dard error of a 2-minute target measurement, or even higher than the 1-second precision. The fact that the mean offsets are on the same order of magnitude as the expected uncertainties indicates that these uncertainties listed in Table 5 properly envelope the uncertainties of the ﬁeld measurements. The slightly larger deviation from the CO assigned value for one of the two working standards with the ﬁxed response correc- tion is because this standard has a higher concentration than the reference standard (237.2 against 97.8 ppb) and therefore the uncertainty due to the bias correction is larger. The other target standard, which has a concentration of 119.1 ppb, is closer to the reference standard and only shows a deviation of −0.6 ppb. For CO2 it is visible that using the ﬁxed response curve gives a bias up to 0.2 ppm in the target measurements, which is not visible when using the hourly response curve. This bias is due to the fact that the response did change in the ﬁrst 10 days of the period, after which it became stable. We could not relate the changing response curve to any param- eter such as temperature or pressure. www.atmos-meas-tech.net/9/5293/2016/ L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 1 and 2 (they are all un- coated aluminum cylinders), but cylinder nos. 1 and 2 were stored with a pressure of 25 and 40 bar, which is lower than 130 bar for cylinder no. 3. We cannot conﬁrm if the drift in the cylinders is related with the cylinder pressure; however, it is the only difference that we were able to ﬁnd. Experience with cylinders over the past 15 years at NOAA indicates that COS in Aculife treated aluminum cylinders is typically much more stable than untreated aluminum cylinders. A potential method to improve COS calibrations is to calibrate these us- ing a ppb-level standard accurately diluted to a range of de- the measurement uncertainty. However, a re-calibration in November 2015 showed a decrease of 18.2 and 24.1 ppt in these cylinders, a decrease with a rate of 2.3 and 3 ppt per month, while cylinder no. 3 only changed by 1.9 ppt. Cylin- der no. 3 is not different from nos. 1 and 2 (they are all un- coated aluminum cylinders), but cylinder nos. 1 and 2 were stored with a pressure of 25 and 40 bar, which is lower than 130 bar for cylinder no. 3. We cannot conﬁrm if the drift in the cylinders is related with the cylinder pressure; however, it is the only difference that we were able to ﬁnd. Experience with cylinders over the past 15 years at NOAA indicates that COS in Aculife treated aluminum cylinders is typically much more stable than untreated aluminum cylinders. A potential method to improve COS calibrations is to calibrate these us- ing a ppb-level standard accurately diluted to a range of de- Our COS measurements are reported on the NOAA-2004 scale and can be compared to the observations from the global network of NOAA/ESRL (e.g., Montzka et al., 2007). The same holds for CO2 and CO on the WMO-X2007 CO2 scale and WMO-X2004 CO scale. To test the accuracy of the measurements against the NOAA or WMO scale we an- alyzed the measurements of one/two target standards after www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O For the QCLS two measurements were done: in April 2015 (orange) and in January 2016 (blue). The GC-MS measurement at NOAA/ESRL was performed in May 2015. For the QCLS and GC-MS instrument comparison an overview of the measurements is given in Fig. 9, where the ﬂask pair measurements are averaged and are shown as the deviation from the assigned cylinder value (see also Sect. 2.6). The precision within the ﬂask pair is shown by the error bars. The comparison demonstrates the uncertainties as- sociated with the transfer to the NOAA scale (see Table 5). For the ﬁrst measurement at the QCLS (orange), three of the four ﬂask pairs are within 1.5 ppt of the assigned value and one ﬂask pair deviates by 7 ppt. This is similar to the GC-MS measurements at NOAA where one of the four ﬂask pairs deviates further from the assigned values. However, the de- viating ﬂask pair measured by the GC-MS is not the same as the one deviating at the QCLS. For the second QCLS measurement (blue) one ﬂask pair has drifted on average by 12 ppt, where other ﬂasks remained stable within 2.5 ppt. It is unclear why the two ﬂasks have drifted as all ﬂasks were ﬁlled, measured and stored in the same way; however, we have kept the pair to monitor the potential drift in the fu- ture and to ﬁnd out what has caused the drift. Moreover, the consistency of the ﬂask measurements between April 2015 and January 2016 is dependent on the stability of the refer- ence standard which was used for all ﬂask measurements. Although we observed that COS can drift in cylinders, we did not ﬁnd indications that the particular reference standard used for this analysis drifted over the 9-month period. The ﬂask pairs have a mean deviation from the assigned cylin- der values of +3.0 ppt for the QCLS (1.7 ppt when excluding the drifting ﬂask pair) and −3.0 ppt for the GC-MS measure- ments. Although these deviations are within the QCLS mea- surement uncertainty of 6.9 ppt (including transfer of scales and the measurement precision), the GC-MS measurements are consistently lower than those from the QCLS, with an av- erage difference of 5.1 ppt (excluding the drifting ﬂask pair). We were not able to ﬁnd an explanation for this bias. p The next comparison in Fig. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O G G G G Cylinder no. Deviation from assigned value [ppt] G G G G Deviation from assigned value [ppt] NOAA no. 1NOAA no. 2 Cyl no. 2 Cyl no. 3 −10 −5 0 5 10 QCLS, Groningen G G Apr'15 Jan'16 G G G G Cylinder no. NOAA no. 1 NOAA no. 2 Cyl no. 2 Cyl no. 3 −10 −5 0 5 10 GC−MS, NOAA/ESRL G May'15 Figure 9. Flask sample measurements at the QCLS in Groningen (left) and GC-MS by NOAA/ESRL (right). Four paired ﬂasks were ﬁlled with dry air from the two NOAA/ESRL calibration standards and two working standards, which were calibrated as in Sect. 2.2.2. The ﬂask pair measurements are averaged and shown as the deviation from the assigned cylinder value. The ﬁrst two cylinders are those calibrated at NOAA/ESRL and have an assigned value of 447.8 ppt (NOAA no. 1) and 486.6 ppt (NOAA no. 2). The last two cylinders were used as working standards in Lutjewad and have assigned values of 467.6 ppt (no. 3) and 455.5 ppt (no. 4). The error bars show the precision. For the QCLS two measurements were done: in April 2015 (orange) and in January 2016 (blue). The GC-MS measurement at NOAA/ESRL was performed in May 2015. G G G G Cylinder no. Deviation from assigned value [ppt] G G G G Deviation from assigned value [ppt] NOAA no. 1NOAA no. 2 Cyl no. 2 Cyl no. 3 −10 −5 0 5 10 QCLS, Groningen G G Apr'15 Jan'16 G G G G Cylinder no. NOAA no. 1 NOAA no. 2 Cyl no. 2 Cyl no. 3 −10 −5 0 5 10 GC−MS, NOAA/ESRL G May'15 Cylinder no. Cylinder no. Cylinder no. Figure 9. Flask sample measurements at the QCLS in Groningen (left) and GC-MS by NOAA/ESRL (right). Four paired ﬂasks were ﬁlled with dry air from the two NOAA/ESRL calibration standards and two working standards, which were calibrated as in Sect. 2.2.2. The ﬂask pair measurements are averaged and shown as the deviation from the assigned cylinder value. The ﬁrst two cylinders are those calibrated at NOAA/ESRL and have an assigned value of 447.8 ppt (NOAA no. 1) and 486.6 ppt (NOAA no. 2). The last two cylinders were used as working standards in Lutjewad and have assigned values of 467.6 ppt (no. 3) and 455.5 ppt (no. 4). The error bars show the precision. 3.2 Measurement comparison Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 3.2 Measurement comparison In this section we make different measurement comparisons. First, we compare COS ﬂask measurements using the QCLS with those analyzed with the GC-MS at NOAA/ESRL. Sec- ond, we compare QCLS in situ and ﬂask measurements of COS in order to test if the water vapor correction that we de- termined in Sect. 2.3 sufﬁciently removes the effect of water vapor on calculated mole fractions. Third, we compare the in situ CO2, CO and H2O measurements from the QCLS with those of a collocated CRDS to assess the calibration strategy and water vapor correction presented in this study. Based on the previous sections, we applied the following corrections to the in situ QCLS data: (1) before 25 March 2015, the TDL- WINTEL water vapor correction was applied with broaden- ing coefﬁcients 1.5 for COS and CO and 1.7 for CO2. On top of this correction we applied a linear water correction curve for COS, CO2 and CO as obtained with the TDLWINTEL correction turned on. After 25 March 2015, the TDLWIN- TEL correction was turned off and we applied the correction curve from Fig. 3 as obtained with the TDLWINTEL correc- tion off. (2) The calibration correction curves as obtained in Sect. 2.2.1, Fig. 2, were applied to transfer the data to the WMO scales for CO2 and CO. (3) A bias correction was ap- plied to remove instrument drift and to calibrate the data with the NOAA scale for COS; here we used the same, and single, reference standard over the whole measurement period. −0.6 ppb. For CO2 it is visible that using the ﬁxed response curve gives a bias up to 0.2 ppm in the target measurements, which is not visible when using the hourly response curve. This bias is due to the fact that the response did change in the ﬁrst 10 days of the period, after which it became stable. We could not relate the changing response curve to any param- eter such as temperature or pressure. We did notice that the response curve was changing only in the period after the in- strument was transported. A potential reason for the change could therefore be that the instrument still had to be stabi- lized after transportation. We did not ﬁnd indications that the response curve for CO2 changed outside the period be- www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 5308 M. J. www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5309 G G G G G G G G G −15 −10 −5 0 COS [ppt] 25 Mar 30 Mar 04 Apr 09 Apr 14 Apr 19 Apr 24 Apr COS in situ − flask [ppt] GGGGGGGGGGGGG GGG G GG GGGGGG G GG GGGGGGGGGGGG GGG GGGGGGGGG GGGGGGG GG GGGG GGGGGG GGGGGGG GGG GGG G GGGGGG GGGGG G G GGG G GG GG GGGG GGGGGGG GGGG GG GGGGGGGGGGGGG GGGGGGGGGGGGG G GGGGGG GGGGGGGGGGGGG GGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG G GGGGGGG GGGG GG GGG G GGG GGGGGG GGGGGGGGGGGG GGGGGG GGGGGGGG G GGGGGGGG GGGG GGGGGGGGGGGGG GGGGGGGGGGGGGG GGGGGGGG GGGG GGGGG GG GGGGGGGGGGGGG GG G G G G G GGGGGG G G G GGG GGGGGGG GGGG GG GGGGGGG GGGGGGG GGGGGGGGGGGGGGGGGGG G GGGGG GGGGGGG GGG GGG GGGGG GGGG GGGG G GG GGGG GG G GG G G GGGG GG G GGGGG GGGGGGG GG G G G G GGG GG G G G G GGGG GGG G GG G G GG GGG G GG GGGG GG GGGGGGGG GG G GGGG G G GGGG G GG G GG G GG GG GGGGGGGGGGG GGGG GG GGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG G GGGGGGGG GGG G GGGGGGGG G GGGGGGGGGGGGGGG G GGGG GGGGGG GGGGG G G GG G GGGG G G G G GGGG GGG GGGGGGGG GG GGGG G GGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGG GGGGGGGGGGG G GGGGG GG GGGGG GGGGGG G G G GGGGG GGGGGGGGGG GGG GG GGGGG GGGGGGG GGGG G GGGGGGGG GGGGGGG GG GGG GGGG GGG GGGGGG GGGGGGGGGG GGG GGGGGGG GGGGGG G G GG GG GGGGG GGGGG GGGG G GG GGGGGGGGGG GGGG GG G G G G G GG GGGG G G G GGGG GGGGGG G G GGGGGG G G GGGGG GGGGGGG GGGGGGGGGGGGGGGGGGGGGGG GGG GG GGGG GGGGGGG GGGGGG GGGGGGG GGGGGG GG G GGGGGGG GGG GGGGGGG GGGGGGGG GG GG GGGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGG GGGGGGG GGG G G G GGGGG GG GG G GGGG GGGGGG GG G GG G GGGGGGGGGGG GGGG GGGGG G G GGG G GGG GG G GGG GGGG G GG G G G G GGG G G G G G G GGG GGGGGGG G GGGG GG GG GGG GG GG G GG GG GGG GG GG G GGGG GGG GG G G GG G GGGGG GGG GGGGG GG G GGGGG G GGGGGGG GGGGGG GGGGGGG GGGGG GGGG GGGG GGGGG G GGGG GG GGG GG GG G GG GG GGG GG G G G G GG GGGG GG GG GG GGG GGGGGGGGGGGGGGGGG GGGG G G GG GG G GG G GG GGGG GGGGGG GG GGGG GGGGG GG GGG GGGGGGGGG GGGGG GG GG G GGGG G GGGG G G G GG G GG GG G GGGGGG G GGG G G GG GGGGGGG GG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GG GGGGGGG GGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGG G GGG GGGGG G GG GGG GGG G GG GGGG G GG G GGGG G GGGG GGGGG G GGGGGGGGGGGGGGGGGGGGG GGGGGG G GGG GG GGGG GG G GGGGG GGGGG G GGG GG GG G G G G G G G GGG G G GGGG GGGG GG GGG GG GG GGG G GG G GG GGGG G G GG GG G GGG G G G G GGGGG G GGGGGGGG GGGG G G G GGGG GG G GGG GGG GG G G G GGGGGG GGG GGGGGGGGG G GGGGGGGGG GGG GGG GG GG GGGGGG G G GG G G GG GGG GG G G GG G GG GGGGGG GGG GGGGGGGG GGGGGGGGG G G G G G G GGGGGGG GGG GGGGGGGGGG G GGG G GG G G GGG G G GGG GGG G GGGGG G GGGG G GG GGG G GG GGG GG G G G G GGG G G GGG G G G G GGGGGG GGG G GGGGGGG GGGGGG GGGGGGG G GGG G GGGGGGGG GGGGGG GGG GGG G GGGGGG G GGGGGGGG GG GG GGGGGGG GGGGGG GG GGGG GGG G GGGG GGGGGGGGG GGG G GGG GG G GGGG GGG G GGG GGGGGGGG GG G G GG G GGG GGGGGGGGGGGGG GG GGGGG GGGGGGGGGGGG GGG G GGGGGGG GGGGGGG GGG GG GGGG GGG GGG GGG G GGG G GG G GGG G G GGG GGG G G GG G G G GG GG GGG G G G GGGG GG GGGG GGGGGG GGGG GGG GGGG GG G GGGGG G G GGGGG G GG GGGGGGGGGG GG G G GGGGG GG GG GGGGG GGGGG GG GG G GGGGG GGGGG G GGGGGG GG GGGGG GGGGG GGGGGGGG GGGGGGGGGGGGGGG G GGG G G GGGG G GGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGG GGGGGG GGGGGGG GGGGGGG GGGGGGGGGGGG GGGGGGGG GGGGG G GGGGGG GGGGGG GGGGGG G GGGGGGGGGGG GG GGG GG GG GGG GGGGG G GGGGG GG G G G G GG G G G GGGGGG G G G GGGG GG GGGGGG GG G G G GG G GGG G GGGGGG GGGG GGG GG GG GG GGGGGGG GGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGGGGGGGGG GG G GGGG GGG GGG G G GGGG GG G G GG GGGG GG GGGG G G G GG GG G GGGGGGGGG GG GGG GGG GGGGG GGGG G GG G G G G GG G G G GGGGGGGG GGGGG GGG G GGG GGGGGGGGGGGGG GGGGGG GGGG GGGGGGGGG G GG GGG G G G GGGG G G G G GGGGGGG G G GGGGGGG GG GGGG GG G GG GG G G G G GGG GG GGGG GGGGGGGGG GGGG GGGG G G G GG GG GGGGGGGGGGGGGGG GGG GG G GGG G GGG GGGGGG GGGGGGG GGGG GG GG GGG GG GGGGGG GGGGGG G GGGGGGGGGGGGG G GGGGGGGGGGGG GG GGG GGGGG G G G GG G G G G GGGGGG G GG GG G G GGGGGGGGG GG G G G G GGGGG GGGGGG G GGGGGG G GGGG G GG GGGGGGGGGGG GGGGGGG G GGGGGGGGGG G G GGG GGG G GG G GG G GGGGG G GGGGGGG GGGGGGGGGGGGGGGG GGGGG GGGGG GGGGGG G GGGGGG GGGGGGGGGG G G G G GG G GG GGGGGGGGG G G G G G G G GGGG GGGG G GGG GGGGGGGG GGGGGGGG GG GGG GGG GGGG G GGG GGGGG GGGG G GG GG GG GGGG GG GGGGGGG GG G GGGGGG GGG G GGG GGG GGG GGG G GG GGGG GGG GGGGG GGGGG GGGGGGG GG GGGGGGG GG GGGG GGGGGGG GGGGGGGGGGGGG GGGG GGGGGGGGGGGGGGGGG GGGGGG GGGGG GG GGGG GGGGGGG G GGG G GGGGGGGGG GG G GGG G GGGGGGGGGG GG GG GGGGGGGGGGGGGGGGGGG G G GGGG GG GGG GGGGGG GGGGGGGGGGGGGG GG GGGG GGG G GGG GG GGGG GGGGG GGG GG GGGG GG GG G G GGG GGG GGG GG GG GGG GGG GGGGGGGGGGGGG GGGGGGGG G GGG G G GGGGGG GG GG G G G G G G G GG GGG G G G GGGGGGGGGGGG G GGGGGGG GGGGGG GGGGGGGGGGGGGGGGG GGGG GG GGGGG GG GGGG GGGGG G G G GGGGGGG GG G G G G G GGG GGG GGG G G GGGGGGG G GGG G G GGG GG G G GGGG G G GGGGGGGGGGGGGGG GG GGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGG GGGGGG GG GG GGGG G GG G G G GG G GGG GGG GGG GGGGGGGGGG GGGGGG GGGG GGGGGGGGGGGGG GGGG GGGGGG G GG GGGGGGGG GGGGGGGGGGGGGG GGGG GGGGGG GGGGGGGGGGGGG GGGGGGGGG GGG GGGG GG GG G G G GGGG G GGG G G G G G GGG GG GGG G G GGG G G G G G GG GGG GG GG GG G GG G G G G GG G GG G G G GG GG GGGGG GGGGGGGGGGGGGGG GGGGGGG GGGGGGGGGGGGGGGGGGGG G G G G GG GGG G G GG G GG G G G GGGG G G GGGGGGG G G GGGG GG G G GGG G G GG G GGG GGG GG GGGGGGGGG GGGGGG GGGG GGGGG G G GGGGGGGGGGGGG G GGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGG GG GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGG GGGGGGGGG GGGG GGGGGGGGG GGGGGGGGGGGGGG GGGGGGGG GGGGGG GGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGG G GG G GG GGGGG GGGGGGGG G G GG GGG GGG GGGGGGGGGG GG GGGG GGGGGGGG GGGGG GGGG GGG GGGGG G GGGGGGG GGGGGG GGGGGGG GG GG G G G GG GGG G GG G GG G GG G GGGG GGGGG G G GGGG GGGGG GG G G GGGGGGGG GGG G GGGG GGG G GGG GGGGGGG GG GGG G GG G GGGGG G GG GG GG GGG GG GG GGGGGG G G G GGGG G G GGGG GGGGGGGGGGGGGGGGGGGGGGG GGG GGGGG G G GGGGGG GGGGGGGG GGGGG GGGGGGGGGGGGGGGGGGGG GGG GGG GGGGGGGG G G GG G G GG G G G G GG G GGG G GG GGGG G G GGG GGGGGGGG GGG G G G GGGGGG G GG GG G G GG G G GGG GGGGGG GG GG GGG GGGGGG GG GG G GG GGGGGG GGGG GGGGGGGGGGGGGGGGG GG GG G G GGGGGG GG GGGGG GGGG GGG G GGGGGGG G GGG GGGGGG GGGGG G G GG GG GGGGGG G GGG G G GG GGG GGG GGG G G GG G GGGGG GG G G G G GG G G GG GGG GG G GG G G GG G GGG GGGG GG GGGGGGG GGGGG G G GGGGGGGGGGGGG GGGG G G GGG G GG GGG GGGGGGGG G G G G GGGG GGGG GGG GGGG GGG GG GGG GGG G G GGG GG GGGG GGGGG GGGG G GG G GGG GG GG GGGGGG G G GG GGGG GGG GG G G GGG G G G GG GG GG GGGGG G GGGG G G GG GGGGGGGGGGG GG G GGG GGGGGGGGG GGGGG GGGGGGGGG GGGG GGGGG G GGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGG G G G GG G G GG GGGG GGG GG GG GG GGG G G G GGGG GGG G GGG G G G G GGGGG G GGGGGGGG G GGG G G GG GGGG G G G GG GGGG GGGGGG GGGGGGG GGGGG GG GGG GGG G G G G G G G G GG GG GGGGG G GG G GGGGGGGGGGGGG GGG GGGGGGGG GGGGGGGGG G G GG G G GG G G GG GGG GG GG GGG G GGG G GG GG G G GG GG G GGG GGGG G GGGG G G G G G G G G GGG G G G G GGGG GGGGGG G GG GG GG GGG G GGG G G G GG GG GGG G G G GGG GGG G G G G G GGGGGGGG G GG G GGG GGGGGG G GG GGG G G G GG GG G GGGGGGGGGGGG G GGGGGG GGG G GG GGGGG GG GGGGGGGGGGG GGGG G GGG GG G GGGGGGGGGGGGG GGGG G GG G G GGG GGG G G G G G G G G G GGG GGG G G GGGG G GG GG GG GGG GGG G GG GGGGGGGGG G G GG GGGG GGGGG G G GG GGGG GGGGGG GGGG GG G G GGGG G GGGG GGGGGGGG GGGGG GG GGGGGG GGGGGGGGGGGGGGGGGGGGG GG GGGGGGGG GGGG GGGGGGGG GGG GGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG GGGGG GGGGGGGGGGGGGGG GGGGGGG GG GGG G GGG GGGG G GGG GG GG G GG G G GGGGGG GGG G G GGGGG G GGG GGGG G GGGGGG G GGGGGGGGGGG G G GG GG GGGGG GG GG GG GGGGGG GGGGGGG GG GG GGGG G GG GG GG GG G GGGGG G G GG GGGG GGGGGGGGGGGGGGGG GGGGGGGG GG GG G G G G G GG G GG G G G G G G GGG GG GG G GG G GG G GG GG GGG GGGGGGGGGGGG G G GGGGGG GGG GG G G GG G G G GGGG G GG GG GGGG G G G GG G G G G GGGG GGGG GGGGG G GG GGGG GGGG GG GGGG GGGGGGG G GGGGG GGGGGGG GGGGGGG GGGGGG G G G GG G G GGGGGG G G G GGGGGGGGGG GGGGGGG G GGGGG GGGGGGG GGGGGG GGGGGGGGGGGGG GGGGGGG GGGGGGGGGGG GGGGGGGGGGGGGG GG GGGGG GGGGGGG GGGG GG GGGG GGGGGGGGGG G GGGGGG G G G GGG G GG G G GG G G G G G G GG G G G GGG GGGG GGGGGGGGGG G GG GGGGGG G GGGG G G GGG GGGG GG G G GG GGGG GG G GGGG G G GG GGGGGGG GG G GGG G GGGG GGGGGG GGG G GG G G GGGG G G G GG G G G G GGGGG GGG GG G G GG GG GG G GG G G GG G G G GG G GGGGGGG GGGGGGG GG GGGG GGGGG GG GGGGGG GGG G GGGGGGGG G GGGGG G G GGGGG GGGGGGGGGGG GGGGGGGGG GGGG GG GGGG GGG GG GGG G GGGG G G G GG GGGG GG G GGGGGGGGGG G GGGGGG G GGGGG G G GGGG G GGGGGG GGGGGGG GG GGG G GGGGGG G G G GG GG G G G GGG G G G G GG G GGGGGGG G G GG GG G GGGGGGGGGGGG G GGGG GG G GGG GG GG GGGGG GG GGGG G G G G G GG GG G G GGGGG GGGGGGG GG G G G G GGGG GGG G GG G GG GGG G G GGGGG G G G G G G G GGGGGGGG GGGGGGGGGG G GGGGGGG GGGGGGGGGGGGG GGG G GGGG G G G GG G GG GGG GGG GG G GGGGGGGGGG G G G G GG GGG G G GGGGGGGGGGG GGG GGGGGGG GG G G G G G G GG G GG G G GGG GG GG GGG G GGG GGG G GGGG GG GGG GGGGGG GG G G GGGGGGG GGGGGG GG G GG GG G G G GG GGGGGGGGG G GGG GGG G GGGG GGG G GG G G GG GG G GG G G GG GGGG G G G GG G GGGGGGGGGG GG GGG G GGGGGGG G GGGGG G G G GGGG GG GG GG GGGGGGGGGGGGG G GGGGGG G GGGGG GGG GG GG GG G GGG GGG GGGGGGGGGGGGG GGGG GGGGGG G GGGGGG GGGGGGGGGGG GG G GG GG G GGGGGGG GGGGGG G GGGGGGGGGGGG GG GGGG G GGGGG G GGGGG G G G GG G GG G GGGGGG GGGGGGGGGGG G G GG GGGGGGGGGGG GGGGGG G GGGGG G G G G GGG GGGGGGGGGGGGGGGGGGGG GGGGG G GGG GG GGG GGGGGGGGGGGG GGGGGGGG GGGGGGGGGGG GGGGGGGGGGGGG GGGGGGG GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGGGGGG GG GGGGGGGGGGGGG G GGGGGG G G GGGGGGGGGGG GGGGGGGGGGGGG GGGGGG G G GGGGGGGGGGG GGGGGGGGGGGGG G GG GG GGGGG GGGG GGGG G GGGGGGG G GGGGGG GG GGGGGG GG G GGGGGG G GG GGG G GGGGGGGG GGGGGGG GG GGG G GGGGGG G GGG GGGGGGGGGGGGGGGG GGGGGGG GG GGG G GGGGGGG GG GGGG GGGGGGG GGG GGG GG GGGG GGGGGGG GG G GGGGGGGG G G GGGGGG G GG GGG G GGG G GGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGG GGG GGGGGGG GGG GGG GGGGGGGG GGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGG GGGGG G G GGGG G G G GGG GG GGGG GGGG GGGGG G G G G GGGG G G GG G G G G G G G G GGG G GG G G G GG G G G GG GG G G G GG GGGG G G GG GGGGG G GG GG G G GG GGG GG G GGGGGGG GGGGGG G GGGGGGGGGGGGG GGGGGGGGGGGGG GGGG GG GGGGGGG GGGGGGGGGGGGG GGGGGG GGGGGGGGGGGGG G GGGGGGGGGGGG G G G GGG G G GGG G G GG GGG G GGGGGGGGGGGGGG GGGGGG GGGGGG G G GGG G G GGGGGGG GGGGGG GGGG GG G G G G G G G G G G GG G G G GGGGG G GGGGGG GGGGGGGG G GGGGGG GGGG GGGGGGG GGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGG GGGGGGGGGGGG GGG G G GGG GGGGGG GGGG GGG GGGGGGGGG GGGGG GGGGGGG GGGGGGG GGGGGGG GGGGGGG GGGG GGG GGGGGGG GGGGGGGGGGG GG G GGGG GGG GGGGGGG GGGGGGG GGGGG G G GGGG G GG GGGGGGG GG G GGG G GG GGGGG GGG G GGGGGGGGGG GGGGGG G GGGGGGGGG GGGGGGGGGGGG GGGGGGG GG GGGGG GGGGGGGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGG G G GGGGGGGGGGGGGG G GGGGGG GGGGGG GGG GGGGGGGGG GGGGGGGGGG GGGGGGG GGGGGG G GGGGGGG GGGGGGGGGGGGGGGGGGGGG GG G GGGG GGGGGGGGG GGG GG G GG GGGG GGGGG G G GGGGGGG GGGGGGG GGGGGG G GGGGGGG GGGG G G G GGGGG GG GGGG GGG GGGGGGGGGGGGGGGG GGG G G GGGGGGG G GG G G GG G GG GG G G GGG GGGGGGGGGGGGGG G G GG G G GGGG G GGGGGGG G GGGGGG G G GG GG G GGG G G G G GGGGG G G GGG GGGGGGGGG G G GG GG GGG GGGG GGG GGGG G G G G G GGG GG GG G GGG G G GGGGGG GGGG G GG GGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGGG G GG GGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGG GGGG GG GG GG G G G GGGG GGG GGGG G G G GG GGGGG G GGGGGG GGGGG GG GG GGGGG G GGGG GG GGG GGG GG GGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGG GGGGGG G GG G GGGG GGGGGGGGGGGG GG G G G GGGG GG GG GGG GGGGG GG GGGGG GGGG GGGG GGGGGGGG GGG GGGG GGGGGGG GG GGGGG GGGGGGG GGGG GGG GGGGGGG GGGGGGGGGG GGGG GG G GG GGGG G GGG G G G GG G G GGGGGGGG G GGGGG G GGGGGGG G GGG G GGGG GG G G G GG G G GGG G GG G GG GG GGGGGGGGGGGGG GGGGGG G GGGGGGG GGGGGGGG GGGGGGGGGG GGG GGGGGGG G GGGGGG GGGGGGG GGGGGGGGGGG GGG GG G GGGG GG GG GGGGGGGGGG GGG GGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGG GGGGGG GGG GG GG GGGGG G GG G GGGGG G GGG GGGG GG G GGGG G GGGGGG GG GGG GGGGG GGGG G GGG G G GG GGG GG GG G G G GG G G GG G GGG G GGGG G G GG GG G GG GG GGGGGG G G G GGG GG G GG GG GGG GGGGGG GG GGG GG GGGGG GG G GG GG GGGGG GGGGGGGG GGGGGG GGGGG G GGG GGGGGGG GGGGGGG GGGGG GG G GGGGGGGG GGGGG GGGGGG G G GGG G GG GGGG GGG GG GGGGG GG G G G G GG G GG GG G GGGG GGG G G GG GGGGGGGG G GGGGG G GG GG G GG GG GGGGGG G GGGG GGGGGGGGGGG GGGG G G G G G G GGG GGG GGG G GGG G G GG GGGGGG GGG G GG G GG G G G G GGG GG G GG GG G GGG GGGGGGGGGGGGGG GG GGG GGGGGGG GGGGG G GGGG G GG G GGGG GGGG G GG GG GG GGGGGGGGGGGGGG G G G GGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGG G GGG GGG GGGGGGG GGG G GGG GG G GGGG G G GGG GG G G G GG G G GGGGGGG GGGGGGGGGGGGGG GGGGGG GGGGGGGGGGGGGG GGGGGGGG GG GGGGG GGGG GGGGGGGGGG GG G GGGG G GG G G G GGG G GG G GGG GGG GG GGGGG GG GG G G G GG G GG G G GG GGGGG GG G GGGG GG G GGGGGG GGGG GGG G GGGGG GGGGGGGG GGGGG G G G GGGGGG GGGGG G G GG GGGGG G G G G G GG GGGGGGGGG GGGGGGGGGGGG G GG GG GG GGGGGGGGG G GGGG G G G GG GGGGGGG GGGGG GGGG G GGGGGG GGGGG GGGGGGGG GGGGGGG GGGGGGGGGGGGG G GGGGGGG GGGGG GG GG GGG GG G GGGG GG G GG GG GG G GGGGGG GGGG GG G G GG GG G G GGG GG G G G GGG GG G G G G G G GGGG G GGG GGGGGGG G GGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGG G G GG GGGGGGGGG GGGG G G G G G G GGGGGGG GGGGGGG GGGGGG G GG GG G G GGG GGGGG GGGGGG G G GGGGGGGGGGGGG GGG GG GG G G G GGG G G GGGG GG G GGG G GG GGGGGGGGGG G G GGGGGGGGGG GGGGGGGG G GGGG GGGGGGGGGGGGGG GG GGG G G G GG G G GG GG G GGG G G G GGGG G G GG GG G G GGGG G G G GG GGGG GGGGGGGG GG G GGGG GGGG GGG GGGG GGG GGGG G GG GG G GGGG G GGGG GG G GGGGG GGG GG G GG GG G GG GG GG GGGGGGGGG GG G GGGGGGGG GGGGGG GGG GGGG GGGGGG GGGG GG GG GGGG GGGGGGGGGG GGGGGGG GGGGGGGGGGGG GGGGGGGGG GGGGGGGGGGG GG G GGGGG GGGGGGGGG GGGGGGGGGGGGGGGGGGG GGGGGGGG GGGGGGGGG GG GG G GGGGGGGGG GGGGGGG GGGGGGG G GGGGGGGGGGGG GGG GGGG GGG G GGG GGGGGGG G GG GGGGGGGGGG G G GG GGG GGGG GG GG GGGGGGG GGGG G G GG GGGGG GGGGGGG G GGG GGGGGG GG G G G G GG G GG G GGGGGGG GG G GGG G G GG G G GG GG GGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGG GG G G GG GGGGGG GGGGGGGGG GGGGG G GGG GGGG G GGGGG G GG GGG GGG G GG G G G G G GGGG GGGGGGG G GGGGGGG GGGG GGGG GGGG GGG GG G G G G GG GGGGG GGGGGGGGGGGGGG G G G G GGGGGGGGGG GGGGGGG GGGGGGGGG GGGG G G GGGGG G GGGGGGG GGGG GG G G G GG GGG G GG GG GG G GG GGGG G GG GG G G G GGG G G G G G GG GGGGGG G G GG GGG GGGG GGGGGGGGGG GGGG GGGGGG G GGG GGG GGGG GGGGGGG G G GGGGGGGG G GGGGGGGG G GGGGG GGG GGGG GGG GGGG G GG G GGGGGGGGGG GGGGGGGGGGGGGGGG G GGGG GGGGGG G GGGG GGGGGG GG GG G GGGGGG G GG GG G G G G G G GGGGGGGGG G GG GGGG G GGGGGGG GGGGGG G GGGGGGG G GGG G GG GGGG GGG G GGGGG G GGG GGGGG GGGGGGGGGGGG G GGGGGGGG G GGGGGGGGG G GG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGGG GGGGGGG GGGGGGG GGGGGGG G G GG GGG G GGGG G GGGGG G GG GG G GG GG G GG G GG G GGG GGGG GGGGGGG GGGG G GG GGG GGGG GG GG G GG G G GGG G G GGGGGGG GGGGGGG GGGGGGG GGG GGGG GGGGGGG G GGG G G GGGGGGG G G GGGG G G GGGGGGG GGGGGG G GG GGGGG GGGGGGGGGG GGGG GGG GG GGGGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG G GG GGGG G G GGGG GGGGGGGGG GG G GG GGG G GGG GG GGGG GGG GGGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGGG G GGGG GGGGGGGGGGG GGGGGGGGGGGGG GGGGGGG GGG GGG G G G GGG GG G GG G GG G G G G GG G G G G G G GG G GG G GGGG GGG GGGG GGGGGG GGGGGG G GGG GGGGG G GG GGGG GGGGGGG GGGGGGGGGG G G G G GG GG GGGGGGGGGG GGGGGGG GGGG GGG GGGGGG G GG GGGGG GGGGGGGG GGG GGG GG GG G GGG G G G G G GGG G GGGG GG G GGG GGGGGGGG G G G GG GG G G GGGG GGG G GG G GGG G GGG G GGG G GG G GGGGGGG G GGGG GGGGGGGGG GGGGGGGGGGGG GGGGGGGGGGGGGGGGGGG GGGGGG GGGGG GGGGGGGGGGGGGG G GG GG G G GGG G GGG G G GG GGG GGG G G GG G GGG GGG GGGGGGGGG GGGGGGGGGGGG GGGGGGGGGGGGGG GGGGGG G GGGGGGG GGGGGGG GGGGGGG GGGGGGG GGGGGGG GGGGGGGGGGGGG GGGGGGGGGGGGGGG GGGG GG G G GGG G G G G GGG G G G GGG G G GG G GG GGGG GGGGGGG GGGGGG G G G G G G G G G G G G G G GG G G G GG G G G G GGG G G G GG G GG GGGGGGGGGGGGG G GGGGGGG GGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGG G GGG G GG G G G GGGGGGGGGGGGGGGGGGG G G GGG GG GG GGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGG GGGG GGGGGGG GGGGGGGGGGGGGGGGGG GGG GGGGGGG GGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGGG GGGGGG GGGGG GGGGGGGGG GGGGGGGGGG GGGG GGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGG GGG GGGG GGGGG GG GGGGGGG GGG GGGGGGGGGGGGGGGGGGGGGGGGG GGG GGGGG GGGGGG G GGGGGGGGG GG GGGG G GGGGGGGGGGGGGG GG G G G GGGGGGGG GGGGGGGGGGGGGGG GGGGGG G GGGGGGGGGGG GGGGGGGGG GG GG G G GG GGG G GG GGGGG GG G G GG GG G G G GG GGGGGGG G GGGGGGG GGGGGGGGGG GGG GG G G GG GG G G G GGG G GGG G G GGG GG G G G G GG G GG G GGGGG G G GGGGGG G GGGGG GG G GG GG GGG GGGGG GGGG GGG GGGGGGGGGGGGGG G GGGGGGGGGGGGG G GGGG G GGG GGG G GGGGGG GG GG G GGGGGGGGGGGGGGGGGG G GGG GG GG GGGGGG GG G G G G G G G GGG G G G G G G GGG GG G G G G G G G G GGGG G G G GG GG GGGGGG G GGGGGGG G G GG G G GG G G G GG G GGGGGG G GG GGGGG GGGG GG G G GGG GG G GGG GGG GGGGGGG GGGGGGG GG GG GG G G GGG G G GGGGGG GGGG GG GGG G GG G GG G G GG G GG GGGGG G GG G G GG GGG G GGG G G G G G G GGGG G G G GGG G G GGG GGGG GGGGGGGGGGGG G GGGGGGG GG GG G G GGGGG G G GG G G G G G G GGGGG GGGGGGGGGGGGGGGG G GG GG GGGGGGGGGGGGGGGGGGGG GG G G GG G GGG GGGGG GGGG GGG G G GGG G G GG G GG G G GG G G G G G G G G GGGGGG GGG G GG G G G GG GG GG GG GGGG GG GGGG GGGGG GG G G G GGGGG GGGGGGGGG GGGGG GGG GGGGGG GGG GG GGGGGGG GGGG GGGG GGGG G G GGGGG GG GG G GGGGG GGGGGG GGG GG GGGGGG GGGGGGGGGG GGGGGGG G GGG G G G GGGGGGG GGGG GGGGGGGGGGGGGGGGG GGGGGGGGGGGGGG G GG G G GGG G GG G GG G GG GGGG GG GGG GG GGGGG G GGGG G G GG G GG GGG G GGGG GGG G GGG GGG G G GG G G G GGGG GG GGGGGGGG GGGGGGGGGGGGGG GGGGGGGG GGG G GG G GGG GGG G G GGGGG G G GGGG G GG GG GGG G GGGGGGGGGGGGG GGGG GGG G GGG GGGG GG G G G G GG G G G GG GGGG G G G GG G GG G G G GGGG GG GG GG G GG GGG GGGG G GGG GGGGG G GG G G G G GGG GG G G GGGG G GGGGGGGGGGG GG G GGGG G GG GGGGG G G G G GG G GGG G GG GG GG G GGGG G GGGGG GG GGGG G GG GGGGGGGGGGGGGG GGGGG GGGG GG GGGGG G G GGGGGGGGGGGG GGGGGGGGGGGGG GGGG GGGGGG GG G GG GGG GGGG GGGGGGGGGGGG GGG G GG G G G GG G GG GGGG G GG G G G G GG GG G G GGGGG G G GGG G GGGGGGG GGGGGGG GGGG GG GGGG GG GGG G GGG G G G GG GG G GGG G GGGG G GG GGG G GGGGGG GGGGGGGG GGGG GGGGG G GGGG GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGG GGG GGGGGG GGGGG GGGGG GGGGGGG GGG GG GGG GGGGGGGG GGG GGGG GGGGGGGGGGGGGG GGGGGGG GGGGGGG GGGG GGG GGGGGGGGGGGGG G GGGGG GG GGGGGGG G GGGGGG GG G GG GG G GGGGGG G G GGGGG G GGGG GG GGG G GGG GG GG GGG GGGGG G G GGGGGGG GGGGGGGGGGGGGG GG GG GGG GGG G G GG G G G G G G G G GGG GG GG GG G G GG G GG G G GGGGGG G G G GG GG GGG GGG GG G G G GGGGG G GGGGGGGGG GGGGG GGG GGGG GG G GGGG GGGG G GG GG GG G G G G G GGG G G GG G G G G GGGGGGGGG G GGGGG G GGG GGG GG G GGG GGG GGGGG G G GG G G G G GGGG G G GGGGGGG GGGGG G G G GG GG GG G GGGGGG GGG GGGG GG G GG G G GGGG GG G GGG GG GGG GGGGGG G GGGGGG GGGGGGG G GGGGGG G G G G G G GG GGGGGGG GGGG GGGGGGGGGG GGGGGGGGGG GG GGGGG GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGG GGGGG GG GG GGGGGGGGGGG GG GGGGGG G G GGGG G GGGGGGG GG GGGGG G GGGGGGGGGGGG GGGGG G GGG GGG G GG G G GG G G G G GG GGG G G G GGG GGGG GGG GG G GG G GG G GGGGG GGGGGGGGGGGGGGGGGGG G G GG GGGGGG GGGGGGGG GGGG G G G GG GGGG GGGG GGGG GGGG GG G GGG GG G GGG GGG GG GGG GGGG GGG G GGG G GG G GG G GGGG G GG G G GG G GG GGGGGGG GG G G G GG G GGG G GGGGGGGG G G GGGGG G GGGGGG G GGG GGGG G G GGGG G GGG GGG G G GG GG G G GGGGGGG GGGGGGGGGGGG G GG G G GGGGGGGG GGG G GG GGG GGG GGG G G G G G GGGG GGGGGGGGGGGGGGGG GG GG G GGG GGGG GGGGG G G GGG GG GGGG GGGGGGGGG GGG GGGGG GG GGGG G G G GG G G G G G GGGG G G G GG G GGGGGGGGGGG GG GGG GG G G GGG G GGGGGGGG GGGGGGGG GGGGGG GGGGGGG GGG GG G GGGG G GG G G G G GG G G GGGGGGGGGGGG GG G G GGG G GGG G G G G G GGGGGGG GGGG GGG G G G G GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GGGGG GG G G G G GGG GG G G GGGG GGGGGGGG GGGG G GGGG GGG GGGGG GGGG G GG G G G GG GG GGG GGGG G G GG G G G G G G G G G GG GGGG GG GGG G GGG GGG GG GG GGGGG G GG G G G G GGGGGGGG G G G GGG GGGG GG G G GGGGGG G G G G G G GGGGGGGG G GG G G G G GG GG GGG G GGGGG G GG GGG GG G GG GG GGGGG G G GG GG GG GGG G GGG GGG G GGGGGGGGGGGGG GGGGGGG GGGGGGG GGGGGGG GGGGGGG GGGG G GGGG G G GGG G GGGGGGGG G GG G GG G GGG GG GGGGG GGGGGGGGG GGGGGGG GGGGGGGGG GGGGGGGGGGGG GGG G G G GGGGGGGGGGGGGGG GGGG GGGGG GG GGGGGGGGGGGGGGGGGGGG GGGGG GGGG GG GG GG GGGGGGGGGG GGG GGGG G GGGGGGG GGGGGGGGGGGGG GGGGGGG G GGG GGG GGGGGGGGGGGGGG GGGGGGGGGGGGGG GG GGGGG G G GGGGGGGGG GGGGGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGG GGGGGGG GGGGGGG GGGGGGG GGGGGGGGGGGGGG GGGGGGGGGGGGGGG GG GGGG GGGGGGG GGGGGGGGG G GGGG G G G G G GGGGGGGGG GGGGGGG GGGGGGGGGGGGGGGGGGGG GGGGGG GGGGGGGGGGG GGGG G GGGGGGG GGGGGG G GGGGGGG G GGG G G GGGG GGGG G GGGGGG GG GG G G G GGG G G G G GGGGGGGGGGG GG G GGGGGG G G GGGGGG GGG GGGG GGGGGGGGGGGG GG GGGGGGGGGGGGGGGG G G G GGGGGGGGGGGGGG G GGGGGGGGGGG G GG GGGGG GGG GGGGGGGGGGGGGGGGG GGGGGG G G G GGG GG G GG GGGGGGGGGGGGGGGGGGGGGGG GG GGGGGGGGGGGGGGGGGGGGGGGG G G GGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGG G GGGGG GGG GG GGGGGGGG GG G G G GGGGG G G GG GGG G G G GG G GGG GG G G GGG GG GGG GGGG G GG G G G GGGG GG G GGG G G G G GG G GGG GG G G GG G GGG G G G G G GGGGGGGG G GG G G G G GG GGGGG GG G G GG GGG G GG G GGG GG GGGGG GG G GGG GG GGGGGGG G GGGGGGG GGG GGG GGGGGGG GG GGGGGGGG GGGGG GGGGGGGG G GGGG GGGGG GG GGGGGGG GGGGGGG GG GGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGG GGG G GGGG GGGGGGGGGGGGGGGGGGGGGGGG GGGG G GGG GG G GGG G GG G GGGGG G G GGGGGG G GG G GG G GGG GGGGGGGGGGGGG G GGG GG GG GGGGG GGGG GGGGGGGGGG G GGGG GGGGGGGGG GGGGG GG GGGGGGGGGGGGGG GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGG GGG GG G 400 450 500 550 600 650 qclrt$COS G G G G G G G G G G G COS [ppt] 25 Mar 30 Mar 04 Apr 09 Apr 14 Apr 19 Apr 24 Apr 29 Apr G G Flask QCLS In situ QCLS COS [ppt] Figure 10. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O The average difference of −9.7 ppt could be caused by the uncertainty of the ﬂask and the in situ measurements (where the latter includes the uncertainty of the water vapor correction), the sampling bias of both in situ and ﬂask measurements, as well as the storage effect of air in the ﬂasks. In Table 5 we quantiﬁed the uncer- tainty related to the transfer of the scale to working standards and measurements, the water correction and the measure- ment precision. The uncertainty related to the repeatability of the NOAA scale cancels out for the ﬂask–in situ compar- ison. Combining these uncertainties gives an uncertainty of 6.6 ppt for the ﬂasks and 7.2 ppt for the in situ measurements (with the uncertainty of the water vapor correction included). The combined uncertainty of both methods is 9.8 ppt, which is very similar to the −9.7 ppt difference that we found. We were not able to test the effect of biases resulting from the sampling of ﬂasks. We ﬁnd that after applying the water cor- rection to the in situ data, the ﬂask–in situ difference does not show a correlation with H2O (R2 = 0.01; slope = 3.9 ppt COS per % H2O), whereas without water vapor correction there is a correlation with H2O (R2 = 0.36; slope = 26.3 ppt COS per % H2O), which demonstrates that our laboratory- derived water vapor correction properly corrects for the ef- fect of water vapor. Last, we compare the minute-averaged QCLS measure- ments for CO2, CO and H2O with those made by a collo- cated CRDS in Fig. 11. The air samples were taken from the same height but through a different inlet. The CRDS mea- surements of CO2 and CO were performed on humid air and were corrected for water vapor dilution and interference ef- fects based on a set of instrument-speciﬁc correction factors determined in the laboratory of the Center for Isotope Re- search before ﬁeld deployment (Chen et al., 2010, 2013). The comparison was only made from 7 January onwards be- cause before that period there was the problem of nitrogen leaking into the sample air for which we used the CRDS data to correct for the dilution. There are no temporal pat- terns visible in the difference plot on the right of Fig. 11. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O In Table 5 we quantiﬁed the uncer- tainty related to the transfer of the scale to working standards and measurements, the water correction and the measure- ment precision. The uncertainty related to the repeatability of the NOAA scale cancels out for the ﬂask–in situ compar- ison. Combining these uncertainties gives an uncertainty of 6.6 ppt for the ﬂasks and 7.2 ppt for the in situ measurements (with the uncertainty of the water vapor correction included). The combined uncertainty of both methods is 9.8 ppt, which is very similar to the −9.7 ppt difference that we found. We were not able to test the effect of biases resulting from the sampling of ﬂasks. We ﬁnd that after applying the water cor- rection to the in situ data, the ﬂask–in situ difference does not show a correlation with H2O (R2 = 0.01; slope = 3.9 ppt COS per % H2O), whereas without water vapor correction there is a correlation with H2O (R2 = 0.36; slope = 26.3 ppt COS per % H2O), which demonstrates that our laboratory- and 60 m concentrations. For the moments when the ﬂask and in situ measurements are compared, the difference is −4.2 ± 3.4 ppt. We neglected one ﬂask–in situ pair for which the 40 and 60 m difference was larger than 10 ppt. For the re- maining data we do not expect a large bias associated with including ongoing results from the 40 m height in the av- erages. Peaks and valleys in COS mole fractions shown in Fig. 10 are well covered by both sampling techniques (ﬂasks and in situ); for example, the peak up to 620 ppt at 11 April is clearly visible in both the in situ and ﬂask measurements. The average difference between the in situ and ﬂask mea- surements (in situ–ﬂask) is −9.7 ± 4.6 ppt. Besides neglect- ing one ﬂask measurement where the 40–60 m difference was larger than 10 ppt, we neglected one ﬂask sample where the in situ COS measurement over 15 min showed large variation (standard deviation of 17.5 ppt where the average standard deviation of the other periods is only 4.7 ppt) and thereby in- troduced an error in the comparison. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O Minute-averaged measurements of COS at the Lutjewad measurement tower (60 m) between 25 March and 29 April 2015 as measured in situ by the QCLS (blue), compared with dry air ﬂasks samples measured at the QCLS (orange). The left plots show the time series; the right plots show the difference between the in situ and ﬂask measurement. G G G G G G G G G −15 −10 −5 0 COS [ppt] 25 Mar 30 Mar 04 Apr 09 Apr 14 Apr 19 Apr 24 Apr COS in situ − flask [ppt] Figure 10. Minute-averaged measurements of COS at the Lutjewad measurement tower (60 m) between 25 March and 29 April 2015 as measured in situ by the QCLS (blue), compared with dry air ﬂasks samples measured at the QCLS (orange). The left plots show the time series; the right plots show the difference between the in situ and ﬂask measurement. and 60 m concentrations. For the moments when the ﬂask and in situ measurements are compared, the difference is −4.2 ± 3.4 ppt. We neglected one ﬂask–in situ pair for which the 40 and 60 m difference was larger than 10 ppt. For the re- maining data we do not expect a large bias associated with including ongoing results from the 40 m height in the av- erages. Peaks and valleys in COS mole fractions shown in Fig. 10 are well covered by both sampling techniques (ﬂasks and in situ); for example, the peak up to 620 ppt at 11 April is clearly visible in both the in situ and ﬂask measurements. The average difference between the in situ and ﬂask mea- surements (in situ–ﬂask) is −9.7 ± 4.6 ppt. Besides neglect- ing one ﬂask measurement where the 40–60 m difference was larger than 10 ppt, we neglected one ﬂask sample where the in situ COS measurement over 15 min showed large variation (standard deviation of 17.5 ppt where the average standard deviation of the other periods is only 4.7 ppt) and thereby in- troduced an error in the comparison. The average difference of −9.7 ppt could be caused by the uncertainty of the ﬂask and the in situ measurements (where the latter includes the uncertainty of the water vapor correction), the sampling bias of both in situ and ﬂask measurements, as well as the storage effect of air in the ﬂasks. www.atmos-meas-tech.net/9/5293/2016/ L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 10 considers the in situ COS measurements between 25 March and 29 April 2015 and 11 dry air ﬂask samples measured by the same QCLS. The ﬂasks are ﬂushed for an hour before closing, but because of mixing in the ﬂask we assume that the ﬂask sample represents the last 15 min; therefore we average the in situ measurements over these 15 min. The ﬂask samples have an inlet at 60 m height, but because the in situ 60 m measurements only cover the period between 9 and 15 min before ﬂask closure we also include 40 m measurements to cover the last 9 min before the ﬂask is closed. The average difference of COS mole fractions between the 40 and 60 m level is 0.7 ± 9.7 ppt (40–60 m) over the measurement period in March and April. Although the mean difference of 0.7 ppt is not signiﬁcant, the 9.7 ppt stan- dard deviation points to actual differences between the 40 www.atmos-meas-tech.net/9/5293/2016/ www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 3.3 Continuous COS, CO2, CO and H2O observations from Lutjewad peak-to-peak amplitude is estimated to be 96 ppt based on the two-harmonic ﬁt. Kettle et al. (2002) showed that vegetative uptake is the ﬂux with the largest seasonal cycle on the NH, and Montzka et al. (2007) showed that the seasonal ampli- tude of COS depends on the degree to which the sampled air is inﬂuenced by terrestrial ecosystems. It is therefore likely that the seasonal variation of COS observed at the Lutjewad site is inﬂuenced by vegetative uptake. In Fig. 13 we com- pare COS mole fractions from Lutjewad with that from three other sites as measured from ﬂask samples with a GC-MS by NOAA/ESRL Montzka et al. (2007). The ﬂask samples cover data between 2000 and 2015 for Wisconsin, USA (LEF), and Mauna Loa, USA (MLO), and between 2001 and 2015 for Mace Head, Ireland (MHD). These data are an update of those presented in Montzka et al. (2007) (data available at ftp://ftp.cmdl.noaa.gov/hats/carbonyl_sulﬁde/). The ﬂask measurements in Fig. 13 are plotted as function of time of the year. The high-altitude MLO site is less directly inﬂuenced by terrestrial ecosystems and therefore shows only small sea- sonal variation, in contrast to the LEF site, which is largely inﬂuenced by (forested) continental air. The Lutjewad COS mole fractions are most consistent with measurements from MHD, which can be expected since both stations are coastal sites and are located at similar latitudes. The seasonal ampli- tude of COS at MHD and Lutjewad is in between that of LEF and MLO, most likely because both sites are not solely inﬂu- enced by marine or continental air but by both types of air masses. The COS mole fraction has a minimum in Septem- The COS, CO2, CO and H2O data record obtained at the 60 m level of the Lutjewad tower is presented in Fig. 12. The data corrections used for these data were summarized in the previ- ous section. Besides of these corrections we corrected for the dilution of nitrogen due to a leaking solenoid valve before 7 January 2015. We determined a dilution factor by comparing CO2 measurements from the QCLS and a CRDS from the same location and height, under the assumption that without dilution the two analyzers measure the same concentrations (see Sect. 3.2). The percentage of dilution was calculated for CO2 and was typically between 0.4 and 4.9 %. These dilution factors were then applied to all species. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O The mean differences (QCLS−CRDS) are −0.12 ± 0.77 ppm for CO2, −0.9 ± 3.8 ppb for CO and −0.01 ± 0.09 % for H2O. For H2O there were problems with water vapor in the sample lines in March and April, so we only calculated the differ- ence for H2O over the month January. A correlation of the CRDS and QCLS data of both CO2 and CO (slope = 0.98) indicates that there is no concentration-dependent offset. Also, no correlation was found between water vapor and the difference between CRDS and the QCLS CO2 and CO data (CO2: slope = 0.16, R2 = 0.003; CO: slope = 0.48, R2 = 1.9 × 10−5). These results give conﬁdence in the cal- ibration strategy and water vapor correction presented in this study. For the period up to 29 January, the TDLWIN- TEL water vapor correction was turned on, and on top of this correction we applied the linear water correction curves as obtained with the TDLWINTEL correction turned on. If we would not apply this linear correction curve on top of the TDLWINTEL correction, then the median difference be- tween the QCLS and CRDS measurements for this period would be −0.59 ppm for CO2 and −1.1 ppb for CO (against −0.02 ppm and −0.9 ppb for speciﬁcally this period when the linear correction curve is applied). Atmos. Meas. Tech., 9, 5293–5314, 2016 www.atmos-meas-tech.net/9/5293/2016/ L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5310 Figure 11. Minute-averaged measurements of CO2, CO and H2O from the Lutjewad measurement (60 m) tower between 7 January and 29 April 2015 as measured by the QCLS (blue), compared with measurements from a CRDS (orange). The left plots show the time series and the right plots show the difference between CRDS and QCLS measurements together with the frequency distribution of the difference, which is plotted as a histogram. Figure 11. Minute-averaged measurements of CO2, CO and H2O from the Lutjewad measurement (60 m) tower between 7 January and 29 April 2015 as measured by the QCLS (blue), compared with measurements from a CRDS (orange). The left plots show the time series and the right plots show the difference between CRDS and QCLS measurements together with the frequency distribution of the difference, which is plotted as a histogram. www.atmos-meas-tech.net/9/5293/2016/ 3.3 Continuous COS, CO2, CO and H2O observations from Lutjewad COS [ppt] GG G G GG GGG G GG GGG G G G GG G GG GGG GGG G G GG G GG G G G G G G G G G G GG G G G G GGG GGG G G GGG GG GG GG G G G G G G G G GG G G GG G G G G G GG G G G G GG GGGG G G GGGG GGG G G GG GG G G GG GG G G G G GG GGG GGG G GG G G G GGG G G G G GGG G GG G GGGG Jan Apr Jul Oct Jan 300 350 400 450 500 550 600 G LEF MLO MHD LUT GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G Figure 13. COS seasonal cycle of four sites: Wisconsin, USA (LEF); Mauna Loa, USA (MLO); Mace Head, Ireland (MHD); and Lutjewad, the Netherlands (LUT); the data of the latter site are pre- sented in this study. COS mole fractions for the LEF, MLO and MHD sites were measured from ﬂask samples at a GC-MS by NOAA/ESRL (Montzka et al., 2007). The NOAA/ESRL data are shown as ﬂask pair means from individual sampling events. All NOAA measurements are plotted as function of time of the year and cover a period between 2000 and 2015 for LEF, MLO and MHD. In situ COS measurements with the QCLS at the Lutjewad site during 2014–2015 are shown as daily averages (black). A two-harmonic seasonal cycle is ﬁt through the data. COS [ppt] GG G G GG GGG G GG GGG G G G GG G GG GGG GGG G G GG G GG G G G G G G G G G G GG G G G G GGG GGG G G GGG GG GG GG G G G G G G G G GG G G GG G G G G G GG G G G G GG GGGG G G GGGG GGG G G GG GG G G GG GG G G G GG GGG GGG G GG G G G GGG G G G G GGG G GG G GGGG Jan Apr Jul Oct Jan 300 350 400 450 500 550 600 G LEF MLO MHD LUT GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G Figure 13. 3.3 Continuous COS, CO2, CO and H2O observations from Lutjewad COS seasonal cycle of four sites: Wisconsin, USA (LEF); Mauna Loa, USA (MLO); Mace Head, Ireland (MHD); and Lutjewad, the Netherlands (LUT); the data of the latter site are pre- sented in this study. COS mole fractions for the LEF, MLO and MHD sites were measured from ﬂask samples at a GC-MS by NOAA/ESRL (Montzka et al., 2007). The NOAA/ESRL data are shown as ﬂask pair means from individual sampling events. All NOAA measurements are plotted as function of time of the year and cover a period between 2000 and 2015 for LEF, MLO and MHD. In situ COS measurements with the QCLS at the Lutjewad site during 2014–2015 are shown as daily averages (black). A two-harmonic seasonal cycle is ﬁt through the data. Figure 12. Hourly-averaged measurements of COS, CO2, CO and H2O in ambient air at the 60 m level of the Lutjewad measurement tower as measured by the QCLS. Data are shown with corrections as described in the text. ber and October and is a few weeks later than the minimum of the CO2 mole fraction. Montzka et al. (2007) and Blon- quist et al. (2011) also observed a COS minimum later than that of CO2. They reasoned that this difference is due to the fact that at the end of the growing season COS mole fractions keep decreasing due to vegetative uptake without at the same time having a source of COS, whereas during this time of year respiration is beginning to offset assimilation in deter- mining the ambient CO2 mole fractions. surement site. However, as we have not seen indications that the response curve for CO2 changed outside this period, and also taking into account logistical reasons (use of cylinder gases), we suggest calibrating ﬁeld data with a ﬁxed response curve for CO2 and CO as determined once with calibration standards. Second, we investigated the needed frequency of background and reference measurements. Based on labora- tory tests we have shown that background measurements ev- ery 6 h with reference measurements every 30 min (for re- moval of instrument drift) are sufﬁcient to keep the standard deviation of working standard measurements within 5.3 ppt for COS, 0.09 ppm for CO2 and 0.3 ppb for CO over a pe- riod of 19 h. We characterized the water vapor dependence of COS, CO2 and CO from laboratory experiments. 3.3 Continuous COS, CO2, CO and H2O observations from Lutjewad Hourly-averaged measurements of COS, CO2, CO and H2O in ambient air at the 60 m level of the Lutjewad measurement tower as measured by the QCLS. Data are shown with corrections as described in the text. COS [ppt] GG G G GG GGG G GG GGG G G G GG G GG GGG GGG G G GG G GG G G G G G G G G G G GG G G G G GGG GGG G G GGG GG GG GG G G G G G G G G GG G G GG G G G G G GG G G G G GG GGGG G G GGGG GGG G G GG GG G G GG GG G G G G GG GGG GGG G GG G G G GGG G G G G GGG G GG G GGGG Jan Apr Jul Oct Jan 300 350 400 450 500 550 600 G LEF MLO MHD LUT GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G Figure 13. COS seasonal cycle of four sites: Wisconsin, USA (LEF); Mauna Loa, USA (MLO); Mace Head, Ireland (MHD); and Lutjewad, the Netherlands (LUT); the data of the latter site are pre- sented in this study. COS mole fractions for the LEF, MLO and MHD sites were measured from ﬂask samples at a GC-MS by NOAA/ESRL (Montzka et al., 2007). The NOAA/ESRL data are shown as ﬂask pair means from individual sampling events. All NOAA measurements are plotted as function of time of the year and cover a period between 2000 and 2015 for LEF, MLO and MHD. In situ COS measurements with the QCLS at the Lutjewad site during 2014–2015 are shown as daily averages (black). A two-harmonic seasonal cycle is ﬁt through the data. 3.3 Continuous COS, CO2, CO and H2O observations from Lutjewad The location of the Lutjewad station along the coast of the province of Groningen in the Netherlands allows the mea- surement of marine background air during northerly winds and continental air during southerly winds (van der Laan et al., 2009). Daytime CO2 concentrations are typically corre- lated with elevated CO concentrations, indicating the inﬂu- ence of local and regional fossil fuel emissions (van der Laan et al., 2010). Even though the data do not cover a full year cy- cle, it can be seen that the seasonal amplitude of CO2 mole fractions is approximately 15 ppm with a minimum around the end of August. The seasonal variation of CO2 for the Lut- jewad measurement station is analyzed in detail by van der Laan-Luijkx et al. (2010) and Van Leeuwen et al. (2015). The seasonal cycle of COS is captured well by the QCLS and the www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O L. M. J. 3.3 Continuous COS, CO2, CO and H2O observations from Lutjewad Kooijmans et al.: High-precision concentration m G G G GG G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G G G G G G G G GG G G G GG G G G G G G G GG G G GG G G G G GG G GGG G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G GG G G GG G G G G GGG G G G G GG G G G G G G G G G GG G G G G G GG G G G G G G G G G G G GG G G G G G GG G G G GG GG G GG G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG GG GG G G G G G G G G G G GG GG G G G G GG G G G G G GG G G G G G GG G G G G G G G G G G GGG G G GG G G GG GG G GGGG GG G G GG G G G G G G G G G GG GG G G G G G G G GG GG GG G GG GG G GG GG G G G GG G G G G G G G G G G G GG G G G G GG GG G GG G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G GG G G G G GG GG G G G G G G G GG G G GG G G G G G G G G G G G G G G G G GG G G G G G G GG G G G GG G G G G G GG GG GG GG G G GG GG G G G G G GG G G G GG G G G G GG GG GG G G GG GG GG G G GG GG G G G G G GG G G G G GG G G GGGG G GG G G G GG G G GG G G G G G G G GG GG G G GG GGG G G G G G G G G GG GGG GG G G G G G G GG G G G GG G G G G G G G G G GG G G G G G G GG G G G G G G G G GG G GG G G GG G G GG G G G G G G G G G G G G G G G G G G GG G G G GGG G GGGG GG GG G GG G G G G G G G G G G GG G GG G G G GG G G G G G G G G G G GG G G GG G G G G G G GGG GG GG GG GG G G G G G GG G G G GG G G G G G G G G G GG GG G G G G G GG GG GG G G GGG G G GGG G G G G G G GG G G G G G G GG GGG G G GG G G GG G G GGG GG G G GG G G G GG GG GG G G G G GG GGG GG GG G G G G G G G GG GG G G G G G G G G GG G G GG GGG G G G G GG GG GG G G GGG GG G G G GGGGG G G G G G G G G G GG G G G G G GG GG GG GGG GG G G GG G G G GG G G GG G G G G GG G G G G G G G G G G G GG G GG G G G G G G G G G G G GG GG G G G G G GG G G G G GG G G GG G G G G G G GG G G G G G GG G G G G G G G GG G G G G G GG GG GG G G G G GGG G GG G GG GG G G GG G GG GG G G G G GG G GG G G G G G G G G G G G GG GG GG G G G GG GG G G GG G G G G GGG G G G G G G G G GG GG G G G G G G G G G GG G G G G G G G G G G G G G G GG GG GG G G GG G G GG G GG GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G G GG G GG G G G G GG GG G G G G G GG G G GG GG G GG GG G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G GG GG G G G G GG GG G GG G G GG GG G GG GG G G G G G G G G G G GG G G G G GGG G G G G G G G G GGG G G GG GG GG GG GG G GG GG GG G G G GG G G G G GG GG GG GGG G G G G GG G GGGG G G G G G G G G GG GG G G GG G G G G G GG G G GG G G G GG G G G G G G GG G G G GG G G GG G G G G G G G G G G G G G GG G GG GG G G G G G G G G G GGG G G G G G G G GG G G GG GGG G G G G G G G G G G G G G G G G GG G G G G G G G GG G G GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G GG G G G G G G G G G G GG G G G G G G G G G G G GG G GG G G GG G G GGG GG G G G G G G G G G G G GG G G G G GG GG G GG G G G G G G G G G G G GG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G GG GG G G G G G G G G G G G GG G GG G G G G G G G G G G GGG GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G GG G G G G GG G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G GG G G G GG GG G GG GG GG G G GG G G G G G G G G G G G G G G GG G G G G G G G GG GG GG G G G G GG G G G G G GG G G G G G G GG G G G G G G G GG G G G G G G G G G GGG G G G G G G GG G G G GG G G G G G G G G G G G G GG G G G G GG GG G G G GG G G G G GGGG G GG GG G G G G G G G G G G G GG G G G GG G G G GGG G G G G GG G G GGG G G G G G G G G GG GG G GG G G G GG G G G G G G G GG G G G GG G G G G G G G G G GG GG G G G G GG G G G G G G G GG G G GG G G G G G GG G G G G G G G G G G G G G GG G GG G G GG GG G G G G G G GG G G GG G GG G G G G G G G G G G GGG G G G G GG G G G G G GG G G G G G G G G G G G G G GG GG G GG GG G G G G G GG G G G GG GG GG G G G GG G G G G GG G G G G G G G G G GG G G GGG G G GG G G GG G G GGG GG G G G G G GG G G G G G G G GG GG G G G G G G G G G GG G G GG G G GGG G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG GG G G G GG G G G G GG GG G G G G G G G G G G GG GG G G G G GG GGG GG GG G G GG GG GGG G GG G G G GG GGG G G G G G G GG GG G G G G G GG G G G G GG G G G G G G G G G G G G G G G GG G G G G GG GGG GG G G G G G G GGGG G GG G G G G GG G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G GG G GG GG G G G G GG G G G G G G G G G GG GG G G GG G GG GG G G GG GG G G G G G G G GG GG G G GGG G G G G GG G G G G G G G G G G G GG G GG G G GG GGGGG G GGG G G G G GG GG G G G G G G GG G G G G G GG G G GG GG G GG GG G G G G GG G G G G G GG GG G GG GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G GG G G GGG G G G G GG G G G G GG G G GG G G GG G G G G G G G G G G G GG G G G GG GG G G G G G G GG GG G G GG GG G GGG GG G G G G G GG G G G G G GGG G GG GGG G G G G GGGGGGGG G GGGG GG G G GG G G G G G G GG G G G G GG G GGGGG GG G G G G G G GG GGG GG G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G GG G G G G GG GG GG G G GG G G G G G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G GG GG GG G G G GG G G G G G G G GG G G G G G G G G G G G G G GG GG GG G G G G G G G GG G G G G GG G GG G G GG GG G G G G G G G G G GG G G G G GG GG GG G GG G G G G G G G GG GG GG G G G G GG G G G GG G G G G GG G GG G G G G G G GG G G G GG G G GG G G GG G G G GG GG G G GG G G G GG G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G G G GG GG G G G G GG G G G GG GG G G G G G G G G G G GG G G G G G G G G G GG G G G G GG G GG G G GG G G GG G G G G G GG G G GG G G G G G G G G G GG GG GG G G G G G G G GG GG G G G G G G G GG GG GG G G G GG G G GG G G G G G G G G G G G G G G GGG G G GG G G GG G G G G G GG G G GG G G G G G GG GG GG G G G G G G G GG GG GG GG GGG GGG G G G G G G G G G GG G G GG G G GG GG G G G G G GG G G G G G GG GG G G GG GG GGG GG G G GG G G G G GGG G G G G G G G G G GGGG GG G GGG G GGG G GGGG GGGGG GGGGGG GGG G G GGGG GG GGGGGGGGG G G GGG GGG G GGGG GGGGG G GGG GGGGGGGG G G G G GG GG G GGG GGG G GGGGGGGG GGGGG G GG G G G GGG GGG GGGG G G G G G GG G GG G GG G GG G G GGGGGGGGG G G G GGGGGGGG G GGGG G GG G G G GGGG G G GGGG G GG GGGGGG GGGGG GG GGG G G G G G G G G G G GGGG GG G G GG G G G G GGG G GG GGGGG GG GGG GG G G G G G GG G G GG G G G G G G G GG G G GGGGGG GGGGG G G G GGGGG GG GG GGG GGG GG G G GG GGGGGGG GGGG G G G GG G G GG GG GG G G G G G GGGGG GGGGG G GG GGGG GG G GGGG G G G GGGGGG G GG GG G GG GG GGG GGGGGGGG G G G G G G GG G GG GG G G GGG GG GGGG G G G G GG G GGG GGG GGG GG G GG GGGGG G GGGGGG GGGGGGGGGGGGGG GGG G GGGGG GGG GGGGGGGGGGGG GG G 350 450 550 COS [ppt] G G G GG G GG G G G G G G GG G GG GG G G G G G G G G G G G GG GG G G GGG GG G G G G G G G G G G G G G G G GG GG GGG GG G G G G G G G GG G G G G GG GG GG GG G G G G GG GG GG GGG GG GG GG GG GG GG G G G G G G G G G G G G GG GG GG G G GG G G G G G G G G G GG G G G G G G G GG GG GG GGG GG GG GG GG G GG G G G G G G GGG G G G GG G G G G G GG G G GG GG GGG GG GG G G GG G G G G G GG G G G G G G GGG G G GG G G GG G GG GG GG G G GG GGG GG GG GG G G GGG G G G G G G G G G G GGG GG GG GG G G G G G G G G G G G G G GG GGG GG GG GG G G G G G G G G G G G G G G G GGG GG GGG GG GG GG GG G GG GG GG GG GG GG G GG G G G G G G G G G GG GG GG GG GG GGG GG G G GG GG GG G G G GG G G GG GG GGG GG GG GG G G G GG G G G G G G G G GG GGG G G GG GG GG GGG GG G G GGGG GG GGG GG G G G G G G G GG G G GG G G GG GG GGG GG G G G G GGG G G G G GG G G G G G GG GG GG G G G G G G G G G G G G G G G GG G GG GG GG G G G G G GG G G GG G G G G GGG GG GG G G GG G G G G G GG G G G G G G G G G GG GG GG GG GGG GG GG GG GG GGG GGG GG GG GG GG G G G G G G G G G G G G G G G G G GGG G G G G G G G GG GG G G G G G G GGG GG GG G G G G G G G GG GG GG G G GG G GG GG GG G G G G G G G G G G GG G G G GG GG G GGG G GGGGGG G G GG GG GG GG GGG GG GG GG GG GGG GG GG G G G G GG G GG GG GG G G G G G G G G G G G G G G G G GG GG G G G G G G GG G G G GG G G G GG G G G G G G G G G G G G GG GG G G GG GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG GG GG GG GG GGG G G G G G G GG G G G G G G G G G GG GG GGG G G GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GG GG G G GG GGG GG GG GGGG G G GG GG GG GG GGG GG GG GG GG GG GG GGG GG GG G GGGGG GG G GG GG G G G GG GG GG GG G G GG GGG GG G G GG GG G GG G G G G G G G G G G G G G G G GG GG G G G GG G G GG G G G G G G G G G G G GG GG G G G G G GG G G GG GG G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG GGG G G G G G G GG G G G G GG G GG G G GG GGG G G G G G G GG GGG GG GG GG G G GG GGG G G GG GG G G G GG G G GG G G GG GG GGG GG G G GG G G G G GG G G G G G G G GG GG GG G G G G G G G G G GG G G G G G GG GG GG G G GGG G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G GG G G GG G G G G G G G G G G G G G G G GG GG GG G G G GG G G G G GGG GG GG GG GG GGG GG G G G GG GG G G G G G G G G G GG GG G G G G G G G G G G G GG GG G GG G G G G G G G GG GG GG GG GG GGG GG GG GG GG GG G G G GG G G GG G G GG GG G G GG G G G G G GG G G G G GGG G G G G GG GG G G G G G GG GG GG GG GG G GG G G G G GG G G G GG G G GG G G G G GG G G GG GG GG G G G GG GG G G G G G G G G G GG G G G G G G G GG G G GG GG G GG G G G G GG GG G GG G G G G G G GG GG G G GGG G G G G G GG G G G GG G G GG GGG GG GG GG GGG G G G G GG G G G G G G G G GG G G G G G G G G G GG G G G G G G GG G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G GG G G GGG G G GG G G G G G G G G G GG G G G G GG GGG GG GGG G GG GG G G GGGG G G G GG G G GG G G G GG G G G G G G GG GG G GG GG GG GG GGG GG GG GG GG GGG GG GG GG GG GG G G G GGGG G G G G G G G GG G G GG GG GG GGG GG G G G G G G GG G G G GG GG GG GG G GG GG G G GG G G G G G GG GG G G G G GGG G GG GGG GG G G G G G G G GG GG GG GG G G G G G GG G G G G G G G G G G GGG G G GG G G G G G G G GG G G G G G G G G G G G GG GGGG GG G GGGGGGGGGG G G G GGGG G G G G G GG G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG GG G G GG G G G G G G GG G G G GG G G GGG G G GG G G G G G GG GG G G G G GG GGG GG GG GG GG GGG G G GG GG GG GG G GG GG GG G G G G G G G G G G G G GG G G G G G G G G G G G GG G GG G G G G GG GG GG GGGG G G G G GG G G G G G GG G G G G G G G G G G G G G G G G GG G GGG GG GG G G G G GG G G GG G GG GGG G G G G G GG G G G GG G G G G G GGG G G G GG G G GG G G G G G GG GGG G GG GG GG G G G GG GG G G G G G G G G G G G G G G G G G GG G G G GG GG GG GG G G G G G G G G G GG GG G G G GG GG G G GGG GG GGGG G G GG G G G G G GG G G GG G GG GG G G GG GG GG G GG GG GG G G GG G G G G G G G G G G G G G G GG GG G G G G GG G GG GG G G G GGG GG GG GG G G GG G G G GG G G GG GGG G G GG GG GG GG G G G G G GG GG GG GG GGG G G G GGG G GG G G G G GGG G G GG GG GG G G GGG GG GG GG GG GGG GG GG G G GGGG GGG GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GG G GG G G G G GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GGGG GG GG GGG GG GG GG G G GG GGG GG GG GG GG G GG GGG GG GG GG GG GGG GG GG GG GG GGG GG GG GG G G GGGGG GG GG GG G G G G G G G GG G G G G G G G G G GG GG G G G G G GGG G GG G G G G GG G G G G G GG GG GGG G GG G G GGGGG GG GG GG GGG GG GG GG G G GG GGG GG GG G G G G G GG GG GG GG GG GG G GG GG G G GG GGG GG GG GGG G GG G GG G G G G G G GG GGG GGG GG G GG GG GG G G GG GGG GG GG GG G G G G GGG GG GG GG GGG GG GG GG GG G GG GG GG G GG GG GG GG GGG GG GG GG GG GG GGG GG GG GG G G GGG GG GG GG G G GG GG GG GG G G G G G G G GG GG G G G G GG G G G GG GG G G G G G GGG G GGGGGGGGGG GGG GG G GGG G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GG GG G G GGG GG GG G G GG GGG GG G G GG G G G G GG GG GG GG GG G G G GG GG GG GG GG G GG GG G G GG GGGG G G G G GG GG G G G G G GG GG G G G G GG G G G GG G G GGGG GGG GG G G GG GG GGG G G G G GG GG GG GGG GG G G GG G G G G G GG GG G G G G G G G G G G G GG G G G G GG GG GG G G G G G G G G G GG GGG GG G G G G GG G G G G G G G G G G G GG G G G G G G G G G GGG GG GG GG GG G G GG G G G G G GG G G GG G GG GG GG GG GG G G G GG GG GG GG GG GG G GG GG GG GG GGG GG GG G G G G G G GG G GG G G GG GG G G G G G G G G GG GG GG GG GGG GG GG GG GG GGG GG GG G G G G G G G GG GG GG G G G G GG GG G GG GG GG GGG GG GG GG G G GG GGG GG GG G G GG G GG G G G G G G GG G G G G G G G G G G GG G G GGG GG GG G G GG G G GG G G G GG GG GG G G G G G GG GG GG G G G GG GG GG GG GG GGG G GG G G G G GG G G G GG G G GG G G G G GGG GG GG GG G G GGGGG GG G G GG GG GGG G G GG GG GG GG GGG GG GG GG GG G G GGG GG G GGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GG GGGGGGGGGG G GG G G G G GGGGGGG GGGG GGGGGGGG GGGG GGGGG G G G G G G G G G GGG GGG GGG G GG G G G G G G G G G GG GGGGGG G G GGG G GGG G GG G G G GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G G G G G G GG G G G G G GG GGG GGG GG G G G G GG G G G G G GGG G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G G G G GG G G GGGGG G G G G G GG GG G G G G GG G G G G G G GG GGG G G GG G G GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G 380 400 420 440 460 CO2 [ppm] G GG GG G G G G G GG G G GG GGG GG G G G G G G G GG GG GG GG GG G GG GG G G GG GG G GG G G G G GG GG GG GGG GG GG GG GG G G G G G G G GG GG GG GGGG GG GG G G GG GGG GG GG GG GG GG GG G GG G G GG G G GGG GG G G G G G G GG GGG GG GG GG GG G G G G G G G G G GG GG GGG GG GG GG GG G GG GG GG GG G GG G G G GG G G GG G GG G G GG GG GGG GG GG G G G G G GG G G G G G G G G GG G GG G G G G G G G G G G G GG GG GG GG GGG G G GG GG GG GGG GG GG GG GG GG GGG G G GG GG GG G G G GGGG GG GG GG G GG GG GG GG GG GG GG G G G G G G G GG GGG GG GGG GG GG GG GGG G G GG GG GG GG GGG GG G G G G G G GGG GG GG GG GG GG G GG G G GG G G GG GGG GG GG G G GG GG G GG GG GG GG GG GGG GG G G GG GG GG GGG GG GG GG GG GGG GG GG GGGG GG GGG G G GG GG GG GGG GG GG GG GG G G G GG GG G GGG GG G GG GG GG GG G GG GG GG GG GG GG GGG GG GG GG GG GGG GG GG GG GG GGG GG GG GG G G G G GGG GG GG G G G G G GG GG GG GG G G GG GGG GG GG GG GG GGG GG GG GG GGGGG GGG GG GG GG GG GG G G G GG G G G G G G G G G G G GG G G G G G G G G G G G GG G G GG G GG GG G G G G GG GGG G G GG GG GGGG GGG GG GG GG GG GGG GG G GG G G G GG GG GGGG G G GG GGG G G GG GG GG GG GGG GG GG GG GG GGG G G GG GG GG GGG GG GG GG GG G G G G G GG G G G G G G G GG GG GG GG G G GG GG G GG GG G G G G G G G G GG GG GGG G G GG GG GG G G GGG GG GG G G GG G G G G G G G G GG GG G G GG G G G GG GG GG GG GGG G G G G GG GG GGG GG GGGG GG G G GGG GG GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG GG GGGG GG GG GG GG GG G GG GG GG GG GG GG GGG GG GG G GGGGG GG GGG GG GG GGG GG GG GG G G GG GGG GG GG GG G G GGG GG G G G G G G G G G G G G G G G G G GG GG G GG GG G G G G G G G G G GG GG G G G G G G G G G G G G G G G GG G G GG G G G G G GG G G G G G G G G G G G GG G G GG GGG G G GG G G G G G GG G G G G GG GG G G GGG G G GG GG GG GGG GG GG GG G G GG GGG GG G G GG G G GGG GG GG G G GG G G GGG GGG G GG GG GG GGGG G GG GG GG GG G G G G G GG GG G G G G GG G G G G G GG GG G G G GG G G GG G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G GG GG G G GG GG GG G G G G G GG G G GG GG G G G GG GG G G G G GG GGG GG GG GG GG GGG GG GG G G G GG GG G G GGG GG GG GG GG G G G G G G G GG GG G G G G G G G GG GG G GG GG GG GG GG GGG GG GG GG GG GG G G G GG G G GG GG GG GG G G GG G G G GG GG GG G G GG G G G G G G G G G GG G G G G G G G G G GG GG G GG GG G G GG GG GGG GG GG GG G GGGG G G G G G GG GGG G G G G G G G G G G G GG G G G G GG G G G G G G G G G G GG G G G G G G GG GG GGG GG GG GG G G GG G GGGGG G G GG GG G G GGG G G GG GGG GG G G GG GGG GG G G GG G G GG G G GG GG G G G G GG G G G GG G G GG GG GG GG G G G G G GG GG G G G G G G G GG G G G G G G G GG G G G G GG GGG G G GG GG GGGG GGG G G G G GG GG GGG GG GGG G GG GG GG GGG G G GG GG GG GG G G G G G GG G G G G GG GG G G G GG GG GG GGG GG GG GG GG GGG GG GG GG GG GG G G G GGG G G G GG G G G GG G G G G G G GG G GG GG G G G G GG G G G GG GG G G G G GG G GG GG GG G G G G GG G G G GG G G G G GG G GGG GG G GG G GGG G GG G G GG G G G G G G G GG GG G G G G G G G G G G GGG G G GG G G G G G G G GG G G G G GG GG G GG GG GGGG G G G GGGGGGGGGGG GG GGGGG G G G GG GG G G G G G GG G G GG G G G G G GG G G GG GG GGG GG G GGGGGG G G G GG GG G GG G G G G GG G G G GG GG GG G G G G G GG GG GG G G GG GGG G G G G GG GG GGG GG GG GG GG GGG GG GG GG G G G G GG G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G GG G G GG G G G G GG GG GG G GG G G GG G G G G GG G G G G G G G GG GGG G G GG G G G G GG G G G G G GG GG G GG G G G G GG G G GG G G G G GGGG G G G G G G G G G GG G G G G G GGG G GG G G GG G G G GG GG GG G G G G GG G G G G G G G GG G G G G G GG GG G G GG G G G G G G G G G G G G G G G G G G GG G G G G G GG GG G G G G G GG G G G G G G G G GG GGG GG G G G G GG GG GG G GG GG G G GG G G G G G GG GG GG G G G GG GG GG G G GGG GG G G G G G GGG GG GG GG G G G G G G G G G G G GG G GG G G GG GG GG GG G G G G G G G G G GG G G GGG G G G G GG G GG GG G G GGG G G G G GG GG GG GGG GG GG GG GG G GG GG GG G G G GGG GGG GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG G G GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GGGG GG GG GGG GG GG GG GG G G GGG GG GG GG GG G G G GGG G G GG GG GG GGG GG GG GG GG G GG GG GG GG G G GGGGG GG G G GG G G G G G G G GG G G G G G G G GG G G G G G G G G G G GG G G G G G G G GG G G G G G GG GG GGG G GG G G GG GGG GG GG G G GG G GG GG GG G G G G GGG GG GG G G G G G GG GG GG GG G G G G G GG GG G G G G GGG G G GG GGG GGG G G G G G G G G G G G G G G GGG GG G G G G G G G G G G G G GG GG GG G G GG G G GGG GG GG GG GGG GG GG GG GG G GG GG GG G GG G G G G GG GGG GG GG GG GG GG GGG GG GG GG G G G G G GG GG GG G G GG GG GG GG G G G G G G G G G G G G G G G G G G G G GG GG GG G G G GGG G GGGGGG GGGG G GG G GGGG G G G GG GGGGGGGGGGGG G G GGGGGGGGGGGGGG GGG GGGG GGGGG GGGGGGG G GG GGGG GGGGG GGGGGG GG G GGGGG G G G G GG G G GGG GG G G GG GG GGG GG G G GG G G G G GG GG G G GG GG G GG GG GG G G G G G G G GG GG G G GG G G GG G G G G GG GG GG G G G GG GG G G GG GG G G G G G G G G G GG G GG G G G G GG G G GGG G G GG GG G G GG GGG G G G G GG G G GG G GG GG G G G G GG G G G GG GG G G G G G G G G G G G G G G G G G GG GG G G G GG G G G G G G G G G G G G G G G G G G G GG G GG G G G G G GG G G GG GG GG G G GG G G G G G G GG G GG G G G G G G G GG G G G G G GG GG GG GG GG GG G GG GG GG GG G GG GG G G G G G G G G GG G GG G G G G G G G G G G G GG G GG G G GG GG GGG GG GG GG GG GGG GG GG G G G G G G G GG G G G G G G G G G G GG G GG G G GG G GG GG GG GG G G GG GGG GG GG G G G G G GG G G G G G G GG G G G G G G G G G G G G G G GG G GG GG G G GG GG GG G G G G G G G GG G G G G G GG GG GG G G G G G GG G G G G GG G GG GGG GG G G G G G G G G G G G G G GG G G GGG G G GG GG G G GGGGG GG GG G G GG G GG G G G G GG GG GG GGG GG GG GG G G G G GGG G G GG GGGGGGG GGG G GGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGG G GGGGGGGGGG GGGGGGGGGGGGGG G G G G G G G G GGG G GG GGGG GGGGGGGGGG GGGGGGG G G G G GG G G G GGGGGGGGG GGG G G G G G G G G GG G GGGGGGG G GGG G G G G G GG GG G G G G GG G G G G G GGGGGGGGG GGGGGGGGGGGGGGG GG GG G GGGG G G G G G G GG GGGGGGGGGGG G G G G G G GGG G GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G G G G G GG G GG G G GG G GG G G G G GG G GG G G G G G G G G G G GGG GG GG G GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG G 100 200 300 CO [ppb] G G GG G G G G G G G G G G G G G G G G G GG G G G G GGG G G GG G G GG G G GGG G G G G G G G G G GG GG G G G G GG GG GGG GG GG GG G G G G G G G GG G G G G G G GGG G G G G G GG GG G GG GG G G G G GG GG G G G G G G G GG G G GGG G G G G G G G G GG G G G G G GG GG GG GG G GG G G G G G G G G GGG G G GG G G G G G GG GG G G G G GG G G G G GG GG GG G GG GG G G G G G G G G G GG G G GG GGG G G GG GG GG G G GG G GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G GG G G G G G G GG G GG G G G G GG G G G GGGG G G G G G G G G G GG G G G G G G G G GG G GG G G G G GG G G G GG GG G G G G G G G GGG G G G G G G G G G G G G G G G G G G G GG GG G GG GG GG GG G G G G G G G G G G G GG GGG G G GG G G G G GG G G G G G G G G G GG GGG GG GG G G GG G G G GG GG G G GG G G GGG GG GG G GGG G G G G G G G GG GG GG GGG G G GG GG G G G G G G G GG G GG G GGGGG G G G G G G GG G G G G G G G G G GG G G G GG GG G G G G G GG GG GG GG G G GGG GG G G G G G G G G G G G GG G G G G G G GGG GG GG GG G G G G G GG G G GG G G G G G G G GG G G GG G G G GG GGG GG GG GG G G G GG GG G G G G G G GG GGG G GGGGG GG GGG GG GG GG GG GG GGG GG G G G G GG GG G GG GG G G GGG G G G G G G G G G G GG G G G G G GG G G G G GG G G G G G G G G GG G G GG G GG GG GG G G G G GGG GG GG GG GG G G G G G GG GG G G GGG G G G G GG G G GG G G G G G G G GG G G GGG GG G G G G G G GG GGG GG G G GGG G GG G G G G GG G GG GG GG GG GG GG GGG GG GG GG GGG GG GG GG G GG GG GG GG G GG GG GG GG GG GGG GG GG G G GG G G G GG GG GG GG GG GG G G G G G G G G G G GG G G GG G G GG G G G G G GG G G G G G G G G G GG G G G G G G GG GGG G G GG G G GG GGG GG GG GG GG G G G G GG G G G G G G G G G G G G G GG G G GG GGG GG GG G G G GGG GG GGG G G GG GG G GG G G G G GG GG G GG G G G G G G GG GGG GG G G G G G G G GG GG GG GG GG GG GGG GG G G GG G G G GG GG GG GG GG GG GGG GG GG GG GG GGG G GG G G G G G G G G GGG G GG GG GG GG G G GG G GG GG GG GG GG GGG G GG G GG G G GG G G G GG G G GG G G G G G G G G G G G GG G G G GG GG G G G G G G GGG GG GG GG GG GG GG G GG G G G G GG GG GG GG G G G G G GG G G G G GG G GG G G GG GG G G G GG G G GG G G GG GGG G G GG GG G G G G G GG GG GG GG GG GGGGG GG GG GG GGG GG GG GG GG GGG G GG G G G G G G G G G GG GG GG GGG GG GG GG GG GGG G G G G GG GG GGG GG GG GG G GG GG GG GGG G G G G GG G G GGG GG G G GG GG G G G GG G G GG GG G G G GG G G G G G G G GG GG GG GG G G G G G GG GG GG GG G G GG GG G G G G GGG G G G G G G G G GGG GG GG GG GG GG GG G G G G G GG GG GGG GG GG GG GG GG GGG G G G G GG G G GGG G GG GG GG G G G G G GG G G G G G GG G G G G G G GGG GG GG G G GG G G G G G GG G G GG G G G GG GG GG GG GG GG G GGGGG GG G G G G G G GGG G G GG GGGG G G G G GG G G G GG G G G G GG GGG G G G GG GG GG G G GG G G G GG GG G G GG GGG G GGGGG GG GG G GG GG GG GG GG G G G GG GG GG GG GGG GG GG GG GGGG GGG GG GG GG GG GGG GG GG G G GG GG G G GG G G G GG GG GG GG GG G GG GG G G GG GG G GG G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G GG G G G GG GG GG GGG GG GG GG GG GG GG G GG GG GG GG GG G GG GG GG GG GG G GG GG G G G G G G G G G G G GG GG G G G GG GGG GGG GGGGG GGGG G G GG GG GG GG G GG GG GG G G GG GG GG GGG GG GG G G G G G G GG G G G G G GG G G GGG GG GGGG G G G GG G G G G G GG GGGG GGGGGG G G G GG GG GGGGGG GG GG G G G GG GG G G GG GG GGG GG GGGG GGG GG G G G G G G G G GGGG GG GG GG GGG GG GG GG GG G G G GG GG GG G G GG G G G G G GG GG GGG GG GG GG GG GG G GG G G GG GG G G G GG G G G G G GG G G GG GG GGG G G GG GGG GG GG GG GG GG GG GG G G G G GG GG GGG GG G G GG GG GG G G G G G G G G GGGG GG GG GG GG GG GG GG GG G G G GGG G G G G G G G G GG GG GG G G GGGGG GG G G G G GG GGG GG GG GGG G GG GG GG GG G GG GG GG G G GG GG G GG GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG G G GG GGG G G G GGG GG GGG G G G G GG GG G G GGG GG GG G G G G GG G GG GG GG G G GG GG G GG GG GG GG GGG G G G G G G GG GGG GG GG G G G G GG GG G G G G GG GGG GG GG G G GG GG G GG GG GG GG GG GG G GG GG GG GG GG GGG GG G G GG GG GG GG GGGG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG G G GGGG GGG GG GG GG GGG GG GG GG GG GG GGG GG G G G G GG GGG GG G G GG GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GGGG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GGG GG GG GG GG GGG GG GG GG GG GGG G G GG GG G G GGG GG GG GG GG GG G GG GG GG G G GG G G GGG GG GG GG GG GGGGG GG GG GG GG G GG G G GG G G GGG G GG G G G GGGG GG GG GG G G G GG G G GG G G GG G GG GG GG GG G G G G G GG G G GG GG GG G GG G G G G GG GGG GG GG GGGGGG GGG GG GG G G GG GG GGG G G G G GG GG GGG GG GG GG G G GGG G G G G G G G G GG GGG GG GG GG GGG GG GG GG GG G GG GG GG GGG GG GG GG GGG GG GG GG GG GG GGG GG GG G G GG GGG GG GG GG G G GG GGGG GG GG GG GGG GG GG GG GG GG G G G GG G G GG GG G GGGGG GGGGGGGG GG G G GG GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGG GGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG G G GG GG GG GGG G G G G GG GG GG G GG GG GG G G GG G G G G G G GG GG G GG GG GG GG GG G G G GG GG GG GG GGGG GG G G GG GG GG G G G GG GG GG GG GG G GG GG G GGGGG GGG GG G G GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GG GG GGG G GG GG GG GG G GG GG GG GG GGG GG GG G G GG GG G GG GG GG GG GG GGG GG G G GG GGG GG GG GG GG GG G G G G G GG GGGG GG GGG GG GG GG GG GG GGG GG GG G G GG GG G GG GG GG GG GGG GG GG GG GG GG GGG GG GG G G GG GGG GG GG G GG GG GG GG GGG GG G G G G G G G GG GG GG G G G G GG G GG GG GG GG GG GG G G G GG GG GG GGG GG GG GG G G G G GGG G G GG GG G G GGG GG GG G G G GG G G GG GG GG G GG GG GG GG GGG GG GG GG GG G G GGG G G G G GG GG GGG GG GG GG GG G G GGG G G GG GG GG GGG G G G GG GG G G G G GGG GG GG GG GG GGG GG GG GG G G G G G G G GG GG GG GG GGG GG GG GG GG G G GGG GG GG GG GG G G GG G G G G G GGG GGGGGGGGGGGGGGGGGGG GGG G G GGGGGG GGGGGG GGGGGGGGG G G GGGGG GGGGGGGGG GG GG G G G GG G GGGGGG G GG GGGG GGGGGG G G GG GGGGG GG GG GGGGGGGG GG G G G GG GGGGGGGGGGGGGGGGGGGGGG GGGGGGGGG G GG GG GGGGGGG G GG G G G GG GG GGGG GGGGGGGGGG G GGGGG GGGGGG G G GGGGG G G G GG GGGG GGG G G G GGGGGGG G G G G GGGGGGGGGG G G G GG GG GGGGGGGG G GGG GGGGG G G G GGGGGGGGG G GGGGGGGGGG G G G GG G GGGG G GGGGGGGGGGGGG GGGGGGGG G G GGGGGGGG GGGGGGGGGGGG G GGGGG GGGG G G G G GGGGGGGGGG G GGG G GGGGGG G G G GGGG GG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GGGGG G G G GGGGG G GGGGGGG G G GG G G G GGGGGG G GG GG G GGG GG G G 0.0 0.5 1.0 1.5 2.0 H2O [%] Aug'14 Sep'14 Oct'14 Nov'14 Dec'14 Jan'15 Feb'15 Mar'15 Apr'15 May'15 Figure 12. 3.3 Continuous COS, CO2, CO and H2O observations from Lutjewad Based on an assessment of the TDLWINTEL water correction we determined optimal broadening coefﬁcients for the use of the water correction within TDLWINTEL. Besides that, we pre- sented an alternative water correction based on linear depen- dence of the wet air mole fractions with H2O concentration. Furthermore, we demonstrate that a small H2O peak close to the COS peak has caused a water-vapor-dependent concen- tration error that is larger than the direct water vapor dilu- tion effect. This water vapor interference can be minimized by careful adjustments to the software ﬁtting parameters and 3.3 Continuous COS, CO2, CO and H2O observations from Lutjewad Hourly-averaged measurements of COS, CO2, CO and H2O in ambient air at the 60 m level of the Lutjewad measurement tower as measured by the QCLS. Data are shown with corrections as described in the text. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5311 L. M. J. 3.3 Continuous COS, CO2, CO and H2O observations from Lutjewad Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5311 G G G GG G G G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G G G G G G G G GG G G G GG G G G G G G G GG G G GG G G G G GG G GGG G G G G G G G G G G G G G G G G G GG G G G G G G G GG G G G G G G G GG G G GG G G G G GGG G G G G GG G G G G G G G G G GG G G G G G GG G G G G G G G G G G G GG G G G G G GG G G G GG GG G GG G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG GG GG G G G G G G G G G G GG GG G G G G GG G G G G G GG G G G G G GG G G G G G G G G G G GGG G G GG G G GG GG G GGGG GG G G GG G G G G G G G G G GG GG G G G G G G G GG GG GG G GG GG G GG GG G G G GG G G G G G G G G G G G GG G G G G GG GG G GG G G G G GG G G G G G G G GG G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G GG G G G G GG GG G G G G G G G GG G G GG G G G G G G G G G G G G G G G G GG G G G G G G GG G G G GG G G G G G GG GG GG GG G G GG GG G G G G G GG G G G GG G G G G GG GG GG G G GG GG GG G G GG GG G G G G G GG G G G G GG G G GGGG G GG G G G GG G G GG G G G G G G G GG GG G G GG GGG G G G G G G G G GG GGG GG G G G G G G GG G G G GG G G G G G G G G G GG G G G G G G GG G G G G G G G G GG G GG G G GG G G GG G G G G G G G G G G G G G G G G G G GG G G G GGG G GGGG GG GG G GG G G G G G G G G G G GG G GG G G G GG G G G G G G G G G G GG G G GG G G G G G G GGG GG GG GG GG G G G G G GG G G G GG G G G G G G G G G GG GG G G G G G GG GG GG G G GGG G G GGG G G G G G G GG G G G G G G GG GGG G G GG G G GG G G GGG GG G G GG G G G GG GG GG G G G G GG GGG GG GG G G G G G G G GG GG G G G G G G G G GG G G GG GGG G G G G GG GG GG G G GGG GG G G G GGGGG G G G G G G G G G GG G G G G G GG GG GG GGG GG G G GG G G G GG G G GG G G G G GG G G G G G G G G G G G GG G GG G G G G G G G G G G G GG GG G G G G G GG G G G G GG G G GG G G G G G G GG G G G G G GG G G G G G G G GG G G G G G GG GG GG G G G G GGG G GG G GG GG G G GG G GG GG G G G G GG G GG G G G G G G G G G G G GG GG GG G G G GG GG G G GG G G G G GGG G G G G G G G G GG GG G G G G G G G G G GG G G G G G G G G G G G G G G GG GG GG G G GG G G GG G GG GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G G GG G GG G G G G GG GG G G G G G GG G G GG GG G GG GG G G G GG G G G G G G G G G GG G G G G G G G G G G G G G G G GG G G G G G G G G G G GG GG G G G G GG GG G GG G G GG GG G GG GG G G G G G G G G G G GG G G G G GGG G G G G G G G G GGG G G GG GG GG GG GG G GG GG GG G G G GG G G G G GG GG GG GGG G G G G GG G GGGG G G G G G G G G GG GG G G GG G G G G G GG G G GG G G G GG G G G G G G GG G G G GG G G GG G G G G G G G G G G G G G GG G GG GG G G G G G G G G G GGG G G G G G G G GG G G GG GGG G G G G G G G G G G G G G G G G GG G G G G G G G GG G G GG G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG GG G G G GG G G G G G G G G G G GG G G G G G G G G G G G GG G GG G G GG G G GGG GG G G G G G G G G G G G GG G G G G GG GG G GG G G G G G G G G G G G GG G G G G G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G G G GG G G GG GG G G G G G G G G G G G GG G GG G G G G G G G G G G GGG GG G G G G G G G G G G G G G G G G G G G GG G G G G G G G G G G G G G G G G G G G G GG GG G G G G G G G G G GG G G G G GG G G G G G G G G GG G G GG G G G G G G G G G G G G G G G G G G G G G G G G G G G GG G G G GG G G G G G G GG G G G GG GG G GG GG GG G G GG G G G G G G G G G G G G G G GG G G G G G G G GG GG GG G G G G GG G G G G G GG G G G G G G GG G G G G G G G GG G G G G G G G G G GGG G G G G G G GG G G G GG G G G G G G G G G G G G GG G G G G GG GG G G G GG G G G G GGGG G GG GG G G G G G G G G G G G GG G G G GG G G G GGG G G G G GG G G GGG G G G G G G G G GG GG G GG G G G GG G G G G G G G GG G G G GG G G G G G G G G G GG GG G G G G GG G G G G G G G GG G G GG G G G G G GG G G G G G G G G G G G G G GG G GG G G GG GG G G G G G G GG G G GG G GG G G G G G G G G G G GGG G G G G GG G G G G G GG G G G G G G G G G G G G G GG GG G GG GG G G G G G GG G G G GG GG GG G G G GG G G G G GG G G G G G G G G G GG G G GGG G G GG G G GG G G GGG GG G G G G G GG G G G G G G G GG GG G G G G G G G G G GG G G GG G G GGG G G G G G G G G G G G G G G G G G G G G GG G G G G G G GG GG G G G GG G G G G GG GG G G G G G G G G G G GG GG G G G G GG GGG GG GG G G GG GG GGG G GG G G G GG GGG G G G G G G GG GG G G G G G GG G G G G GG G G G G G G G G G G G G G G G GG G G G G GG GGG GG G G G G G G GGGG G GG G G G G GG G G G G G G G G G G G GG G G G G G G G G G G G G G G G G GG G G G G GG G GG GG G G G G GG G G G G G G G G G GG GG G G GG G GG GG G G GG GG G G G G G G G GG GG G G GGG G G G G GG G G G G G G G G G G G GG G GG G G GG GGGGG G GGG G G G G GG GG G G G G G G GG G G G G G GG G G GG GG G GG GG G G G G GG G G G G G GG GG G GG GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G G G GG G G GGG G G G G GG G G G G GG G G GG G G GG G G G G G G G G G G G GG G G G GG GG G G G G G G GG GG G G GG GG G GGG GG G G G G G GG G G G G G GGG G GG GGG G G G G GGGGGGGG G GGGG GG G G GG G G G G G G GG G G G G GG G GGGGG GG G G G G G G GG GGG GG G G G G G G G G G G GG GG G G G G G G G G G G G G G G G G GG G G G G GG GG GG G G GG G G G G G G G GG G G G G G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G GG GG GG G G G GG G G G G G G G GG G G G G G G G G G G G G G GG GG GG G G G G G G G GG G G G G GG G GG G G GG GG G G G G G G G G G GG G G G G GG GG GG G GG G G G G G G G GG GG GG G G G G GG G G G GG G G G G GG G GG G G G G G G GG G G G GG G G GG G G GG G G G GG GG G G GG G G G GG G G G G GG G G G G G G G G G G G G G G G G G G G G GG G G G G G GG GG G G G G GG G G G GG GG G G G G G G G G G G GG G G G G G G G G G GG G G G G GG G GG G G GG G G GG G G G G G GG G G GG G G G G G G G G G GG GG GG G G G G G G G GG GG G G G G G G G GG GG GG G G G GG G G GG G G G G G G G G G G G G G G GGG G G GG G G GG G G G G G GG G G GG G G G G G GG GG GG G G G G G G G GG GG GG GG GGG GGG G G G G G G G G G GG G G GG G G GG GG G G G G G GG G G G G G GG GG G G GG GG GGG GG G G GG G G G G GGG G G G G G G G G G GGGG GG G GGG G GGG G GGGG GGGGG GGGGGG GGG G G GGGG GG GGGGGGGGG G G GGG GGG G GGGG GGGGG G GGG GGGGGGGG G G G G GG GG G GGG GGG G GGGGGGGG GGGGG G GG G G G GGG GGG GGGG G G G G G GG G GG G GG G GG G G GGGGGGGGG G G G GGGGGGGG G GGGG G GG G G G GGGG G G GGGG G GG GGGGGG GGGGG GG GGG G G G G G G G G G G GGGG GG G G GG G G G G GGG G GG GGGGG GG GGG GG G G G G G GG G G GG G G G G G G G GG G G GGGGGG GGGGG G G G GGGGG GG GG GGG GGG GG G G GG GGGGGGG GGGG G G G GG G G GG GG GG G G G G G GGGGG GGGGG G GG GGGG GG G GGGG G G G GGGGGG G GG GG G GG GG GGG GGGGGGGG G G G G G G GG G GG GG G G GGG GG GGGG G G G G GG G GGG GGG GGG GG G GG GGGGG G GGGGGG GGGGGGGGGGGGGG GGG G GGGGG GGG GGGGGGGGGGGG GG G 350 450 550 COS [ppt] G G G GG G GG G G G G G G GG G GG GG G G G G G G G G G G G GG GG G G GGG GG G G G G G G G G G G G G G G G GG GG GGG GG G G G G G G G GG G G G G GG GG GG GG G G G G GG GG GG GGG GG GG GG GG GG GG G G G G G G G G G G G G GG GG GG G G GG G G G G G G G G G GG G G G G G G G GG GG GG GGG GG GG GG GG G GG G G G G G G GGG G G G GG G G G G G GG G G GG GG GGG GG GG G G GG G G G G G GG G G G G G G GGG G G GG G G GG G GG GG GG G G GG GGG GG GG GG G G GGG G G G G G G G G G G GGG GG GG GG G G G G G G G G G G G G G GG GGG GG GG GG G G G G G G G G G G G G G G G GGG GG GGG GG GG GG GG G GG GG GG GG GG GG G GG G G G G G G G G G GG GG GG GG GG GGG GG G G GG GG GG G G G GG G G GG GG GGG GG GG GG G G G GG G G G G G G G G GG GGG G G GG GG GG GGG GG G G GGGG GG GGG GG G G G G G G G GG G G GG G G GG GG GGG GG G G G G GGG G G G G GG G G G G G GG GG GG G G G G G G G G G G G G G G G GG G GG GG GG G G G G G GG G G GG G G G G GGG GG GG G G GG G G G G G GG G G G G G G G G G GG GG GG GG GGG GG GG GG GG GGG GGG GG GG GG GG G G G G G G G G G G G G G G G G G GGG G G G G G G G GG GG G G G G G G GGG GG GG G G G G G G G GG GG GG G G GG G GG GG GG G G G G G G G G G G GG G G G GG GG G GGG G GGGGGG G G GG GG GG GG GGG GG GG GG GG GGG GG GG G G G G GG G GG GG GG G G G G G G G G G G G G G G G G GG GG G G G G G G GG G G G GG G G G GG G G G G G G G G G G G G GG GG G G GG GG G G G G G G G G G G G G G G G G GG G G G G G G G G G G GG GG GG GG GG GGG G G G G G G GG G G G G G G G G G GG GG GGG G G GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GG GG G G GG GGG GG GG GGGG G G GG GG GG GG GGG GG GG GG GG GG GG GGG GG GG G GGGGG GG G GG GG G G G GG GG GG GG G G GG GGG GG G G GG GG G GG G G G G G G G G G G G G G G G GG GG G G G GG G G GG G G G G G G G G G G G GG GG G G G G G GG G G GG GG G G G G G G G G G G G GG G G G G G G G G G G G G G G G GG GGG G G G G G G GG G G G G GG G GG G G GG GGG G G G G G G GG GGG GG GG GG G G GG GGG G G GG GG G G G GG G G GG G G GG GG GGG GG G G GG G G G G GG G G G G G G G GG GG GG G G G G G G G G G GG G G G G G GG GG GG G G GGG G G G G G G G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G GG G G GG G G G G G G G G G G G G G G G GG GG GG G G G GG G G G G GGG GG GG GG GG GGG GG G G G GG GG G G G G G G G G G GG GG G G G G G G G G G G G GG GG G GG G G G G G G G GG GG GG GG GG GGG GG GG GG GG GG G G G GG G G GG G G GG GG G G GG G G G G G GG G G G G GGG G G G G GG GG G G G G G GG GG GG GG GG G GG G G G G GG G G G GG G G GG G G G G GG G G GG GG GG G G G GG GG G G G G G G G G G GG G G G G G G G GG G G GG GG G GG G G G G GG GG G GG G G G G G G GG GG G G GGG G G G G G GG G G G GG G G GG GGG GG GG GG GGG G G G G GG G G G G G G G G GG G G G G G G G G G GG G G G G G G GG G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G GG G G GGG G G GG G G G G G G G G G GG G G G G GG GGG GG GGG G GG GG G G GGGG G G G GG G G GG G G G GG G G G G G G GG GG G GG GG GG GG GGG GG GG GG GG GGG GG GG GG GG GG G G G GGGG G G G G G G G GG G G GG GG GG GGG GG G G G G G G GG G G G GG GG GG GG G GG GG G G GG G G G G G GG GG G G G G GGG G GG GGG GG G G G G G G G GG GG GG GG G G G G G GG G G G G G G G G G G GGG G G GG G G G G G G G GG G G G G G G G G G G G GG GGGG GG G GGGGGGGGGG G G G GGGG G G G G G GG G GG G G G G G G G G G G G G G G G G G G GG G G G G G G G G G GG GG G G GG G G G G G G GG G G G GG G G GGG G G GG G G G G G GG GG G G G G GG GGG GG GG GG GG GGG G G GG GG GG GG G GG GG GG G G G G G G G G G G G G GG G G G G G G G G G G G GG G GG G G G G GG GG GG GGGG G G G G GG G G G G G GG G G G G G G G G G G G G G G G G GG G GGG GG GG G G G G GG G G GG G GG GGG G G G G G GG G G G GG G G G G G GGG G G G GG G G GG G G G G G GG GGG G GG GG GG G G G GG GG G G G G G G G G G G G G G G G G G GG G G G GG GG GG GG G G G G G G G G G GG GG G G G GG GG G G GGG GG GGGG G G GG G G G G G GG G G GG G GG GG G G GG GG GG G GG GG GG G G GG G G G G G G G G G G G G G G GG GG G G G G GG G GG GG G G G GGG GG GG GG G G GG G G G GG G G GG GGG G G GG GG GG GG G G G G G GG GG GG GG GGG G G G GGG G GG G G G G GGG G G GG GG GG G G GGG GG GG GG GG GGG GG GG G G GGGG GGG GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GG G GG G G G G GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GGGG GG GG GGG GG GG GG G G GG GGG GG GG GG GG G GG GGG GG GG GG GG GGG GG GG GG GG GGG GG GG GG G G GGGGG GG GG GG G G G G G G G GG G G G G G G G G G GG GG G G G G G GGG G GG G G G G GG G G G G G GG GG GGG G GG G G GGGGG GG GG GG GGG GG GG GG G G GG GGG GG GG G G G G G GG GG GG GG GG GG G GG GG G G GG GGG GG GG GGG G GG G GG G G G G G G GG GGG GGG GG G GG GG GG G G GG GGG GG GG GG G G G G GGG GG GG GG GGG GG GG GG GG G GG GG GG G GG GG GG GG GGG GG GG GG GG GG GGG GG GG GG G G GGG GG GG GG G G GG GG GG GG G G G G G G G GG GG G G G G GG G G G GG GG G G G G G GGG G GGGGGGGGGG GGG GG G GGG G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GG GG G G GGG GG GG G G GG GGG GG G G GG G G G G GG GG GG GG GG G G G GG GG GG GG GG G GG GG G G GG GGGG G G G G GG GG G G G G G GG GG G G G G GG G G G GG G G GGGG GGG GG G G GG GG GGG G G G G GG GG GG GGG GG G G GG G G G G G GG GG G G G G G G G G G G G GG G G G G GG GG GG G G G G G G G G G GG GGG GG G G G G GG G G G G G G G G G G G GG G G G G G G G G G GGG GG GG GG GG G G GG G G G G G GG G G GG G GG GG GG GG GG G G G GG GG GG GG GG GG G GG GG GG GG GGG GG GG G G G G G G GG G GG G G GG GG G G G G G G G G GG GG GG GG GGG GG GG GG GG GGG GG GG G G G G G G G GG GG GG G G G G GG GG G GG GG GG GGG GG GG GG G G GG GGG GG GG G G GG G GG G G G G G G GG G G G G G G G G G G GG G G GGG GG GG G G GG G G GG G G G GG GG GG G G G G G GG GG GG G G G GG GG GG GG GG GGG G GG G G G G GG G G G GG G G GG G G G G GGG GG GG GG G G GGGGG GG G G GG GG GGG G G GG GG GG GG GGG GG GG GG GG G G GGG GG G GGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GG GGGGGGGGGG G GG G G G G GGGGGGG GGGG GGGGGGGG GGGG GGGGG G G G G G G G G G GGG GGG GGG G GG G G G G G G G G G GG GGGGGG G G GGG G GGG G GG G G G GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G G G G G G GG G G G G G GG GGG GGG GG G G G G GG G G G G G GGG G G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G G G G GG G G GGGGG G G G G G GG GG G G G G GG G G G G G G GG GGG G G GG G G GGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G 380 400 420 440 460 CO2 [ppm] G GG GG G G G G G GG G G GG GGG GG G G G G G G G GG GG GG GG GG G GG GG G G GG GG G GG G G G G GG GG GG GGG GG GG GG GG G G G G G G G GG GG GG GGGG GG GG G G GG GGG GG GG GG GG GG GG G GG G G GG G G GGG GG G G G G G G GG GGG GG GG GG GG G G G G G G G G G GG GG GGG GG GG GG GG G GG GG GG GG G GG G G G GG G G GG G GG G G GG GG GGG GG GG G G G G G GG G G G G G G G G GG G GG G G G G G G G G G G G GG GG GG GG GGG G G GG GG GG GGG GG GG GG GG GG GGG G G GG GG GG G G G GGGG GG GG GG G GG GG GG GG GG GG GG G G G G G G G GG GGG GG GGG GG GG GG GGG G G GG GG GG GG GGG GG G G G G G G GGG GG GG GG GG GG G GG G G GG G G GG GGG GG GG G G GG GG G GG GG GG GG GG GGG GG G G GG GG GG GGG GG GG GG GG GGG GG GG GGGG GG GGG G G GG GG GG GGG GG GG GG GG G G G GG GG G GGG GG G GG GG GG GG G GG GG GG GG GG GG GGG GG GG GG GG GGG GG GG GG GG GGG GG GG GG G G G G GGG GG GG G G G G G GG GG GG GG G G GG GGG GG GG GG GG GGG GG GG GG GGGGG GGG GG GG GG GG GG G G G GG G G G G G G G G G G G GG G G G G G G G G G G G GG G G GG G GG GG G G G G GG GGG G G GG GG GGGG GGG GG GG GG GG GGG GG G GG G G G GG GG GGGG G G GG GGG G G GG GG GG GG GGG GG GG GG GG GGG G G GG GG GG GGG GG GG GG GG G G G G G GG G G G G G G G GG GG GG GG G G GG GG G GG GG G G G G G G G G GG GG GGG G G GG GG GG G G GGG GG GG G G GG G G G G G G G G GG GG G G GG G G G GG GG GG GG GGG G G G G GG GG GGG GG GGGG GG G G GGG GG GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG GG GGGG GG GG GG GG GG G GG GG GG GG GG GG GGG GG GG G GGGGG GG GGG GG GG GGG GG GG GG G G GG GGG GG GG GG G G GGG GG G G G G G G G G G G G G G G G G G GG GG G GG GG G G G G G G G G G GG GG G G G G G G G G G G G G G G G GG G G GG G G G G G GG G G G G G G G G G G G GG G G GG GGG G G GG G G G G G GG G G G G GG GG G G GGG G G GG GG GG GGG GG GG GG G G GG GGG GG G G GG G G GGG GG GG G G GG G G GGG GGG G GG GG GG GGGG G GG GG GG GG G G G G G GG GG G G G G GG G G G G G GG GG G G G GG G G GG G G G G G G G G G G G GG G G GG G G G G G G G GG G G G G G GG GG G G GG GG GG G G G G G GG G G GG GG G G G GG GG G G G G GG GGG GG GG GG GG GGG GG GG G G G GG GG G G GGG GG GG GG GG G G G G G G G GG GG G G G G G G G GG GG G GG GG GG GG GG GGG GG GG GG GG GG G G G GG G G GG GG GG GG G G GG G G G GG GG GG G G GG G G G G G G G G G GG G G G G G G G G G GG GG G GG GG G G GG GG GGG GG GG GG G GGGG G G G G G GG GGG G G G G G G G G G G G GG G G G G GG G G G G G G G G G G GG G G G G G G GG GG GGG GG GG GG G G GG G GGGGG G G GG GG G G GGG G G GG GGG GG G G GG GGG GG G G GG G G GG G G GG GG G G G G GG G G G GG G G GG GG GG GG G G G G G GG GG G G G G G G G GG G G G G G G G GG G G G G GG GGG G G GG GG GGGG GGG G G G G GG GG GGG GG GGG G GG GG GG GGG G G GG GG GG GG G G G G G GG G G G G GG GG G G G GG GG GG GGG GG GG GG GG GGG GG GG GG GG GG G G G GGG G G G GG G G G GG G G G G G G GG G GG GG G G G G GG G G G GG GG G G G G GG G GG GG GG G G G G GG G G G GG G G G G GG G GGG GG G GG G GGG G GG G G GG G G G G G G G GG GG G G G G G G G G G G GGG G G GG G G G G G G G GG G G G G GG GG G GG GG GGGG G G G GGGGGGGGGGG GG GGGGG G G G GG GG G G G G G GG G G GG G G G G G GG G G GG GG GGG GG G GGGGGG G G G GG GG G GG G G G G GG G G G GG GG GG G G G G G GG GG GG G G GG GGG G G G G GG GG GGG GG GG GG GG GGG GG GG GG G G G G GG G G G G G G G GG G G G G G G G G G GG G G G G G G G G G G G GG G G GG G G G G GG GG GG G GG G G GG G G G G GG G G G G G G G GG GGG G G GG G G G G GG G G G G G GG GG G GG G G G G GG G G GG G G G G GGGG G G G G G G G G G GG G G G G G GGG G GG G G GG G G G GG GG GG G G G G GG G G G G G G G GG G G G G G GG GG G G GG G G G G G G G G G G G G G G G G G G GG G G G G G GG GG G G G G G GG G G G G G G G G GG GGG GG G G G G GG GG GG G GG GG G G GG G G G G G GG GG GG G G G GG GG GG G G GGG GG G G G G G GGG GG GG GG G G G G G G G G G G G GG G GG G G GG GG GG GG G G G G G G G G G GG G G GGG G G G G GG G GG GG G G GGG G G G G GG GG GG GGG GG GG GG GG G GG GG GG G G G GGG GGG GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG G G GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GGGG GG GG GGG GG GG GG GG G G GGG GG GG GG GG G G G GGG G G GG GG GG GGG GG GG GG GG G GG GG GG GG G G GGGGG GG G G GG G G G G G G G GG G G G G G G G GG G G G G G G G G G G GG G G G G G G G GG G G G G G GG GG GGG G GG G G GG GGG GG GG G G GG G GG GG GG G G G G GGG GG GG G G G G G GG GG GG GG G G G G G GG GG G G G G GGG G G GG GGG GGG G G G G G G G G G G G G G G GGG GG G G G G G G G G G G G G GG GG GG G G GG G G GGG GG GG GG GGG GG GG GG GG G GG GG GG G GG G G G G GG GGG GG GG GG GG GG GGG GG GG GG G G G G G GG GG GG G G GG GG GG GG G G G G G G G G G G G G G G G G G G G G GG GG GG G G G GGG G GGGGGG GGGG G GG G GGGG G G G GG GGGGGGGGGGGG G G GGGGGGGGGGGGGG GGG GGGG GGGGG GGGGGGG G GG GGGG GGGGG GGGGGG GG G GGGGG G G G G GG G G GGG GG G G GG GG GGG GG G G GG G G G G GG GG G G GG GG G GG GG GG G G G G G G G GG GG G G GG G G GG G G G G GG GG GG G G G GG GG G G GG GG G G G G G G G G G GG G GG G G G G GG G G GGG G G GG GG G G GG GGG G G G G GG G G GG G GG GG G G G G GG G G G GG GG G G G G G G G G G G G G G G G G G GG GG G G G GG G G G G G G G G G G G G G G G G G G G GG G GG G G G G G GG G G GG GG GG G G GG G G G G G G GG G GG G G G G G G G GG G G G G G GG GG GG GG GG GG G GG GG GG GG G GG GG G G G G G G G G GG G GG G G G G G G G G G G G GG G GG G G GG GG GGG GG GG GG GG GGG GG GG G G G G G G G GG G G G G G G G G G G GG G GG G G GG G GG GG GG GG G G GG GGG GG GG G G G G G GG G G G G G G GG G G G G G G G G G G G G G G GG G GG GG G G GG GG GG G G G G G G G GG G G G G G GG GG GG G G G G G GG G G G G GG G GG GGG GG G G G G G G G G G G G G G GG G G GGG G G GG GG G G GGGGG GG GG G G GG G GG G G G G GG GG GG GGG GG GG GG G G G G GGG G G GG GGGGGGG GGG G GGGGGGGGGGG GGGGGGGGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG GGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGGGG G GGGGGGGGGG GGGGGGGGGGGGGG G G G G G G G G GGG G GG GGGG GGGGGGGGGG GGGGGGG G G G G GG G G G GGGGGGGGG GGG G G G G G G G G GG G GGGGGGG G GGG G G G G G GG GG G G G G GG G G G G G GGGGGGGGG GGGGGGGGGGGGGGG GG GG G GGGG G G G G G G GG GGGGGGGGGGG G G G G G G GGG G GGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G G G G G G G GG G GG G G GG G GG G G G G GG G GG G G G G G G G G G G GGG GG GG G GGGGGGGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GG G 100 200 300 CO [ppb] G G GG G G G G G G G G G G G G G G G G G GG G G G G GGG G G GG G G GG G G GGG G G G G G G G G G GG GG G G G G GG GG GGG GG GG GG G G G G G G G GG G G G G G G GGG G G G G G GG GG G GG GG G G G G GG GG G G G G G G G GG G G GGG G G G G G G G G GG G G G G G GG GG GG GG G GG G G G G G G G G GGG G G GG G G G G G GG GG G G G G GG G G G G GG GG GG G GG GG G G G G G G G G G GG G G GG GGG G G GG GG GG G G GG G GG GG G G G G G G G G G G G G G G G G G G G G G G G G G G GGG G G GG G G G G G G GG G GG G G G G GG G G G GGGG G G G G G G G G G GG G G G G G G G G GG G GG G G G G GG G G G GG GG G G G G G G G GGG G G G G G G G G G G G G G G G G G G G GG GG G GG GG GG GG G G G G G G G G G G G GG GGG G G GG G G G G GG G G G G G G G G G GG GGG GG GG G G GG G G G GG GG G G GG G G GGG GG GG G GGG G G G G G G G GG GG GG GGG G G GG GG G G G G G G G GG G GG G GGGGG G G G G G G GG G G G G G G G G G GG G G G GG GG G G G G G GG GG GG GG G G GGG GG G G G G G G G G G G G GG G G G G G G GGG GG GG GG G G G G G GG G G GG G G G G G G G GG G G GG G G G GG GGG GG GG GG G G G GG GG G G G G G G GG GGG G GGGGG GG GGG GG GG GG GG GG GGG GG G G G G GG GG G GG GG G G GGG G G G G G G G G G G GG G G G G G GG G G G G GG G G G G G G G G GG G G GG G GG GG GG G G G G GGG GG GG GG GG G G G G G GG GG G G GGG G G G G GG G G GG G G G G G G G GG G G GGG GG G G G G G G GG GGG GG G G GGG G GG G G G G GG G GG GG GG GG GG GG GGG GG GG GG GGG GG GG GG G GG GG GG GG G GG GG GG GG GG GGG GG GG G G GG G G G GG GG GG GG GG GG G G G G G G G G G G GG G G GG G G GG G G G G G GG G G G G G G G G G GG G G G G G G GG GGG G G GG G G GG GGG GG GG GG GG G G G G GG G G G G G G G G G G G G G GG G G GG GGG GG GG G G G GGG GG GGG G G GG GG G GG G G G G GG GG G GG G G G G G G GG GGG GG G G G G G G G GG GG GG GG GG GG GGG GG G G GG G G G GG GG GG GG GG GG GGG GG GG GG GG GGG G GG G G G G G G G G GGG G GG GG GG GG G G GG G GG GG GG GG GG GGG G GG G GG G G GG G G G GG G G GG G G G G G G G G G G G GG G G G GG GG G G G G G G GGG GG GG GG GG GG GG G GG G G G G GG GG GG GG G G G G G GG G G G G GG G GG G G GG GG G G G GG G G GG G G GG GGG G G GG GG G G G G G GG GG GG GG GG GGGGG GG GG GG GGG GG GG GG GG GGG G GG G G G G G G G G G GG GG GG GGG GG GG GG GG GGG G G G G GG GG GGG GG GG GG G GG GG GG GGG G G G G GG G G GGG GG G G GG GG G G G GG G G GG GG G G G GG G G G G G G G GG GG GG GG G G G G G GG GG GG GG G G GG GG G G G G GGG G G G G G G G G GGG GG GG GG GG GG GG G G G G G GG GG GGG GG GG GG GG GG GGG G G G G GG G G GGG G GG GG GG G G G G G GG G G G G G GG G G G G G G GGG GG GG G G GG G G G G G GG G G GG G G G GG GG GG GG GG GG G GGGGG GG G G G G G G GGG G G GG GGGG G G G G GG G G G GG G G G G GG GGG G G G GG GG GG G G GG G G G GG GG G G GG GGG G GGGGG GG GG G GG GG GG GG GG G G G GG GG GG GG GGG GG GG GG GGGG GGG GG GG GG GG GGG GG GG G G GG GG G G GG G G G GG GG GG GG GG G GG GG G G GG GG G GG G G G G G G G G G G G G G G G G G GG G G G G G GG G G G G G G GG G G G GG GG GG GGG GG GG GG GG GG GG G GG GG GG GG GG G GG GG GG GG GG G GG GG G G G G G G G G G G G GG GG G G G GG GGG GGG GGGGG GGGG G G GG GG GG GG G GG GG GG G G GG GG GG GGG GG GG G G G G G G GG G G G G G GG G G GGG GG GGGG G G G GG G G G G G GG GGGG GGGGGG G G G GG GG GGGGGG GG GG G G G GG GG G G GG GG GGG GG GGGG GGG GG G G G G G G G G GGGG GG GG GG GGG GG GG GG GG G G G GG GG GG G G GG G G G G G GG GG GGG GG GG GG GG GG G GG G G GG GG G G G GG G G G G G GG G G GG GG GGG G G GG GGG GG GG GG GG GG GG GG G G G G GG GG GGG GG G G GG GG GG G G G G G G G G GGGG GG GG GG GG GG GG GG GG G G G GGG G G G G G G G G GG GG GG G G GGGGG GG G G G G GG GGG GG GG GGG G GG GG GG GG G GG GG GG G G GG GG G GG GG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG G G GG GGG G G G GGG GG GGG G G G G GG GG G G GGG GG GG G G G G GG G GG GG GG G G GG GG G GG GG GG GG GGG G G G G G G GG GGG GG GG G G G G GG GG G G G G GG GGG GG GG G G GG GG G GG GG GG GG GG GG G GG GG GG GG GG GGG GG G G GG GG GG GG GGGG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG G G GGGG GGG GG GG GG GGG GG GG GG GG GG GGG GG G G G G GG GGG GG G G GG GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GGGG GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GGG GG GG GG GG GGG GG GG GG GG GGG G G GG GG G G GGG GG GG GG GG GG G GG GG GG G G GG G G GGG GG GG GG GG GGGGG GG GG GG GG G GG G G GG G G GGG G GG G G G GGGG GG GG GG G G G GG G G GG G G GG G GG GG GG GG G G G G G GG G G GG GG GG G GG G G G G GG GGG GG GG GGGGGG GGG GG GG G G GG GG GGG G G G G GG GG GGG GG GG GG G G GGG G G G G G G G G GG GGG GG GG GG GGG GG GG GG GG G GG GG GG GGG GG GG GG GGG GG GG GG GG GG GGG GG GG G G GG GGG GG GG GG G G GG GGGG GG GG GG GGG GG GG GG GG GG G G G GG G G GG GG G GGGGG GGGGGGGG GG G G GG GGGGGG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGG GGGGGGG GGGGGG G GGGGGGGGGGGGGGGGGGGGGGGGGGG GGGG G G GG GG GG GGG G G G G GG GG GG G GG GG GG G G GG G G G G G G GG GG G GG GG GG GG GG G G G GG GG GG GG GGGG GG G G GG GG GG G G G GG GG GG GG GG G GG GG G GGGGG GGG GG G G GG GG GGG GG GG GG GG GG GGG GG GG GG GG GGG GG GG GG GG GG GGG GG GG GGG G GG GG GG GG G GG GG GG GG GGG GG GG G G GG GG G GG GG GG GG GG GGG GG G G GG GGG GG GG GG GG GG G G G G G GG GGGG GG GGG GG GG GG GG GG GGG GG GG G G GG GG G GG GG GG GG GGG GG GG GG GG GG GGG GG GG G G GG GGG GG GG G GG GG GG GG GGG GG G G G G G G G GG GG GG G G G G GG G GG GG GG GG GG GG G G G GG GG GG GGG GG GG GG G G G G GGG G G GG GG G G GGG GG GG G G G GG G G GG GG GG G GG GG GG GG GGG GG GG GG GG G G GGG G G G G GG GG GGG GG GG GG GG G G GGG G G GG GG GG GGG G G G GG GG G G G G GGG GG GG GG GG GGG GG GG GG G G G G G G G GG GG GG GG GGG GG GG GG GG G G GGG GG GG GG GG G G GG G G G G G GGG GGGGGGGGGGGGGGGGGGG GGG G G GGGGGG GGGGGG GGGGGGGGG G G GGGGG GGGGGGGGG GG GG G G G GG G GGGGGG G GG GGGG GGGGGG G G GG GGGGG GG GG GGGGGGGG GG G G G GG GGGGGGGGGGGGGGGGGGGGGG GGGGGGGGG G GG GG GGGGGGG G GG G G G GG GG GGGG GGGGGGGGGG G GGGGG GGGGGG G G GGGGG G G G GG GGGG GGG G G G GGGGGGG G G G G GGGGGGGGGG G G G GG GG GGGGGGGG G GGG GGGGG G G G GGGGGGGGG G GGGGGGGGGG G G G GG G GGGG G GGGGGGGGGGGGG GGGGGGGG G G GGGGGGGG GGGGGGGGGGGG G GGGGG GGGG G G G G GGGGGGGGGG G GGG G GGGGGG G G G GGGG GG G GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG G G GGGGG G G G GGGGG G GGGGGGG G G GG G G G GGGGGG G GG GG G GGG GG G G 0.0 0.5 1.0 1.5 2.0 H2O [%] Aug'14 Sep'14 Oct'14 Nov'14 Dec'14 Jan'15 Feb'15 Mar'15 Apr'15 May'15 Figure 12. www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5312 was virtually eliminated with corrections as demonstrated in Fig. 3. 2. time series of COS used for the Allan deviation plots in Fig. 7. The QCLS was set up for continuous in situ measurements at different heights at the tower of the Lutjewad monitoring station. Hourly target measurements were used to assess the accuracy and precision of the measurements. After applica- tion of a calibration response curve for CO2 and CO, and a single bias correction for removal of instrument drift and to calibrate the COS measurements to the NOAA scale, the target measurements showed a mean difference with the as- signed cylinder value within 3.3 ppt for COS, 0.05 ppm for CO2 and 1.7 ppb for CO over a period of 35 days. One- second precisions during reference gas ﬂow were typically 4.3 ppt for COS, 0.04 ppm for CO2 and 0.9 ppb for CO, but substantial variations in the instrument precision were ob- served during the 7-month ﬁeld campaign. We improved the temperature stability of the QCLS by applying an additional insulation layer that is controlled by the same thermoelec- tric chiller as the one used for cooling the laser and detector. However, improvement of the temperature stability of the in- strument did not show a consistent relation with instrument precision. We showed that variations in the precision were largely driven by mirror alignment. The different uncertainty contributions for measurements of COS, CO2 and CO were summarized and the overall uncertainty was determined to be 7.5 ppt for COS, 0.23 ppm for CO2 and 3.3 ppb for CO. We discussed the relevance of different uncertainty contributions for different types of analyses (e.g., differences across sites and interpreting proﬁle measurements). Our setup with the QCLS provides sufﬁcient accuracy and precision to detect gradients larger than 6.0 ppt for COS, 0.13 ppm for CO2 and 1.1 ppb for CO. For the current setup in Lutjewad, the day- time gradients were too small compared to the measurement uncertainty to be able to use the ﬂux-gradient method. QCLS measurements were compared with independent CRDS mea- surements for CO2 and CO and with dry air ﬂask sam- ples at the QCLS for COS, which showed a mean differ- ence of −9.7 ± 4.6 ppt for COS, 0.12 ± 0.77 ppm for CO2, −0.9 ± 3.8 ppb for CO and −0.01 ± 0.09 % for H2O. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O After correction for water vapor there was no correlation between the offsets and water vapor for all species. The measurement record over the 7-month period was presented and compared with NOAA/ESRL ﬂask measurements for COS at other sites in the NH. The peak-to-peak amplitude of COS in ambient air at the Lutjewad monitoring station was estimated to be 97 ppt, which is comparable to other coastal sites at similar latitudes in the NH. Other data from laboratory tests are available upon request (huilin.chen@rug.nl). The ﬂask measurements made by NOAA/ESRL as part of a COS monitoring network (Montzka et al., 2007) are avail- able at ftp://ftp.cmdl.noaa.gov/hats/carbonyl_sulﬁde/. References Allan, D. W.: Time and Frequency (Time-Domain) Characteriza- tion, Estimation, and Prediction of Precision Clocks and Oscilla- tors, IEEE T. Ultrason. Ferr., 34, 647–654, 1987. Asaf, D., Rotenberg, E., Tatarinov, F., Dicken, U., Montzka, S. A., and Yakir, D.: Ecosystem photosynthesis inferred from mea- surements of carbonyl sulphide ﬂux, Nat. Geosci., 6, 186–190, doi:10.1038/ngeo1730, 2013. Belviso, S., Schmidt, M., Yver, C., Ramonet, M., Gros, V., and Launois, T.: Strong similarities between night-time de- position velocities of carbonyl sulphide and molecular hy- drogen inferred from semi-continuous atmospheric observa- tions in Gif-sur-Yvette, Paris region, Tellus B, 65, 20719, doi:10.3402/tellusb.v65i0.20719, 2013. Berkelhammer, M., Asaf, D., Still, C., Montzka, S., Noone, D., Gupta, M., Provencal, R., Chen, H., and Yakir, D.: Constraining surface carbon ﬂuxes using in situ measurements of carbonyl sul- ﬁde and carbon dioxide, Global Biogeochem. Cy., 28, 161–179, doi:10.1002/2013GB004644, 2014. The Supplement related to this article is available online at doi:10.5194/amt-9-5293-2016-supplement. Acknowledgements. We would like to thank B. A. M. Kers, M. de Vries, H. G. Jansen and H. A. Been for their help in preparing the system for installation in the ﬁeld and for maintenance of the instrumentation in Lutjewad. We thank D. Paul for the valuable discussions and suggestions, H. A. Scheeren for providing the CO2 and CO calibrations of our working standards and J. J. Spriensma for her help in sorting out ﬂask samples. We also acknowledge the preparation of gravimetric standards and working standards at NOAA by B. Hall and the technical assistance of C. Siso. Finally, we would like to thank the reviewers for their valuable comments and suggestions. This work was supported by the Dutch Sector Plan Physics. Edited by: M. von Hobe Reviewed by: two anonymous referees Edited by: M. von Hobe Reviewed by: two anonymous referees Edited by: M. von Hobe Reviewed by: two anonymous referees 4 Conclusions In this study we have tested a QCLS for its suitability for making accurate and high-precision measurements of COS, CO2, CO and H2O. First, the instrument response was char- acterized using calibration standards and to transfer raw data to the NOAA or WMO scale. Unfortunately, the range of mole fractions in the calibration standards did not al- low the COS response to be accurately determined over the entire range of measured mole fractions. Based on an analysis of different calibration methods, however, we con- cluded that the measurements and working standards were best calibrated using a single bias correction for COS. From hourly paired measurements of working standards we ob- served changes in the response curve for CO2 over a pe- riod of 10 days after transporting the instrument to the mea- www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O A., and Meijer, H. A. J.: A Computer-Controlled Continuous Air Drying and Flask Sampling System, J. Atmos. Ocean. Tech., 21, 651–659, doi:10.1175/1520- 0426(2004)021<0651:ACCADA>2.0.CO;2, 2004. Commane, R., Herndon, S. C., Zahniser, M. S., Lerner, B. M., McManus, J. B., Munger, J. W., Nelson, D. D., and Wofsy, S. C.: Carbonyl sulﬁde in the planetary boundary layer: Coastal and continental inﬂuences, J. Geophys. Res.-Atmos., 118, 8001– 8009, doi:10.1002/jgrd.50581, 2013. Novelli, P. C., Masarie, K. A., Lang, P. M., Hall, B. D., Myers, R. C., and Elkins, J. W.: Reanalysis of tropospheric CO trends: Effects of the 1997–1998 wildﬁres, J. Geophys. Res.-Atmos., 108, 4464, doi:10.1029/2002JD003031, 2003. Commane, R., Meredith, L. K., Baker, I. T., Berry, J. A., Munger, J. W., Montzka, S. A., Templer, P. H., Juice, S. M., Zahniser, M. S., and Wofsy, S. C.: Seasonal ﬂuxes of carbonyl sulﬁde in a midlatitude forest, P. Natl. Acad. Sci. USA, 112, 14162–14167, doi:10.1073/pnas.1504131112, 2015. Protoschill-Krebs, G. and Kesselmeier, J.: Enzymatic Path- ways for the Consumption of Carbonyl Sulphide (COS) by Higher Plants*, Bot. Acta, 105, 206–212, doi:10.1111/j.1438- 8677.1992.tb00288.x, 1992. Protoschill-Krebs, G., Wilhelm, C., and Kesselmeier, J.: Consump- tion of carbonyl sulphide (COS) by higher plant carbonic anhy- drase (CA), Atmos. Environ., 30, 3151–3156, doi:10.1016/1352- 2310(96)00026-X, 1996. LaFranchi, B., Bambha, R., Schrader, P., and Michelsen, H.: Char- acterization of continuous OCS, CO and CO2 measurements at a tower site in Livermore, CA USA, 18th WMO/IAEA Meeting on Carbon Dioxide, Other Greenhouse Gases, and Related Mea- surement Techniques (GGMT), La Jolla, 13–17 September 2015, A10, 2015. Rella, C. W., Chen, H., Andrews, A. E., Filges, A., Gerbig, C., Hatakka, J., Karion, A., Miles, N. L., Richardson, S. J., Stein- bacher, M., Sweeney, C., Wastine, B., and Zellweger, C.: High accuracy measurements of dry mole fractions of carbon diox- ide and methane in humid air, Atmos. Meas. Tech., 6, 837–860, doi:10.5194/amt-6-837-2013, 2013. Kettle, A., Kuhn, U., von Hobe, M., Kesselmeier, J., and An- dreae, M.: Global budget of atmospheric carbonyl sulﬁde: Tem- poral and spatial variations of the dominant sources and sinks, J. Geophys. Res.-Atmos., 107, 4658, doi:10.1029/2002JD002187, 2002. Rothman, L. S., Gordon, I. E., Babikov, Y., Barbe, A., Chris Benner, D., Bernath, P. F., Birk, M., Bizzocchi, L., Boudon, V., Brown, L. R., Campargue, A., Chance, K., Cohen, E. A., Coudert, L. H., Devi, V. M., Drouin, B. J., Fayt, A., Flaud, J.-M., Gamache, R. R., Harrison, J. J., Hartmann, J.-M., Hill, C., Hodges, J. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5313 ple Preconcentration and GC/MS Detector System for in Situ Measurements of Atmospheric Trace Halocarbons, Hydrocar- bons, and Sulfur Compounds, Anal. Chem., 80, 1536–1545, doi:10.1021/ac702084k, 2008. Blonquist, J. M., Montzka, S. A., Munger, J. W., Yakir, D., Desai, A. R., Dragoni, D., Grifﬁs, T. J., Monson, R. K., Scott, R. L., and Bowling, D. R.: The potential of carbonyl sulﬁde as a proxy for gross primary production at ﬂux tower sites, J. Geophys. Res.- Biogeo., 116, G04019, doi:10.1029/2011JG001723, 2011. Montzka, S., Aydin, M., Battle, M., Butler, J., Saltzman, E., Hall, B., Clarke, A., Mondeel, D., and Elkins, J.: A 350-year atmo- spheric history for carbonyl sulﬁde inferred from Antarctic ﬁrn air and air trapped in ice, J. Geophys. Res.-Atmos., 109, D22302, doi:10.1029/2004JD004686, 2004. g Campbell, J. E., Carmichael, G. R., Chai, T., Mena-Carrasco, M., Tang, Y., Blake, D. R., Blake, N. J., Vay, S. A., Collatz, G. J., Baker, I., Berry, J. A., Montzka, S. A., Sweeney, C., Schnoor, J. L., and Stanier, C. O.: Photosynthetic Control of Atmospheric Carbonyl Sulﬁde During the Growing Season, Science, 322, 1085–1088, doi:10.1126/science.1164015, 2008. Montzka, S. A., Calvert, P., Hall, B. D., Elkins, J. W., Conway, T. J., Tans, P. P., and Sweeney, C.: On the global distribution, seasonality, and budget of atmospheric carbonyl sulﬁde (COS) and some similarities to CO2, J. Geophys. Res.-Atmos., 112, D09302, doi:10.1029/2006JD007665, 2007. Chen, H., Winderlich, J., Gerbig, C., Hoefer, A., Rella, C. W., Crosson, E. R., Van Pelt, A. D., Steinbach, J., Kolle, O., Beck, V., Daube, B. C., Gottlieb, E. W., Chow, V. Y., Santoni, G. W., and Wofsy, S. C.: High-accuracy continuous airborne measure- ments of greenhouse gases (CO2 and CH4) using the cavity ring- down spectroscopy (CRDS) technique, Atmos. Meas. Tech., 3, 375–386, doi:10.5194/amt-3-375-2010, 2010. Nelson, D. D., McManus, B., Urbanski, S., Herndon, S., and Zah- niser, M.: High precision measurements of atmospheric nitrous oxide and methane using thermoelectrically cooled mid-infrared quantum cascade lasers and detectors, Spectrochim. Acta. A., 60, 3325–3335, doi:10.1016/j.saa.2004.01.033, 2004. Chen, H., Karion, A., Rella, C. W., Winderlich, J., Gerbig, C., Filges, A., Newberger, T., Sweeney, C., and Tans, P. P.: Accurate measurements of carbon monoxide in humid air using the cavity ring-down spectroscopy (CRDS) technique, Atmos. Meas. Tech., 6, 1031–1040, doi:10.5194/amt-6-1031-2013, 2013. Neubert, R. E. M., Spijkervet, L. L., Schut, J. K., Been, H. 5 Data availability Berry, J., Wolf, A., Campbell, J. E., Baker, I., Blake, N., Blake, D., Denning, A. S., Kawa, S. R., Montzka, S. A., Seibt, U., Stim- ler, K., Yakir, D., and Zhu, Z.: A coupled model of the global cycles of carbonyl sulﬁde and CO2: A possible new window on the carbon cycle, J. Geophys. Res.-Biogeo., 118, 842–852, doi:10.1002/jgrg.20068, 2013. The following data are available as a Supplement to this pa- per: 1. water vapor experiment data used in Fig. 3 Atmos. Meas. Tech., 9, 5293–5314, 2016 www.atmos-meas-tech.net/9/5293/2016/ L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O 5314 by deposition velocities normalized to the uptake of carbon diox- ide (CO2), Biogeosciences, 2, 125–132, doi:10.5194/bg-2-125- 2005, 2005. van der Laan-Luijkx, I. T., Neubert, R. E. M., van der Laan, S., and Meijer, H. A. J.: Continuous measurements of atmospheric oxygen and carbon dioxide on a North Sea gas platform, Atmos. Meas. Tech., 3, 113–125, doi:10.5194/amt-3-113-2010, 2010. Santoni, G. W., Lee, B. H., Goodrich, J. P., Varner, R. K., Crill, P. M., McManus, J. B., Nelson, D. D., Zahniser, M. S., and Wofsy, S. C.: Mass ﬂuxes and isoﬂuxes of methane (CH4) at a New Hampshire fen measured by a continuous wave quantum cas- cade laser spectrometer, J. Geophys. Res.-Atmos., 117, D10301, doi:10.1029/2011JD016960, 2012. Van Leeuwen, C.: Highly precise atmospheric oxygen measure- ments as a tool to detect leaks of carbon dioxide from Carbon Capture and Storage sites, PhD thesis, University of Groningen, the Netherlands, 2015. White, M. L., Zhou, Y., Russo, R. S., Mao, H., Talbot, R., Varner, R. K., and Sive, B. C.: Carbonyl sulﬁde exchange in a temperate loblolly pine forest grown under ambient and elevated CO2, At- mos. Chem. Phys., 10, 547–561, doi:10.5194/acp-10-547-2010, 2010. Seibt, U., Kesselmeier, J., Sandoval-Soto, L., Kuhn, U., and Berry, J. A.: A kinetic analysis of leaf uptake of COS and its relation to transpiration, photosynthesis and carbon isotope fractionation, Biogeosciences, 7, 333–341, doi:10.5194/bg-7-333-2010, 2010. Stimler, K., Nelson, D., and Yakir, D.: High precision measurements of atmospheric concentrations and plant exchange rates of car- bonyl sulﬁde using mid-IR quantum cascade laser, Glob. Change Biol., 16, 2496–2503, doi:10.1111/j.1365-2486.2009.02088.x, 2010a. Wohlfahrt, G., Brilli, F., Hoertnagl, L., Xu, X., Bingemer, H., Hansel, A., and Loreto, F.: Carbonyl sulﬁde (COS) as a tracer for canopy photosynthesis, transpiration and stomatal conduc- tance: potential and limitations, Plant Cell Environ., 35, 657– 667, doi:10.1111/j.1365-3040.2011.02451.x, 2012. Stimler, K., Montzka, S. A., Berry, J. A., Rudich, Y., and Yakir, D.: Relationships between carbonyl sulﬁde (COS) and CO2 during leaf gas exchange, New Phytol., 186, 869–878, doi:10.1111/j.1469-8137.2010.03218.x, 2010b. Xiang, B., Nelson, D. D., McManus, J. B., Zahniser, M. S., Wehr, R. A., and Wofsy, S. C.: Development and ﬁeld testing of a rapid and ultra-stable atmospheric carbon dioxide spectrometer, At- mos. Meas. Tech., 7, 4445–4453, doi:10.5194/amt-7-4445-2014, 2014. van der Laan, S., Neubert, R. E. M., and Meijer, H. A. J.: Methane and nitrous oxide emissions in The Netherlands: ambient mea- surements support the national inventories, Atmos. Chem. L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O T., Jacquemart, D., Jolly, A., Lamouroux, J., Le Roy, R. J., Li, G., Long, D. A., Lyulin, O. M., Mackie, C. J., Massie, S. T., Mikhailenko, S., Müller, H. S. P., Nau- menko, O. V., Nikitin, A. V., Orphal, J., Perevalov, V., Per- rin, A., Polovtseva, E. R., Richard, C., Smith, M. A. H., Starikova, E., Sung, K., Tashkun, S., Tennyson, J., Toon, G. C., Tyuterev, V. G., and Wagner, G.: The HITRAN2012 molecu- lar spectroscopic database, J. Quant. Spectrosc. Ra., 130, 4–50, doi:10.1016/j.jqsrt.2013.07.002, 2013. Maseyk, K., Berry, J. A., Billesbach, D., Campbell, J. E., Torn, M. S., Zahniser, M., and Seibt, U.: Sources and sinks of carbonyl sulﬁde in an agricultural ﬁeld in the Southern Great Plains, P. Natl. Acad. Sci. USA, 112, 14162–14167, doi:10.1073/pnas.1319132111, 2014. McManus, J. B., Zahniser, M. S., Nelson, D. D., Jr., Shorter, J. H., Herndon, S., Wood, E., and Wehr, R.: Application of quantum cascade lasers to high-precision atmospheric trace gas measure- ments, Opt. Eng., 49, 111124, doi:10.1117/1.3498782, 2010. Meredith, L. K., Commane, R., Munger, J. W., Dunn, A., Tang, J., Wofsy, S. C., and Prinn, R. G.: Ecosystem ﬂuxes of hydrogen: a comparison of ﬂux-gradient methods, Atmos. Meas. Tech., 7, 2787–2805, doi:10.5194/amt-7-2787-2014, 2014. Sandoval-Soto, L., Stanimirov, M., von Hobe, M., Schmitt, V., Valdes, J., Wild, A., and Kesselmeier, J.: Global uptake of car- bonyl sulﬁde (COS) by terrestrial vegetation: Estimates corrected Miller, B. R., Weiss, R. F., Salameh, P. K., Tanhua, T., Gre- ally, B. R., Mühle, J., and Simmonds, P. G.: Medusa: A Sam- www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 www.atmos-meas-tech.net/9/5293/2016/ Atmos. Meas. Tech., 9, 5293–5314, 2016 L. M. J. Kooijmans et al.: High-precision concentration measurements of COS, CO2, CO and H2O Phys., 9, 9369–9379, doi:10.5194/acp-9-9369-2009, 2009. Zhao, C. L. and Tans, P. P.: Estimating uncertainty of the WMO mole fraction scale for carbon dioxide in air, J. Geophys. Res.- Atmos., 111, D08S09, doi:10.1029/2005JD006003, 2006. van der Laan, S., Karstens, U., Neubert, R. E. M., Van der Laan- Luijkx, I. T., and Meijer, H. A. J.: Observation-based esti- mates of fossil fuel-derived CO2 emissions in the Netherlands using Delta 14C, CO and 222Radon, Tellus B, 62, 389–402, doi:10.1111/j.1600-0889.2010.00493.x, 2010. Zhao, F. and Zeng, N.: Continued increase in atmospheric CO2 seasonal amplitude in the 21st century projected by the CMIP5 Earth system models, Earth Syst. Dynam., 5, 423–439, doi:10.5194/esd-5-423-2014, 2014. Atmos. Meas. Tech., 9, 5293–5314, 2016 www.atmos-meas-tech.net/9/5293/2016/
https://openalex.org/W1640734465	https://ojs.unud.ac.id/index.php/jumpa/article/download/15209/10069	Indonesian	null	PERKEMBANGAN DESA WISATA JATILUWIH SETELAH UNESCO MENETAPKAN SUBAKNYA SEBAGAI BAGIAN DARI WARISAN BUDAYA DUNIA	JUMPA	2,015	cc-by	5,880	JUMPA 2 [1] : 61 - 78 JUMPA 2 [1] : 61 - 78 ISSN 2406-9116 PERKEMBANGAN DESA WISATA JATILUWIH SETELAH UNESCO MENETAPKAN SUBAKNYA SEBAGAI BAGIAN DARI WARISAN BUDAYA DUNIA Dewa Ayu Diyah Sri Widari Email: dewaayusriwidari@gmail.com Abstract This article identifies the social, cultural, and economic development of Jatiluwih tourism village after subak (rice field and its irrigation cultural system) in this village was decreed by UNESCO as a World’s Cultural Heritage in 2012, under the label Cultural Landscape of Bali Province: the Subak System as a Manifestation of the Tri Hita Karana. It also discusses tourists’ perception to the rice field tourist attraction. The data were collected through observation, interview, and documentary study, and then descriptively and qualitatively analyzed. The informants were selected using the purposive sampling technique and the accidental sampling technique. The study shows that the mutual cooperation, the rituals related to the farming activities, and rice planting system were still well maintained by farmers of the subak member. The traditional arts were conserved and developed, and the social organization were also well maintened. The economic aspects which includes job and business opportunities, income and investment have improved. The local people’s participation in the planning and controlling phases were manipulative and functional. In the development phase, the local people participated actively. While tourist’s perception on subak is good, but that is not the case on the poor condition of public facilities and information service by local management. Keywords: tourism village, subak, world cultural heritage, tri hita karana. 1. Pendahuluan Tahun 2012 merupakan salah satu tonggak penting dalam dunia kepariwisataan Bali dilihat dari konteks pengembangan pariwisata budaya. Dikatakan penting karena pada tahun 2012, Pemprov Bali menegaskan kembali pariwisata budaya sebagai landasan pengembangan kepariwisataan Bali lewat penetapan Peraturan Daerah Provinsi Bali Nomor 2 Tahun 2012 61 JUMPA Volume 2 Nomor 1 Juli 2015 Dewa Ayu Diyah Sri Widari tentang Kepariwisataan Budaya Bali. Untuk pertama kali, Perda Pariwisata Budaya ditetapkan tahun 1974, lalu direvisi tahun 1991, dan kemudian tahun 2012, rangkaian revisi atau penyempurnaa ketiga kali yang menunjukkan betapa sentralnya budaya dalam kepariwisataan. Selain itu, pada bulan Juni 2012, Bali mendapat pengakuan istimewa dari UNESCO dengan penetapan subak sebagai warisan budaya dunia dengan label Cultural Landscape of Bali Province: the Subak System as a Manifestation of the Tri Hita Karana Philosophy (Lanskap budaya Bali: Sistem Subak sebagai Manifestasi Filosofi Tri Hita Karana). Tri Hita Karana adalah filosofi Hindu yang berarti tiga penyebab kesejahteraan yaitu harmonisnya hubungan antara manusia dengan Tuhan (parhyangan), dengan manusia (pawongan), dan manusia dengan alam (palemahan). Dalam penetapan lanskap budaya Bali atau subak sebagai warisan budaya dunia, maka perhatian public di Bali dan Indonesia tertuju pada kawasan persawahan luas di Desa Jatiluwih. Desa Jatiluwih merupakan desa wisata yang berlokasi di Kecamatan Penebel, Kabupaten Tabanan. Desa wisata ini memiliki panorama alam yang indah dengan keunikan sawah berundak- undak yang berada di kaki Gunung Batukaru. Udaranya relatif sejuk karena berada pada ketinggian 700 meter dari permukaan laut. Pola hidup masyarakat di Desa Wisata Jatiluwih bersifat agraris religius (Lihat Foto 1). Tujuan umum penataan dan pengembangan desa wisata adalah memberikan model pembangunan pariwisata yang mungkin berdampingan dengan dan bahkan mengembangkan budaya setempat, dan mendukung konsep pembangunan yang berkesinambungan, memberikan pemerataan kesempatan berusaha, serta peningkatan daya dukung desa dengan pranata sosial kepariwisataan dan peningkatan kualitas pengelolaan (Universitas Gadjah Mada, 1992:3). Berdasarkan tujuan umum yang ingin dicapai dalam penataan dan pengembangan desa wisata tersebut, ternyata tidak sesuai dengan realita yang terjadi di Desa Wisata Jatiluwih. Menurut penelitian yang dilakukan oleh Prasiasa (2010) pengembangan pariwisata di Desa Wisata Jatiluwih kurang memberikan manfaat ekonomi bagi masyarakat lokal, menimbulkan komodifikasi pada bangunan tradisional saka roras, semakin meningkatnya penjualan tanah di sekitar pura dan hutan sebelah utara Pura Luhur Petali. Kunjungan wisatawan ke Desa Wisata Jatiluwih setelah penetapan subak sebagai Warisan Budaya Dunia, mengalami peningkatan. Pada tahun 2012 jumlah kunjungan wisatawan sebanyak 97.909 orang, sedangkan tahun 2013 sebanyak 101.560 orang, terjadi peningkatan sebesar 37,28%. 1. Pendahuluan Tahun 2014 jumlah kunjungan wisatawan sebanyak 165.158. Jika jumlah kunjungan wisatawan tahun 2013 dibandingkan dengan jumlah kunjungan wisatawan tahun 2014, terjadi peningkatan sebesar 62,62% (Diparda Provinsi Bali, 2015). 62 JUMPA Volume 2 Nomor 1, Juli 2015 Perkembangan Desa Wisata Jatiluwih Setelah UNESCO Menetapkan Subak Sebagai Bagian dari Warisan Budaya DARMA PUTRA Foto 1: Hamparan sawah luas berundak di Desa Jatiluwih. Aktivitas bertani menambah daya tarik wisata ini. DARMA PUTRA Foto 1: Hamparan sawah luas berundak di Desa Jatiluwih. Aktivitas bertani menambah daya tarik wisata ini. Pendapatan masyarakat yang diterima secara langsung dari berkembang nya lapangan kerja dan kesempatan berusaha akibat perkembangan desa wisata juga meningkat. Namun, peningkatan tersebut belum memberikan manfaat ekonomi yang optimal bagi masyarakat lokal. Perkembangan Desa Wisata Jatiluwih juga memberi dampak pada berkembangnya kehidupan so- sial budaya masyarakat setempat. Di sisi lain, juga menimbulkan pergeseran pada budaya pertanian yang diterapkan oleh masyarakat lokal. Mengacu pada tujuan pembangunan pariwisata yang berbasis pada pemberdayaan masyarakat serta International Council of Monument and Site (ICOMOS), masyarakat harus dilibatkan dalam pengembangan pariwisata. Menurut Prasiasa (2010) masyarakat Jatiluwih kurang berpartisipasi dalam pengembangan desanya. Di samping itu, menurut Heny Urmila Dewi dkk. (2013) partisipasi masyarakat Jatiluwih pada tahap perencanaan, implementasi, dan pengawasan masih minim. Perkembangan suatu desa wisata memerlukan kerjasama dan keterlibatan seluruh masyarakat di desa tersebut untuk memajukannya. Dengan dibentuknya Badan Pengelola Daya Tarik Wisata, masyarakat Desa Jatiluwih sudah mulai aktif dalam mengelola desa wisata. Partisipasi masyarakat dalam pengembangan desa wisata, baik pada pengelolaan fasilitas pariwisata maupun keterlibatan pada badan pengelola. Namun, dalam perencanaan dan pengawasan desa wisata, masyarakat kurang dilibatkan. Untuk menciptakan pengelolaan sumber daya di bidang pariwisata yang 63 JUMPA Volume 2 Nomor 1 Juli 2015 Dewa Ayu Diyah Sri Widari mampu melibatkan masyarakat lokal dalam menjaga keseimbangan dan keharmonisan lingkungan hidup, sumber daya dan kepuasan wisatawan guna keberlanjutan sistem sosial, budaya, dan ekonomi, pengembangan pariwisata di Bali tidak bisa dilepaskan dari kearifan lokal Tri Hita Karana. Untuk itulah perlu mengkaji persepsi wisatawan terhadap perkembangan Desa Wisata Jatiluwih setelah penetapan subak sebagai Warisan Budaya Dunia dari aspek parhyangan, pawongan, dan palemahan. 1. Pendahuluan Berdasarkan latar belakang, rumusan masalah dalam artikel ini adalah (1) bagaimana perkembangan sosial budaya dan ekonomi Desa Wisata Jatiluwih setelah penetapan subak sebagai Warisan Budaya Dunia dari aspek Tri Hita Karana, (2) bagaimana partisipasi masyarakat dalam perkembangan Desa Wisata Jatiluwih setelah penetapan subak sebagai Warisan Budaya Dunia dari aspek Tri Hita Karana, dan (3) bagaimana persepsi wisatawan terhadap perkembangan Desa Wisata Jatiluwih setelah penetapan subak sebagai Warisan Budaya Dunia dari aspek Tri Hita Karana. JUMPA Volume 2 Nomor 1, Juli 2015 2. Teori dan Metode Teori yang digunakan untuk menganalisis permasalahan dalam artikel ini adalah teori siklus hidup destinasi, teori konflik, teori dampak, teori partisipasi, dan teori persepsi. Adapun uraian dari masing-masing teori tersebut sebagai berikut. Tujuan utama dari penggunaan model siklus hidup destinasi adalah sebagai alat untuk memahami evolusi dari produk dan destinasi pariwisata. Siklus hidup pariwisata mengacu pada pendapat Butler (1980) tentang tourism area life cycle dengan tahapan yaitu exploration, involvement, development, consolidation, stagnation, dan decline atau rejuvenation. Teori Konflik menolak anggapan bahwa masyarakat ada dalam situasi stabil dan tidak berubah. Sebaliknya, masyarakat selalu dilihat dalam suatu kondisi tidak seimbang atau tidak adil, dan keadilan/keseimbangan dapat dicapai dengan penggunaan kekuatan revolusi terhadap kelompok- kelompok yang memegang kekuasaan (Pitana dan Gayatri, 2005:20). Menurut Higo Tugiman (dalam Musoman, dkk., 2013) dampak merupakan pengawasan internal sangat penting, dengan mudah dapat diubah menjadi sesuatu yang dipahami dan ditanggapi secara serius oleh manajemen. Pengembangan pariwisata dapat menimbulkan berbagai dampak, baik pengaruh positif maupun negatif. Cooper, dkk. (1999:133), menyebutkan bahwa pengeluaran wisatawan memiliki efek bertingkat di seluruh perekonomian penduduk setempat. Timothy (1999) menyebutkan ada dua perspektif untuk melihat partisipasi masyarakat dalam pariwisata. Kedua perspektif tersebut adalah partisipasi masyarakat lokal dalam proses pengambilan keputusan, dan berkaitan dengan manfaat yang diterima masyarakat dari pembangunan pariwisata. 64 JUMPA Volume 2 Nomor 1, Juli 2015 Perkembangan Desa Wisata Jatiluwih Setelah UNESCO Menetapkan Subak Sebagai Bagian dari Warisan Budaya Persepsi didefinisikan sebagai proses yang dilalui seseorang dalam memilih, mengatur, dan menginterpretasikan masukan-masukan informasi untuk menciptakan gambaran yang berarti tentang dunia (Cravens, 1997:131). Konsep yang digunakan dalam artikel ini adalah perkembangan, desa wisata, subak, Warisan Budaya Dunia, dan Tri Hita Karana. Artikel ini merupakan hasil penelitian dengan pendekatan penelitian kualitatif. Data dikumpulkan melalui observasi, wawancara dan studi dokumen. Data yang terkumpul dianalisis secara deskriptif kualitatif. Teknik penentuan informan dengan purposive sampling dan accidental sampling. Berdasarkan tujuan serta pertimbangan peneliti, informan yang diambil terdiri atas Kepala Desa, Bendesa Adat, Pekaseh, Kelian Dusun, Pengelola Desa Wisata, Pengusaha Pariwisata, masyarakat dan wisatawan. 3. Jatiluwih setelah Subak menjadi Warisan Budaya Dunia 3.1 Perkembangan Sosial Budaya Perkembangan sosial budaya Desa Wisata Jatiluwih setelah penetapan subak sebagai Warisan Budaya Dunia meliputi aspek kegotongroyongan, budaya pertanian, kesenian tradisional, dan organisasi sosial. Kegotongroyongan. Kegotongroyongan dalam hal pelaksanaan adat dan suka duka, di kalangan masyarakat Desa Jatiluwih masih terjaga bahkan semakin baik. Pada saat dilaksanakan upacara adat seperti pujawali/ piodalan di pura kahyangan tiga, ngusaba desa, dan pura-pura lain yang tergolong pura dang kahyangan, masyarakat mempersiapkan segala keperluan upacara tersebut secara gotong-royong. Kegotongroyongan dalam kaitannya dengan suka duka juga masih terpelihara. Apabila ada salah satu keluarga yang mengalami musibah kematian, warga masyarakat lainnya akan membantu dalam proses upacara yang dilaksanakan sehubungan dengan terjadinya kematian tersebut. Kebersamaan berupa kegotongroyongan di kalangan masyarakat Desa Jatiluwih yang masih dijaga, mencerminkan implementasi positif aspek pawongan dari konsep Tri Hita Karana. Rasa kekeluargaan dan solidaritas yang tinggi membuat terjalinnya harmonisasi di antara masyarakat setempat. Hal ini perlu terus dijaga dan dipertahankan agar kehidupan bermasyarakat di Desa Jatiluwih dapat berjalan dalam suasana yang harmonis. Budaya Pertanian. Budaya pertanian yang diterapkan oleh warga subak terus dilaksanakan, yaitu upacara yang berkaitan dengan aktivitas pertanian, antara lain upacara mapag toya, kempelan, upacara ngendag tanah carik, upacara ngurit, upacara ngerasakin, upacara pangawiwit (nuwasen), upacara ngekambuhin, upacara pamungkah, upacara penyepian, upacara pangarestitian nyegara gunung, upacara mesaba, ngadegang Batari Sri (Batari Nini), upacara nganyarin dan manyi. Semua upacara tersebut terus 65 JUMPA Volume 2 Nomor 1 Juli 2015 Dewa Ayu Diyah Sri Widari diselenggarakan oleh masing-masing warga subak sebagai suatu kebiasaan (tradisi) yang memang harus dilaksanakan (hasil wawancara dengan Pekaseh Subak Jatiluwih dan Kepala Dusun di Desa Jatiluwih tanggal 2 Januari 2015). Selain upacara yang berkaitan dengan aktivitas petani di sawah, sistem tanam padi juga terus diterapkan oleh warga subak. Subak Jatiluwih membagi sistem tanam padi menjadi dua periode. Periode pertama, penanaman padi Bali (lokal) pada bulan Januari, dengan masa panen antara bulan Mei dan Juni. Periode kedua, penanaman padi varietas (Serang, Mangsur, Sigelis, PB dan IR) pada bulan Agustus, dengan masa panen pada bulan Nopember. Sistem tanam padi tersebut sudah diterapkan sejak tahun 1977/1978. Penerapan sistem tanam padi Bali (lokal) sebagai upaya untuk mempertahankan ketersediaan beras merah cendana, beras putih (ketan), dan beras hitam (injin) sebagai perlambang tri datu. Tiga warna tersebut disakralkan oleh umat Hindu yaitu merah, putih dan hitam (hasil wawancara dengan Pekaseh Subak Jatiluwih tanggal 2 Januari 2015). Budaya pertanian dalam pelaksanaan upacara yang berkaitan dengan aktivitas pertanian dan sistem tanam padi memang terus dilaksanakan. JUMPA Volume 2 Nomor 1, Juli 2015 3. Jatiluwih setelah Subak menjadi Warisan Budaya Dunia 3.1 Perkembangan Sosial Budaya Akan tetapi, dari pola tanam padi yang diterapkan oleh warga subak telah terjadi pergeseran. Pergeseran terjadi sebagai akibat adanya arus modernisasi dalam bidang pertanian yang melanda petani-petani di wilayah Bali termasuk di Desa Jatiluwih. Modernisasi dalam bidang pertanian, yaitu penggunaan traktor untuk membajak sawah dan penggunaan pupuk anorganik (urea) untuk menyuburkan tanah dan tanaman pertanian. Kesenian Tradisional. Kesenian tradisional yang ada di Desa Jatiluwih antara lain Baris Memedi (Baris Keraras), Baris Tombak, Topeng Sidakarya, Topeng Perembon, Joged Bumbung, Wayang, Angklung, Sekaha Gong Lanang, dan Sekaha Gong Istri. Dengan dijadikannya Desa Jatiluwih sebagai Desa Wisata dan ditetapkannya subak sebagai Warisan Budaya Dunia oleh UNESCO, keberadaan kesenian tradisional semakin dilestarikan dan dikembangkan. Organisasi Sosial. Seiring dengan perkembangannya, keberadaan desa adat di Desa Jatiluwih masih tetap dihormati sebagai organisasi sosial tradisional. Menurut Bendesa Adat Jatiluwih dan Bendesa Adat Gunungsari, hal-hal yang berkaitan dengan adat serta permasalahan mengenai adat, ditangani dan diatur oleh desa adat. Dari leluhur terdahulu, apabila ada masyarakat yang menghadapi permasalahan adat akan diselesaikan bersama antara pengurus desa adat dan warga desa, hal ini masih berlaku sampai saat ini. Selain desa adat sebagai organisasi sosial yang mengurus tentang adat, di Desa Jatiluwih terdapat organisasi tradisional dalam bidang kesenian. Organisasi bidang kesenian (sekaha) yang dimiliki Desa Jatiluwih yaitu JUMPA Volume 2 Nomor 1, Juli 2015 66 Perkembangan Desa Wisata Jatiluwih Setelah UNESCO Menetapkan Subak Sebagai Bagian dari Warisan Budaya DARMA PUTRA Foto 2. Wisatawan menikmati pesona sawah Jatiluwih dengan bersepeda. Foto 2. Wisatawan menikmati pesona sawah Jatiluwih dengan bersepeda. Sekaha Angklung, Sekaha Gong Lanang, Sekaha Gong Istri, dan Sekaha Shanti. Organisasi bidang kesenian (sekaha-sekaha) tersebut terus dilestarikan dan dikembangkan. Sekaha Angklung, Sekaha Gong Lanang, Sekaha Gong Istri, dan Sekaha Shanti. Organisasi bidang kesenian (sekaha-sekaha) tersebut terus dilestarikan dan dikembangkan. Disamping keindahan terasering sawah sebagai daya tarik wisata, keberadaan subak juga merupakan potensi lain yang dimiliki Desa Wisata Jatiluwih. Subak sebagai organisasi tradisional yang mengatur pengairan, merupakan organisasi sosial yang diwarisi secara turun-temurun oleh masyarakat setempat. Berkembangnya jaman tidak mempengaruhi eksistensi Subak Jatiluwih, terlebih lagi dengan ditetapkannya subak sebagai Warisan Budaya Dunia. Masyarakat Jatiluwih mayoritas bermata pencaharian sebagai petani, semakin termotivasi untuk menjaga keberadaan subak. Desa adat sebagai organisasi sosial yang mengatur hubungan antar warga Desa Jatiluwih mengenai hal-hal yang berkaitan dengan adat. Subak sebagai organisasi yang mengatur segala sesuatu yang berhubungan dengan pertanian milik warga desa. Kedua organisasi tersebut tetap dijaga keberadaannya, sebagai wujud penerapan aspek pawongan di kalangan masyarakat setempat. 3.2 Perkembangan Ekonomi Perkembangan ekonomi Desa Wisata Jatiluwih setelah penetapan subak sebagai Warisan Budaya Dunia meliputi lapangan kerja dan kesempatan berusaha, pendapatan serta investasi. Lapangan Kerja dan Kesempatan Berusaha. Perkembangan 67 JUMPA Volume 2 Nomor 1 Juli 2015 Dewa Ayu Diyah Sri Widari pariwisata di Desa Jatiluwih memberi dampak terhadap lapangan pekerjaan. Sejak ditetapkan sebagai desa wisata dan daya tarik wisata, jumlah kunjungan wisatawan ke Desa Wisata Jatiluwih mengalami peningkatan (Lihat Foto 2). Meningkatnya jumlah kunjungan wisatawan, diiringi juga dengan peningkatan fasilitas/sarana pariwisata yang diperlukan untuk menunjang perkembangan pariwisata. Setelah penetapan subak sebagai warisan Budaya Dunia, jumlah tenaga kerja yang terserap dari usaha pariwisata mengalami peningkatan. Jumlah tenaga kerja langsung pada usaha pariwisata di Desa Wisata Jatiluwih sebelum penetapan subak sebagai Warisan Budaya Dunia sebanyak 48 orang. Setelah penetapan meningkat menjadi 63 orang, dari jumlah tersebut 68,25% atau 43 orang berasal dari Desa Jatiluwih dan 31,75% atau 20 orang berasal dari luar Desa Jatiluwih. Pendapatan. Pendapatan yang diterima oleh masyarakat Desa Wisata Jatiluwih berasal dari usaha-usaha pariwisata maupun pengelolaan desa wisata. Dari usaha pariwisata, pendapatan yang diterima oleh pengusaha pariwisata mengalami peningkatan. Pengusaha pariwisata mengemukakan bahwa setelah penetapan subak sebagai Warisan Budaya Dunia, pendapatan dari usaha yang ditekuni meningkat sebesar 20% sampai 40%. Setelah penetapan subak sebagai Warisan Budaya Dunia, dari pendapatan kotor (bruto) yang diterima oleh Badan Pengelola Daya Tarik Wisata Jatiluwih, setelah dikurang asuransi maka diperoleh pendapatan bersih (netto). Selanjutnya, pendapatan netto didistribusikan untuk manajemen operasional 15%, pengembangan 10%, promosi 5%, dan badan pengelola 10%. Kemudian, dari sisa pendapatan tersebut didistribusikan lagi untuk Pemerintah Daerah Kabupaten Tabanan 45% dan untuk desa 55%. Pendapatan yang diterima oleh desa, didistribusilkan lagi, yaitu untuk Desa Dinas 25%, Desa Adat Jatiluwih 30%, Desa Adat Gunungsari 20%, Subak Jatiluwih 21%, Subak Abian Jatiluwih 2%, dan Subak Abian Gunungsari 2%. Pendapatan yang diperoleh Badan Pengelola Daya Tarik Wisata Jatiluwih berupa pendapatan kotor (bruto) tahun 2011 sebesar Rp 466.593.000,00. Pada tahun 2012 pendapatan kotor sebesar Rp 804.010.400,00 terjadi peningkatan 72,31% dari tahun 2011. Tahun 2013 jumlah pendapatan kotor sebesar Rp 1.509.795.500,00 terjadi peningkatan 87,78% dari tahun 2012. Tahun 2014 (sampai Nopember) pendapatan kotor sebesar Rp. 3.001.059.500,00 terjadi peningkatan sebesar 98,77% dari tahun 2013. Pendapatan yang diterima dari pengelolaan desa wisata oleh desa adat, dan subak dipergunakan untuk membiayai upacara keagamaan. Dengan adanya pembagian pendapatan untuk desa adat dan subak, masyarakat Desa Jatiluwih tidak perlu lagi membayar iuran wajib pada saat ada upacara keagamaan. JUMPA Volume 2 Nomor 1, Juli 2015 3.2 Perkembangan Ekonomi Bagi masyarakat setempat, hal ini dapat meringankan biaya yang harus dikeluarkan untuk upacara keagamaan. Secara tidak langsung, 68 JUMPA Volume 2 Nomor 1, Juli 2015 Perkembangan Desa Wisata Jatiluwih Setelah UNESCO Menetapkan Subak Sebagai Bagian dari Warisan Budaya DEWA AYU DIYAH SRI WIDARI Foto 3. Fasilitas ATM di Desa Jatiluwih. DEWA AYU DIYAH SRI WIDARI Foto 3. Fasilitas ATM di Desa Jatiluwih. masyarakat juga mendapat manfaat dari pengelolaan desa wisata. Investasi. Berkembangnya Desa Wisata Jatiluwih, mendorong masyarakat untuk berinvestasi secara langsung dalam usaha-usaha pariwisata. Usaha-usaha pariwisata seperti homestay, restoran, warung makan, dan penyewaan ATV (All Train Vehicle), sebagian besar modalnya berasal dari masyarakat setempat. Di samping investasi dalam usaha-usaha pariwisata yang dimiliki oleh masyarakat, setelah penetapan subak sebagai Warisan Budaya Dunia, terjadi penambahan investasi. Di sebelah selatan Puskesmas yang berdekatan dengan kantor Badan Pengelola Daya Tarik Wisata Jatiluwih, dibangun fasilitas Anjungan Tunai Mandiri (ATM) Bank Rakyat Indonesia. ATM tersebut dapat dipergunakan untuk bertransaksi mempergunakan VISA, Master Card, Cirrus, Link, dan ATM Bersama. Di areal parkir kendaraan (sebelah timur Billy’s Terrace Cafe) juga sudah dibangun fasilitas toilet umum bagi wisatawan dengan kondisi kebersihan yang baik. Fasilitas ATM Bank Rakyat Indonesia di Desa Wisata Jatiluwih seperti Foto 3. 4. Partisipasi Masyarakat 4.1 Partisipasi pada Tahap Perencanaan 4.1 Partisipasi pada Tahap Perencanaan Dalam pengembangan desa wisata, masyarakat Desa Jatiluwih kurang dilibatkan dalam proses pengambilan keputusan. Seperti pada kasus kontrak Billy’s Terrace Cafe, pada awal pendiriannya kafe ini diberi nama 69 JUMPA Volume 2 Nomor 1 Juli 2015 Dewa Ayu Diyah Sri Widari Cafe Jatiluwih. Cafe Jatiluwih milik masyarakat Desa Jatiluwih karena pembangunannya atas inisiatif masyarakat desa dan dibangun di atas tanah pelabe (milik) Pura Dalem. Pada awalnya pengelolaan dilakukan oleh masyarakat desa setempat. Karena pengelolaannya kurang maksimal, keuntungan sulit dihasilkan, akhirnya kafe ini dikontrakkan oleh Desa Adat Jatiluwih. Namun pada proses pengambilan keputusan mengenai besarnya biaya sewa, jangka waktu kontrak, serta hal-hal lain yang berkaitan dengan perjanjian kontrak, masyarakat tidak dilibatkan. Mengacu pada tipologi Pretty (dalam Mowforth dan Mount, 2000:146), partisipasi masyarakat pada tahap perencanan ini tergolong sebagai partisipasi manipulatif. Partisipasi manipulatif yaitu partisipasi orang- orang yang mewakili organisasi resmi, tetapi mereka tidak terpilih dan tidak memiliki kekuasaan. Digolongkan demikian karena pada tahap ini masyarakat setempat memang dilibatkan dalam sosialisasi pengembangan Desa Wisata Jatiluwih, tetapi masyarakat tidak memiliki kekuasaan untuk menolak program-program yang sudah ditetapkan. Implementasi aspek pawongan dari konsep Tri Hita Karana dalam partisipasi masyarakat pada tahap perencanaan tidak berjalan dengan baik. Masyarakat Desa Jatiluwih tidak dilibatkan pada proses pengambilan keputusan dalam pembangunan desanya. Masyarakat diikutsertakan dalam sosialisasi program-program kepariwisataan, namun tidak memiliki kemampuan untuk menolak program-program yang sudah ditetapkan. Hal ini menyebabkan terjadinya ketidakharmonisan hubungan antara masyarakat dengan badan pengelola, desa adat, dan pemerintah. 4.2 Partisipasi pada Tahap Pengembangan 4.2 Partisipasi pada Tahap Pengembangan Berdasarkan wawancara dengan pengusaha pariwisata (tanggal 2-3 Januari 2015), terdapat peningkatan jumlah usaha-usaha pariwisata yang terdapat di Desa Wisata Jatiluwih. Sebelum penetapan subak sebagai Warisan Budaya Dunia, jumlah usaha pariwisata yang ada sebanyak sembilan, yaitu Warung Makan Nadi, Warung Makan Sari Rasa, Galang Kangin Inn dan Warung Makan, Warung Terrace Subak, Warung Dhea, Warung Krisna, Billy’s Terrace Cafe, Billy’s Villas, dan Padi Bali. Setelah penetapan subak sebagai Warisan Budaya Dunia, terjadi penambahan pendirian usaha pariwisata yaitu restoran yang diberi nama J’Terrace, Warung Makan Wayan, dan Warung Terrace Subak menambah fasilitas rumah makan dengan fasilitas penginapan (berubah nama menjadi Warung Terrace Subak and Homestay). Partisipasi masyarakat secara langsung dalam pengembangan desa wisata juga diwujudkan dengan terlibat pada pendirian dan operasional badan pengelola. Sekretariat Badan Pengelola Daya Tarik Wisata Jatiluwih seperti Foto 2. Pendirian badan pengelola sebagai upaya untuk dapat 70 JUMPA Volume 2 Nomor 1, Juli 2015 Perkembangan Desa Wisata Jatiluwih Setelah UNESCO Menetapkan Subak Sebagai Bagian dari Warisan Budaya mengembangkan desa wisata secara profesional. Sebelum dibentuk badan pengelola, pengembangan desa wisata dikelola oleh desa dinas. Setelah pendirian badan pengelola, masyarakat lokal yang terlibat dan berpartisipasi dalam pengelolaan desa wisata semakin meningkat. Sumber Daya Manusia (SDM) yang dilibatkan dalam operasional dan manajemen diambil dari masing-masing dusun yang terdapat di Desa Jatiluwih. Implementasi aspek pawongan dalam partisipasi masyarakat pada tahap pengembangan berjalan dengan baik. Masyarakat berpartisipasi dalam pengelolaan usaha-usaha pariwisata dan dilibatkan pada operasionalisasi badan pengelola daya tarik wisata. Dengan adanya pelibatan masyarakat tersebut, membuat terjalinnya keseimbangan hubungan antara masyarakat dengan badan pengelola. 4.3 Partisipasi Pada Tahap Pengawasan 4 3 p p g Partisipasi yang dilakukan oleh masyarakat Desa Jatiluwih, pada tahap pengawasan untuk aspek operasional pengembangan desa wisata lebih bersifat praktis. Dalam hal ini, masyarakat yang bekerja pada badan pengelola daya tarik wisata melakukan pengawasan terhadap pelaksanaan operasionalisasi kegiatan desa wisata. Masyarakat melakukan pengawasan secara aktif dan langsung berdasarkan pada tugasnya masing-masing yang sudah ditentukan oleh badan pengelola. Pengawasan umumnya dilakukan oleh masyarakat yang bekerja sebagai petugas tiket, parkir, keamanan, dan pemandu wisata trekking. Pengecekan ulang pembayaran retribusi yang dilakukan oleh petugas tiket sebagai bagian dari karyawan pada manajemen operasional Badan Pengelola Daya Tarik Wisata Jatiluwih, merupakan salah satu kewenangan yang diberikan oleh desa adat kepada karyawan bagian tiket. Petugas tiket memiliki wewenang untuk melakukan pengecekan ulang jumlah karcis yang terjual. Selain pengawasan melekat yang dilakukan secara langsung oleh masyarakat lokal yang bekerja pada badan pengelola, untuk memantau perkembangan desa wisata, pengawasan juga dilakukan pada tingkat manajerial. Setelah penetapan subak sebagai Warisan Budaya Dunia, dibentuk Badan Pengelola Daya Tarik Wisata Jatiluwih berdasarkan Peraturan Bupati Tabanan Nomor 84 Tahun 2013 (Foto 4). Bidang Pengawas bertugas untuk melakukan monitoring dan evaluasi terhadap perkembangan Desa Wisata Jatiluwih. Monitoring dilakukan terhadap program-program pengembangan desa wisata, baik mencakup aspek perencanaan dan pelaksanaan. Pelibatan pihak Pemerintah Kabupaten Tabanan (pihak eksternal) dalam pengawasan sebagai upaya untuk dapat memberikan arahan, bimbingan, dan kontrol dari kegiatan yang dilaksanakan. Namun, setelah kepengurusan berjalan selama kurang 71 JUMPA Volume 2 Nomor 1 Juli 2015 Dewa Ayu Diyah Sri Widari Dewa Ayu Diyah Sri Widari DEWA AYU DIYAH SRI WIDARI Foto 4. Kantor badan pengelola daya tarik wisata Jatiluwih. DEWA AYU DIYAH SRI WIDARI Foto 4. Kantor badan pengelola daya tarik wisata Jatiluwih. lebih satu tahun, pengawasan yang dilakukan berjalan kurang efektif. Para pengurus khususnya pada Bidang Pengawas, memiliki kegiatan di kantor masing-masing sehingga jarang melaksanakan tugas sebagai pengawas pada Badan Pengelola Daya Tarik Wisata Desa Jatiluwih. Meningkatnya jumlah kunjungan wisatawan setelah penetapan subak se- bagai Warisan Budaya Dunia, berimplikasi pada peningkatan fasilitas pari- wisata yang diperlukan. Kondisi ini menyebabkan terjadinya pembangunan fasilitas restoran dan rumah makan yang tidak terkendali serta tidak sesuai dengan aturan tata ruang. Fasilitas restoran dan rumah makan dibangun di atas lahan yang merupakan bagian dari Warisan Budaya Dunia, yang se- harusnya tidak boleh ada pendirian bangunan. Pemilik lahan beranggapan bahwa mereka berhak dan bebas melakukan pembangunan karena lahan tersebut milik mereka sendiri. JUMPA Volume 2 Nomor 1, Juli 2015 4.3 Partisipasi Pada Tahap Pengawasan Bendesa Adat dan Pekaseh Subak Jatiluwih yang mengetahui tentang ketidaksesuaian tersebut lebih bersikap diam. Pa- dahal seharusnya berhak melakukan pengawasan dan kontrol dari setiap pembangunan di daerahnya. Pengawasan terhadap zonasi Warisan Budaya Dunia juga dilakukan oleh Pekaseh Subak Jatiluwih. Namun pengawasan tersebut mengalami kendala, selain disebabkan oleh kurangnya keterlibatan masyarakat dalam pembuatan perencanaan pengembangan desa wisata juga diakibatkan oleh 72 JUMPA Volume 2 Nomor 1, Juli 2015 Perkembangan Desa Wisata Jatiluwih Setelah UNESCO Menetapkan Subak Sebagai Bagian dari Warisan Budaya belum adanya Peraturan Daerah yang secara khusus mengatur tentang zonasi di kawasan Warisan Budaya Dunia. Pekaseh beserta masyarakat tidak mengetahui apa yang direncanakan dan apa yang perlu diawasi. Masyarakat setempat tidak memiliki kompetensi untuk melakukan pengawasan. Claiborne (2010:42) menyatakan, jelas yang ideal bagi masyarakat untuk memutuskan bersama-sama bentuk dan fungsi perkembangan pariwisata dan memiliki kontrol penuh atas setiap skema pariwisata di lokasi mereka. Partisipasi masyarakat pada pengawasan menyebabkan terjadinya konflik antara badan pengelola, pemerintah, dan masyarakat. Konflik ini merupakan indikasi tidak berjalannya implementasi pawongan. Harapan yang ditujukan kepada pemerintah daerah untuk lebih aktif melakukan pengawasan, pembinaan, dan pendampingan pada pengelolaan daya tarik wisata, juga tidak efektif. Hal ini menimbulkan ketidakpuasan masyarakat, dan memicu hubungan yang kurang harmonis antara pihak-pihak yang terkait. 5. Persepsi Wisatawan Persepsi wisatawan terhadap perkembangan Desa Wisata Jatiluwih Setelah Penetapan Subak Sebagai Warisan Budaya Dunia mencakup implementasi pada aspek parhyangan, pawongan, dan palemahan. 5.1 Persepsi Wisatawan terhadap Aspek Parhyangan 5 p p p y g Wisatawan yang berkunjung ke Desa Wisata Jatiluwih selain melihat terasering sawah, terdapat dua orang yang mengunjungi Pura Desa tempat penyelenggaraan upacara Ngusaba Desa. Anton dan Pieter Soder (wisatawan asal Belanda) mengungkapkan bahwa kondisi tempat suci (pura) terpelihara dengan baik. Kebersihan pura terjaga serta disediakannya tempat sampah di areal pura. Begitu memasuki pura, wisatawan diminta untuk mengenakan selendang sebagai persyaratan untuk masuk ke tempat suci. Wisatawan asal Belanda, dengan senang hati mentaati peraturan tersebut karena sudah mengetahui tata cara untuk masuk ke tempat suci dan sebagai bentuk penghormatan kepada masyarakat setempat (hasil wawancara tanggal 19 Maret 2015). Wisatawan yang berkunjung menikmati terasering sawah seperti Pierre, Yuelle, Susan Allen, dan Lasandra Smith menyatakan palinggih yang ada pada tiap-tiap sawah milik petani kondisinya bagus. Bangunan palinggih pada sawah milik petani Desa Jatiluwih tergolong unik, karena atapnya ada yang menggunakan ijuk, mencerminkan bangunan tradisional Bali (hasil wawancara tanggal 19 Maret 2015). Persepsi wisatawan terhadap perkembangan Desa Wisata Jatiluwih dari kondisi tempat suci yang terpelihara dengan baik, arsitektur palinggih yang unik sebagai apresiasi terhadap pelestarian budaya lokal, dan wisatawan 73 JUMPA Volume 2 Nomor 1 Juli 2015 Dewa Ayu Diyah Sri Widari mendapatkan pengalaman baru, merupakan bentuk implementasi aspek parhyangan yang masih terpelihara dengan baik. 5.2 Persepsi Wisatawan terhadap Aspek Pawongan Pelayanan yang diberikan oleh petugas tiket memang ramah dan sopan, namun wisatawan tidak mendapatkan informasi yang diperlukan. Dari delapan orang wisatawan yang diwawancarai mengungkapkan seharusnya pada saat membeli tiket masuk, wisatawan diberikan informasi mengenai penggunaan uang retribusi sehingga wisatawan mengetahui pemanfaatan dari uang retribusi yang mereka bayarkan. Informasi lain yang berkaitan dengan daya tarik wisata juga tidak didapatkan oleh wisatawan. Padahal wisatawan memerlukan informasi mengenai sejarah, potensi yang dimiliki, dan hal-hal lain yang berkaitan dengan keberadaan subak sebagai Warisan Budaya Dunia. Dari hasil wawancara, semua wisatawan menyatakan merasa aman pada saat melakukan kunjungan ke Desa Wisata Jatiluwih. Meskipun pada saat menikmati pemandangan terasering sawah, wisatawan memasuki areal persawahan tanpa ditemani oleh guide, wisatawan tetap merasa aman. Sesuai dengan persyaratan desa wisata, menurut Nuryanti (1993) bahwa keamanan desa tersebut harus terjamin. Dari segi keamanan, wisatawan memberikan apresiasi yang baik, tetapi dari faktor kenyamanan tidak mendapatkan respons yang bagus. Semua wisatawan menyatakan tidak nyaman pada saat melakukan perjalanan baik dari arah Desa Senganan maupun dari arah Batukaru menuju Desa Wisata Jatiluwih yang disebabkan oleh jalan rusak. Wisatawan asal Kanada, Susan Allen (berprofesi sebagai guru), secara kritis menyarankan perlu adanya kerja sama antara Badan Pengelola Daya Tarik Wisata Jatiluwih dan pemerintah untuk mengetahui perasaan serta pemahaman para petani tentang apa yang terjadi di Desa Wisata Jatiluwih dalam kaitannya dengan penetapan subak sebagai Warisan Budaya Dunia (hasil wawancara tanggal 19 Maret 2015). Dalam pengembangan desa wisata, bukan hanya menjaga keharmonisan hubungan antara wisatawan dengan masyarakat (aspek pawongan), namun perlu juga menjaga keharmonisan antara petani dengan pengelola daya tarik wisata, dan pemerintah. JUMPA Volume 2 Nomor 1, Juli 2015 5.3 Persepsi Wisatawan terhadap Aspek Palemahan Wisatawan yang menikmati panorama terasering sawah Jatiluwih, menyatakan kekagumannya dengan pemandangan alam yang dimiliki. Di samping mengagumi keindahan terasering sawah, wisatawan juga berpendapat pemandangan di sekitar areal persawahan masih alami. Hijaunya hamparan sawah dengan latar belakang Gunung Batukaru yang JUMPA Volume 2 Nomor 1, Juli 2015 74 Perkembangan Desa Wisata Jatiluwih Setelah UNESCO Menetapkan Subak Sebagai Bagian dari Warisan Budaya masih asri, sungai-sungai dengan airnya yang jernih, pepohonan yang masih banyak dijumpai sepanjang perjalanan, membuat wisatawan merasakan suasana pedesaan yang masih alami. Selain persepsi positif wisatawan terhadap keunikan dan keindahan terasering sawah Jatiluwih, rata-rata wisatawan mengatakan bahwa kebersihan di sekitar daya tarik wisata sangat baik. Tidak ditemukan sampah di sekitar areal persawahan, sepanjang jalur trekking maupun pada fasilitas restoran yang dikunjungi. Penyediaan tempat sampah yang dipisahkan antara sampah organik dan anorganik, sudah sesuai dengan standar kebersihan. Dari segi jumlah, tempat sampah masih diperlukan penambahan. Terasering sawah sebagai daya tarik wisata ditempatkan pada posisi utama sehingga wisatawan mudah untuk menikmati pemandangan. Jarak antara tempat parkir dengan lokasi terasering sawah yang cukup dekat, membuat wisatawan mudah untuk mengakses daya tarik wisata. Semua wisatawan berpendapat bahwa fasilitas pariwisata yang disediakan kurang mencerminkan arsitektur tradisional Bali. Beberapa fasilitas pariwisata ada yang memakai arsitektur tradisional Bali, namun ada juga yang tidak. Kondisi ini tidak sesuai dengan konsep desa wisata, bukan hanya lingkungan yang mencerminkan suasana lokal, fasilitas pariwisata yang ditawarkan juga sebaiknya bernuansa tradisional (hasil wawancara tanggal 19 Maret 2015). Fasilitas souvenirshop merupakan bagian dari fasilitas pokok yang disediakan bagi wisatawan. Sebagian besar wisatawan tidak mengetahui cendera mata khas Desa Wisata Jatiluwih. Tidak dijumpai fasilitas souvenirshop yang secara khusus menjual barang kerajinan dan makanan atau minuman khas hasil karya masyarakat Desa Jatiluwih. Wisatawan hanya mendapat informasi dari guide, bahwa produk beras merah dan teh beras merah sebagai cendera mata khas desa wisata tersebut. Dalam perkembangan Desa Wisata Jatiluwih dari aspek palemahan, diperlukan upaya-upaya agar terjaga keharmonisan antara manusia (masyarakat Desa Jatiluwih dan wisatawan) dengan lingkungan alam. Natori (dalam Ardika, 2015:60) bahwa paradigma pengembangan pariwisata berkelanjutan yang berbasis masyarakat mensyaratkan terpeliharanya mutu sumber daya alam dan budaya, meningkatnya kesejahteraan masyarakat lokal, dan terwujudnya kepuasan wisatawan. 6. Simpulan dan Saran Perkembangan sosial budaya dan ekonomi Desa Wisata Jatiluwih setelah penetapan subak Jatiluwih sebagai Warisan Budaya Dunia dari aspek Tri Hita Karana berdasarkan tujuh aspek. Aspek tersebut yaitu kegotongroyongan, budaya pertanian, kesenian tradisional, organisasi sosial, lapangan kerja 75 JUMPA Volume 2 Nomor 1 Juli 2015 Dewa Ayu Diyah Sri Widari dan kesempatan berusaha, pendapatan, serta investasi. Kegotongroyongan, upacara yang berkaitan dengan aktivitas petani, dan sistem tanam padi semakin terpelihara. Kesenian tradisional semakin dilestarikan dan dikembangkan, organisasi sosial semakin dijaga keberadaannya. Terjadi pergeseran penggunaan sapi/kerbau untuk membajak sawah ke penggunaan traktor, penggunaan pupuk organik ke pupuk anorganik. Lapangan kerja dan kesempatan berusaha, pendapatan serta investasi mengalami peningkatan. Dari aspek parhyangan, pelaksanaan upacara yang berkaitan dengan aktivitas petani tetap dilaksanakan dengan baik. Aspek palemahan diimplementasikan dengan mempertahankan sistem tanam padi, namun berjalan kurang baik pada pola tanam yang diterapkan oleh petani. Dari aspek pawongan, implementasinya cukup baik dengan adanya organisasi sosial, peningkatan lapangan kerja dan kesempatan berusaha, pendapatan serta investasi. Partisipasi masyarakat dalam perkembangan Desa Wisata Jatiluwih setelah penetapan subak sebagai Warisan Budaya Dunia dari aspek Tri Hita Karana dikaji dari tiga tahap yaitu perencanaan, pengembangan dan pengawasan. Partisipasi masyarakat pada tahap perencanaan dan pengawasan bersifat manipulatif dan fungsional. Pada tahap pengembangan, masyarakat berpartisipasi aktif dalam pengelolaan fasilitas pariwisata dan operasionalisasi badan pengelola daya tarik wisata. Implementasi aspek pawongan berjalan kurang baik pada tahap perencanaan dan pengawasan, pada tahap pengembangan berjalan dengan baik. Persepsi wisatawan pada implementasi aspek parhyangan baik. Persepsi wisatawan terhadap implementasi aspek pawongan dan palemahan kurang baik dalam hal pemberian informasi, kenyamanan, dan penyediaan fasilitas umum. Wisatawan memberikan apresiasi positif dari segi keamanan, pemandangan yang masih alami, dan kebersihan daya tarik wisata. Saran yang dapat direkomendasikan yaitu untuk Pemerintah Kabupaten Tabanan, hendaknya terus memberikan dukungan, pendampingan, dan sosialisasi. Untuk Badan Pengelola Daya Tarik Wisata Jatiluwih, produk barang kerajinan (souvenir) sebagai ciri khas Desa Wisata Jatiluwih perlu dikembangan, perlu meningkatkan kemampuan kerja masyarakat lokal dalam sektor pariwisata. Untuk Masyarakat Desa Jatiluwih, harus mampu mempertahankan pengakuan Warisan Budaya Dunia, menjaga keamanan dan ketertiban di wilayah desanya, serta mencegah dan menghindari terjadinya konflik. Ucapan Terima Kasih Ucapan terima kasih disampaikan kepada Prof. Dr. dr. A.A. Raka Sudewi, Sp.S(K) selaku Direktur Program Pascasarjana Universitas Udayana. Prof. Dr. I Nyoman Darma Putra, M.Litt., selaku Ketua Program Studi Magister 76 JUMPA Volume 2 Nomor 1, Juli 2015 Perkembangan Desa Wisata Jatiluwih Setelah UNESCO Menetapkan Subak Sebagai Bagian dari Warisan Budaya Kajian Pariwisata Program Pascasarjana Universitas Udayana. Apresiasi dan ucapan terima kasih juga disampaikan kepada Dr. Ir. Syamsul Alam Paturusi, MSP., sebagai Sekretaris Program Studi Magister Kajian Pariwisata Program Pascasarjana Universitas Udayana. Prof. Dr. Made Antara, M.S. dan Dr. Ir. Syamsul Alam Paturusi, MSP., sebagai Pembimbing I dan Pembimbing II. Para penguji yang terdiri dari Prof. Dr. I Nyoman Sirtha, SH.,MS., Prof. Dr. I Wayan Ardika, M.A., dan Dr. Ir. I Made Adhika, MSP., yang telah memberikan banyak masukan dan koreksi untuk penyempurnaan penelitian ini. Ucapan terima kasih juga disampaikan kepada semua pihak yang telah membantu dalam penelitian ini. Kajian Pariwisata Program Pascasarjana Universitas Udayana. Apresiasi dan ucapan terima kasih juga disampaikan kepada Dr. Ir. Syamsul Alam Paturusi, MSP., sebagai Sekretaris Program Studi Magister Kajian Pariwisata Program Pascasarjana Universitas Udayana. Prof. Dr. Made Antara, M.S. dan Dr. Ir. Syamsul Alam Paturusi, MSP., sebagai Pembimbing I dan Pembimbing II. Para penguji yang terdiri dari Prof. Dr. I Nyoman Sirtha, SH.,MS., Prof. Dr. I Wayan Ardika, M.A., dan Dr. Ir. I Made Adhika, MSP., yang telah memberikan banyak masukan dan koreksi untuk penyempurnaan penelitian ini. Ucapan terima kasih juga disampaikan kepada semua pihak yang telah membantu dalam penelitian ini. Daftar Pustaka Penyusunan Tata Ruang dan Rencana Detail Teknis Desa Wisata Terpadu di Bali. Yogyakarta: Fakultas Teknik. Peraturan Pemerintah Nomor 2 Tahun 2012 tentang Kepariwisataan Budaya Bali. Daftar Pustaka Ardika, I Wayan. 2015. Warisan Budaya Perspektif Masa Kini. Denpasar: Udayana University Press. Butler, Richard. 1980. The Tourism Area Life Cycle: Application and Modifications. Butler, Richard. 1980. The Tourism Area Life Cycle: Application and Modifications. Great Britain: Cromwell Press. Great Britain: Cromwell Press. Claiborne, Petra. 2010. ”Community Participation in Tourism Development and the Great Britain: Cromwell Press. Claiborne, Petra. 2010. ”Community Participation in Tourism Development and the Value of Social Capital”. Tourism and Hospitality Management. Vol. 84:1-79. Value of Social Capital . Tourism and Hospitality Management. Vol. 84:1 79. Cooper, Chris, Jhon Fletcher, David Gilbert, Stephen Wanhill. 1999. Kepariwisataan Prinsip-prinsip dan Pelaksanaan. Bagian Proyek Penerjemahan dan Penerbitan Buku-buku Pariwisata Jakarta Tahun Anggaran 1999/2000. Cravens, David W. 1997. Strategic Marketing. Texas: Irwin.i Cravens, David W. 1997. Strategic Marketing. Texas: Irwin.i Heny Urmila Dewi, Made, Chafid Fandeli, M. Baiquni. 2013. ”Pengembangan Desa Wisata Berbasis Partisipasi Masyarakat Lokal di Desa Wisata Jatiluwih, Tabanan Bali”. Jurnal Kawistara. Vol. 2(2):117-226. Mowforth, Martin dan Ian Maunt. 2000. Pariwisata dan yang Berkelanjutan. Pariwisata Baru di Dunia Ketiga. Jakarta: Bagian Proyek Pengembangan Literatur Pariwisata Tahun Anggaran 2000. Musoman, Ansyari, Fadhilah Septiari, Na’immah Nur’Aini, Wardana Jaya Jasmitha. 2013. Identifikasi Dampak Ekonomi Terhadap Kawasan Wisata Taman Budaya Sentul. available at: http://www.slidersshare.net/ainiaikudou/ ekonomi-dan-bisnis-wisata-identifikasi-dampak-ekonom. diunduh pada tanggal 21 Oktober 2014. Nuryanti, Windu. 1993. Concept, Perspective and Challenges. Naskah Lengkap Laporan Konferensi Internasional mengenai Pariwisata Budaya. Yogyakarta: Gadjah Mada University Press. Pitana, I Gde dan Putu G. Gayatri. 2005. Sosiologi Pariwisata. Yogyakarta: Andi. Prasiasa, Dewa Putu Oka. 2010. ”Pengembangan Pariwisata dan Keterlibatan Masyarakat Di Desa Wisata Jatiluwih Kabupaten Tabanan” (disertasi). Denpasar: Program Pascasarjana Universitas Udayana. Timothy, D. J. 1999. ”Participatory Planning a View of Tourism in Indonesia” Annals Prasiasa, Dewa Putu Oka. 2010. ”Pengembangan Pariwisata dan Keterlibatan Masyarakat Di Desa Wisata Jatiluwih Kabupaten Tabanan” (disertasi). Denpasar: Program Pascasarjana Universitas Udayana. Timothy, D. J. 1999. ”Participatory Planning a View of Tourism in Indonesia” Annals of Tourism Research. Vol. 26(2). 77 JUMPA Volume 2 Nomor 1 Juli 2015 Dewa Ayu Diyah Sri Widari Universitas Gadjah Mada. 1992. Penyusunan Tata Ruang dan Rencana Detail Teknis Desa Wisata Terpadu di Bali. Yogyakarta: Fakultas Teknik. Peraturan Pemerintah Nomor 2 Tahun 2012 tentang Kepariwisataan Budaya Bali. Dewa Ayu Diyah Sri Widari Universitas Gadjah Mada. 1992. Penyusunan Tata Ruang dan Rencana Detail Teknis Desa Wisata Terpadu di Bali. Yogyakarta: Fakultas Teknik. Peraturan Pemerintah Nomor 2 Tahun 2012 tentang Kepariwisataan Budaya Bali. Universitas Gadjah Mada. 1992. Profil Penulis Dewa Ayu Diyah Sri Widari, A.Par adalah dosen tetap di Akademi Pariwisata (AKPAR) Denpasar sejak tahun 2013. Ia menyelesaikan pendidikan pada Program Studi Diploma IV Pariwisata Universitas Udayana tahun 1995. Dari tahun 1996 sampai 2012 pernah menjadi dosen di Sekolah Tinggi Pariwisata Sahid Jakarta. 78 JUMPA Volume 2 Nomor 1, Juli 2015
W1493878442.txt	null	fr		Recepción de la obra de Madame d’Aulnoy en España: traducciones y prólogos	Çedille	2,014	cc-by	8,243	ISSN: 1699-4949 nº 10, abril de 2014 * Artículos Recepción de la obra de Madame d’Aulnoy en España: traducciones y prólogos Carlota Vicens Pujol Universidad de las Islas Baleares cvicens@uib.es Résumé La carrière littéraire de Mme d'Aulnoy a parcouru un chemin assez complexe aussi bien en France qu’à l’étranger. En Espagne ce n’est qu’à partir de la deuxième moitié du XXème siècle que les éditions et réimpressions semblent se multiplier. La presse et les revues spécialisées se sont toutefois peu occupées de cette œuvre: seule l’analyse des préfaces permettra une approche critique à une œuvre aussi controversée. La première partie de cet article vise à établir le répertoire, de façon aussi exhaustive que possible, de toutes les traductions de l’œuvre de l’écrivain parues en Espagne pour aborder ensuite l’étude des différents prologues qui témoignent de l’accueil critique des différents titres. Abstract The literary career of Madame d'Aulnoy followed a rather complex course, both in France and abroad. In Spain, the first translation of one of her volumes dates back to 1838 and it’s not until the second half of the twentieth century that new editions and reprints seem to multiply. Press and trade journals, however, paid little attention to her books: only the analysis of the prologues allow a critical approach to such a controversial work. The aim of the first part of this paper is to establish the repertoire, as comprehensively as possible, of all the translations of Madame d’Aulnoy’s books published in Spain; then we focuse the study of the different prologues as a testimony of the critical reception in Spain of her works. Mots-clé: Mme d’Aulnoy, réception, Espagne, littérature comparée, traductions, prologues. Key words Madame d'Aulnoy, comparative literature, reception, Spain, translations, prologues. * Artículo recibido el 19/01/2014, evaluado el 17/02/2014, aceptado el 17/03/2014. Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol 0. Introducción Como no podía ser de otra manera, en España el nombre de Marie Catherine Le Jumel de Barneville (1651-1715), más conocida como Mme d’Aulnoy, se asocia ante todo a su Relation du voyage d’Espagne fait en 1679 y sus Mémoires sur la cour d’Espagne o también a su novela Histoire d’Hippolyte, comte de Douglas, siendo estas, por otro lado, las obras que le permitieron granjearse en Francia el reconocimiento de sus contemporáneos. Sin embargo, las escasas reseñas biográficas que aparecen en nuestras historias de la literatura, sea francesa o universal, se refieren por lo general a sus Contes des fées y se reducen normalmente a la mínima expresión. Se la define como una «mujer complicada y fantástica», «imitadora de los cuentos de Perrault»1 y parece que no se tuvo conocimiento en España ni del personaje ni de su obra hasta la segunda mitad del siglo XIX, aunque como apunta José F. Montesinos (1980: 159) debían existir ediciones anteriores. La fecha relativamente tardía de entrada de estos cuentos en España podría deberse al hecho de que en nuestro país la moda de los cuentos fantásticos nacidos de la tradición oral no había conseguido implantarse como lo hizo primero en Italia, con Lo Cunto de li Cunti de Gianbattista Basile o en Francia con Charles Perrault y, más adelante, en Alemania con los hermanos Grimm o en Dinamarca con Hans Christian Andersen. Tanto en Francia como en España, las críticas recalcan hoy en día el carácter innovador de esos cuentos de hadas, el primero de los cuales aparece insertado en Histoire d’Hippolyte, comte de Douglas (1690). Dicho cuento, que se extraería después de la novela con el título de L’île de la félicité, inauguraba en Francia la moda de los relatos fantásticos. El público francés del siglo XVIII recibió con entusiasmo la obra de Mme d’Aulnoy en su faceta de cuentista, como lo atestiguan las numerosas reediciones: Contes des Fées se reeditó en 1710, 1711, 1715, 1725, 1742, 1774, 1785 y 1793 (Jasmin, 2008: 70). Por su parte, Contes nouveaux ou les fées à la mode cuenta con reediciones de los años 1711, 1715, 1725, 1742, 1779, 1785 y 1793 (Jasmin, 2008: 18). Al finalizar el siglo, los cuentos de Mme d’Aulnoy fueron recogidos en los 41 volúmenes de Le Cabinet des fées ou Collection choisie des Contes des fées et autres contes merveilleux, ornés de figures, publicados en Amsterdam entre 1785 y 1789 por Charles- Joseph Mayer. Luego, a lo largo del siglo XIX, estos cuentos van a perder su estatus de literatura seria y entrarán a formar parte de la llamada littérature de colportage2. 1 José García Mercadal in Viajes por España. Alianza Editorial, Madrid 1972: 186. Indicaremos a pie de página, y no en la bibliografía, aquellos libros que no nos ofrecen más que un dato. 2 Así, el repertorio de los títulos de la Bibliothèque Bleue elaborado por Lise Andries y Geneviève Bollème cuenta con un total de diecinueve títulos de Mme d’Aulnoy, con un promedio de tres reimpresiones para cada uno de ellos. En los siglos XX y XXI, se vuelven a recuperar esos cuentos con cuatro ediciones: Le cabinet des fées (tomos III, IV y V), Ginebra, Slatkine, 1978; Le Cabinet des fées (el cual reúne la totalidad de los cuentos), Arles, Picquier Poche, 1994; Contes de Madame d’Aulnoy. http://cedille.webs.ull.es/10/22vicens.pdf 368 Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol Que los demás títulos de la escritora tuvieron igualmente cierto éxito lo demuestran las seis reediciones de Mémoires sur la cour d’Espagne entre 1690 y 1693, así como las doce reediciones de la Relation du Voyage d’Espagne entre finales del siglo XVII y el siglo XVIII. Pero ¿se conoce en realidad la obra de Mme d’Aulnoy al sur de los Pirineos? Aunque su Relation du voyage d’Espagne o sus Contes des fées hayan sido objeto de algunos estudios aislados, es preciso elaborar un inventario de los títulos traducidos para poder sacar las primeras conclusiones. Este será nuestro cometido en la primera parte del presente estudio3, mientras que la segunda parte, centrada en el discurso paratextual, estará dedicada al análisis de los distintos prólogos que preceden los textos traducidos: ¿de qué manera se presenta el relato del viaje por España, en el que los estereotipos y las anécdotas más banales se suceden? y ¿cómo fueron acogidos los Contes de fées, cuyas traducciones, como podremos comprobar, son mucho más numerosas? ¿Estamos en presencia de una escritora con talento, innovadora o, al contrario, como se ha dicho con frecuencia, estamos ante una simple imitadora de Perrault y una novelista mediocre? 1. Repertorio de las traducciones: 1830-2010 Las fuentes bibliográficas de las que disponemos para elaborar el repertorio de las traducciones de la obra de Mme d’Aulnoy en España (en castellano o en otras lenguas nacionales) son escasas. Según hemos podido comprobar, su nombre está ausente de los principales estudios sobre recepción de la literatura francesa en España. Para mayor claridad hemos dividido el inventario que, como se ha señalado, se convierte en nuestro primer objetivo, en dos partes: los cuentos de hadas publicados aisladamente o por grupos de tres o cuatro, y el resto de las obras, que conforman una lista mucho más reducida. 1.1. Los cuentos de hadas: Como puede observarse a continuación, se trata de un total de cuarenta y una ediciones o reediciones, dos de la cuales datan del siglo XIX y el resto de los siglos XX y XXI, siendo la segunda mitad del siglo XX más prolífica con la publicación de veintidós títulos frente a los dieciséis de la primera mitad4: Contes I : Les contes des fées et Contes II : Contes nouveaux, París, STEM, 1997-98 y Contes de fées (2 vol.), París, Champion, 2008. 3 Este artículo se incluye en la producción científica del grupo de investigación consolidado «Literatura popular francesa y cultura mediática» (2009-2014 SGR 646), cuya investigadora principal es la Dra. Àngels Santa. 4 Garmendia de Otaola, en Lecturas buenas y malas a la luz del dogma y de la moral (Bilbao, El Mensajero de Cristo, 1953) se refiere a una recopilación que no hemos podido localizar, editada en Barcelona bajo el título Cuentos de d’Aulnoy y de todos los países, a propósito del cual dice que «en general [estos cuentos] son poco adecuados para niños» (p. 30). http://cedille.webs.ull.es/10/22vicens.pdf 369 Çédille, revista de estudios franceses, 10 (2014), 367-383 – – – – – – – – – – – – – – – Carlota Vicens Pujol Bella Bella o el caballero afortunado (1844), Vdª de Brieva, Logroño (Traducción: José Llorente). Cuentos de Madama de Aulnoy (1852), Imprenta de la Biblioteca Universal, Madrid5. Cuentos de la condesa de Aulnoy (1909), Henrich & Cª, Barcelona. (Adaptación española y prólogo: Pedro Umbert). Los cuentos aquí incluidos son: Fortunata, El buen ratoncillo y La bella de los cabellos de oro. El pájaro azul; Riquet el copetudo. Cuentos de la condesa de Aulnoy y de Carlos Perrault (1911), Henrich & Cª, Barcelona (Versión española: Pedro Umbert). Cuentos de Mme d’Aulnoy (1918), Calleja, Madrid (Adaptación: K. Fitz Gerald; Traducción: E. Díez Canedo; Ilustraciones: Thomas Derrik). Los cuentos incluidos en este volumen son: La princesa de los cabellos de oro, El pájaro azul y La gata blanca. La princesa Graciosa y el príncipe florido (1931), Cuentos Clásicos, Barcelona (Adaptación: Jesús Sánchez Tena). El pájaro azul (1931), Juventud, Cuentos Clásicos 20, Barcelona (Adaptación: Jesús Sánchez Tena). La gata blanca (1932), Juventud, Cuentos Clásicos 21, Barcelona (Adaptación: Jesús Sánchez Tena). Deseada, la princesa cierva (1932), Juventud, Cuentos Clásicos 22, Barcelona (Adaptación: Jesús Sánchez Tena). Leandro, el príncipe gnomo (1932), Juventud, Cuentos Clásicos 23, Barcelona (Adaptación: Jesús Sánchez Tena). La princesa Graciosa i el príncep Florind / Riquel del Plomall (1935)6, Mentora, Contes d’ahir i d’avui, serie 1, 19, Barcelona. (Adaptación: Valeri Serra i Bondú). L’ocell blau, (1935), Mentora, Contes d’ahir i d’avui, serie 1, 20, Barcelona. (Adaptación: Valeri Serra i Bondú). La gata blanca (1935), Mentora, Contes d’ahir i d’avui, serie 1, 21, Barcelona. (Adaptación: Lluís Capdevila). Història de Desitjada, la princesa cervola (1935), Mentora, Contes d’ahir i d’avui, serie 1, 22, Barcelona. (Adaptación: Lluís Capdevila). Leandre, el príncep gnom (1935), Mentora, Contes d’ahir i d’avui, serie 1, 23, Barcelona. (Adaptación: Lluís Capdevila). 5 Estos dos primeros libros aparecen citados por Montesinos, pero no se han podido localizar en ninguna biblioteca. No nos es posible, pues, señalar los títulos de los cuentos contenidos en estos volúmenes. 6 Caso cuanto menos curioso es el de este y los demás cuentos que siguen traducidos al catalán, en 1935, por la editorial Mentora, que atribuye la autoría de los cuentos no a Mme d’Aulnoy sino a Hans Christian Andersen. Manuel Llanas y Ramón Pinyol (2001: 32) corrigen el error sin mencionar esta atribución errónea por parte de la editorial. Sí que lo señala, aunque no en todos los casos, Mónica Baró Llambías (2005: 440-442), quien apunta que también la versión española de Leandro, el príncipe gnomo (1932) se atribuye a H.C. Andersen. Por su parte Riquete el copetudo (o Riquet el del plomall en la adaptación catalana) no es ni de Andersen ni de Mme d’Aulnoy, sino de Charles Perrault. http://cedille.webs.ull.es/10/22vicens.pdf 370 Çédille, revista de estudios franceses, 10 (2014), 367-383 – – – – – – – – – – – – – – – – – – – 7 Carlota Vicens Pujol Bella-Bella o el cavaller sortós (1935), Mentora, Contes d’ahir i d’avui, serie 1, 24, Barcelona. (Adaptación: Lluís Capdevila). Cuentos (1942), Maucci, Barcelona (Traducción: H.C Granch; Ilustraciones: Margenot). Este volumen incluye: Graciosa y Percinete, El Príncipe gnomo, La buena ratita, Finette la cenicienta, Afortunada, El Enano amarillo, La Princesa de los cabellos de oro y La Princesa Cierva. El príncipe invisible; La buena ratita (194?), Maucci, col. Reyes Magos, Barcelona. Afortunada y La princesa de los cabellos de oro (ca. 1950), Maucci, Barcelona. La cierva en el bosque: cuento para la juventud (1951), Molino, Barcelona. La gatita blanca (1956), Molino, col. Alfombra Mágica, Barcelona. El pájaro azul (1956?), Paulinas, Madrid. La rama de oro (1957), Molino, Col Alfombra Mágica, Barcelona. [Viaje por España en 1679 y 1680] y Cuentos feéricos (1962), Iberia, Barcelona (2 vol. Traducción: Marta Corominas y Mercedes Villalta; Prólogo: Emiliano M. Aguilera). Los cuentos son los siguientes: La Bella de los cabellos de oro, El Ramo de oro, El Buen ratoncito, El Carnero, Fineta, la Cenicienta, La princesa Rosita, El Pájaro azul y La Gata blanca. El pájaro azul (1963), Bruguera, col. Para la Infancia, Barcelona. El pájaro azul (1967), Bruguera, col. Para la Infancia, Barcelona (Adaptación: J.A. Vidal Sales). El pájaro azul (1969), Bruguera, col. Para la Infancia, Barcelona (Adaptación: Alicia Romero) El gato blanco (1975), Stock, Barcelona. (Adaptación: Carmichael; Ilustraciones: Shirley y Gwen Tourret). Cuentos de hadas (1979), Bruguera, Barcelona (Prólogo de Carmen BravoVillasante). Los cuentos aquí incluidos son: El pájaro azul, La bella de los cabellos dorados, El enano amarillo, El príncipe jabalí, El serpentón verde, La ratita bondadosa, Gatablanca, La corza del bosque, Finita Favillón y La rana bienhechora. Los cuentos de Perrault [seguidos de los cuentos de Madame d’Aulnoye y de Madame Leprince de Beaumont] (1980, 1987), Crítica, Barcelona (Traducción: Carmen Martín Gaite; Introducción Bruno Bettelheim, traducida al francés por Théo Carlier7). Son dos los cuentos de Mme d’Aulnoy: La Cierva del bosque y La Gata blanca. La corza del bosque (1982), Bruguera, Barcelona. (Ilustraciones: Raúl Capitani). El pájaro azul: cuentos de hadas (1982), Bruguera, col. Todolibro, Barcelona (Traducción: José Benito Alique; Ilustraciones: Raúl Capitani). El serpentón verde y La ratita bondadosa (1982), Tocay, D.L. col. Azucena, Barcelona (Adaptación: Mª Luisa Vela, Ilustraciones: María Pascual). El serpentón verde y otros cuentos de hadas (1984), José J. de Olañeta, col Érase una vez. Biblioteca de cuentos maravillosos, Palma de Mallorca (Traducción: José Benito Traducción del volumen Les contes de Perrault, a cargo de Théo Carlier. París, Seghers, 1978. http://cedille.webs.ull.es/10/22vicens.pdf 371 Çédille, revista de estudios franceses, 10 (2014), 367-383 – – – – – – – Carlota Vicens Pujol Alique; Prólogo: Carmen Bravo-Villasante). Los cuentos aquí incluidos son: La bella de los cabellos dorados y La rana bienhechora. La princesa Rosina (1984), Miñón, col. Molinillo de Papel, Valladolid (Traducción: José Benito Alique; Ilustraciones: Jesús Gabán). El pájaro azul y otros cuentos de hadas (1985), José J. de Olañeta, col. Érase una vez. Biblioteca de cuentos maravillosos, Palma de Mallorca, (Prólogo y traducción: José Benito Alique). Los «otros cuentos» son El enano amarillo y Finita Fabillón. El príncipe duende y otros cuentos de hadas (1987). José J. de Olañeta, Palma de Mallorca (Prólogo y traducción: José Benito Alique). Los otros cuentos incluidos en el volumen son: Fortunata y La Ratita bondadosa. El cuarto de las hadas (1991, 1992, 1999, 2005), Siruela, Serie: Las tres edades. Biblioteca de cuentos populares, Madrid. (Introducción: Luis Alberto de Cuenca; traducción: Emma Calatayud). Este volumen incluye: Gata blanca, Finita cenicienta, El pájaro azul, La princesa bella estrella y el príncipe Querido, La cierva del bosque, Serpentino verde, Enano amarillo, Menudencia, La bella de los cabellos de oro y Graciosa y Percinete. La Bella y la Bestia y otros cuentos, Mme Leprince de Beaumont – Mme d’Aulnoy (1992, 2005), Gaviota, col. Trébol, León. (Traducción: Elena del Amo). Los cuentos de Mme d’Aulnoy aquí incluidos son: La Princesa Rosette, La Bella de los cabellos de oro y El pájaro azul. Finette Cendron (1999), Planeta Agostini, Barcelona. (Versión bilingüe abreviada y simplificada. –Curso de lengua Planeta Agostini-). Tres rosas – Título original: Le Prince Lutin- (2009), Novelúa, A Coruña (Version al gallego: Xavier Senín). 1.2. Otras obras: – – – – – – Historia de Hipólito, conde de Douglas (1838), imprenta de Boix, Madrid. (Sin nombre de traductor). Relación que hizo de su viaje por España la señora condesa d’Aulnoy en 1679 (1891), Juan Jiménez, Madrid, (Traducción: Manuel G. Hernández). Relación que hizo de su viaje por España la señora condesa d’Aulnoy en 1679, (1892), Tipografía Franco-Española, Madrid. (Sin nombre de traductor). Un viaje por España en 1679 (1891), Revista contemporánea de Madrid. Traducción y palabras introductorias de Luis Ruiz Contreras8. Memorias de la corte de España (19 ?), Sociedad de ediciones, París. (Versión española: Francisca A. de la Barella, «Notas biográficas» firmadas C. de B.). Un viaje por España en 1679 (1942), La Nave, Madrid. (Traducción y notas: Luis Ruiz Contreras). 8 Este nombre lo propone Palop (apud Romero, 1996: 455): «[…] en la edición de La Nave de 1942 se presenta como una nueva edición revisada de la de 1891: “El autor de esta traducción la publicó por vez primera en la Revista Contemporánea de Madrid en 1891”. […] Pero aunque Palop propone un nombre para el autor de la traducción, este no aparece en el texto». http://cedille.webs.ull.es/10/22vicens.pdf 372 Çédille, revista de estudios franceses, 10 (2014), 367-383 – – – – – Carlota Vicens Pujol Relación del viaje de España in Viajes de extranjeros por España y Portugal, (1952), Aguilar, Madrid. (Traducción y nota previa: José García Mercadal). Nueva edición: Junta de Castilla y León, Valladolid, IV, 1999, 10-281. •Viaje por España en 1679 y 1680 (1962), Barcelona, Gráficas Diamante. Traducción y notas a cargo de Marta Corominas y Mercedes M. Villalta. Prólogo de Emiliano M. Aguilera. Relación del viaje a España (1986), Madrid, Ediciones Akal. (Traducción: García Mercadal, prólogo: Lorenzo Díaz). Relación que hizo de su viaje por España la señora condesa d’Aulnoy en 1679 (1996), Valencia, Librerías París-Valencia. (Sin nombre de traductor) 9. Relación del viaje de España (2000), Madrid, Cátedra. (Traducción: Pilar Blanco y Miguel Ángel Vega; Prólogo: Miguel Ángel Vega). 1.3. Síntesis y comentarios En lo que se refiere a nuestro primer punto, los Cuentos de hadas, destacamos de entrada que veintinueve ediciones pertenecen a colecciones infantiles (el término adaptación sustituye entonces al de traducción) y tres de ellas están catalogadas como «cuentos maravillosos o populares». Un cuarto título, el elegido por la editorial Planeta Agostini, aparece en versión bilingüe francés-español y abreviada para estudiantes. Finalmente debemos señalar que Barcelona es la ciudad que contribuyó en mayor medida a la divulgación de los cuentos de Mme d’Aulnoy y que, aparte de las seis traducciones al catalán, encontramos únicamente una traducción a otra de las lenguas cooficiales del territorio español, el gallego, cuyo título se tradujo de forma bastante libre: Tres rosas por, como vimos anteriormente, Le prince Lutin. No debemos olvidar, sin embargo, que las tres rosas regaladas por el hada juegan un papel esencial en el cuento... El número de títulos traducidos se reduce a dieciocho de un total de veinticinco, doce de los cuales pertenecen a los Contes des fées, y concretamente (nueve de ellos) a los dos primeros volúmenes publicados; cuatro están incluidos en los Contes nouveaux ou les fées à la mode y dos en Suite de Contes nouveaux ou les fées à la mode. L’oiseau bleu es el relato que cuenta con más ediciones (quince) seguido de La Belle aux cheveux d’or, mientras que otros títulos, así como los relatos-marco10 fueron totalmente ignorados por las editoriales. Entre esos títulos, cabe destacar el primer cuento de hadas literario en Francia, L’île de la félicité. Sobre ellos diremos únicamente que a Mme d’Aulnoy la crítica le atribuye el mérito de haber sabido adaptar a su época y a su público, o sea, de haber hecho suyos aquellos cuentos que había recuperado previamente de la tradición oral y folklórica: 9 Reimpresión facsimilar de la edición arriba mencionada de 1892, Tipografía franco-española, Madrid. 10 La edición de Elisabeth Lemirre retoma igualmente los cuentos extrayéndolos de los récits-cadres. http://cedille.webs.ull.es/10/22vicens.pdf 373 Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol L’acclimatation du genre à son public mondain induit en effet la métamorphose d’un conte oral, anonyme, collectif, en conte d’auteur, au terme d’un phénomène d’appropriation et de personnalisation de ce «lieu commun» que constitue le conte merveilleux traditionnel. Par ce travail créateur, Mme d’Aulnoy participe à la naissance du conte de fées féminin français […] (Jasmin, 2008: 20)11. Sobre este mismo aspecto se había pronunciado algunos años antes Mª del Carmen Ramón al afirmar que: Sus cuentos son una reinterpretación aristocrática de la tradición oral aderezada de otras múltiples referencias literarias. Son fundamentalmente historias de amor, casi novelas breves, influidas también por la novela de caballerías, la novela pastoril, el preciosismo, los cuentos de Perrault y otras fuentes. La exuberancia verbal y el delirio imaginativo de esta autora les confieren, no obstante, una especial originalidad (Ramón, 2001: 99). En lo que concierne al resto de obras, la primera traducción aparecida en España es la de Histoire d’Hippolyte, comte de Douglas, en 1838, cuya publicación fue anunciada en la prensa de la época. Así el Diario de Madrid de 27 de junio de 1838 se hace eco de una noticia aparecida en el número 87 de El Progreso [Periódico de la tarde que se publica en esta Corte] que dice: […] la imprenta El Progreso publicará sin interrupción a partir del mes próximo, novelas y otras obras literarias tanto originales como traducidas […]. Los libros que formarán parte de la primera serie serán Hipólito, conde de Douglas por Mme d’Aulnoy; El padre Goriot por Balzac, Los prisioneros de Ab.del.kader por Mr. [A. de] Francia y Mi residencia en Francia, del célebre Fenimore Cooper. Habiéndose ya publicado el Hipólito, los señores suscriptores desde el uno de julio recibirán esta obra, que consta de dos tomos en un solo volumen […]. Todos los libros mencionados se publicaron efectivamente en Madrid, en 1838, en la imprenta de Ignacio Boix. En lo que se refiere a las otras nueve ediciones, la traducción de la Relation du voyage d’Espagne fait en 1679, aparece de forma escalonada hasta en nueve entregas entre 1891 y 2000, lo que no deja de sorprender puesto que el viaje ya había sido traducido al inglés en 1691, al alemán en 1696 y al holandés en 1705. Además, debido a que el autor de la primera traducción se había tomado algunas licencias, resulta que la primera traducción íntegra al español es la de 11 Jasmin señala asimismo que algunos de los cuentos son muy difíciles de encasillar y no están inspirados en ningún cuento tradicional http://cedille.webs.ull.es/10/22vicens.pdf 374 Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol 1962, como destaca en su prólogo Emiliano Aguilera (1962: X-XII). Si nos detenemos en este último título, nos damos cuenta de que dos siglos separan la primera publicación en francés de la primera traducción española. La hipótesis avanzada por María Elena Romero es que esta traducción tiene como punto de partida la edición francesa de Mme B. Carey (1874) en una época en la que el relato de viajes se había convertido en el género por excelencia del siglo XIX y en la que España estaba de moda en Francia, un hecho que queda reflejado no solo en la literatura12 sino también en la pintura, ya que no debemos olvidar que el descubrimiento al otro lado de los Pirineos de los grandes pintores españoles como Velázquez, Murillo, El Greco o Goya, data de la primera mitad del siglo. Además, el relato de Mme d’Aulnoy contiene todos los estereotipos que servirán más adelante para ilustrar «l’Espagne romantique» (Romero, 1999: 453). Finalmente, existe también una traducción de Mémoires de la cour d’Espagne, sin fecha exacta de publicación. Cabe recalcar asimismo que, salvo Histoire d’Hippolyte, comte de Douglas, ninguna otra novela de la escritora, donde se encuentran insertados bastantes cuentos de los arriba citados, se tradujo: ni sus dos novelas españolas, Don Gabriel Ponce de León y Don Fernando de Toledo, ni una tercera novela titulada Le gentilhomme bourgeois13. No se puede descartar, sin embargo, que la sociedad española más cultivada conociese las distintas obras de Mme d’Aulnoy ya que su nombre aparece en varias ocasiones (aunque sin profundizar en ella) en la prensa del siglo XIX y albores del XX. Por poner un segundo ejemplo, El Heraldo de Madrid (24 de marzo de 1928) da cuenta de un artículo publicado por M. Ezio Levi en el semanario italiano Il Marzocco donde se puede leer lo siguiente: Los tres libros de Madame d’Aulnoy, Mémoires de la cour d’Espagne, Relation du voyage d’Espagne et Nouvelles espagnoles, escritos entre 1690 et 1692 marcan el inicio del interés de los franceses sobre las cosas de España. Estas tres obras son el punto de partida de la literatura españolizante que se desarrollará sin interrupción a lo largo de los siglos XVIII y XIX, cuyas obras más célebres son Gil Blas, Le mariage de Figaro y Ruy Blas 14. 12 Nos referimos por ejemplo a títulos como Les aventures du dernier abencérage (Chateaubriand, 1826), Hernani (Victor Hugo, 1830), Contes d’Espagne et d’Italie (Musset, 1830), Mosaïque y Carmen (Mérimée, 1833, 1845) o Le voyage en Espagne (Théophile Gautier, 1843) entre muchos otros. 13 Recordemos aquí las palabras de Alicia Yllera cuando señala que, dejando de lado las traducciones, «durante el período 1700 a 1750, treinta y dos obras de ficción se llaman nouvelle ou histoire espagnole; otras presentan situaciones o personajes españoles o bien son adaptaciones o imitaciones o se dicen traducidas del español (sea cierto o no), mientras que italiana figura en el subtítulo de diez obras e inglesa en el de trece…» (Yllera, 1991: 640). 14 «Como se vivía en Madrid a fines del siglo XVII», El Heraldo de Madrid, 24 /03 /1928, p. 12. http://cedille.webs.ull.es/10/22vicens.pdf 375 Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol 2. Prólogos y prologuistas: ¿contra Mme d’Aulnoy? El análisis de los prólogos permite una aproximación crítica a una obra ausente, con muy pocas excepciones, de las revistas literarias o culturales españolas tanto del siglo XIX como del XX. Utilizamos aquí la palabra prólogo de manera genérica: su uso corriente en español para designar aquel texto que precede el cuerpo de una obra, literaria o no, con el fin último de presentarla a los lectores, junto con la posibilidad de formar derivados (prologuista, prologar, etc.), ha determinado la elección de este vocablo por encima de otros como introducción, prefacio o proemio. De hecho los diferentes textos liminares estudiados se presentan, quizás por un exceso de modestia, como: «A los lectores» (B. Alique), «Notas biográficas» (C. de B.), «Notas prologales» (E. Aguilera), «Introducción» (B. Bettelheim, L.A. de Cuenca y C. Bravo Villasante) o «Prólogo» (Pedro Umbert); la «Advertencia preliminar» que encabeza el texto de Ruiz Contreras se convierte en «Nota preliminar» en el membrete situado en la parte superior de las páginas impares del libro, mientras que el texto de Bravo-Villasante a la edición de Olañeta se anuncia como «Presentación» en el índice pero aparece con un título temático, «Madame de Aulnoy y las hadas a la moda» en el interior del volumen. De lo dicho hasta ahora no es difícil inferir que de las nueve categorías de prólogos establecidas por Gérard Genette (1987: 182-186) todas las mencionadas pertenecen al grupo que este denomina préfaces allographes authentiques, en la que un escritor prologa la obra de otro escritor. Siguiendo siempre el estudio de Genette la función de este tipo de prefacios es doble. Por un lado informan, es decir, «retracent les étapes de la conception, de la rédaction et de la publication et enchaînent logiquement sur une histoire du texte» (Genette, 1987: 268), ofrecen datos biográficos del autor y/o sitúan el texto prologado en el conjunto de la obra del autor. La segunda función, la más importante según Genette, es la de recomendación: Heureusement la fonction de recommandation reste le plus souvent implicite, parce que la seule présence de ce genre de préface est déjà une recommandation. Cette caution est généralement, pour une préface originale, apporté par un écrivain plus consacré (…). Ou, lors d’une traduction, plus connu dans le pays d’importation (…). Ou plus actuel (…) lors de la réédition largement posthume d’un texte classique (…). Ou encore, toute question de notoriété relative mise à part, capable d’ajouter à une œuvre la valeur d’une interprétation et par conséquent d’un statut théorique exemplaire (Genette, 1987: 271). Esta función de recomendación comprende también una dimensión crítica o interpretativa presente, junto con los datos biográficos de la autora y los acontecimientos históricos que marcaron la época, en los prólogos sobre los que a continuación nos centraremos, en un recorrido que nos llevará del autor al texto. http://cedille.webs.ull.es/10/22vicens.pdf 376 Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol Por lo general los datos biográficos se detienen en los aspectos más escabrosos de la vida de la autora. Además de las fechas de nacimiento y muerte de MarieCatherine Jumel-Barneville y de su matrimonio a los 16 años con el barón d’Aulnoy, treinta años mayor que ella, estas reseñas ponen el acento en el complot supuestamente urdido por Mme d’Aulnoy con la ayuda de su madre y del amante de esta para asesinar a su esposo15. A eso se añade un segundo caso de asesinato bastante confuso ocurrido años después relacionado igualmente con el nombre de Mme d’Aulnoy... He aquí material suficiente para adornar la vida de la baronesa de un halo de misterio que lleva a uno de los autores de prólogos a escribir: «Su imagen se mueve entre la luz y la sombra de la legítima ambición y del deseo de prosperar sin ningún escrúpulo» (Vega, 2000: 16), el mismo que la acusa de ser capaz de hacer que cuentos de hadas de terceras personas pasasen a la posteridad como si fueran suyos (Vega, 2000: 16). Así, de forma más o menos implícita, estos autores parecen identificar la obra de la escritora con su vida privada. Luis Alberto de Cuenca, por su parte, hace hincapié en el período histórico del reinado de Luis XIV y en la irrupción, algo paradójica, de los cuentos de hadas en el ambiente absolutista. Se apoya para ello en las palabras de una destacada estudiosa de esos cuentos: Elisabeth Lemirre. También presta especial atención a la época el prólogo de B. de C. a las Memorias de la Corte de España, quien insiste en que Mme d’Aulnoy «traza un cuadro fidelísimo del estado de la corte de España en el momento de la llegada de la princesa de Orleans al lado de Carlos II» (B. de C., sf: VII). Carmen Bravo-Villasante es la autora de dos de estos prólogos, sin lugar a duda los más eruditos16. Establece una clasificación de los distintos cuentos especificando si fueron publicados de manera independiente o si forman parte de una novela marco y, en este último caso, si se encuentran insertados en la novela o si, al contrario, son independientes. La autora analiza la estructura de los cuentos de hadas apoyándose en los estudios de Stith Thomson y de Vladimir Propp; recurre también a Bruno Bettelheim, a Jacques Barchilieu o al Abbé de Villiers quien en 1699 levantó la voz en contra de ese «débordement de la féérie» que, a finales de siglo, parecía imponerse sobre la literatura seria. Conviene recalcar que, en los albores del siglo XVIII, el cuento se conjugaba en femenino, como bien demuestra el hecho de que al lado de Mme d’Aulnoy, autoras como Mlle Lhéritier, Mlle de la Force, Mme de Murat, Mme Deshoulières o Mme de Villedieu contribuyeron a la rápida difusión del género: entre 15 El único de nuestros prologuistas en tomar partido por Mme d’Aulnoy fue E. Aguilera. Según su versión, ella misma hizo que la acusaran para salvar la vida y el honor de su marido (Aguilera p. VIII) 16 Los de 1979, Cuentos de hadas, y de 1984, El serpentón verde y otros cuentos de hadas. Se trata en un 90% del mismo texto. Reconocida especialista en literatura infantil, Carmen Bravo-Villasante nos legó una quincena de títulos, entre los cuales: Antología de la literatura infantil universal (4 vol.), Valladolid, Miñón, 1971; Literatura infantil universal (2 vol.), Madrid, Alameda, 1978 o también Diccionario de autores de la literatura infantil mundial, Madrid, Escuela Española, 1985. http://cedille.webs.ull.es/10/22vicens.pdf 377 Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol 1670 y 1700 se publicaron más de sesenta cuentos, tres cuartas partes de los cuales firmados por mujeres. Por otro lado, la introducción de Bruno Bettelheim se refiere de manera genérica al papel que los cuentos infantiles juegan en la formación del niño, de modo muy similar a cómo lo hiciera en The uses of enchantement, sin entrar en el análisis directo de los cuentos de Mme d’Aulnoy. Sin embargo, la erudición pronto deja paso a los comentarios críticos. BravoVillasante (1984: XXIV) por ejemplo señala que los textos de Mme d’Aulnoy pecan de un tono excesivamente irónico que si bien conviene a «una feminidad consciente de sus locuras», daña la esencia del cuento. José-Benito Alique (1985: XI-XII), en un estilo quizás excesivamente protocolario, califica de ingenuo el hecho de elegir siempre como título de los cuentos el nombre de pila del protagonista y de «delicioso pasatiempo cortesano» esa moda de los cuentos de hadas, moda fadaise, dirá por su parte Vega (2000: 17) escondiéndose detrás de la palabras de Mme Carey17. Como puede observarse, el hecho de recurrir a las palabras de una tercera persona que confiere autoridad a los prólogos (la repetición o redite en palabras de A. Compagnon) es un hecho generalizado entre los prologuistas al igual que la comparación de la cuentista con Charles Perrault. Este último narra de manera simple mientras aquélla añade detalles fruto de una «imaginación desbordante» según afirma Bravo-Villasante (1979: 11), quien volverá sobre la comparación entre los cuentos de Perrault y los de Mme d’Aulnoy en estos términos: A todos estos fondos y temas tradicionales, relatados con cierta sencillez por Perrault, suceden inmediatamente los cuentos de Mme d’Aulnoy, con detalles ingeniosos y picantes, volutas decorativas de su imaginación e inventiva desbordadas, con esta sutileza de medios para sorprender y admirar […]. Mas, aunque parezca sucesora de Perrault, algún crítico agudo, como Marc Soriano, aventura en sus investigaciones que como los Cuentos de la madre Oca son la colección más difundida, la más célebre, Perrault se convierte en el inventor de un género que Madame d’Aulnoy y Mademoiselle l’Héritier ya habían practicado antes que él (Bravo-Villasante 1984: XVII). De hecho, si se la sitúa siempre por detrás de Perrault es: Porque la condesa desvirtúa frecuentemente la ilusión producida por su relato burlándose a lo mejor de sus ficciones, al paso que Perrault refiere con la formalidad del que ha visto lo que cuenta, comunicando tal carácter de ingenuidad a sus narraciones que sin hipérbole puede decirse que se coloca al nivel del niño que palpitante de interés se entera de las extraordinarias 17 Mme Carey se encargó de la edición de 1874 titulada La cour et la ville de Madrid vers la fin du XVIIème siècle. Relation du voyage d’Espagne par la comtesse d’Aulnoy. Édition nouvelle, revue et annotée par Mme B. Carey. París, Ed. Plon et Cie. http://cedille.webs.ull.es/10/22vicens.pdf 378 Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol aventuras y fantásticos sucesos que ocurren en los países de ensueño que gobiernan las hadas (Umbert 1909: 10-11)18. Emiliano Aguilera (1962: XI) se presenta como el más firme defensor de la baronesa, asegurando en varias ocasiones que sus relatos fantásticos son más dignos de admiración que los de Perrault. Por último, los mayores elogios hacia su producción cuentística provienen de los prólogos cuyos autores se muestran bastante críticos con... las otras obras de la escritora prologadas por ellos mismos, en particular la Relation du voyage d’Espagne fait en 1679. En comparación con la Relation, de la que subrayan la antipatía de la escritora hacia el pueblo español, las fábulas parecen muy agradables de leer y dignas de admiración; son páginas hermosas, dicen, que la escritora tuvo el mérito de añadir a la literatura francesa. El resto de la obra de la escritora también generó bastante polémica. La novelística, en primer lugar, calificada de «amalgama indigesta de hechos históricos y definiciones novelescas» (Umbert, 1909: 10), y más concretamente Histoire d’Hyppolyte, comte de Douglas calificada de «execrable novela sentimental» (De Cuenca, 1991: 11); en segundo lugar, la obra pseudo-histórica: Miguel Ángel Vega se refiere así a la producción antiespañola de la escritora, a saber: Relation du voyage en Espagne, Mémoires de la cour d’Espagne y las Nouvelles espagnoles (Vega, 2000: 24). Y como, según se ha visto, parece obligado recurrir a las palabras de una tercera persona, será Taine el que se encargará de cantar las excelencias del viaje por España. Mme d’Aulnoy, asegura Ruiz Contreras19 siguiendo a Taine, escribe con precisión, pero sin exagerar y, sin llegar a ofrecer una obra de arte (ni ambicionarlo), reúne todas la cualidades de un buen escritor (Ruiz, 1891: 8). Si se le replica que sí, que la autora / narradora exagera, el prologuista responde que, precisamente por el hecho de exagerar, el carácter español se perfila con más nitidez a lo largo de esas páginas (Ruiz, 1891: 9). La intertextualidad es doble si tenemos en cuenta el prólogo de Aguilera, quien cita a Ruiz citando a Taine… Una vez más se apunta al carácter precursor del viaje de Mme d’Aulnoy en cuanto a la imagen de la España romántica, ya que su viaje incluye todos los tópicos y lugares comunes de los que más adelante se harán eco otros escritores. En ello insiste B. de C. al afirmar que Víctor Hugo se inspiró para 18 Del mismo parecer es Auguste Jal quien en su Dictionnaire critique de biographie et d’histoire afirma: «ses livres obtinrent un assez grand succès; la plupart sont oubliés; le seul qui ait survécu à leur auteur est le petit recueil de ses Contes de fées, ingénieux et spirituels récits auxquels manque la naïveté qui, surtout, recommande ceux de Perrault » (p.1306). Años después el Grand Dictionnaire Larousse du XIXème siècle subraya que si los cuentos se parecen demasiado a sus novelas es porque «elle a mieux réussi dans les ouvrages de pure imagination. Son roman Histoire d’Hippolyte, comte de Douglas, où se sent l’imitation de Mme de Lafayette, est néanmoins d’une lecture assez agréable. Ses Contes de fées offrent un mélange de finesse et de naïveté qui les rend fort attrayants» (p. 943). 19 Ruiz Contreras realizó una encomiable labor como traductor del francés al español (cf. Pilar Estelrich, in F. Lafarga y L. Pegenaute, eds., Diccionario histórico de la traducción en España, Madrid, Gredos, 2009, 987-988). http://cedille.webs.ull.es/10/22vicens.pdf 379 Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol Ruy Blas «de los relatos de la condesa d’Aulnoy, que tan bien describe y con tanto acierto el carácter y las intrigas de la corte de España» (B. de C., sf: VIII). Por otra parte este aspecto del Viaje parece aceptado por estudios más eruditos como el ya citado de Mª Elena Romero (1999: 451-452): Testigo inteligente de una España en decadencia y que se estaba jugando sus destinos, intérprete exagerada pero cierta, contribuyó a crear una imagen española que iba a ser casi definitiva en Europa. Es evidente que sus exageraciones, su falta de fidelidad histórica y el acarreo literario con que sobrecargó sus memorias, desfiguran en buena medida esta imagen. Pero son precisamente estas limitaciones, a las que tan ágil y divertidamente servía su pluma, las que dibujaron un perfil de España que los visitantes posteriores completarán con las inevitables variables del paso del tiempo. Muchos de los tópicos, que constituyeron la masa de la información sobre España y a los que la fantasía de Madame d’Aulnoy hizo resaltar con los matices precisos para popularizarlos están aquí. La ironía ácida que envuelve el prólogo de Miguel Ángel Vega (el más crítico), el tono de burla que se trasluce en cada una de sus palabras, –compara, por ejemplo, los chismorreos de Mme d’Aulnoy con los cotilleos de la revista Oh La!– o las intrusiones demasiado frecuentes del autor20, alejan este prólogo del tono científico que debería tener. Palabras como «habladurías» o «estereotipos» se repiten una y otra vez y se vuelve a insistir sobre el supuesto desprecio de Mme d’Aulnoy hacia el arte o la literatura española y su preferencia por el arte de la galantería, los juegos amorosos y las desavenencias familiares... Si hemos visto que algunos prologuistas no parecen valorizar demasiado el texto que prologan, en el caso de M. A. Vega se subvierten claramente las convenciones del género y sin ningún apuro se pasa de la recomendación a la no-recomendación… Queda una última cuestión, ya clásica: ¿traduttore, traditore? Nuestra intención no es hacer aquí una comparación en profundidad de las distintas traducciones con el texto original, pero en lo que se refiere a Contes de fées, es fácil ver cómo, cuanto más se dirigen a un público infantil, más se simplifican los textos y más se alejan del cuento original21. 20 M.A. Vega abre un paréntesis, por ejemplo, para saludar la llegada a los quioscos de la revista Oh La!, versión francesa de la revista española Hola, alegrándose de la gran influencia de la cultura española en la francesa. 21 Este mismo fenómeno se observa en Inglaterra, como bien señala Glen Regard : «Comme tous les contes de Madame d’Aulnoy ils ont connu en traduction de nombreuses réécritures textuelles et péritextuelles ; détachés de leur cadre d’origine, ils ont été manipulés à souhait. […] Ces réécritures successives s’accompagnent d’une réorientation générique : à l’origine destinés aux adultes, les contes de Mme d’Aulnoy sont progressivement adaptés au marché naissant de la littérature enfantine en Angle- http://cedille.webs.ull.es/10/22vicens.pdf 380 Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol Solo dos de los prologuistas son a la vez traductores del libro y autores del prólogo: Luis Ruiz Contreras y José Benito Alique. Ninguno de ellos hace sin embargo referencia a los problemas lingüísticos o socio-culturales derivados del proceso de traducción. El primero confiesa, no obstante, haberse tomado la libertad de presentar las cartas del viaje a España sin discontinuidad; así es como, suprimiendo los encabezados, los saludos y las alusiones «desprovistas de interés», saca el viaje de la baronesa del género epistolar. A esto debemos añadir la supresión de las historias insertadas y de «algunos cuentos productos de la fantasía, bastante largos y pesados» (Ruiz, 1891: 10)… José-Benito Alique olvida también en el texto liminar su trabajo de traductor limitándose a hacer unas observaciones sobre la autora y la temática de cada uno de los cuentos traducidos, comentarios que no aportan nada nuevo ni a una mejor comprensión de la autora ni de su obra, como tampoco al proceso de traducción. Mención aparte merece quizás Carmen Martín Gaite, quien incluye una simple «Nota de la traductora», para aclarar tan solo que ha dejado las moralejas en su idioma original «ya que sería difícil conservar el extraño humor y tono un poco arcaico de estos poemitas» refiriéndose, sin embargo, a los cuentos de Perrault incluidos en el volumen. 3. Algunas conclusiones Las conclusiones que se desprenden de este estudio parecen a priori algo contradictorias. Al leer los prólogos nadie duda de que en España la obra de Mme d’Aulnoy no tuvo gran éxito. Frente a los datos biográficos de la escritora, el prologuista se convierte en juez y moralista y, generalmente, condena. Su exceso de ambición, su feminismo avant-la-lettre, su rechazo de las convenciones sociales, su espíritu ácido... son muchos los reproches que se le hacen, sobre todo desde el momento en estos pretendidos defectos dejan su huella en la obra literaria22. Esto se debe a dos razones: en lo que se refiere a los cuentos de hadas, la razón es que esta falta de ingenuidad (la palabra se repite), esta facundia excesiva, desvirtúan la esencia misma de la fantasía y, por ende, del cuento; en lo referido a Relation du voyage d’Espagne es debido a que la autora cae en la exageración y falsea tanto los datos históricos como la imagen del país... que se transmitirá no obstante a los viajeros del siglo XIX. En ambos casos los juicios críticos parecen bastante reductores. Ninguno de los prologuistas llega a detenerse en los principios estéticos que rigen la obra o en el funcionamiento narrativo de los cuentos, a excepción de nuestra especialista en literatura infantil, Carmen Bravo-Villasante. En cuanto a las novelas, son descalificadas sin terre». «The Fairy Tales of Madame d’Aulnoy (1892) : la traduction des Contes de fées et de Contes Nouveaux de Marie Catherine d’Aulnoy par Annie Macdonell et Elizabeth Lee » in Fééries. URL: feeries.revues.org. Consultado el 28 de diciembre de 2013. 22 Según Nadine Jasmin esta huella es precisamente la que distingue y caracteriza los cuentos de hadas de Mme d’Aulnoy. http://cedille.webs.ull.es/10/22vicens.pdf 381 Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol concesión alguna. Muchas de las observaciones hechas por Alicia Yllera (2007) en su estudio de las traducciones al francés de novelas españolas de los siglos XVII y XVIII podrían servir para nuestras conclusiones. En lo que se refiere a las traducciones (ediciones y reediciones), observamos que las de los cuentos de hadas superan con creces a las del resto de la obra de Mme d’Aulnoy, y sorprende ver que de un libro que podría ser de interés para España, como es el caso de Mémoires de la cour d’Espagne23, no se haya encontrado más que una traducción, por otro lado bastante reciente. En cambio, la edición de los cuentos de hadas se multiplicó a lo largo del siglo XX, pudiendo constatar una manipulación de los textos indisociable de esa tendencia hacia una infantilización de los cuentos, por lo demás generalizada en Europa. Advertía ya Bettelheim contra este hecho en su Introducción de 1978: Pero en último término un nuevo peligro (…) ha venido a amenazar estos cuentos (…). Este peligro que puede desposeer a los cuentos de hadas de su más profundo y esencial significado, no es otro que la degradación y banalización a la que se los viene sometiendo para ponerlos al alcance del gran público (Bettelheim, 1980: 22). A ello se une que el traductor se adjudica el deber de mejorar el original, haciéndolo menos pesado y más acorde con la sensibilidad ya sea de la época ya del nuevo público infantil al que se destinan. Invento mundano, los cuentos de hadas pasan así a encasillarse, con el correr del tiempo, en el subgénero de la literatura infantil y juvenil. Este fenómeno no es, sin embargo, nuevo y en este sentido parece interesante señalar que ya los hermanos Grimm van a infantilizar para su edición de 1819 los cuentos recopilados por ellos mismos (con fines más filológicos) en 1812, y que la primera traducción de los cuentos de Grimm al francés saldrá con el título de Contes choisis des Grimm à l’usage des enfants, en 183624. Ahora bien, la publicación muy cuidada de la editorial Siruela en su colección de Cuentos Populares y sus cuatro reediciones en un período de tiempo relativamente corto parecen indicar una revalorización de la literatura popular e infantil. Pensamos sin embargo que cualquier conclusión en este sentido sería a día de hoy arriesgada. REFERENCIAS BIBLIOGRÁFICAS AULNOY, Mme d’ (2008): Contes des fées. París, Champion (2 vols). 23 Por contraste indicamos que de Mémoires de la cour d’Angleterre (1695) se encuentran ya traducciones al inglés a finales de este mismo siglo XVII. Vid. nota 21. 24 C. François: «La première traduction critique des Grimm en français », Acta Fábula. Disponible en: fabula.org/revue. Consultado el 28 de diciembre de 2013. http://cedille.webs.ull.es/10/22vicens.pdf 382 Çédille, revista de estudios franceses, 10 (2014), 367-383 Carlota Vicens Pujol ANDRIES, Lise & Geneviève BOLLEME, (2003) : La Bibliothèque bleue. Littérature de colportage. París, Robert Laffont. BARÓ LLAMBÍAS, Mónica (2005): Les edicions infantils i juvenils de l’editorial Joventud. 19231969. Tesis Doctoral, Universidad de Barcelona. BÖHM, Roswitha (2010): «Femme de lettres – femme d’aventures : Marie-Catherine d’Aulnoy et la réception de son œuvre». Œuvres & Critiques XXXV (1), 135-146. DÍAZ BORQUE, José María (1975): La sociedad española y los viajeros del siglo XVII. Madrid, Sociedad General Española de Librerías. GENETTE, Gérard (1987): Seuils. París, Éditions du Seuil. GUENTER, Melissa (2010): «La mirada de una francesa: la Relación del viaje de España (1691), de madame d’Aulnoy». Revista de teoría de la literatura y de literatura comparada 2, 127-136. JASMIN, Nadine (2002): Naissance du conte féminin. Mots et merveilles : les Contes des fées de Madame d’Aulnoy. París, Champion. JASMIN, Nadine (2008) : «Naissance du conte féminin : Mme d’Aulnoy», Introduction à Mme d’Aulnoy, Contes des fées. París, Champion, 9-81. LAFARGA, Francisco (2006): «Sobre actitudes de traductores y editores de relatos franceses. El Viaje a España de Mme d’Aulnoy y de Dumas», in José M. Oliver & al. (eds), Escrituras y reescrituras del viaje, Berna, Peter Lang, 321-331. LLANAS, Manuel & Ramon PINYOL (2001): «Prosistas y poetas franceses (hasta el siglo XVIII) traducidos al catalán en el siglo XX», in Francisco Lafarga y Antonio Domínguez (eds.), Los clásicos franceses en la España del siglo XX. Estudios de traducción y recepción. Barcelona, PPU, 23-37. MONTESINOS, José F. (1980): Introducción a una historia de la novela en España en el siglo XIX. Seguida del esbozo de una bibliografía española de traducciones de novelas, 18001850). Madrid, Castalia. RAMÓN DÍAZ, Mª del Carmen (2001): «Las hadas modernas en el cuento clásico francés escrito por mujeres: ¿personaje o autor?». Thélème. Revista complutense de estudios franceses 16, 95-107. ROMERO ALFARO, Mª Elena (1999): «Reflexiones sobre la traducción de la Relation du voyage d’Espagne de madame d’Aulnoy», in Relaciones culturales entre España, Francia y otros países de lengua francesa. Cádiz, Publicaciones de la Universidad de Cádiz, 447458. VICENS PUJOL, Carlota (2012): «Los Cuentos de hadas de Mme d’Aulnoy y su recepción en España». Nerter 17-18 (Àngels Santa, ed., La literatura popular francesa en España: historia de una recepción), 13-16. YLLERA, Alicia (2007): «Cuando los traductores desean ser traidores», in Mª Luisa Donaire & Francisco Lafarga (ed.), Traducción y adaptación cultural España-Francia. Oviedo, Publicaciones de la Universidad de Oviedo, 639-655. http://cedille.webs.ull.es/10/22vicens.pdf 383
https://openalex.org/W4285269552	https://ieeexplore.ieee.org/ielx7/4563994/9795343/09815511.pdf	English	null	Deep Learning: A Rapid and Efficient Adaptive Design Approach for Phase-Type Optical Needle Modulators	IEEE photonics journal	2,022	cc-by	7,856	Deep Learning: A Rapid and Efﬁcient Adaptive Design Approach for Phase-Type Optical Needle Modulators Simo Wang , Jiangyong Zhang, Fanxing Li , Jupu Yang, Jixiao Liu, and Wei Yan [8], and lithography [9]. The optical needle can be obtained near the focal point by focusing a radially polarized incident beam with a combination of an optical needle modulator and a high numerical aperture (NA) focusing objective. At present, there are three main types of optical needle modulators [10]–[15], and the phase-type optical needle modulators are slightly better due to their high conversion efﬁciency and low processing difﬁculty.Moreover,theﬁve-ringbandphase-typeopticalneedle modulator is a trade-off between the depth of focus, resolution and optical efﬁciency of the optical needle [15]. Therefore, the design of a ﬁve-ring band phase-type optical needle modulator is used as an example in this paper. For the convenience of expression, all the modulators refer to ﬁve-ring band phase-type optical needle modulators in the following. Abstract—The radially polarized beams are modulated by phase-type optical needle modulators can be tightly focused to cre- ate needle-like focused beams, which are called optical needles. The use of optical needles with different resolutions and focal depths as direct writing heads for laser direct lithography enables periodic, cross-scale processing of high aspect ratio micro-nano structures with different line widths. The design of the phase-type optical needle modulators is the key to obtain optical needles with different resolutions and focal depths. However, the existing conventional methods for designing phase-type optical needle modulators rely on the physical model for generating optical needles and the de- ﬁned ﬁtness function, which makes their design time long and not adaptive. Based on the deep learning, a novel phase-type optical needle modulator design (PONMD) approach is proposed in this paper. The results show that the PONMD method takes 0.5526ms to design a phase-type optical needle modulator, and the similarity between the designed and target values is 96.73%. Compared with the conventional methods, the time consumption is reduced by about 8 orders of magnitude, and the similarity is improved by 11.19%. The PONMD approach has the advantages of adaptabil- ity, more efﬁcient, less time-consuming, and less computational resource-consuming. The optical needle can be regarded as the extension of the subwavelength focused spot along the axial direction, which is a symmetric structure. The structure of the optical needle can be constrained by the intensities along the z = 0 and r = 0 directions. 2741409 2741409 IEEE PHOTONICS JOURNAL, VOL. 14, NO. 4, AUGUST 2022 Deep Learning: A Rapid and Efﬁcient Adaptive Design Approach for Phase-Type Optical Needle Modulators Moreover, the depth of focus (DOF) and resolution (FWHM) are used to measure the length of the optical needle along the axial direction and the size along the radial direction, which are respectively deﬁned as the full width at half maximum of the normalized intensity proﬁles along the z = 0 and r = 0 directions. At present, the conventional methods represented by simulated annealing algorithm and genetic algorithm are the mainstream methods to design modulators. Such methods usu- allyincludethedeﬁnitionofﬁtnessfunction,parameterscanning and other processes [13], [15]–[18], which are not only time- consuming and computational resource-consuming, but also not adaptive. Index Terms—Optical needles, phase-type optical needle modulators, deep learning, laser direct lithography, high aspect ratio micro-nano structures. Jiangyong Zhang is with the 54th Research Institute of CETC, Shijiazhuang 050000, China (e-mail: zhangjiangyong18@mails.ucas.ac.cn). er a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ Fanxing Li and Wei Yan are with the Institute of Optics and Electronics, Chinese Academy Sciences, Chengdu, Sichuan 610209, China (e-mail: lifanx- ing15@mails.ucas.ac.cn; yanwei@ioe.ac.cn). This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecomm Simo Wang, Jupu Yang, and Jixiao Liu are with the Institute of Optics and Electronics, Chinese Academy Sciences, Chengdu, Sichuan 610209, China, and also with School of Electronic, Electrical and Communication Engineer- ing, University of Chinese Academy of Sciences, Beijing 100049, China (e-mail: wangsimo18@mails.ucas.ac.cn; yangjupu20@mails.ucas.ac.cn; liujix- iao20@mails.ucas.ac.cn). IEEE PHOTONICS JOURNAL, VOL. 14, NO. 4, AUGUST 2022 2741409 Fig. 1. Schematic diagram of the principle of generating optical needle. The numerical aperture of the focusing lens is NA = 0.95, α = arcsin(NA) is the maximum convergence angle of the focusing objective, n is the refractive index of the immersion liquid (n = 1), A is the coefﬁcient of power (A = 1), k = 2πn/λ is the wavenumber, J0 and J1 denote Bessel functions, and λ is the wavelength of the incident beam. Considering that Er(r,z) and Ez(r,z) are orthogonal, the total electric energy density can be expressed as It = \|Er(r,z)\|2 + \|Ez(r,z)\|2. The electric ﬁeld distribution of the Bessel Gaussian beam at the entrance pupil can be expressed as [15]: Fig. 1. Schematic diagram of the principle of generating optical needle. l(θ) = exp −β2 sin θ sin α 2 J1 2γ sin θ sin α (3) (3) generator to achieve the selection of initial parameters. The method is essentially the same as the existing conventional methods, it only improves the selection of the initial parameters using artiﬁcial neural networks and still relies on the physical model that generates the optical needles. The modulator can be considered as a multi-ring band binary optical element with alternating phases of 0 and π. The trans- mittance function of the ﬁve-ring band phase-type optical needle modulator can be expressed as: The modulator can be considered as a multi-ring band binary optical element with alternating phases of 0 and π. The trans- mittance function of the ﬁve-ring band phase-type optical needle modulator can be expressed as: Inspired by this, we proposed a rapid, efﬁcient and adap- tive phase-type optical needle modulator design method named PONMD method in this paper. The PONMD method built a PONMD model based on deep learning, which established a mapping relationship between the normalized intensities of the optical needle to the structural parameters of the phase-type optical needle modulator. That is, the PONMD method can design the required structural parameters of the modulator based on the normalized intensities of the target optical needle. It is found that the PONMD method takes only 0.5526ms to design a modulator, and the similarity between the design and target values can reach 96.73%. In addition, considering practical applications, the mathematical model of the ideal optical needle is established by ignoring the side lobes of the optical needle. B. Design Methods of Phase-Type Optical Needle Modulators Fig. 2 is a ﬂow chart of the comparison between the PONMD method and the conventional methods to design modulators. After the PONMD model training is completed, the structural parameters of the modulator can be designed only by inputting the normalized intensities of the target optical needle into the model. In contrast, the design process of conventional methods requires steps such as initialization, deﬁnition of ﬁtness func- tions, parameter sweeps, and calculation of physical models for generating optical needles. I. INTRODUCTION O [5] [6 O PTICAL needles play an important role in many applica- tions such as particle acceleration [1]–[4], optical trapping [5], [6], optical data storage [7], super-resolution microscopy Manuscript received 6 April 2022; revised 15 June 2022; accepted 30 June 2022. Date of publication 5 July 2022; date of current version 25 July 2022. This work was supported in part by the Instrument Development of Chinese Academy of Sciences under Grant YJKYYQ20180008, YJKYYQ20200060, YJKYYQ20210041, and YJKYYQ20180006 and in part by the Sichuan Sci- ence and Technology Program under Grants 2020JDJQ0007, 2021JDRC0089, 2021JDRC0084, 2022YFG0249, 2022YFG0223 and 2022102. (Corresponding author: Dr. Wei Yan). Machine learning has been widely used as a powerful tool to study the interaction between complex light ﬁelds and matter, and it has been used to design structures such as nano-photonic devices [19], metasurfaces [20], and chiral metamaterials [21]. Recently, some scholars have also applied deep learning to the optimal design of optical needles. In 2020, Min Gu et al. introduced artiﬁcial neural networks into the design of optical needles for the ﬁrst time [22]. They reported a two-layer fully connected neural network with advanced performance in solving nonlinear problems based on the dipole array principle, enabling the design of optical needles with a central zero intensity point. In 2021, a physically assisted artiﬁcial neural network scheme was proposed by Rui Chen et al [23]. They built a four-layer fully connected neural network as a transmission coefﬁcient Simo Wang, Jupu Yang, and Jixiao Liu are with the Institute of Optics and Electronics, Chinese Academy Sciences, Chengdu, Sichuan 610209, China, and also with School of Electronic, Electrical and Communication Engineer- ing, University of Chinese Academy of Sciences, Beijing 100049, China (e-mail: wangsimo18@mails.ucas.ac.cn; yangjupu20@mails.ucas.ac.cn; liujix- iao20@mails.ucas.ac.cn). Fanxing Li and Wei Yan are with the Institute of Optics and Electronics, Chinese Academy Sciences, Chengdu, Sichuan 610209, China (e-mail: lifanx- ing15@mails.ucas.ac.cn; yanwei@ioe.ac.cn). Digital Object Identiﬁer 10.1109/JPHOT.2022.3188303 IEEE PHOTONICS JOURNAL, VOL. 14, NO. 4, AUGUST 2022 IEEE PHOTONICS JOURNAL, VOL. 14, NO. 4, AUGUST 2022 The normalized intensities of the ideal optical needle is fed into the trained PONMD model, and the desired structural design of the modulator is also achieved. Compared with conventional methods, the PONMD method proposed in this paper is applica- ble to the rapid and efﬁcient design of modulators required for optical needles with different resolutions and focal depths. It is important for achieving periodic, cross-scale processing of high aspect ratio micro-nano structures with different line widths. T(θ) = 1, for 0 ≤θ < θ1, θ2 ≤θ < θ3, θ4 ≤θ < α, −1, for θ1 ≤θ < θ2, θ3 ≤θ < θ4. (4) (4) θi (i = 1–4) are the structural parameters of the ﬁve-ring phase optical needle modulator, and θi are normalized and expressed as Qi = θi/α. It is apparent that the structural parameters of each modulator correspond to a set of Q parameters, and a set of Q parameters corresponds to an optical needle of a speciﬁc DOF and FWHM. However, an optical needle with a speciﬁc DOF and FWHM may correspond to multiple sets of Q parameters, and it is difﬁcult to directly obtain the Q parameters from the optical needle. In addition, the number of combinations of Q parameters is quite large, and it is almost an impossible task to traverse all combinations under the current computational conditions. The following is a description of the methods for designing modulators. WANG et al.: DEEP LEARNING: A RAPID AND EFFICIENT ADAPTIVE DESIGN APPROACH FOR PHASE-TYPE OPTICAL NEEDLE 2741409 Fig. 2. The contrastive ﬂowchart of design process of PONMD method and conventional method. r = 0 and z = 0 directions and the structural parameters of the modulator required for each optical needle to build the data set. Weﬁrstusedthe“rand”functioninMATLABtogenerate50,000 sets of random matrices representing the Q parameters of the modulator, and then input these matrices into the physical model shown in Fig. 1 to calculate the corresponding focused spot and save the normalized intensities of the focused spot along the r = 0 and z = 0 directions. We ﬁltered 29,984 sets of data from all the data as a dataset, and all of them were in the shape of optical nee- dles. The dataset was divided by the hold-out method. Among them, the training set and test set account for 20000 and 9984, respectively. The intensities of the optical needles along the r = 0 and z = 0 directions are extracted and expressed as IDOF_ln and IFWHM_ln, respectively. l is the lth set of data. n is the nth sam- pling point of each set of data, and the value range of n is [1, 256]. The intensities of the optical needle along the r = 0 and z = 0 di- rections are stitched by rows and represented by Ilj = [IDOF_ln, IFWHM_ln], and the value range of j is [1, 512]. The structural parameters of the phase-type optical needle modulators in the dataset are denoted by Qli, where i denotes the i-th sampling point of each data set, and the value range of i is [1], [4]. At this point, the collection of the data set is completed, and the inputs and outputs of the PONMD model are Ilj and Qli, respectively. Fig. 2. The contrastive ﬂowchart of design process of PONMD method and conventional method. the modulator can be completed after one iteration. Therefore, the PONMD approach can signiﬁcantly improve efﬁciency. From the perspective of data collection, the PONMD ap- proach is based on a data-driven, which requires collecting a large amount of data that can describe the problem in advance, and then using some of the data as a training set to train the established PONMD model. B. Test and Comparison After the training of the PONMD model is completed, Eq. (5) is used to evaluate the PONMD model. It represents the mean absolute error between the designed and target values of the Q parameters of the modulator over the range of the test set. MAE_Q = L l=1 4 i=1 \|Qli−Q′ li\| 4 L , i = 1, 2, 3, 4 (5) (5) Where Qli and Q’li are the target and designed value of the i-th sampling point of the l-th sample, respectively. L is the total number of samples in the test set. The similarity function S_Q as shown in (6) is deﬁned to measure the similarity between the designed Q parameters and the target Q parameters. A higher similarity S_Q indicates that the designed value is closer to the target value. Intermsofself-adaptability,theconventionalmethodsdepend on the ﬁtness function deﬁned according to the characteristics of the target optical needle such as FWHM and DOF, which relies on the physical model for generating optical needles. The ﬁtness function directly determines the optimization result. When the target optical needles are different, the designer needs to change the ﬁtness function and re-optimize the design. Therefore, the conventional methods are not adaptive. However, the PONMD approach is data-driven. Once the model is trained, it can be adapted to the design of modulators required for optical needles with different FWHM and DOF. S_Q = 1 1 + MAE_Q × 100% (6) (6) MAE_Q and S_Q can characterize the overall level of the model in designing modulators. However, it is not clear how small MAE_Q should be smaller to meet the design require- ments. Therefore, we established the test and comparison pro- cess shown in Fig. 3 to verify whether the designed modulator meets the requirements. In summary, the PONMD method is highly competitive in periodic, cross-scale laser direct writing lithography with the optical needle as the direct writing head, which is important for realizing the processing of high aspect ratio micro-nano structures with different line widths and depths. The testing and comparison process is divided into three steps. The ﬁrst step is to use the trained PONMD model to design the Q parameters of the modulator through the intensities of the target optical needle. In the second step, the designed Q parameters are substituted into the physical model according to the Richards-Wolf vector diffraction theory to obtain intensities of the designed optical needle. B. Test and Comparison In the third step, the intensities of the target and designed optical needle are compared. WANG et al.: DEEP LEARNING: A RAPID AND EFFICIENT ADAPTIVE DESIGN APPROACH FOR PHASE-TYPE OPTICAL NEEDLE In contrast, conventional methods passively generate data during iterative optimization and then determine the strategy for generating new data in the next iteration by means of a deﬁned ﬁtness function, and do not end the generation of new data until an approximate optimal solution is found. Therefore, the data collection of the PONMD method is a one-time cost, while the data collection of the conventional methods is a continuous cost. A. Physical Model for Generating Optical Needles The schematic diagram of the radially polarized Bessel Gaus- sian beam tightly focused to generate an optical needle after modulation by a modulator is shown in Fig. 1. According to the Richards-Wolf vector diffraction theory, the radial and longitu- dinal electric ﬁeld components of the optical needle generated near the focal point of the focusing objective can be expressed as (1) and (2), respectively [15]. In the following, we compared the two methods in terms of time and computational resources consumption, data collection, and self-adaptability. From the perspective of time-consuming and computational resource-consuming, the integral calculation introduced by the physical model can take up a lot of computa- tional time and computational resources. Conventional methods require hundreds or even thousands of parameter sweeping, while the number of integration calculations in the physical model is T times the number of parameter sweeping, where T is the number of sampling points for the normalized intensities of the optical needle. On the contrary, the PONMD method only needs to focus on the design target, and the structural design of Er(r, z) = A α 0 cos1/2θ sin(2θ)l(θ)T(θ)J1(kr sin θ) exp(ikz cos θ)dθ (1 Ez(r, z) = 2iA α 0 cos1/2θsin2θl(θ)T(θ)J0(kr sin θ) Er(r, z) = A α 0 cos1/2θ sin(2θ)l(θ)T(θ)J1(kr sin θ) Er(r, z) = A α 0 cos1/2θ sin(2θ)l(θ)T(θ)J1(kr sin θ) exp(ikz cos θ)dθ ( α (1) Ez(r, z) = 2iA α 0 cos1/2θsin2θl(θ)T(θ)J0(kr sin θ) exp(ikz cos θ)dθ (2) A. Data Gathering The training of the PONMD model requires the collection of the normalized intensities of the optical needles along the IEEE PHOTONICS JOURNAL, VOL. 14, NO. 4, AUGUST 2022 2741409 Fig. 3. The test and comparison process. Fig. 4. The structural of the FCNN model. Fig. 5. FCNN Method: The comparison of the intensity distributions along (a) r = 0 and (b) z = 0 directions. Fig. 4. The structural of the FCNN model. Fig. 4. The structural of the FCNN model. Fig. 3. The test and comparison process. Fig. 3. The test and comparison process. Fig. 4. The structural of the FCNN model. Fig. 5. FCNN Method: The comparison of the intensity distributions along (a) r = 0 and (b) z = 0 directions. The mean absolute error of the intensities of the optical needle is deﬁned as: MAE_I = L l=1 512 j=1 \|Ilj−I′ lj\| 512 L , j = 1, 2, . . . , 512 (7) Where Ilj and I′ lj are the target and designed value of the j-th sampling point of the l-th sample, respectively. Similarly, deﬁne the similarity function S_I as shown in (8) to calculate the similarity of the normalized intensities of the designed optical needle and the target optical needle. Fig. 5. FCNN Method: The comparison of the intensity distributions along (a) r = 0 and (b) z = 0 directions. S_I = 1 1 + MAE_I × 100% (8) (8) IV. MODEL BUILDING AND TRAINING optical needle corresponding to the designed modulator by using the physical model for generating optical needle, which greatly reduces the calculation time and resource consumption. In ad- dition, Network2 is a six-layer fully connected neural network whose input is the normalized intensities of the optical needle and the output is the Q parameters of the modulator. The number of neurons in the ﬁve hidden layers are 1024, 666, 512, 200 and 100, respectively. This section introduced the building, training, evaluating, and testing and comparing process of the PONMD models. Before that, in order to illustrate the design idea of the PONMD model, the FCNN model based on fully connected neural network built by ourselves was introduced ﬁrst. Both models were built under a Dell laptop. The image processor, central processor, and oper- ating system are GTX1060 Max-Q, Core i5-8300H, and Ubuntu 16.04, respectively. The integrated development environment is Pycharm, and the framework is Pytorch. The experiment found that the MAE_I of Network1 on the test set is 0.0364, and the S_I is 96.49%. The MAE_Q of Network2 on the test set is 0.2773 and the S_Q is 78.29%. The S_Q of Network2 characterizes the ability of the FCNN model to design modulators. B. PONMD Model Based on the above research, we found that the design of the modulator can be regarded as a regression problem, which maps a vector of size 1 × 512 to a vector of size 1 × 4. Convolutional neural networks have local correlation and stronger expres- siveness. We ﬁrst proposed a CNN model with convolutional neural network as the basic element to design the structure of modulators, whose network structure and hyperparameters are shown in Fig. 6. The input and output are the normalized intensities of the optical needle along the r = 0 and z = 0 directions and the Q parameters of the modulator, respectively. The experiment found that the MAE_Q of the CNN model on the test set is 0.0772, and the S_Q is 92.83%. The encoder of NetworkA and NetworkB are cascaded using the partial stacking strategy [26]. The input of NetworkB is the output of the encoder of NetworkA, which is a vector of size 1 × 26. NetworkB consists of a convolution module and a fully connected module. The role of the convolutional layer in the convolution module is to capture the key data content from the input data, and the role of the pooling layer is to compress the convolutional results to obtain more important features and effectively control overﬁtting. The Batch Normalization layer is added to solve the problem of the internal covariate shift of the convolutional neural network. The Flatten layer stretches the feature map obtained from the last convolutional layer into a vector, and the fully connected module completes the feature- to-target mapping. The output of NetworkB is the Q parameters of the modulator, which is a 1 × 4 vector. Therefore, the Net- workB is a dimensionality reduction process. Slowly reducing the dimensionality by increasing the number of network layers helps to retain valid information. However, too many layers not only lead to information redundancy, but are also harmful to the training of the model. Therefore, under the premise of The CNN model was tested using the intensity of the target optical needle in Fig. 5 as input. The results of the comparison between the intensity of the optical needle corresponding to the phase-type optical needle modulator designed by the CNN model and the intensity of the target optical needle are shown in Fig. 7. WANG et al.: DEEP LEARNING: A RAPID AND EFFICIENT ADAPTIVE DESIGN APPROACH FOR PHASE-TYPE OPTICAL NEEDLE 2741409 Fig. 6. The structural of the CNN model. The hyperparameters p, f, s and c denote the convolution kernel size, step size, padding and number of channels, respectively. The number on the fully connected layer represents the number of neurons in that layer. Fig. 6. The structural of the CNN model. The hyperparameters p, f, s and c denote the convolution kernel size, step size, padding and number of channels, respectively. The number on the fully connected layer represents the number of neurons in that layer. Fig. 7. CNN Method: The comparison of the intensity distributions along (a) r = 0 and (b) z = 0 directions. proposed, whose network structure and hyperparameters are shown in Fig. 8. The hyperparameters tuning process of the network is not the focus of this paper, so it will not be repeated here. Only the structure and function of the network will be introduced below. NetworkA is a convolutional autoencoder [24], [25], which consists of an encoder and a decoder. It enables feature down- scaling and feature extraction, and uses the extracted high-level abstract features as input to NetworkB to improve the perfor- mance of the whole deep learning model. The input and output of NetworkA are the normalized intensities of the target and designed optical needles, respectively. The encoder maps the original data to the hidden layer feature space and compresses the hidden feature data into a low-dimensional encoding space to extract the most relevant information. Among them, the convo- lutional layer can extract the characteristics of the input data and reduce the parameters of the network. The max pooling layer can achieve spatial dimensionality reduction and prevent over-ﬁtting to a certain extent. The decoder decompresses the encoded data and reconstructed the same output as the encoder input. The encoder in NetworkA not only summarizes the inherent characteristics of high-dimensional input data, but also provides low-dimensional input data for NetworkB. Fig. 7. CNN Method: The comparison of the intensity distributions along (a) r = 0 and (b) z = 0 directions. data equally, which is contrary to the characteristics of the input data. Therefore, the FCNN model cannot learn the relationship between the data. A. Related Work The test and comparison process shown in Fig. 3 is used to determine whether this FCNN model meets the design require- ments of the modulator. As shown in Fig. 5, a randomly selected optical needle from the test set is input to the trained FCNN model as the target optical needle, and the intensity distributions of the optical needle corresponding to the designed modulator are shown as the red dashed line. The results show that the FCNN model has a large error in the design of modulators. That is, the FCNN model with MAE_Q of 0.2773 does not satisfy the design requirements of modulators, which mainly because the expres- sive power of fully connected neural networks is limited when the number of input parameters is too large. For the input data, the connection between each sample point and its surrounding neighboring sample points is relatively tight, and the connection with the far away sample points is small. However, the fully connected neural network connects one neuron to all neurons in the previous layer, and treats each sampling point of the input Fully connected neural networks are widely used because of their simple structure and advanced performance in solving nonlinear problems. Inspired by the research of Min Gu and Rui Chen et al. [22], [23], we ﬁrst established a fully connected neural network (FCNN) model as shown in Fig. 4 to realize the design of the modulator. The FCNN model is composed of Network1 and Network2, and the optimal network structure of the FCNN model is ﬁnally obtained through tuning hyperpa- rameters. Network1 is used to replace the physical model for generating the optical needle, and its input and output are the Q parameters of the modulator and the normalized intensities of the optical needle along the r = 0 and z = 0 directions, respectively. Network1 is a three-layer fully connected neural network with 512 neurons in both hidden layers. Since a modulator corre- sponds to only one optical needle, the network with a one-to-one mapping relationship is easy to converge in the training process. Network1 avoids the integral operation when calculating the WANG et al.: DEEP LEARNING: A RAPID AND EFFICIENT ADAPTIVE DESIGN APPROACH FOR PHASE-TYPE OPTICAL NEEDLE B. PONMD Model The results show that the phase-type optical needle modulator designed by the CNN model has some improvement in similarity compared with the FCNN model, but the predicted optical needle still has a certain gap compared with the target optical needle. To further improve the design performance of the model, a PONMD model consists of NetworkA and NetworkB is IEEE PHOTONICS JOURNAL, VOL. 14, NO. 4, AUGUST 2022 2741409 2741409 IEEE PHOTONICS JOURNAL, VOL. 14, NO. 4, AUGUST 2022 Fig. 8. The structure of PONMD Model. ensuring design similarity, we reduced the number of network layers of NetworkB. Finally, three convolutional layers and six hidden layers are used for the convolutional module and the fully connected module, respectively. 1) Training and Evaluation: NetworkA and NetworkB in the PONMD model needed to be individually trained with the same batch size of 16 for 500 iterations, which employed an Adam optimization [27] with a learning rate of 0.001. In order to reduce overﬁtting, we used the dropout to increase regularization, with initial values set at 0.5. After every 50 iterations, the dropout factor, and learning rate were reduced by 0.5, and 0.2, respec- Fig. 9. PONMD Method: The comparison of the intensity distributions along (a) r = 0 and (b) z = 0 directions. Fig. 8. The structure of PONMD Model. Fig. 8. The structure of PONMD Model. ensuring design similarity, we reduced the number of network layers of NetworkB. Finally, three convolutional layers and six hidden layers are used for the convolutional module and the fully connected module, respectively. p y 1) Training and Evaluation: NetworkA and NetworkB in the PONMD model needed to be individually trained with the same batch size of 16 for 500 iterations, which employed an Adam optimization [27] with a learning rate of 0.001. In order to reduce overﬁtting, we used the dropout to increase regularization, with initial values set at 0.5. After every 50 iterations, the dropout factor, and learning rate were reduced by 0.5, and 0.2, respec- tively. The smooth L1 loss was use as the cost function in the experiment. The loss functions of NetworkA and NetworkB are shown in (9) and (10), respectively. Fig. 9. PONMD Method: The comparison of the intensity distributions along (a) r = 0 and (b) z = 0 directions. TABLE I LIST OF THE DESIGNED PHASE-TYPE OPTICAL NEEDLE MODULATORS BY PONMD METHOD IRadial = exp −(r −bR)2 2c2 R (11) Iaxial3 = ⎧ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎪ ⎪ ⎩ exp − z+bA cA 2 , z < −bA 1, −bA ≤z ≤bA exp − z−bA cA 2 , z < bA (12) (11) was trained, three target optical needle with different DOF and FWHM were arbitrarily taken in the test set to design the structure of the required modulators. The normalized intensity distributions of the target optical needles along the r = 0 and z = 0 directions are shown by the solid blue line. Through the testing and comparison process shown in Fig. 3, the corresponding optical needles of the modulators designed by the PONMD method are shown in Fig. 10(a4)–(c4), and the normalized intensity distributions of each optical needle along the r = 0 and z = 0 directions are indicated by the red dashed lines. Comparing the red dashed lines with the blue solid lines, it can be found that the modulators designed by the PONMD model can meet the requirements of the optical needles that with different DOF and FWHM. In addition, it also shows that the PONMD method is adaptive. Applying the PONMD method to laser direct lithography with an optical needle as the direct writing head can improve the processing speed and efﬁciency of periodic, cross-scale, high aspect ratio micro-nano structures. (12) Where bR and bA represent the center coordinates of the peak of the Gaussian curve. cR and cA are the width of the Gaussian curve, which directly affects the slope of the edge curve. bR, bA, cR, and cA are kept in the same units as z and r, and they are all quantiﬁed in wavelength λ. B. Design Example of a Modulator Corresponding to an Ideal Optical Needle V. DESIGN OF MODULATORS CORRESPONDING TO IDEAL OPTICAL NEEDLE The required for the optical needle in the test set is designed above, but the speciﬁc intensity distributions of the target optical needle in the actual design is usually unknown. Therefore, in this section, the optical needle after ignoring the side lobes is regarded as an ideal optical needle. The mathematical model of the ideal optical needle is established, and the required mod- ulator is designed with the normalized intensities of the ideal optical needle as the input of the PONMD model. Fig. 10. The design of modulatorscorresponding to differentDOFand FWHM of the optical needles. A. Mathematical Model of the Ideal Optical Needle The intensities of the ideal optical needle along the z = 0 direction can be approximated as a ﬁrst-order Gaussian function, and the intensities of the ideal optical needle along the r = 0 direction can be regarded as a piecewise function with a ﬂat top and a ﬁrst-order Gaussian function on both sides. Therefore, the mathematical model of the ideal optical needle is deﬁned as (11) and (12) to ﬁt the normalized intensity distributions along the z = 0 direction and the r = 0 direction, respectively. WANG et al.: DEEP LEARNING: A RAPID AND EFFICIENT ADAPTIVE DESIGN APPROACH FOR PHASE-TYPE OPTICAL NEEDLE WANG et al.: DEEP LEARNING: A RAPID AND EFFICIENT ADAPTIVE DESIGN APPROACH FOR PHASE-TYPE OPTICAL NEEDLE 2741409 Fig. 10. The design of modulatorscorresponding to differentDOFand FWHM of the optical needles. with a high effective numerical aperture has a great inﬂuence on the optical needle. Therefore, the inﬂuence of Q4 on the optical needle is the greatest, and the inﬂuence of Q1 on the optical needle is the smallest. V. DESIGN OF MODULATORS CORRESPONDING TO IDEAL OPTICAL NEEDLE B. PONMD Model and NetworkB are updated and corrected according to the dif- ference using the back propagation algorithm, and the process is repeated until the model meets the demand. The experiment found that the MAE_I of NetworkA on the test set is 0.0012, and the S_I is 99.88%. The MAE_Q of NetworkB on the test set is 0.0338 and the S_Q is 96.73%. The S_QofNetworkBcharacterizestheabilityofthePONMDmodel to design phase-type optical needle modulators. Loss_I = Ij −I′ j −0.5, Ij −I′ j > 1 0.5(Ij −I′ j)2, Ij −I′ j < 1 (9) Loss_Q = \|Qi −Q′ i\| −0.5, \|Qi −Q′ i\| > 1 0.5(Qi −Q′ i)2, \|Qi −Q′ i\| < 1 (10) (9) (10) 2) Test and Comparison: The test and comparison process shown in Fig. 3 is also used to determine whether the PONMD model meets the design requirements of the modulator. The normalized intensities of the target optical needle shown in Fig. 5 was used as the input, and the results were shown in Fig. 9. It is easy to ﬁnd that the PONMD model can accurately design the structure of the modulator required by the target optical needle. Ij and Ij’ denote the intensity of the optical needles, Qj and Qj’ denote the structural parameters of phase-type optical needle modulators. Ij and Qj are the actual values, and Ij’ and Qj’ are the predictions values of the data labels. Based on the deﬁned loss function, the difference between the model prediction and the data label value can be obtained. Then, the weights of the convolutional, pooling, and fully connected layers in NetworkA 3) Design of Modulators Corresponding to Optical Needles With Different DOF and FWHM: After the PONMD model IEEE PHOTONICS JOURNAL, VOL. 14, NO. 4, AUGUST 2022 The FCNN model based on fully connected neural network treats all sampling points equally and ignores the local connection relationship between individual sampling points. While the PONMD model based on convolutional neural network is more concerned with the extraction of local fea- tures, it can synthesize the local information at a higher level to get the global information. Therefore, the PONMD model ﬁts the intensity distribution characteristics of optical needles more closely. In addition, Network A, consisting of encoders and decoders, can extract high-level abstract features from the input data and use the high-level abstract features as another deep learning model input, ﬁnally improving the performance of the whole deep learning model. In conclusion, both the optical physical properties of the optical needle and the experimental results demonstrate that the PONMD method proposed in this paper can achieve rapid, efﬁcient and adaptive design of the modulators. Fig. 11. The comparison of the intensity distributions along r = 0 and z = 0 directions. (a) Design target, (b) Conventional method, (c) PONMD method. Fig. 12. Comparison of three design methods in terms of design time, number of iterations, mean squared error, and similarity. VI. CONCLUSION In this paper, a new method named PONMD is proposed to design phase-type optical needle modulators. The PONMD model enables the design of phase-type optical needle modula- tors required for optical needles with different focal depths and resolutions. In addition, ignoring the side lobes of the optical needle, the mathematical model of the ideal optical needle is established. It is found by experiments that the PONMD model can also realize the design of the phase-type optical needle modulator required by the ideal optical needle. Com- pared with the conventional method and the FCNN method, the PONMD method proposed in this paper has the advantages of rapidity, high efﬁciency and self-adaptation. It only takes 0.5526ms to design a phase-type optical needle modulator re- quired for a target optical needle, and the similarity between the design value and the target value is as high as 96.73%. In addition, the PONMD method can be applied to periodic and cross-scale laser direct writing lithography with optical needles as direct writing heads, which provides a powerful technology to achieve rapid and efﬁcient processing of high aspect ratio micro-nano structures with different line widths and depths. Fig. 12. Comparison of three design methods in terms of design time, number of iterations, mean squared error, and similarity. are [Q′ 1, Q′ 2, Q′ 3, Q′ 4] = [0.1715, 0.3468, 0.5169, 0.6862], and the normalized intensity distributions of the corresponding op- tical needle are shown in the red curves in Fig. 11(c1)–(c2). Obviously, the designed optical needle obtained by the PONMD method is closer to the target ideal optical needle than the conventional method. In order to more intuitively illustrate the advantages of the PONMD method proposed in this paper. As shown in Fig. 12, we compared the conventional method, the FCNN method and the PONMD method from four aspects: design time, iterations, mean absolute error and similarity. It can be seen that the time to design a modulator by the PONMD method is nearly 8 orders of magnitude faster than that of the conventional method. In addition, compared with the conventional method and the FCNN method, the mean absolute error of the PONMD method is the smallest, and the similarity is the highest. [1] D. Richard, R. D. Romea, and W. D. Kimura, “Modeling of inverse erenkov laser acceleration with axicon laser-beam focusing,” Phys. Rev. D Part Fields, vol. 42, no. 5, pp. B. Design Example of a Modulator Corresponding to an Ideal Optical Needle In the following, an ideal optical needle is used as the tar- get optical needle, and the structure of the required modula- tor is designed by the conventional method and the PONMD method, respectively. The intensity distributions along r = 0 and z = 0 directions of the target ideal optical needle are shown in Fig. 11(a1)–(a2). The Q parameters of the modulator designed by the conventional method are [Q′ 1, Q′ 2, Q′ 3, Q′ 4] = [0.0691, 0.3035, 0.477, 0.6527]. Substituting the designed Q pa- rameters of the modulator into the physical model for generating optical needle, the intensity distributions shown by the green dotted line in Fig. 11(b1) and (b2) are obtained. The results show that there is a 0.84λ deviation of the DOF between the designed and target optical needles. Using the same design target, the Q parameters of the modulator designed by the PONMD method From the comparison in Table I, it can be found that the designed values Q’2–Q’4 of the modulators are very close to the target values, while Q’1 may have certain errors in some cases. However, it does not affect the designed optical needle. This is mainly because the effective numerical aperture is proportional to the radius of the pupil at the exit of the objective lens. From the structure of the modulator, Q1 is close to the center of the focusing objective, while Q4 is close to the edge. Therefore, the effective numerical apertures corresponding to Q1, Q2, Q3, and Q4 are increased sequentially. The incident beam at a position IEEE PHOTONICS JOURNAL, VOL. 14, NO. 4, AUGUST 2022 IEEE PHOTONICS JOURNAL, VOL. 14, NO. 4, AUGUST 2022 2741409 Fig. 11. The comparison of the intensity distributions along r = 0 and z = 0 directions. (a) Design target, (b) Conventional method, (c) PONMD method. intensity of the optical needle. In contrast, the inputs to both the FCNN model and the PONMD model are 512 neurons, with each neuron representing one sampling point. The ﬁrst 256 sampling points and the last 256 sampling points represent the normalized intensity distribution of the optical needle along the z = 0 and r = 0 directions, respectively. From the physical meaning of the sampling points, it can be found that the curve formed by these sampling points can be regarded as a continuous function, and the closer the distance between two sampling points, the stronger the connection. The FCNN model based on fully connected neural network treats all sampling points equally and ignores the local connection relationship between individual sampling points. While the PONMD model based on convolutional neural network is more concerned with the extraction of local fea- tures, it can synthesize the local information at a higher level to get the global information. Therefore, the PONMD model ﬁts the intensity distribution characteristics of optical needles more closely. In addition, Network A, consisting of encoders and decoders, can extract high-level abstract features from the input data and use the high-level abstract features as another deep learning model input, ﬁnally improving the performance of the whole deep learning model. In conclusion, both the optical physical properties of the optical needle and the experimental results demonstrate that the PONMD method proposed in this paper can achieve rapid, efﬁcient and adaptive design of the modulators. intensity of the optical needle. In contrast, the inputs to both the FCNN model and the PONMD model are 512 neurons, with each neuron representing one sampling point. The ﬁrst 256 sampling points and the last 256 sampling points represent the normalized intensity distribution of the optical needle along the z = 0 and r y p g = 0 directions, respectively. From the physical meaning of the sampling points, it can be found that the curve formed by these sampling points can be regarded as a continuous function, and the closer the distance between two sampling points, the stronger the connection. VI. CONCLUSION 1807–1818, 1990, doi: 10.1103/phys- revd.42.1807. This is mainly because the conventional method relies on an adaptationfunctiondeﬁnedaccordingtotheresolutionanddepth of focus of the target optical needle, without a constraint on the [2] S. C. Tidwell, G. H. Kim, and W. D. Kimura, “Efﬁcient radially polarized laser beam generation with a double interferometer,” Appl. Opt., vol. 32, no. 27, pp. 5222–5229, 1993, doi: 10.1364/AO.32.005222. WANG et al.: DEEP LEARNING: A RAPID AND EFFICIENT ADAPTIVE DESIGN APPROACH FOR PHASE-TYPE OPTICAL NEEDLE 2741409 2741409 [3] J. R. Fontana and R. H. Pantell, “A high-energy, laser accelerator for electrons using the inverse Cherenkov effect,” J. Appl. Phys., vol. 54, no. 8, pp. 4285–4288, 1983, doi: 10.1063/1.332684. [16] J. Yu, C. Zhou, and W. Jia, “Transverse superresolution with ex- tended depth of focus using binary phase ﬁlters for optical stor- age system,” Opt. Commun., vol. 283, no. 21, pp. 4171–4177, 2010, doi: 10.1016/j.optcom.2010.06.055. [4] L. Cicchitelli, H. Hora, and R. Postle, “Longitudinal ﬁeld components for laser beams in vacuum,” Phys. Rev. A, vol. 41, no. 7, pp. 3727–3732, 2008, doi: 10.1103/PhysRevA.41.3727. [17] C. K. Shi, Z. Q. Nie, Y. T. Tian, C. Liu, Y. C. Zhao, and B. H. Jia, “Super-resolution longitudinally polarized light needle achieved by tightly focusing radially polarized beams,” Optoelectron. Lett., vol. 14, no. 1, pp. 1–5, 2018, doi: 10.1007/s11801-018-7162-6. [5] Q. Zhan, “Trapping metallic Rayleigh particles with radial po- larization,” Opt. Exp., vol. 12, no. 15, pp. 3377–3382, 2004, doi: 10.1364/OPEX.12.003377. g y p p pp. 1–5, 2018, doi: 10.1007/s11801-018-7162-6. [18] J. Lin, H. Zhao, Y. Ma, J. Tan, and P. Jin, “New hybrid ge- netic particle swarm optimization algorithm to design multi-zone bi- nary ﬁlter,” Opt. Exp., vol. 24, no. 10, pp. 10748–10758, 2016, doi: 10.1364/OE.24.010748. [6] Q. Zhan, “Trapping metallic Rayleigh particles with radial polarization: Reply to comment,” Opt. Exp., vol. 20, no. 6, pp. 6058–6059, 2012, doi: 10.1364/OE.20.006058. [7] Y. Zhang and J. Bai, “Improving the recording ability of a near-ﬁeld optical storage system by higher-order radially polarized beams,” Opt. Exp., vol. 17, no. 5, pp. 3698–3706, 2009, doi: 10.1364/OE.17.003698. [19] W. Ma, Z. Liu, Z. A. Kudyshev, A. Boltasseva, W. Cai, and Y. Liu, “Deep learning for the design of photonic structures,” Nat. Photon., vol. 15, no. 2, pp. 77–90, 2021, doi: 10.1038/s41566-020-0685-y. [8] K. S. Youngworth and T. G. Brown, “Focusing of high numerical aperture cylindrical-vector beams,” Opt. Exp., vol. 7, no. 2, pp. 77–87, 2000, doi: 10.1364/OE.7.000077. VI. CONCLUSION [20] Z. Liu, D. Zhu, S. P. Rodrigues, K. T. Lee, and W. Cai, “Generative model for the inverse design of metasurfaces,” Nano Lett., vol. 18, no. 10, pp. 6570–6576, 2018, doi: 10.1021/acs.nanolett.8b03171. [9] M. Meier, V. Romano, and T. Feurer, “Material processing with pulsed radially and azimuthally polarized laser radiation,” Appl. Phys. A, vol. 86, no. 3, pp. 329–334, 2007, doi: 10.1007/s00339-006-3784-9. [21] W. Ma, F. Cheng, and Y. Liu, “Deep-learning-enabled on-demand design of chiral metamaterials,” ACS Nano, vol. 12, no. 6, pp. 6326–6334, 2018, doi: 10.1021/acsnano.8b03569. [22] W. Xin, Q. Zhang, and M. Gu, “Inverse design of optical needles with central zero-intensity points by artiﬁcial neural networks,” Opt. Exp., vol. 28, no. 26, pp. 38718–38732, 2020, doi: 10.1364/oe.410073. [10] Y. Zha, J. Wei, H. Wang, and F. Gan, “Creation of an ultra-long depth of focus super-resolution longitudinally polarized beam with a ternary optical element,” J. Opt., vol. 15, no. 7, 2013, Art. no. 075703, doi: 10.1088/2040-8978/15/7/075703. [23] Z.-Y. Chen, Z. Wei, R. Chen, and J.-W. Dong, “Focus shaping of high numerical aperture lens using physics-assisted artiﬁcial neu- ral networks,” Opt. Exp., vol. 29, no. 9, pp. 13011–13024, 2021, doi: 10.1364/oe.421354. [11] C.-C. Sun and C.-K. Liu, “Ultrasmall focusing spot with a long depth of focus based on polarization and phase modulation,” Opt. Lett., vol. 28, no. 2, pp. 99–101, 2003, doi: 10.1364/OL.28.000099. [12] K. Kitamura, K. Sakai, and S. Noda, “Sub-wavelength focal spot with long depth of focus generated by radially polarized, narrow- width annular beam,” Opt. Exp., vol. 18, no. 5, pp. 4518–4525, 2010, doi: 10.1364/OE.18.004518. [24] E. Kodirov, T. Xiang, and S. Gong, “Semantic autoencoder for zero-shot learning,” in Proc. 30th IEEE Conf. Comput. Vis. Pattern Recognit ., 2017, pp. 4447–4456, doi: 10.1109/CVPR.2017.473. [25] J. Deng, Z. Zhang, F. Eyben, and B. Schuller, “Autoencoder-based un- supervised domain adaptation for speech emotion recognition,” IEEE Signal Process. Lett., vol. 21, no. 9, pp. 1068–1072, Sep. 2014, doi: 10.1109/LSP.2014.2324759. [13] K. Huang, P. Shi, X. Kang, X. Zhang, and Y. Li, “Design of DOE for generating a needle of a strong longitudinally polarized ﬁeld,” Opt. Lett., vol. 35, no. 7, pp. 965–967, 2010, doi: 10.1364/OL.35.000965. pp [14] L. Jie et al., “Achievement of longitudinally polarized focusing with long focal depth by amplitude modulation,” Opt. Lett., vol. 36, no. 7, pp. 1185–1187, 2011, doi: 10.1364/OL.36.001185. [26] D. H. Wolpert, “Stacked generalization,” Neural Netw., vol. 5, no. 2, pp. VI. CONCLUSION 241–259, 1992, doi: 10.1016/S0893-6080(05)80023-1. [27] D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” in Proc. BT-3rd Int. Conf. Learn. Representations, San Diego, CA, USA, May 7–9, 2015. [Online]. Available: http://arxiv.org/abs/1412.6980 pp [15] H. Wang, L. Shi, B. Luk‘yanchuk, C. Sheppard, and C. T. Chong, “Creation of a needle of longitudinally polarized light in vacuum using binary optics,” Nature Photon., vol. 2, no. 8, pp. 501–505, 2008, doi: 10.1038/npho- ton.2008.127.
https://openalex.org/W3205884569	https://europepmc.org/articles/pmc8564359?pdf=render	English	null	The Correlation Between Olfactory Test and Hippocampal Volume in Alzheimer's Disease and Mild Cognitive Impairment Patients: A Meta-Analysis	Frontiers in aging neuroscience	2,021	cc-by	7,771	SYSTEMATIC REVIEW published: 20 October 2021 doi: 10.3389/fnagi.2021.755160 Received: 08 August 2021 Accepted: 22 September 2021 Published: 20 October 2021 Conclusion: The correlation appears to be more predominant in the cognitive disorder group (including MCI and AD) than in the non-cognitive disorder group. Age is an independent factor that affects the severity of the correlation during disease progression. The mildness of the correlation suggests that olfactory tests may be more accurate when combined with other non-invasive examinations for early detection. Citation: Su M-W, Ni J-N, Cao T-Y, Wang S-S, Shi J and Tian J-Z (2021) The Correlation Between Olfactory Test and Hippocampal Volume in Alzheimer’s Disease and Mild Cognitive Impairment Patients: A Meta-Analysis. Front. Aging Neurosci. 13:755160. doi: 10.3389/fnagi.2021.755160 https://inplasy.com/, identiﬁer INPLASY Keywords: olfactory deﬁcits, hippocampus, mild cognitive impairment, Alzheimer’s disease, meta-analysis The Correlation Between Olfactory Test and Hippocampal Volume in Alzheimer’s Disease and Mild Cognitive Impairment Patients: A Meta-Analysis Ming-Wan Su 1†, Jing-Nian Ni 2†, Tian-Yu Cao 1, Shuo-Shi Wang 1, Jing Shi 2* and Jin-Zhou Tian 2* 1 Dongzhimen Hospital, Beijing University of Chinese Medicine, Beijing, China, 2 Department of Neurology, Dongzhimen Hospital, Beijing University of Chinese Medicine, Beijing, China Background: Previous studies have reported that olfactory identiﬁcation deﬁcits may be the earliest clinical features of Alzheimer’s disease (AD). However, the association between odor identiﬁcation and hippocampal atrophy remains unclear. Edited by: Ping Wu, Fudan University, China Edited by: Ping Wu, Fudan University, China Objective: This meta-analysis quantiﬁed the correlation between odor identiﬁcation test scores and hippocampal volume in AD. Reviewed by: Carlos Ayala Grosso, Instituto Venezolano de Investigaciones Cientiﬁcas, IVIC, Venezuela Liping Fu, China-Japan Friendship Hospital, China Method: A search of the PUBMED, EMBASE, and WEB OF SCIENCE databases was conducted from January 2003 to June 2020 on studies with reported correlation coefﬁcients between olfactory identiﬁcation score and hippocampal volume in patients with amnestic AD or mild cognitive impairment (MCI). The quality of the studies was assessed using the Newcastle-Ottawa quality assessment scale (NOS). Pooled r-values were combined and computed in R studio. *Correspondence: Jing Shi shijing87@hotmail.com Jin-Zhou Tian jztian@hotmail.com Results: Seven of 627 original studies on AD/MCI using an olfactory identiﬁcation test (n = 902) were included. A positive correlation was found between hippocampal volume and olfactory test scores (r = 0.3392, 95% CI: 0.2335–0.4370). Moderator analysis showed that AD and MCI patients were more profoundly correlated than normal controls (AD: r = 0.3959, 95% CI: 0.2605–0.5160; MCI: r = 0.3691, 95% CI: 0.1841–0.5288; NC: r = 0.1305, 95% CI: −0.0447–0.2980). Age difference and patient type were the main sources of heterogeneity in this analysis. †These authors have contributed equally to this work and share ﬁrst authorship Received: 08 August 2021 Accepted: 22 September 2021 Published: 20 October 2021 Systematic Review Registration: 202140088. October 2021 \| Volume 13 \| Article 755160 Frontiers in Aging Neuroscience \| www.frontiersin.org Olfactory Function and Hippocampal Volume Su et al. Data Extraction The coeﬃcient r between olfactory test scores and hippocampal volume (either calculated using the Pearson or Spearman method) were extracted in eligible studies, which could be either in total (left and right hippocampal volume) or bilaterally (left or right hippocampal volume). In some studies, the r-values were tabulated directly. For others in which these values were absent, SPSS 22.0 software (IBM, Inc., Chicago) was used to calculate the Pearson correlation coeﬃcient if the raw data was obtainable. However, the r-value usually does not follow a normal distribution. Since the variance strongly depends on the correlation, it usually cannot be directly synthesized. The bias from these sample correlations could be partially eliminated through correction of the Fisher estimator (Berry and Mielke, 2000). Thus, an r to Z transformation—Fisher’s z transformation—was introduced. The correlation was converted to Fisher’s z-scale to obtain a normal distribution. Quality Assessment The methodological quality of the included studies was assessed using the Newcastle-Ottawa Quality Assessment Scale (NOS) (Wells et al., 2013) by two independent researchers (M-WS and T-YC). Quality evaluation was applied to assess non-randomized studies. The NOS scale contains four domains including patient selection, comparability, and ascertainment of exposure or outcome of interest for case-control or cohort studies. The scale is assigned from 0 to 9 points, with studies scoring ≥7 points being considered high quality. Hippocampal atrophy and volumetric measurements are included among the biomarkers of neuronal injury in MCI and AD diagnosis (Albert et al., 2011). In recent years, the link between olfactory identiﬁcation performance and hippocampal atrophy has been recognized in some cross-sectional and longitudinal studies (Murphy et al., 2003; Kjelvik et al., 2014; Marigliano et al., 2014; Hagemeier et al., 2016). These positive results suggest that olfactory deﬁcits may be a potential biomarker of hippocampal function. The aim of this systematic review and meta-analysis was to examine whether olfactory deﬁcits correlate quantitatively with hippocampal atrophy, and to provide a comprehensive overview of the circumstances under which this correlation may be prominent due to diﬀerent moderation factors. INTRODUCTION INPLASY202140088; doi: 10.37766/inplasy2021.4.0088) Two researchers (M-WS, S-SW) separately conducted an online search for papers from the PUBMED, EMBASE, and WEB OF SCIENCE databases from January 2003 to June 2020 using the MESH terms “Alzheimer’s disease” and free words “olfactory” and “hippocampus OR hippocampal” (in the title/abstract). A complementary search of “Mild cognitive impairment” (free words in the title/abstract) substituting “Alzheimer’s disease” was repeated. Among the results, we read through the abstract to include the studies that could potentially meet the criteria, then screened the full article for further veriﬁcation, as well as relevant articles from the references in the full text for Supplementary Material. Alzheimer’s disease (AD) is an insidiously progressive neurodegenerative disease that primarily causes dementia. It is estimated that 44 million people live with this condition (Lane et al., 2018). Mild cognitive impairment (MCI) is a transitional stage between normal cognitive functioning and dementia (Albert et al., 2011). Approximately 15% to 20% of people aged ≥65 years have MCI and are susceptible to dementia, with a higher conversion rate (Roberts and Knopman, 2013). AD is characterized by memory decline, which is related to pre-mature atrophy of the hippocampus, entorhinal cortex, and other medial temporal lobe structures (Hatashita and Yamasaki, 2013). Alteration in olfactory function often coincides with clinical symptoms and may even precede it (Hawkes, 2003). Olfactory dysfunction (OD) typically occurs in the prodromal stage of AD and can progress to the disease. Since early detection is crucial to prevent and slow progression, OD has been considered as a potential clinical marker for AD prediction, severity, and progression (Servello et al., 2015; Zou et al., 2016b). Olfactory structures, such as the entorhinal cortex, amygdala, hippocampus, caudate, and other medial temporal lobes have been discovered (Kovács et al., 1999; Karas et al., 2003) to contain classic pathological features, such as neuroﬁbrillary tangles and amyloid-β plaques, which are also observed in olfactory regions in early stage AD and MCI patients, including the olfactory bulb and tract and anterior olfactory nucleus (Hyman et al., 1991). Studies have suggested that aggregation of Aβ and tau proteins occurs in the olfactory neuroepithelium. Nevertheless, the central olfactory structures play a more important role in olfactory dysfunction. Impaired odor identiﬁcation during lifetime was found to be robustly related to increased density of tangles in the entorhinal cortex and CA1/subiculum region of the hippocampus, but unrelated to other cortical sites after death (Wilson et al., 2007). Study Selection Studies were included if they met the following criteria: (1) participants with clinical diagnosis of amnestic AD or MCI were involved, with or without a health control; (2) both olfactory testing and hippocampal volumetric counting from MRI images were conducted from both hemisphere; (3) the correlation coeﬃcient could be extracted directly or through calculation from the raw data; (4) studies in English published in peer- reviewed journals from 2003 onwards; (5) study type was a cohort study, case-control or cross-sectional study. The results were ﬁltered to include only those written in English and conducted on living humans. Olfactory structures, such as the entorhinal cortex, amygdala, hippocampus, caudate, and other medial temporal lobes have been discovered (Kovács et al., 1999; Karas et al., 2003) to contain classic pathological features, such as neuroﬁbrillary tangles and amyloid-β plaques, which are also observed in olfactory regions in early stage AD and MCI patients, including the olfactory bulb and tract and anterior olfactory nucleus (Hyman et al., 1991). Studies have suggested that aggregation of Aβ and tau proteins occurs in the olfactory neuroepithelium. Nevertheless, the central olfactory structures play a more important role in olfactory dysfunction. Impaired odor identiﬁcation during lifetime was found to be robustly related to increased density of tangles in the entorhinal cortex and CA1/subiculum region of the hippocampus, but unrelated to other cortical sites after death (Wilson et al., 2007). INTRODUCTION Alzheimer’s disease (AD) is an insidiously progressive neurodegenerative disease that primarily causes dementia. It is estimated that 44 million people live with this condition (Lane et al., 2018). Mild cognitive impairment (MCI) is a transitional stage between normal cognitive functioning and dementia (Albert et al., 2011). Approximately 15% to 20% of people aged ≥65 years have MCI and are susceptible to dementia, with a higher conversion rate (Roberts and Knopman, 2013). AD is characterized by memory decline, which is related to pre-mature atrophy of the hippocampus, entorhinal cortex, and other medial temporal lobe structures (Hatashita and Yamasaki, 2013). Alteration in olfactory function often coincides with clinical symptoms and may even precede it (Hawkes, 2003). Olfactory dysfunction (OD) typically occurs in the prodromal stage of AD and can progress to the disease. Since early detection is crucial to prevent and slow progression, OD has been considered as a potential clinical marker for AD prediction, severity, and progression (Servello et al., 2015; Zou et al., 2016b). Association Between Olfactory Tests Score and Hippocampal Volumes There was a positive correlation between olfactory test scores and hippocampal volume (r = 0.3392, 95% CI: 0.2335–0.4370, p < 0.0001) (Figure 2). Egger’s regression test revealed an overall reporting bias (p = 0.029). A trim-and-ﬁll funnel plot showed a weak positive correlation (r = 0.2074, 95% CI: 0.0876–0.3214, p < 0.0001). Further, an inﬂuential analysis identiﬁed that no outliers in the included studies could reverse the analytical results using the leave-one-out method (Figure 3). Moreover, there was moderate heterogeneity in the sample of all included studies (I2 = 57%, p < 0.01). A total of seven studies were included in the meta-analysis (Table 1). Five of the seven studies were considered high- quality (Table 2). Follow-up research was performed in a pilot study (Marigliano et al., 2014) which contains a baseline TDI score, hippocampal volume, and 12-month follow-up data. We computed the Pearson correlation coeﬃcient r from the baseline data, since the baseline participants were all clinically conﬁrmed aMCI participants. A cohort study (Devanand et al., 2010) initially enrolled 1,092 participants, 571 of whom had undergone hippocampal volume measurement with olfactory data. Description of Included Studies Our search strategy initially identiﬁed 627 citations (Figure 1). After removing 47 duplicates, 575 studies were excluded by viewing the abstract for the animal model (n = 218) or non- relevance (n = 351). Eleven papers met the inclusion criteria (Murphy et al., 2003; Devanand et al., 2008, 2010; Wang et al., 2010; Lojkowska et al., 2011; Kjelvik et al., 2014; Marigliano et al., 2014; Vasavada et al., 2015; Hagemeier et al., 2016; Wu et al., 2019; Yu et al., 2019), among which four studies were excluded by screening the full article for speciﬁc reasons: the correlation in one study (Devanand et al., 2008) cannot be calculated or extracted through proper methods due to incomplete records; another (Kjelvik et al., 2014) presented a coeﬃcient in a linear regression model; and two studies demonstrated the hippocampal volume either in an fMRI activated form (Wang et al., 2010) or volume changes in a 24-month follow-up study (Lojkowska et al., 2011). FIGURE 1 \| Flow chart of study selection. Search Strategy Our meta-analysis was prepared according to the PRISMA guidelines and checklist (http://www.prisma-statement. org/PRISMAStatement/Checklist) and was registered with insplay.com. (Systematic Registration Number: October 2021 \| Volume 13 \| Article 755160 Frontiers in Aging Neuroscience \| www.frontiersin.org 2 Olfactory Function and Hippocampal Volume Su et al. Su et al. In each study, the eﬀect size was transformed into z through the equation z’ = 0.5 [ln (1 + r) – ln (1 – r)]. Then, the syntheses of z were performed in the meta-analysis. FIGURE 1 \| Flow chart of study selection. Statistical Analysis Meta-analysis was conducted in R language with “meta” package in R-studio Version 1.3.959 (https://rstudio.com/), where random and ﬁxed eﬀect models were applied according to the heterogeneity test. The I2 statistic was calculated to assess the heterogeneity between studies. We attempted to ﬁt a ﬁxed eﬀect model when the I2-value is <50%. An I2-value >50% or p-value < 0.05 suggests a rather heavy inconsistency and high heterogeneity, so we chose a sensitivity and subgroup analysis to render it and further discuss the potential sources. Subgroups were divided into the following categories: (1) participants, patients/normal; (2) sides, left/right/both; and (3) age groups with a diﬀerence of 5 years. Frontiers in Aging Neuroscience \| www.frontiersin.org Moderator Effects To investigate potential sources of heterogeneity, we performed a subgroup analysis with several moderator variables, including patient type, age, hemisphere, and olfactory tests. The following results revealed that patient type and age might be the two possible sources of heterogeneity. All seven studies yielded 22 eﬀect sizes and 902 participants. The participants were clinically diagnosed with MCI/AD or normal controls. In four studies (Murphy et al., 2003; Hagemeier et al., 2016; Wu et al., 2019; Yu et al., 2019), the correlation coeﬃcients were computed bilaterally according to hippocampal volume measurements on each side. In the three remaining studies (Devanand et al., 2010; Marigliano et al., 2014; Vasavada et al., 2015), r was calculated from the double-sided volume in total. A signiﬁcant diﬀerence in the correlation between the AD, MCI, and NC groups was discovered. The moderator analysis for patient type was signiﬁcant (Q = 17.64; p = 0.0014), suggesting that this variable may contribute to heterogeneity. Subgroups of AD (r = 0.3959, 95% CI: 0.2605–0.5160, k = 6), MCI (r = 0.3691, Frontiers in Aging Neuroscience \| www.frontiersin.org October 2021 \| Volume 13 \| Article 755160 3 Olfactory Function and Hippocampal Volume Su et al. Su et al. TABLE 1 \| Demographic data and relevant parameters. Moderator Effects M/F, Male/Female; MMSE, Mini-Mental State Examination; UPSIT, University of Pennsylvania Smell Identiﬁcation Test; CSIT, Chinese sm Test; SDOIT, San Diego Odor Identiﬁcation Test; MCI, mild cognitive impairment; NC, normal control; AD, Alzheimer’s disease. 95% CI: 0.1841–0.5288, k = 5), and NC (r = 0.1305, 95% CI: −0.0447–0.2980, k = 6) were not signiﬁcant in heterogeneity (AD: I2 = 27%, p = 0.12; MCI: I2 = 36%, p = 0.18; I2 = 0%, p = 0.53). The diﬀerences were not signiﬁcant between AD and non-AD (MCI + NC) (AD: r = 0.4222, 95% CI: 0.2372–0.5776, k = 6; non-AD: r = 0.2728, 95% CI: 0.1494–0.3879, k = 14; p = 0.1735), AD and MCI (p = 0.8072), but signiﬁcant in AD and NC (p = 0.0154) and AD. The correlation was signiﬁcantly stronger in the patient group than in the control group (p = 0.0121) and in the AD group than in the MCI group, indicating a pathology-dependent penetrance (Figure 4). age of all the participants was 75.20 years (range from 66.86 to 80.6 years). For a mean age diﬀerence of 5 years, the moderator analysis was statistically signiﬁcant (Q = 17.14, p = 0.0002). The moderator analysis for hemisphere was not signiﬁcant (Q = 5.02, p = 0.0811), suggesting that lateralization of odor memory might not contribute to the observed heterogeneity. Moreover, no obvious hemispheric dominance was found in olfaction (left: r = 0.35, 95% CI: 0.2318–0.4615, I2 = 53%, p = 0.03; right: r = 0.31, 95% CI: 0.1905–0.4268, I2 = 27%, p = 0.20). We further investigated the lateralization among patient groups and subgroup eﬀects in the left hippocampus group. The hemispheric parameters in patients were not signiﬁcant. The olfactory deﬁcits were found to be most correlated in the age range of 70.6–75.6 years old (r = 0.5113, 95% CI: 0.3181– 0.6637, k = 7) showing a low risk of heterogeneity (I2 = 46%, p = 0.08), and more predominantly than the 65.6–70.6 years group (r = 0.2698, 95%CI: 0.1376–0.3926, k = 11) and the 75.6–80.6 years group (r = 0.2591, 95% CI: 0.0809–0.4211, k = 3). The mean In all seven studies, odor identiﬁcation scores were obtained using various methods: the University of Pennsylvania Smell Identiﬁcation Test (UPSIT) in four studies (Vasavada et al., 2015; Hagemeier et al., 2016; Marin et al., 2018; Yu et al., 2019), the Chinese smell identiﬁcation test (CIST) in Wu et al. Moderator Effects References Subject N Age Sex (M/F) Olfactory test MMSE Correlation r Side(s) Devanand et al., 2010 MCI 571 – – UPSIT – MCI + NC 0.16 Double NC Hagemeier et al., 2016 aMCI 19 73.6 ± 11 9/10 22.9 ± 8.6 – AD 0.394 Right AD 42 76 ± 9 18/24 21.1 ± 7.9 aMCI 0.675 NC 19 69.4 ± 2.9 6/13 30.0 ± 6.7 NC −0.185 UPSIT AD 0.364 Left aMCI 0.438 NC −0.132 Vasavada et al., 2015 MCI 21 73.2 ± 9.0 10/11 24.2 ± 8.6 26.5 ± 1.9 AD + MCI + N 0.55 Double AD 16 71.9 ± 11.9 5/10 15.5 ± 8.4 18.9 ± 5.4 AD + MCI 0.33 NC 27 69.5 ± 10.4 12/15 34.0 ± 4.2 28.5 ± 1.5 UPSIT Wu et al., 2019 MCI 27 68.04 ± 7.58 13/14 CSIT 26 (25, 28) AD 0.242 Right AD 37 66.86 ± 10.27 17/20 16.03 ± 4.04 MCI 0.231 NC 30 67.23 ± 6.71 11/19 29 (28, 30) NC 0.167 TOTAL 0.512 AD 0.323 Left MCI 0.088 NC 0.326 TOTAL 0.512 Yu et al., 2019 MCI 31 65.9 ± 7.9 14/17 UPSIT – MCI + NC 0.42 Right NC 9 66.44 ± 7.05 3/6 MCI + NC 0.55 Left Marigliano et al., 2014 aMCI 18 68.05 ± 3.5 9/9 SSET – 0.508 Double Murphy et al., 2003 AD 13 73.08 ± 2.19 8/5 SDOIT 22.85 ± 1.04 AD 0.54 Right NC 22 72.45 ± 1.78 10/12 29.68 ± 0.12 NC 0.23 AD 0.85 Left NC 0.17 M/F, Male/Female; MMSE, Mini-Mental State Examination; UPSIT, University of Pennsylvania Smell Identiﬁcation Test; CSIT, Chinese smell identiﬁcation test; SSET, Snifﬁn Sticks Extended Test; SDOIT, San Diego Odor Identiﬁcation Test; MCI, mild cognitive impairment; NC, normal control; AD, Alzheimer’s disease. TABLE 1 \| Demographic data and relevant parameters. M/F, Male/Female; MMSE, Mini-Mental State Examination; UPSIT, University of Pennsylvania Smell Identiﬁcation Test; CSIT, Chinese smell identiﬁcation test; SSET, Snifﬁn Sticks Extended Test; SDOIT, San Diego Odor Identiﬁcation Test; MCI, mild cognitive impairment; NC, normal control; AD, Alzheimer’s disease. M/F, Male/Female; MMSE, Mini-Mental State Examination; UPSIT, University of Pennsylvania Smell Identiﬁcation Test; CSIT, Chinese smell identiﬁcation test; SSET, Snifﬁn Sticks Extended Test; SDOIT, San Diego Odor Identiﬁcation Test; MCI, mild cognitive impairment; NC, normal control; AD, Alzheimer’s disease. Moderator Effects (2019), Frontiers in Aging Neuroscience \| www.frontiersin.org October 2021 \| Volume 13 \| Article 755160 Frontiers in Aging Neuroscience \| www.frontiersin.org 4 Sco 7 7 Comparability Exposure n of ls Deﬁnition of controls Control for important factor Ascertainment of exposure Same method of ascertainment for cases and controls Non- response rate – – – – – Comparability Outcome nment sure Demonstration that outcome of interest was not present at start of study Comparability of cohorts on the basis of the design or analysis Assessment of outcome Enough follow-up of cohorts Adequacy of follow-up of cohorts – – d exposure categories. A maximum of two stars can be given for comparability. Olfactory Function and Hippocampal Volume Su et al. Su et al. FIGURE 2 \| Forest plot summarizing the overall correlation between odor identiﬁcation score and hippocampal volume across all studies and their 95% interval for each study. (Random effects model selected. NC, normal control; AD, Alzheimer’s disease; MCI, mild cognitive impairment; D, hippocampal volume measurement in double sides; R, hippocampal volume measurement in right side; L, hippocampal volume measurement in left side). FIGURE 2 \| Forest plot summarizing the overall correlation between odor identiﬁcation score and hippocampal volume across all studies and their 95% interval for each study. (Random effects model selected. NC, normal control; AD, Alzheimer’s disease; MCI, mild cognitive impairment; D, hippocampal volume measurement in double sides; R, hippocampal volume measurement in right side; L, hippocampal volume measurement in left side). could considerably aﬀect the precision of the statistical results. The moderator analysis showing patient types and age were the two main variables that might be most likely to account for heterogeneity. In addition, half of the sample size was due to one study alone whose r-value was nearly negligible (r = 0.157), but stronger relationships tended to be observed in smaller samples. Although no outliers were identiﬁed, the study of Devanand et al. (2010) has inﬂuenced the overall eﬀect size to a greater extent for those with a heavier weight. Typically, sample sizes are reciprocal to the precision of the estimated eﬀects (Sedgwick and Marston, 2015), and studies with larger sample sizes are given for more weight in analysis. Therefore, sample size is considered to aﬀect heterogeneity, and thus studies with larger sample sizes are necessary for further validation. Additional unpublished papers and non-English results should also be involved to further reduce heterogeneity. Moderator Effects the Sniﬃn Sticks Extended Test (SSET) in Marigliano et al. (2014), and the San Diego Odor Identiﬁcation Test (SDOIT) in Murphy et al. (2003). The subgroup analysis revealed that the diﬀerence between the types of olfactory identiﬁcation tests was not signiﬁcant (Q = 3, p = 0.3916). Given the lack of demographic ﬁgures for gender information, the pooled r-value categorized by sex was unable to be detected. The subgroup analysis revealed that part of the heterogeneity was due to subject type and age. Frontiers in Aging Neuroscience \| www.frontiersin.org DISCUSSION Our meta-analysis explored the relationship between odor identiﬁcation decline and hippocampal atrophy in AD and MCI patients with normal controls. The main result obtained from our meta-analysis showed a signiﬁcant positive correlation (r = 0.3392, 95% CI: 0.2335–0.4370, p < 0.0001) between olfactory identiﬁcation deﬁcits and hippocampal atrophy. A prominent diﬀerence was noted in the MCI/AD group, with a stronger correlation than the control group (p = 0.0121). In addition, the association in the AD group was stronger than in the MCI group, suggesting that odor identiﬁcation decline could be detected early in the MCI stage and followed the disease progression. Patient type was an independent factor in OD. Olfactory identiﬁcation deﬁcits were more prominently correlated with hippocampal atrophy in the AD group than in the MCI group, both of which were consistently stronger than in the normal control group. Previous meta-analyses have validated similar results. Rahayel et al. (2012) conducted a meta-analysis and conﬁrmed that AD has severe detrimental eﬀects on olfactory function across the whole spectrum, but has a stronger eﬀect on odor identiﬁcation than odor detection. Olfactory identiﬁcation was the most impaired among all domains in MCI (Roalf et al., 2017) and AD patients. Kotecha et al. Moderate heterogeneity was detected, suggesting that the overall combination of associations might not be present across all contexts. This may be due to clinical heterogeneity in the variation in participants, and the diversity of participant numbers October 2021 \| Volume 13 \| Article 755160 6 Olfactory Function and Hippocampal Volume Su et al. Su et al. FIGURE 3 \| Sensitive analysis in leave-one-out method. FIGURE 3 \| Sensitive analysis in leave-one-out method. FIGURE 3 \| Sensitive analysis in leave-one-out method. of olfactory-related regions in the elderly (Suzuki et al., 2001; Ferdon and Murphy, 2003). This was in line with a longitudinal study showing an inverse correlation of B-SIT scores before death and post-mortem density of neuroﬁbrillary tangles in the entorhinal cortex, the CA1 subﬁeld of the hippocampus. Our pooled correlation in age was predominant in patients between the ages of 70–75, showing a moderate association (r = 0.5113, 95% CI: 0.3181–0.6637). This result did not explain the progressive trend in olfactory impairment. Thus, we speculate that this is due to the discontinuity of the wide age interval. Frontiers in Aging Neuroscience \| www.frontiersin.org Frontiers in Aging Neuroscience \| www.frontiersin.org DISCUSSION We re-analyzed a 2-year interval in patient and control groups separately, and discovered that the growth of correlation increases with age (66–68: r = 0.2953, 95% CI: 0.1030–0.4664; 70–72: r = 0.2521, 95% CI: 0.0060–0.4694; 72– 74: r = 0.4554, 95% CI: 0.2434–0.6259; 74–76: r = 0.4679, 95% CI: 0.2999–0.6078; Q = 15.18, p = 0.2317). This indicates that aging could be an independent factor for odor identiﬁcation deﬁcits when the magnitude of the disease was ruled out. Thus, we inferred that age-dependent hippocampal volume decrement clouds aﬀect olfactory function physiologically; on the other hand, this physiological function could be worsened under the pathological extension from MCI to AD. (2018) systematically reviewed and concluded that olfaction progressively worsens from MCI to AD, which highlights the potential utility of olfactory identiﬁcation tests as prognostic tools for AD (Sun et al., 2012). Jung et al. (2019) reported similar results, revealing that olfactory identiﬁcation was more profoundly impaired in AD than in MCI; further, Roalf et al. (2017) concluded a more extensively impaired odor identiﬁcation in MCI. The former result is compatible with our ﬁnding that the relationship in AD is higher than in MCI groups (MCI: r = 0.3691; AD: r = 0.3959; p = 0.081). This clear increase in odor identiﬁcation deﬁcits from cognitively normal to MCI and AD has been described in both clinical and epidemiological studies (Graves et al., 1999; Schubert et al., 2008; Devanand et al., 2015). In addition, this increase in correlation with disease progression might indicate that the olfactory cortex (hippocampus as the second olfactory cortex) is compromised through the pathophysiological continuum (Bathini et al., 2019) of sequential events of the pathology of the disease. It is widely accepted that odor identiﬁcation generally declines with normal aging, especially over age 70 (Doty et al., 1984). Signiﬁcant age-related alterations have been observed in odor identiﬁcation tests in various studies. In functional magnetic resonance imaging (fMRI), there is a decrease in the activation Previous studies have suggested that odor memory is lateralized to the right hemisphere (Jones-Gotman and Zatorre, October 2021 \| Volume 13 \| Article 755160 7 Olfactory Function and Hippocampal Volume Su et al. Su et al. FIGURE 4 \| Subgroup analysis in different subject in group of AD MCI NC FIGURE 4 \| Subgroup analysis in different subject in group of AD, MCI, NC. DISCUSSION FIGURE 4 \| Subgroup analysis in different subject in group of AD, MCI, NC. our heterogeneity in the hemispheric moderator on the left side signiﬁcant. We would assume that the current, small numbered, and conﬂicting results require further observation. 1993; Olsson and Cain, 2003). The right hippocampus was found to be larger in the NC and MCI groups, while there was no signiﬁcant diﬀerence in AD in Wolf et al.’s (2001) study. Zou et al. (2016a) concluded that the right hemisphere is predominant in odor hedonic judgment. In contrast, fMRI brain scans of brain activation are generally lateralized to the left hemisphere when received pleasant smell of odors, and unpleasant smells to the right (Henkin and Levy, 2001). However, the controversial hemispheric prominence generally did not include the hippocampus. Our analysis indicated that there were no signiﬁcant hemispheric diﬀerences. One study (Murphy et al., 2003) reported a stronger correlation in the left hippocampus over the right (r = 0.85, p < 0.001), which made It can be aﬃrmed that our results in brain-behavior relationships are congruent with previous meta-analyses that have validated olfactory dysfunction in AD. However, the correlation between hippocampal atrophy and odor identiﬁcation deﬁcits is by far the ﬁrst to be explored, which could be a key explanation for the hypothesis that it is generated from the pathology burden in the medial-temporal lobe. Consequently, olfactory deﬁcits originate in central structures, suggesting that odor identiﬁcation and recognition tests could be beneﬁcial for the early detection of subclinical cases. October 2021 \| Volume 13 \| Article 755160 Frontiers in Aging Neuroscience \| www.frontiersin.org 8 Olfactory Function and Hippocampal Volume Su et al. Su et al. Several clinical studies have observed that OD and cognitive impairment share the same anatomical modiﬁcations of AD- signature cortex decrease (Lian et al., 2019), especially the olfactory cortex and the hippocampus (Al-Otaibi et al., 2020). In recent years, a link between olfactory deﬁcits and AD has been consistently reported. It is commonly recognized that prior to cognitive symptoms (Price et al., 1991; Jellinger and Attems, 2005; Attems and Jellinger, 2006), AD pathology appears in the trans-entorhinal region, entorhinal cortex, hippocampus and successively in olfactory bulb (OB), olfactory tract, and other structures (Ohm and Braak, 1988; Kovács et al., 1999). However, the mechanisms underlying the relationship between odor identiﬁcation (OI) and hippocampal pathology have not been fully elucidated. SUPPLEMENTARY MATERIAL The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fnagi. 2021.755160/full#supplementary-material CONCLUSION This meta-analysis quantiﬁed a positive correlation between olfactory identiﬁcation deﬁcits and hippocampal atrophy. The correlation appears to be more predominant in MCI and AD patients, suggesting that olfactory identiﬁcation deﬁcits appear in the early stages of the continuum. Age is an independent factor that aﬀects the severity of the correlation during disease progression. The mildness of correlation suggests that olfactory tests may be more accurate in early detection when combined with other non-invasive examinations in AD. This meta-analysis quantiﬁed a positive correlation between olfactory identiﬁcation deﬁcits and hippocampal atrophy. The correlation appears to be more predominant in MCI and AD patients, suggesting that olfactory identiﬁcation deﬁcits appear in the early stages of the continuum. Age is an independent factor that aﬀects the severity of the correlation during disease progression. The mildness of correlation suggests that olfactory tests may be more accurate in early detection when combined with other non-invasive examinations in AD. DATA AVAILABILITY STATEMENT The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors. ACKNOWLEDGMENTS Our study has several limitations. First, there is inadequate inclusion of studies aiming at olfactory discrimination and detection threshold, along with studies reporting a correlation between OB and olfactory epithelium deﬁcits and hippocampal atrophy. Odor discrimination and detection thresholds (Mesholam et al., 1998) were not adequately covered in our analysis. Second, according to the subgroup analysis, we could conﬁrm that aging is one of the moderator factors; however, the linear regression could not be drawn from the present We thank all the coauthors for their contribution for this study. We thank all the coauthors for their contribution for this study. FUNDING This work was supported by the National Natural Science Foundation of China (Grant No. 82074362). AUTHOR CONTRIBUTIONS M-WS and J-NN wrote the manuscript. M-WS and T-YC performed the analysis. S-SW and M-WS helped to proofread the literature search. JS and J-ZT supervised the study. All authors contributed to the article and approved the submitted version. M-WS and J-NN wrote the manuscript. M-WS and T-YC performed the analysis. S-SW and M-WS helped to proofread the literature search. JS and J-ZT supervised the study. All authors contributed to the article and approved the submitted version. DISCUSSION Evidence suggests that neuroinﬂammation occurs in Aβ burden structures (Hanzel et al., 2014). A decrease in hippocampal volume is associated with hippocampal-dependent dysfunction in learning and memory (Ziehn et al., 2010), which also correlates with microglial activation, synaptopathy/synaptic loss, and neurodegeneration (Mandolesi et al., 2010; Girard et al., 2014). Soluble Aβ accumulation in the OB is strongly correlated with early olfactory dysfunction in both AD patients and mouse models (Wesson et al., 2010). Further, a recently published meta-analysis by Tu et al. (2020) discovered a weak negative correlation between OI ability and cerebral Aβ PET (r = −0.25, P = 0.008) and CSF tau (r = −0.17, p = 0.006) levels. The speciﬁcity was speculated to be the marginal burden of pathological changes that implicate OI ability. The review concluded that the combination of OI tests and other biologic markers still preserves the predictive value of assessing cognitive decline and progression from MCI to AD. However, this may conversely explain the hypothesis that soluble toxic aggregates of both Aβ and tau can self-propagate and spread throughout the brain by prion-like mechanisms (Goedert et al., 2010; Bloom, 2014), and propagation of proteotoxicity along the olfactory nerve could likely aﬀect olfactory-ERC-hippocampal circuits (Busche et al., 2008; Rey et al., 2018). Oligomeropathy (Forloni and Balducci, 2018), neuroinﬂammation, and the prion-like hypothesis may trigger olfactory dysfunction. discontinuous data. Furthermore, heterogeneity in sample size preserves obvious diﬀerences in the statistical results, which could aﬀect precision. Thus, meticulously designed studies with larger sample sizes are necessary for validation. Frontiers in Aging Neuroscience \| www.frontiersin.org Attems, J., and Jellinger, K. A. (2006). Olfactory tau pathology in Alzheimer disease and mild cognitive impairment. Clin. Neuropathol. 25, 265–271. REFERENCES Al-Otaibi, M., Lessard-Beaudoin, M., Castellano, C. A., Gris, D., Cunnane, S. C., and Graham, R. K. (2020). Volumetric MRI demonstrates atrophy of the olfactory cortex in AD. Curr. Alzheimer Res. 17, 904–915. doi: 10.2174/1567205017666201215120909 Albert, M. S., DeKosky, S. T., Dickson, D., Dubois, B., Feldman, H. H., Fox, N. C., et al. (2011). The diagnosis of mild cognitive impairment due to Alzheimer’s disease: recommendations from the National Institute on Aging- Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement. 7, 270–279. doi: 10.1016/j.jalz.2011.03.008 Albert, M. S., DeKosky, S. T., Dickson, D., Dubois, B., Feldman, H. H., Fox, N. C., et al. (2011). The diagnosis of mild cognitive impairment due to Alzheimer’s disease: recommendations from the National Institute on Aging- Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement. 7, 270–279. doi: 10.1016/j.jalz.2011.03.008 Attems, J., and Jellinger, K. A. (2006). Olfactory tau pathology in Alzheimer disease and mild cognitive impairment. Clin. Neuropathol. 25, 265–271. October 2021 \| Volume 13 \| Article 755160 Frontiers in Aging Neuroscience \| www.frontiersin.org 9 Olfactory Function and Hippocampal Volume Su et al. Bathini, P., Brai, E., and Auber, L. A. (2019). Olfactory dysfunction in the pathophysiological continuum of dementia. Ageing Res. Rev. 55:100956. doi: 10.1016/j.arr.2019.100956 Jones-Gotman, M., and Zatorre, R. J. (1993). Odor recognition memory in humans: role of right temporal and orbitofrontal regions. Brain Cogn. 22, 182–198. doi: 10.1006/brcg.1993.1033 Berry, K. J., and Mielke, P. W. Jr. (2000). A Monte Carlo investigation of the Fisher Z transformation for normal and nonnormal distributions. Psychol. Rep. 87, 1101–1114. doi: 10.2466/pr0.2000.87.3f.1101 Jung, H. J., Shin, I. S., and Lee, J. E. (2019). Olfactory function in mild cognitive impairment and Alzheimer’s disease: a meta-analysis. Laryngoscope 129, 362–369. doi: 10.1002/lary.27399 Bloom, G. S. (2014). Amyloid-β and tau: the trigger and bullet in Alzheimer disease pathogenesis. JAMA Neurol. 71, 505–508. doi: 10.1001/jamaneurol.201 3.5847 Karas, G. B., Burton, E. J., Rombouts, S. A., van Schijndel, R. A., O’Brien, J. T., Scheltens, P.h, et al. (2003). A comprehensive study of gray matter loss in patients with Alzheimer’s disease using optimized voxel-based morphometry. Neuroimage 18, 895–907. doi: 10.1016/S1053-8119(03)00041-7 Neuroimage 18, 895–907. doi: 10.1016/S1053-8119(03)00041-7 Busche, M. A., Eichhoﬀ, G., Adelsberger, H., Abramowski, D., Wiederhold, K. H., Haass, C., et al. (2008). Clusters of hyperactive neurons near amyloid plaques in a mouse model of Alzheimer’s disease. Science 321, 1686–1689. doi: 10.1126/science.1162844 Kjelvik, G., Saltvedt, I., White, L. REFERENCES R., Stenumgård, P., Sletvold, O., Engedal, K., et al. (2014). The brain structural and cognitive basis of odor identiﬁcation deﬁcits in mild cognitive impairment and Alzheimer’s disease. BMC Neurol. 14, 168. doi: 10.1186/s12883-014-0168-1 Devanand, D. P., Lee, S., Manly, J., Andrews, H., Schupf, N., Doty, R. L., et al. (2015). Olfactory deﬁcits predict cognitive decline and Alzheimer dementia in an urban community. Neurology 84, 182–189. doi: 10.1212/WNL.0000000000001132 Kotecha, A. M., Corrêa, A. D. C., Fisher, K. M., and Rushworth, J. V. (2018). Olfactory dysfunction as a global biomarker for sniﬃng out Alzheimer’s disease: a meta-analysis. Biosensors (Basel) 8:41. doi: 10.3390/bios8020041 Kovács, T., Cairns, N. J., and Lantos, P. L. (1999). Beta-amyloid deposition and neuroﬁbrillary tangle formation in the olfactory bulb in ageing and Alzheimer’s disease. Neuropathol. Appl. Neurobiol. 25, 481–491. doi: 10.1046/j.1365-2990.1999.00208.x Devanand, D. P., Liu, X., Tabert, M. H., Pradhaban, G., Cuasay, K., Bell, K., et al. (2008). Combining early markers strongly predicts conversion from mild cognitive impairment to Alzheimer’s disease. Biol. Psychiatry 64, 871–879. doi: 10.1016/j.biopsych.2008.06.020 Lane, C. A., Hardy, J., and Schott, J. M. (2018). Alzheimer’s disease. Eur. J. Neurol. 25, 59–70. doi: 10.1111/ene.13439 Devanand, D. P., Tabert, M. H., Cuasay, K., Manly, J. J., Schupf, N., Brickman, A. M., et al. (2010). Olfactory identiﬁcation deﬁcits and MCI in a multi-ethnic elderly community sample. Neurobiol. Aging 31, 1593–1600. doi: 10.1016/j.neurobiolaging.2008.09.008 Lian, T. H., Zhu, W. L., Li, S. W., Liu, Y. O., Guo, P., Zuo, L. J., et al. (2019). Clinical, structural, and neuropathological features of olfactory dysfunction in patients with Alzheimer’s disease. J. Alzheimers Dis. 70, 413–423. doi: 10.3233/JAD-181217 Doty, R. L., Shaman, P., Applebaum, S. L., Giberson, R., Siksorski, L., and Rosenberg, L. (1984). Smell identiﬁcation ability: changes with age. Science 226, 1441–1443. doi: 10.1126/science.6505700 Lojkowska, W., Sawicka, B., Gugala, M., Sienkiewicz-Jarosz, H., Bochynska, A., Scinska, A., et al. (2011). Follow-up study of olfactory deﬁcits, cognitive functions, and volume loss of medial temporal lobe structures in patients with mild cognitive impairment. Curr. Alzheimer Res. 8, 689–698. doi: 10.2174/156720511796717212 Ferdon, S., and Murphy, C. (2003). The cerebellum and olfaction in the aging brain: a functional magnetic resonance imaging study. Neuroimage 20, 12–21. doi: 10.1016/S1053-8119(03)00276-3 Forloni, G., and Balducci, C. (2018). Alzheimer’s disease, oligomers, and inﬂammation. J. Alzheimers Dis. 62, 1261–1276. doi: 10.3233/JAD-170819 Mandolesi, G., Grasselli, G., Musumeci, G., and Centonze, D. (2010). Cognitive deﬁcits in experimental autoimmune encephalomyelitis: neuroinﬂammation and synaptic degeneration. Neurol. Sci. 31, S255–S259. REFERENCES doi: 10.1007/s10072-010-0369-3 Girard, S. D., Jacquet, M., Baranger, K., Migliorati, M., Escoﬃer, G., Bernard, A., et al. (2014). Onset of hippocampus-dependent memory impairments in 5XFAD transgenic mouse model of Alzheimer’s disease. Hippocampus 24, 762–772. doi: 10.1002/hipo.22267 Marigliano, V., Gualdi, G., Servello, A., Marigliano, B., Volpe, L. D., Fioretti, A., et al. (2014). Olfactory deﬁcit and hippocampal volume loss for early diagnosis of Alzheimer disease: a pilot study. Alzheimer Dis. Assoc. Disord. 28, 194–197. doi: 10.1097/WAD.0b013e31827bdb9f Goedert, M., Clavaguera, F., and Tolnay, M. (2010). The propagation of prion-like protein inclusions in neurodegenerative diseases. Trends Neurosci. 33, 317–325. doi: 10.1016/j.tins.2010.04.003 Graves, A. B., Bowen, J. D., Rajaram, L., McCormick, W. C., McCurry, S. M., Schellenberg, G. D., et al. (1999). Impaired olfaction as a marker for cognitive decline: interaction with apolipoprotein E epsilon4 status. Neurology 53, 1480–1487. doi: 10.1212/WNL.53.7.1480 Marin, C., Vilas, D., Langdon, C., Alobid, I., López-Chacón, M., Haehner, A., et al. (2018). Olfactory dysfunction in neurodegenerative diseases. Curr. Allergy Asthma Rep. 18:42. doi: 10.1007/s11882-018-0796-4 Mesholam, R. I., Moberg, P. J., Mahr, R. N., and Doty, R. L. (1998). Olfaction in neurodegenerative disease: a meta-analysis of olfactory functioning in Alzheimer’s and Parkinson’s diseases. Arch. Neurol. 55:8490. doi: 10.1001/archneur.55.1.84 Hagemeier, J., Woodward, M. R., Raﬁque, U. A., Amrutkar, C. V., Bergsland, N., Dwyer, M. G., et al. (2016). Odor identiﬁcation deﬁcit in mild cognitive impairment and Alzheimer’s disease is associated with hippocampal and deep gray matter atrophy. Psychiatry Res. Neuroimag. 255, 87–93. doi: 10.1016/j.pscychresns.2016.08.003 Murphy, C., Jernigan, T. L., and Fennema-Notestine, C. (2003). Left hippocampal volume loss in Alzheimer’s disease is reﬂected in performance on odor identiﬁcation: a structural MRI study. J. Int. Neuropsychol. Soc. 9, 459–471. doi: 10.1017/S1355617703930116 Hanzel, C. E., Pichet-Binette, A., Pimentel, L. S., Iulita, M. F., Allard, S., Ducatenzeiler, A., et al. (2014). Neuronal driven pre-plaque inﬂammation in a transgenic rat model of Alzheimer’s disease. Neurobiol. Aging 35, 2249–2262. doi: 10.1016/j.neurobiolaging.2014.03.026 Ohm, T. G., and Braak, H. (1988). The pigmented subpeduncular nucleus: a neuromelanin-containing nucleus in the human pontine tegmentum. Morphology and changes in Alzheimer’s disease. Acta Neuropathol. 77, 26–32. doi: 10.1007/BF00688239 Hatashita, S., and Yamasaki, H. (2013). Diagnosed mild cognitive impairment due to Alzheimer’s disease with PET biomarkers of beta amyloid and neuronal dysfunction. PLoS ONE 8:e66877. doi: 10.1371/journal.pone.0066877 Olsson, M. J., and Cain, W. S. (2003). Implicit and explicit memory for odors: hemispheric diﬀerences. Mem. Cognit. 31, 44–50. doi: 10.3758/BF031 96081 Hawkes, C. (2003). REFERENCES Olfaction in neurodegenerative disorder. Mov. Disord. 18, 364–372. doi: 10.1002/mds.10379 Price, J. L., Davis, P. B., Morris, J. C., and White, D. L. (1991). The distribution of tangles, plaques and related immunohistochemical markers in healthy aging and Alzheimer’s disease. Neurobiol. Aging 12, 295–312. doi: 10.1016/0197-4580(91)90006-6 Henkin, R. I., and Levy, L. M. (2001). Lateralization of brain activation to imagination and smell of odors using functional magnetic resonance imaging (fMRI): left hemispheric localization of pleasant and right hemispheric localization of unpleasant odors. J. Comput. Assist. Tomogr. 25, 493–514. doi: 10.1097/00004728-200107000-00001 Rahayel, S., Frasnelli, J., and Joubert, S. (2012). The eﬀect of Alzheimer’s disease and Parkinson’s disease on olfaction: a meta-analysis. Behav. Brain Res. 231, 60–74. doi: 10.1016/j.bbr.2012.02.047 Hyman, B. T., Arriagada, P. V., and Van Hoesen, G. W. (1991). Pathologic changes in the olfactory system in aging and Alzheimer’s disease. Ann. N. Y. Acad. Sci. 640, 14–19. doi: 10.1111/j.1749-6632.1991.tb00184.x Rey, N. L., Wesson, D. W., and Brundin, P. (2018). The olfactory bulb as the entry site for prion-like propagation in neurodegenerative diseases. Neurobiol. Dis. 109, 226–248. doi: 10.1016/j.nbd.2016.12.013 Jellinger, K. A., and Attems, J. (2005). Alzheimer pathology in the olfactory bulb. Neuropathol. Appl. Neurobiol. 31, 203. doi: 10.1111/j.1365-2990.2004.00619.x October 2021 \| Volume 13 \| Article 755160 Frontiers in Aging Neuroscience \| www.frontiersin.org 10 Olfactory Function and Hippocampal Volume Su et al. Wilson, R. S., Arnold, S. E., Schneider, J. A., Tang, Y., and Bennett, D. A. (2007). The relationship between cerebral Alzheimer’s disease pathology and odour identiﬁcation in old age. J. Neurol. Neurosurg. Psychiatry 78, 30–35. doi: 10.1136/jnnp.2006.099721 Roalf, D. R., Moberg, M. J., Turetsky, B. I., Brennan, L., Kabadi, S., Wolk, D. A., et al. (2017). A quantitative meta-analysis of olfactory dysfunction in mild cognitive impairment. J. Neurol. Neurosurg. Psychiatry 88, 226–232. doi: 10.1136/jnnp-2016-314638 Roberts, R., and Knopman, D. S. (2013). Classiﬁcation and epidemiology of MCI. Clin. Geriatr. Med. 29, 753–772. doi: 10.1016/j.cger.2013. 07.003 Wolf, H., Grunwald, M., Kruggel, F., Riedel-Heller, S. G., Angerhöfer, S., Hojjatoleslami, A., et al. (2001). Hippocampal volume discriminates between normal cognition; questionable and mild dementia in the elderly. Neurobiol. Aging 22, 177–186. doi: 10.1016/S0197-4580(00)00238-4 Schubert, C. R., Carmichael, L. L., Murphy, C., Klein, B. E., Klein, R., and Cruickshanks, K. J. (2008). Olfaction and the 5-year incidence of cognitive impairment in an epidemiological study of older adults. J. Am. Geriatr. Soc. 56, 1517–1521. REFERENCES doi: 10.1111/j.1532-5415.2008.01826.x Wu, X., Geng, Z., Zhou, S., Bai, T., Wei, L., Ji, G. J., et al. (2019). Brain structural correlates of odor identiﬁcation in mild cognitive impairment and Alzheimer’s disease revealed by magnetic resonance imaging and a Chinese olfactory identiﬁcation test. Front. Neurosci. 13:842. doi: 10.3389/fnins.2019.00842 Sedgwick, P., and Marston, L. (2015). How to read a funnel plot in a meta-analysis. BMJ 351:h4718. doi: 10.1136/bmj.h4718 Yu, H. L., Chen, Z. J., Zhao, J. W., Duan, S. R., and Zhao, J. K. (2019). Olfactory impairment and hippocampal volume in a Chinese MCI clinical sample. Alzheimer Dis. Assoc. Disord. 33, 124–128. doi: 10.1097/WAD.0000000000000305 Servello, A., Fioretti, A., Gualdi, G., Di Biasi, C., Pittalis, A., Sollaku, S., et al. (2015). Olfactory dysfunction, olfactory bulb volume and Alzheimer’s disease: is there a correlation? A pilot Study1. J. Alzheimers Dis. 48, 395–402. doi: 10.3233/JAD-150232 Ziehn, M. O., Avedisian, A. A., Tiwari-Woodruﬀ, S., and Voskuhl, R. R. (2010). Hippocampal CA1 atrophy and synaptic loss during experimental autoimmune encephalomyelitis, EAE. Lab. Invest. 90, 774–786. doi: 10.1038/labinvest.2010.6 Sun, G. H., Raji, C. A., Maceachern, M. P., and Burke, J. F. (2012). Olfactory identiﬁcation testing as a predictor of the development of Alzheimer’s dementia: a systematic review. Laryngoscope 122, 1455–1462. doi: 10.1002/lary.23365 Zou, L. Q., van Hartevelt, T. J., Kringelbach, M. L., Cheung, E. F. C., and Chan, R. C. K. (2016a). The neural mechanism of hedonic processing and judgment of pleasant odors: an activation likelihood estimation meta-analysis. Neuropsychology 30, 970–979. doi: 10.1037/neu0000292 Suzuki, Y., Critchley, H. D., Suckling, J., Fukuda, R., Williams, S. C., Andrew, C., et al. (2001). Functional magnetic resonance imaging of odor identiﬁcation: the eﬀect of aging. J. Gerontol. A Biol. Sci. Med. Sci. 56, M756–M760. doi: 10.1093/gerona/56.12.M756 Neuropsychology 30, 970–979. doi: 10.1037/neu0000292 Zou, Y. M., Lu, D., Liu, L. P., Zhang, H. H., and Zhou, Y. Y. (2016b). Olfactory dysfunction in Alzheimer’s disease. Neuropsychiatr. Dis. Treat. 12, 869–875. doi: 10.2147/NDT.S104886 Tu, L., Lv, X., Fan, Z., Zhang, M., Wang, H., and Yu, X. (2020). Association of odor identiﬁcation ability with amyloid-β and tau burden: a systematic review and meta-analysis. Front. Neurosci. 14:586330. doi: 10.3389/fnins.2020.5 86330 Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Vasavada, M. M., Wang, J., Eslinger, P. J., Gill, D. J., Sun, X., Karunanayaka, P., et al. (2015). Frontiers in Aging Neuroscience \| www.frontiersin.org October 2021 \| Volume 13 \| Article 755160 REFERENCES Olfactory cortex degeneration in Alzheimer’s disease and mild cognitive impairment. J. Alzheimers Dis. 45, 947–958. doi: 10.3233/JAD-1 41947 Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their aﬃliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Wang, J., Eslinger, P. J., Doty, R. L., Zimmerman, E. K., Grunfeld, R., Sun, X., et al. (2010). Olfactory deﬁcit detected by fMRI in early Alzheimer’s disease. Brain Res. 1357, 184–194. doi: 10.1016/j.brainres.2010.08.018 Wells, G., Shea, B., O’Connell, D., Robertson, J., Peterson, J., Welch, V., et al. (2013). The Newcastle-Ottawa Scale (NOS) for Assessing the Quality of Nonrandomised Studies in Meta-analyses. Available online at: http:// www3.med.unipmn.it/dispense_ebm/2009-2010/Corso%20Perfezionamento %20EBM_Faggiano/NOS_oxford.pdf (accessed May 9, 2021). Copyright © 2021 Su, Ni, Cao, Wang, Shi and Tian. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Wesson, D. W., Levy, E., Nixon, R. A., and Wilson, D. A. (2010). Olfactory dysfunction correlates with amyloid-beta burden in an Alzheimer’s disease mouse model. J. Neurosci. 30, 505–514. doi: 10.1523/JNEUROSCI.4622-09.2010 October 2021 \| Volume 13 \| Article 755160 Frontiers in Aging Neuroscience \| www.frontiersin.org 11
https://openalex.org/W1562270672	https://inria.hal.science/hal-01525096/document	English	null	Mapping the Most Significant Computer Hacking Events to a Temporal Computer Attack Model	IFIP advances in information and communication technology	2,012	cc-by	4,044	To cite this version: Renier Van Heerden, Heloise Pieterse, Barry Irwin. Mapping the Most Significant Computer Hacking Events to a Temporal Computer Attack Model. 10th International Conference on Human Choice and Computers (HCC), Sep 2012, Amsterdam, Netherlands. pp.226-236, 10.1007/978-3-642-33332-3_21. hal-01525096 Mapping the Most Significant Computer Hacking Events to a Temporal Computer Attack Model Renier Van Heerden, Heloise Pieterse, Barry Irwin To cite this version: Renier Van Heerden, Heloise Pieterse, Barry Irwin. Mapping the Most Significant Computer Hacking Events to a Temporal Computer Attack Model. 10th International Conference on Human Choice and Computers (HCC), Sep 2012, Amsterdam, Netherlands. pp.226-236, 10.1007/978-3-642-33332-3_21. hal-01525096 Distributed under a Creative Commons Attribution 4.0 International License HAL Id: hal-01525096 https://inria.hal.science/hal-01525096v1 Submitted on 19 May 2017 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License Mapping the Most Significant Computer Hacking Events to a Temporal Computer Attack Model Renier van Heerden1,2, Heloise Pieterse1 and Barry Irwin2 1Council for Scientific and Industrial Research, Pretoria, South Africa rvheerden@csir.co.za, hpieterse@csir.co.za 2Rhodes University, Grahamstown, South Africa b.irwin@ru.ac.za Abstract. This paper presents eight of the most significant computer hacking events (also known as computer attacks). These events were selected because of their unique impact, methodology, or other properties. A temporal computer at- tack model is presented that can be used to model computer based attacks. This model consists of the following stages: Target Identification, Reconnaissance, Attack, and Post-Attack Reconnaissance stages. The Attack stage is separated into: Ramp-up, Damage and Residue. This paper demonstrates how our eight significant hacking events are mapped to the temporal computer attack model. The temporal computer attack model becomes a valuable asset in the protection of critical infrastructure by being able to detect similar attacks earlier. Keywords: computer attack model, ontology, network attack prediction 1 adfa, p. 1, 2011. © Springer-Verlag Berlin Heidelberg 2011 2.1 Computer Attack Taxonomy and Ontology A detailed taxonomy that describes computer based attacks has the following classes [4]: actor; actor location; aggressor; attack goal; attack mechanism; automation level; effects; motivation; phase; scope; target; and vulnerability. The taxonomy was then used to describe the following scenarios [4]: denial of service (DoS); industrial espio- nage; web deface; spear phishing; password harvesting; snooping for secrets; finan- cial theft; amassing computer resources; industrial sabotage; and cyber warfare. 1 Introduction Computer hacking (also referred as computer cracking) developed in conjunction with the normal usage of computer systems. This paper discusses some of the most signifi- cant hacking events and the features that made them unique. The events listed are considered to be significant because of their unique impact, methodology or other properties. The level of significance is an abstract and relative measure. Other at- tempts to judge the importance of hacking events have been made by Heater [1], Hall [2] and Julian [3]. Research in computer network attack prediction at the Counsel for Scientific and Industrial Research (CSIR) in South Africa has resulted in the development of a Tax- onomy and Ontology of computer network attacks. A temporal attack model was developed with the goal of separating the different stages of a computer network at- tack. The model consists of the following basic stages: Target Identification; Recon- naissance; Attack; and Post Attack. The Attack stage has the following sub-stages: Ramp-up; Damage; and Residue. Research was also organized into strategies for identifying the Reconnaissance and Ramp-up stages. The attack model is a valuable asset in the protection of critical infrastructure as it has the ability to identify attacks at an earlier stage and so improve the responsiveness to incidents. This paper presents the authors’ view on the most important hacking events, and cannot in itself be considered absolute. We chose events based on either the unique- ness of the technique used or their unique impact. The attack model is presented in more detail in Section 2. Section 3 describes the most significant hacking events and their characteristics. Section 4 identifies trends in hacking development. Section 5 maps the hacking events to our temporal attack model. Section 6 focuses on the pro- tection of critical infrastructure. Section 7 discusses mayor future hacking events. 2.2 Temporal Attack Model The Phase class in Section 2.1 was used to build the Temporal Attack Model. The Target Identification stage represents actions undertaken by an attacker in choosing his/her target. Identification of these actions falls outside the scope of the network attack prediction project, but forms part of the overall attack model. The Reconnais- sance stage represents actions undertaken by an attacker to identify potential weak spots. These actions are the earliest indicators of an impending network attack, and occur before any real damage has occurred. Popular reconnaissance actions include network mapping and scanning with tools such as Nmap and Nessus. Google and other search engines can also be used to identify potential weak spots. The Attack stage represents modification of the target system by an attacker. The system can be modified in the following aspects: Confidentiality; Integrity; and Availability. These aspects are also known as the CIA principles. Confidentiality refers to pre- vention of disclosure of information to unauthorized individuals or systems. Integrity means that data in a system cannot be modified undetectably. Availability refers to the availability of information when required by the system to serve its purpose. In computing, e-Business and information security, it is necessary to ensure that data, transactions, communications and documents are genuine. It is also important that authentication validates the identities of both parties involved. In figure 1 the Temporal Attack Model is represented. The Attack stage is subdi- vided into sub-stages. The first sub-stage is the Ramp-up stage. This sub-stage refers to the preparatory actions performed by an attacker before his/her final goal can be attained. The targeted computer network is modified in this stage, but only in prepara- tion for some other goal. This stage typically includes the installation of backdoors and other malware. Fig. 1. Temporal Network Attack Model. Fig. 1. Temporal Network Attack Model. Fig. 1. Temporal Network Attack Model. The Damage sub-stage refers to actions undertaken by an attacker during the achievement of his/her final goal. In this sub-stage the network is damaged according to the Information Security CIA principles. For example when an attacker launches a Distributed Denial of Service (DDoS) attack on a network, the Damage sub-stage is entered as soon as the attack is launched. The action of installing DDoS attack soft- ware falls under the Ramp-up stage. The Residue sub-stage refers to unintended communications and actions by mal- ware after an attack has been completed. 3 Significant Hacking Events We consider the following to be the most significant hacking events. Brain Virus: The world’s first computer virus was created by two brothers, Basit and AmjadFarooqAlvi, in Lahore, Pakistan [5]. This was a boot sector virus since it only affected boot records [6]. The Brain virus marked the area where the virus code was hidden as having bad sectors [7]. It occupied a part of the computer memory and infected any floppy disk that was accessed. It hid itself from detection by hooking into the INT13. When an attempt was made to read the infected sector, the virus simply showed the original sector. This resulted in a change to the volume label. Morris Worm: On 2nd November 1988 a Cornell graduate student, Robert Tappan Morris, unleashed the first computer worm [8]. It started as a benign experiment with a simple bug in a program, but the worm replicated much faster than anticipated [9]. By the following morning it had infected over 6000 hosts [10]. The worm could not determine whether a host had already been infected or not and as a result distributed multiple copies of itself on a single host. The exponential increase in data load even- tually tipped off the system administrators and the worm was discovered. CIH Virus: The CIH virus, also referred to as the Chernobyl or E95.CIH virus, first appeared in June 1998 [7]. It was created by a Taiwanese college student called Chen Ing-Hau [11]. It possessed a destructive payload with the purpose to destroy data. On release, the virus attempted to override a portion of the hard disk as well as the flash ROM of the PC. It infected over a million computers in Korea at the time [7]. I-LOVE-YOU Worm: The I-LOVE-YOU worm first appeared on May 4, 2000 in the form of an e-mail with the subject: I-LOVE-YOU [7]. It was created by a student named Onel de Guzman, and originated from Manila, Philippines. The worm code was written using Visual Basic and processed by the WScript engine [12]. It targeted computers using Internet Explorer and Microsoft's Outlook. Within a few hours it had spread worldwide via e-mail by making use of Outlook addresses of infected users. It exploited human curiosity, enticing people into opening an untrusted email. Code Red Worm: The Code Red worm appeared on July 12, 2001. It exploited a buffer-overflow vulnerability in Microsoft’s IIS web servers [13]. 2.2 Temporal Attack Model For example, computers that have incorrect time settings may attack their target at a later date and/or time than when the original coordinated attack was planned. This is also noticed in DDoS attacks. The Post-Attack Reconnaissance stage refers to scouting and other similar recon- naissance actions performed by an attacker after completion of the Attack stage. The attacker’s goal in this stage is to verify the effects of his/her attack and to assess whether the same methodology can be used again in the future. 3 Significant Hacking Events Upon infection of a machine, it checked whether the date was between the first and the nineteenth of the month. If so, a random list of IP addresses was generated and each machine on the list was probed to infect as many other machines as possible. Proper propagation of the worm failed due to a code error in the random number generator [14]. On 19 July a second version of the Code Red worm appeared that infected computers at a rate of 200 hosts per minute [9]. Estonia Hack Attack: Early in 2007, a series of politically motivated cyber-attacks struck Estonia [16]. The attacks included web defacements and DDoS attacks on Es- tonia government agencies, banks and Internet Service Providers. The attacks fol- lowed the removal of a bronze statue in Tallinn, which commemorated the dead from the Second World War [17]. At the time of the attacks, Estonia was one of the leading nations in Europe with regards to information and communication technologies [16]. This can be considered an example of cyber warfare and its potential effects. Conficker Worm: The Conficker was the first worm to penetrate cloud technology [15], [18]. It first appeared in November 2008 and quickly became one of the most infamous worms to date. The Conficker worm controlled over 6.4 million computer systems and also owned the world’s largest cloud network at the time. As a result of the infrastructure of a cloud, the worm could propagate much faster, infect a broader range of hosts and cause greater damage. Conficker has not been used as an attack weapon since, and it is speculated that it might have been a precursor to Stuxnet. Stuxnet Worm: Stuxnet was one of the most complex threats ever analysed [19]. The primary purpose of Stuxnet was to target industrial control systems such as gas pipelines and power plants with the goal of reprogramming the programmable logic controls (PLCs) of the systems to enable an attacker to control them. Stuxnet was also the first to exploit four zero-day vulnerabilities as well as compromise two digital certificates. As of September 29, 2010, Iran had the greatest number of infected com- puter systems. Stuxnet has shown that direct-attack attempts on critical infrastructures are no longer a myth but a definite possibility. Stuxnet actions can be considered an act of war, but no one has officially claimed responsibility for it. 5 Attack Model Map 5 The following sections describe how the most significant hacking events map to the Attack Model in Section 3. 5.1 Brain Virus ─ Target Identification: Experimentation with 5.25 inch floppy disks. ─ Reconnaissance: Exploring and experimenting with the DOS File Allocation Table (FAT) file system on the floppy disks. ─ Ramp-up: Writing and inclusion of malicious code on a 5.25 inch floppy disk. ─ Damage: Changing the volume label to read either Brain or ashar. ─ Residue: The changed disk label and the message left in infected boot sectors. ─ Post attack: Using the same technique to infect hard disks. 4 Trends 4 Although our selection of significant hacking events is subjective and does not repre- sent a comprehensive list, some interesting trends can be identified. Firstly, the mone- tary impact of each event is shown in figure 2. The vertical scale represents an estima- tion of the effect. Effects are classified as follows: 5 – severe financial impact; 4 – significant financial impact; 3 – major financial impact; 2 – minor financial impact; and 1 – negligible financial impact. On the horizontal scale, the attacks are listed in chronological order. Fig. 2. Monetary impact of hacking events. Fig. 2. Monetary impact of hacking events. Fig. 2. Monetary impact of hacking events. Figure 3 lists the most common countries of origin of hacking events. Most events surprisingly originate from the Philippines. Figure 4 illustrates the number of events per continent, with Europe and Asia at the top of the list. Fig. 3. Countries of origin of hacking events. Fig. 3. Countries of origin of hacking events. Fig. 4. Hacking events per continent. In most hacking events, small malware of between 1,000 and 100,000 bytes were utilized. The significant exception is Stuxnet, with a size of over 1.5 megabytes. The progressive increase in bandwidth and computer memory size will likely lend itself to the use of bigger malware (figure 5). Fig. 4. Hacking events per continent. Fig. 4. Hacking events per continent. Fig. 4. Hacking events per continent. In most hacking events, small malware of between 1,000 and 100,000 bytes were utilized. The significant exception is Stuxnet, with a size of over 1.5 megabytes. The progressive increase in bandwidth and computer memory size will likely lend itself to the use of bigger malware (figure 5). Fig. 5. Infection size in bytes. Fig. 5. Infection size in bytes. Fig. 5. Infection size in bytes. 5.3 CIH Virus ─ Target Identification: Exploring the gaps left in PE (Portable Executable) files. ─ Reconnaissance: Experimenting with PE file formats under Windows 95, 98 and ME for potential vulnerabilities. ─ Ramp-up: Writing the CIH virus code. ─ Ramp-up: Writing the CIH virus code. ─ Damage: Spreading of the CIH virus on computers, and destroying certain PC's BIOS, thus disabling PC use. g ─ Residue: No unintended attacks caused by the virus in this case. ─ Post attack: Verifying the effects of the virus by means of scouting. 5.4 I-LOVE-YOU Worm ─ Target Identification: Exploring and experimenting with the Windows operating system and Microsoft Outlook. ─ Reconnaissance: Using well-known search engines to search for potential weak- nesses in the Windows operating system and Microsoft Outlook. ─ Ramp-up: Writing the I-LOVE-YOU worm code. ─ Ramp-up: Writing the I-LOVE-YOU worm code. ─ Damage: The worm led to an effective DoS attack. ─ Damage: The worm led to an effective DoS attack. ─ Residue: Only hidden files with .mp2 and .mp3 extensions ─ Residue: Only hidden files with .mp2 and .mp3 extensions. ─ Post attack: Searching for other additional weaknesses in the Windows operating system and Microsoft Outlook. 5.5 Code Red Worm ─ Target Identification: Exploring the Microsoft IIS server configurations. ─ Reconnaissance: Using well-known search engines to search for potential vulnera- bilities in the IIS server software. ─ Ramp-up: Writing the Code Red worm code and identifying a buffer overflow vulnerability in the software. ─ Damage: Launching a DoS attack against randomly selected server IP addresses. ─ Residue: The worm used a static seed as its random number generator and so gen- erated identical lists of IP addresses that caused computers to be infected multiple times. ─ Post attack: Searching for additional weaknesses in Microsoft’s IIS servers. 5.2 Morris Worm ─ Target Identification: Experimenting with the ARPANET. ─ Target Identification: Experimenting with the ARPANET. ─ Reconnaissance: Scanning the ARPANET network for flaws and vulnerabilities. ─ Ramp-up: Writing the experimental program which includes the source code for the worm. ─ Damage: The release of the experimental program on the ARPANET g p p g ─ Residue: A machine being infected multiple times rather than only once. ─ Post attack: Experimenting with the possibility of a worm in different environ- ments. 5.7 Conficker Worm ─ Target Identification: Exploring cloud computing. ─ Target Identification: Exploring cloud computing. ─ Reconnaissance: Using well-known search engines to identify possible vulnerabili- ties in a cloud computing system. ─ Ramp-up: Writing the code for the Conficker worm. ─ Ramp-up: Writing the code for the Conficker worm. ─ Damage: Launching the Conficker worm in the cloud, thus making its target re- sources available to the attacker. ─ Residue: No unintended attacks caused by the worm in this case. ─ Post attack: Releasing additional versions of the worm to verify the effects. 5.8 Stuxnet Worm ─ Target Identification: Uranium enrichment infrastructure in Iran. ─ Target Identification: Uranium enrichment infrastructure in Iran. ─ Reconnaissance: Using well-known search engines to identify possible vulnerabili- ties in industrial software and equipment developed by Siemens. ─ Ramp-up: Writing the code for the Stuxnet worm and installing additional malware on targeted computer networks. g p ─ Damage: It physically damaged the Iranian Nuclear enrichment systems. ─ Residue: Infiltration of computer systems other than those in Iran. ─ Post attack: Verifying the effects on Iran’s industrial software systems. 5.6 Estonia Hack Attack ─ Target Identification: The relocation of the Bronze Soldier in Tallinn. ─ Reconnaissance: Using well-known search engines to identify possible weaknesses in the websites of well-known Estonian organizations. ─ Ramp-up: Installation of malware on targeted computer systems. ─ Damage: Government and commercial services (such as banks) became unavaila- ble during the attack. ─ Residue: Russian-language bulletin boards and one defaced website with the phrase: “Hacked from Russian hackers”. ─ Post attack: Scanning of the infected computer networks to determine the effects of the attacks. 7 The goal of the network attack model was to represent the majority of network based attacks. This temporal model was verified by mapping eight significant computer network attacks. The attacks were chosen to represent the most significant computer attacks (hacks) in the authors view. The mapping of these attacks shows the usability of the temporal model in aiding critical infrastructure protection. To prevent or protect against computer attacks, the CSIR are investigating methods to detect the Reconnaissance and Ramp-up stages of an attack. If these stages can be detected, mitigating action can be taken against computer attacks. The attack model is under development and will evolve as the research progress. Future work includes adding new dimensions to the classification of attacks, namely origin and motivation of the attack. Reviewing the reasons of why a network was easily penetrated and fo- cusing on the commonalities of learnt lessons will also be explored. 6 Protection of Critical Infrastructure 6 The protection of critical infrastructure involves the readiness to act against serious incidents threatening the critical infrastructure of a nation. Recently there is an in- creasing need to protect critical infrastructure from terrorist or other physical attacks, including cyber-attacks [20]. The previous sections emphasized this need by review- ing eight of the most significant computer network attacks. Apart from Stuxnet, there have been other instances of infrastructure attacks through computer networks [21]: ─ Maroochy Shire Council’s sewage control system in Queensland, Australia was attacked. ─ A teenager in Worcester, Massachusetts broke into the Bell Atlantic computer system and disabled part of the public switched telephone network using a dial-up modem connected to the system. This attack disabled phone services at the control tower, airport security, the airport fire department, the weather service, and carriers that use the airport. ─ In 2000, the Interior Ministry of Russia reported that hackers seized temporary control of the system regulating gas flows in natural gas pipelines. ─ In August 2005, Zotob worm crashed thirteen of DaimlerChrysler’s U.S. automo- bile manufacturing plants forcing them to remain offline for almost an hour. Plants in Illinois, Indiana, Wisconsin, Ohio, Delaware, and Michigan were also forced down. ─ The Sobig virus was blamed for shutting down train signaling systems throughout the east coast of the U.S. The virus infected the computer system at CSX Corp.’s Jacksonville, Florida headquarters, shutting down signaling, dispatching, and other systems. y ─ The Nuclear Regulatory Commission confirmed that in January 2003, the Mi- crosoft SQL Server worm known as the Slammer worm infected a private comput- er network at the idled Davis-Besse nuclear power plant in Oak Harbor, Ohio, dis- abling a safety monitoring system for nearly five hours. The Attack Model of Section 2.2 is able to map these Infrastructure computer based attacks. The ultimate goal of this research is to prevent such attacks by identify- ing the initial stages early enough for preventative actions. The model is able to pre- sent any type of computer network based attack, since computer based attacks on Infrastructure uses the same techniques and methodologies as traditional computer network attacks. The Reconnaissance and Ramp-up stages for attacking Infrastructure are similar for attacking computer networks. References 1. Heater, B.: Male: A Brief Timeline (2011), http://www.pcmag.com/slideshow/story/261678/malware-a-brief-timeline/ 2. Hall, K.: The 7 worstcyberattacks in history (that we know about) (2012), http://dvice.com/archives/2010/09/7-of-the-most-d.php (2011), 3. Julian: 10 Most Costly Cyber Attacks in History (2011), http://www.businesspundit.com/10-most-costly-cyber-attacks-in-history/ 4. van Heerden, R.P., Irwin B., Burke, I.D.: Classifying Network Attack Scenarios using an Ontology. In: Proceedings of the 7th International Conference on Information Warfare and Security, pp. 331-324 (2012) 5. Desai, P.: Towards an undetectable computer virus, Master’s thesis, San Jose State Uni- versity (2008), http://www.cs.sjsu.edu/faculty/stamp/students/Desai_Priti.pdf 6. Subramanya, S.R., Lakshminarasimhan, N.: Computer viruses. Potential IEEE, 20(4), pp. 16-19 (2001) 7. Blümler, P.: I-LOVE-YOU: Viruses, Trojan Horses and Worms, www.econmr.org/datapool/page/30/virus.pdf 8. Orman, H.: The Morris worm: a fifteen-year perspective. Security & Privacy, IEEE, 1(5), pp. 35-43 (2003) 9. Chen, T.M., Robert J.M.: Worm epidemics in high-speed networks. Computer, 37(6), pp. 48-53 (2004) 10. Cass, S.: Anatomy of malice [computer viruses]. Spectrum, IEEE, 38(11), pp. 56-60 (2004) 11. Bosworth, S., Kabay, M.E.: Computer security handbook. John Wiley & Sons Inc., New York (2002) 12. Bishop, M.: Analysis of the ILOVEYOU Worm (2000), http://nob.cs.ucdavis.edu/classes/ecs155-2005-04/handouts/iloveyou.pdf 13. Moore, D., Shannon, C.: Code-Red: a case study on the spread and victims of an Internet worm. In: Proceedings of the 2nd ACMSIGCOMM Workshop on Internet measurement, ACM, pp. 273-284 (2002) 14. Zou, C.C., Gong, W., Towsley, D.: Code red worm propagation modeling and analysis. In: Proceedings of the 9th ACM conference on Computer and Communications security, ACM, pp. 138-147 (2002) 15. Sarwar, U., Ramadass, S., Budiarto, R.: Dawn Of The Mobile Malware: Reviewing Mo- bile Worms. In: Proceedings of the 4th International Conference on Sciences of Electronic, Technologies of Information and Telecommunications (SETIT2007), pp. 35- 39 (2007) 16. Czosseck, C., Ottis, R., Taliharm, A.M.: Estonia after the 2007 Cyber Attacks: Legal, Stra- tegic and Organisational Changes in Cyber Security. International Journal of Cyber War- fare and Terrorism (IJCWT), 1(1), pp. 24-34 (2011) 17. Davis, J.: Hackers Take Down the Most Wired Country in Europe, Wired Magazine, 9(15) (2007) 18. Sharma, V.: An Analytical Survey of Recent Worm Attacks, In IJCSNS, 11(11), pp. 99 - 103 (2011) 19. Falliere, N., Murchu, L.O., Chien, E.: W32.stuxnet dossier: version 1.4, White paper, Sy- mantec Corp., Security Response (2011), http://www.wired.com/images_blogs/threatlevel/2011/02/Symantec-Stuxnet-Update-Feb- 2011.pdf 19. Falliere, N., Murchu, L.O., Chien, E.: W32.stuxnet dossier: version 1.4, White paper, Sy- mantec Corp., Security Response (2011), http://www.wired.com/images_blogs/threatlevel/2011/02/Symantec-Stuxnet-Update-Feb- 2011.pdf 20. Bradley, F.: Critical infrastructure protection. Electric Energy T and D, 7(2), pp. References 4-6 (2003) 21. Tsang, S.: Cyberthreats, Vulnerabilities and Attacks on SCADA Networks (2009), http://gspp.berkeley.edu/iths/Tsang SCADA%20Attacks.pdf
https://openalex.org/W2054676440	https://ccsenet.org/journal/index.php/cco/article/download/29999/18261	English	null	Outcome of Multimodality Therapy for Elderly Colorectal Cancer Patients	Cancer and clinical oncology	2,013	cc-by	5,765	Cancer and Clinical Oncology; Vol. 2, No. 2; 2013 ISSN 1927-4858 E-ISSN 1927-4866 Published by Canadian Center of Science and Education Cancer and Clinical Oncology; Vol. 2, No. 2; 2013 ISSN 1927-4858 E-ISSN 1927-4866 Published by Canadian Center of Science and Education Received: September 1, 2013 Accepted: October 2, 2013 Online Published: October 13, 2013 doi:10.5539/cco.v2n2p70 URL: http://dx.doi.org/10.5539/cco.v2n2p70 Received: September 1, 2013 Accepted: October 2, 2013 Online Published: October 13, 2013 doi:10.5539/cco.v2n2p70 URL: http://dx.doi.org/10.5539/cco.v2n2p70 Abstract The aim of this study was to analyze patterns of multimodality therapy in elderly patients with advanced colorectal cancer. We enrolled 272 patients with colorectal cancer. All patients received chemotherapy and some patients received secondary cytoreductive surgery and/or radiofrequency ablation. We compared differences between elderly patients (age ≥75 years) and non-elderly patients (age <75 years), especially in relation to multimodality therapy. There were no significant differences in cancer-specific survival between elderly (n = 37) and non-elderly patients (n = 235).Twenty-seven percent of elderly and 35% of non-elderly patients received multimodality therapy, which resulted in prolonged survival. Although the main chemotherapy regimen was the same in both groups who received multimodality therapy, elderly patients who received chemotherapy alone seemed to be under-treated. For elderly patients, prognostic factors were host-related, such as comorbidities, whereas for non-elderly patients prognostic factors were tumor-related. Comorbidities and modified Glasgow Prognostic Score may be prognostic indicators in elderly patients receiving multimodality therapy. In conclusion, chronological age alone should not contraindicate multimodality therapy of colorectal cancer in elderly patients. Appropriate selection criteria for multimodality therapy in elderly patients should include not only tumor characteristics, but also host- and treatment-related factors. Keywords: colorectal cancer, elderly patient, multimodality therapy Outcome of Multimodality Therapy for Elderly Colorectal Cancer Patients Yasuhiro Inoue1, Yuji Toiyama1, Koji Tanaka1, Yasuhiko Mohri1 & Masato Kusunoki1 Correspondence: Yasuhiro Inoue, Department of Gastrointestinal and Pediatric Surgery, Division of Reparative Medicine, Institute of Life Sciences, Mie University Graduate School of Medicine, 2-174 Edobashi, Tsu, Mie 514-8507, Japan. Tel: 81-59-231-5294. Fax: 81-59-232-6968. E-mail: yasinoue@clin.medic.mie-u.ac.jp 1. Introduction Colorectal cancer (CRC), the commonest malignancy worldwide, mainly affects the elderly. The mean age at diagnosis is under 72 years, 40% of cases occurring in patients aged over 75 years (Köhne et al., 2008; Yang et al., 2004; Boyle et al., 2005; Christensen et al., 2009). The geriatric CRC population is a very heterogeneous group that includes patients with excellent health status and those with comorbid conditions, functional dependency, and limited life expectancy (Sanoff et al., 2007), all of which may considerably influence the outcome of CRC treatments. The mainstay of CRC treatment is surgery, however, the role of chemotherapy has expanded considerably over the past 10 years. Modern chemotherapy, including molecular-targeted agents, has increased the survival time of patients with metastatic CRC to more than 2 years (Hurwitz et al., 2004; van Cutsem et al., 2009; Fuchs et al., 2007; Douillard et al., 2013). In addition, cytoreductive surgery for liver, lung, and other metastases has been widely used to achieve cure; however, metastasectomy is appropriate for only a few patients. Multimodality approaches, including surgery, radiotherapy, and chemotherapy, alone or in combination, have been proposed to further prolong survival of patients with recurrent or metastatic CRC. There have been efforts to increase the small proportion of patients receiving multimodality therapy by expanding the indications for it. Many studies have confirmed that chemotherapy can render some originally inoperable liver metastases resectable (Kusunoki et al., 1997; Kopetz et al., 2009; Kozloff et al., 2009; Folprecht et al., 2010; Wong et al., 2011). However, the indications for multimodality therapy in elderly patients with CRC have not been well defined. Elderly patients are more likely to have comorbidities and age-specific deteriorating organ function, which can reduce their tolerance of multimodality therapy, including surgery and modern chemotherapy. Published results for surgical morbidity and mortality rates are conflicting. Some studies show a correlation between age and postoperative 70 Cancer and Clinical Oncology Vol. 2, No. 2; 2013 www.ccsenet.org/cco complications (Colorectal Cancer Collaborative Group, 2000; van Leeuwen et al., 2008; Lee et al., 2007; Grosso et al., 2012), whereas others do not (Schiffmann et al., 2008; She et al., 2013). Whether elderly patients can tolerate multimodality cancer treatment and benefit from it in the same way as younger patients is controversial. 2. Material and Methods This was a retrospective study of all patients (n = 272) who received therapeutic chemotherapy for advanced or recurrent CRC in our Department of Gastrointestinal Surgery, from March 2000 to December 2012. This study included patients with histologically proven unresectable primary CRC, synchronous metastatic CRC, and metachronous metastatic or recurrent CRC. Patients who underwent initial simultaneous primary tumor resection and metastasectomy (e.g., lung and liver) were excluded. Exclusion criteria were any serious major organ dysfunction, a survival expectation of less than 3 months, and any other contraindication to enrollment in the study in the view of the patient’s physician. 1. Introduction The aim of this study was to describe patterns of multimodality therapy in patients with CRC aged ≥75 years, and to compare the outcomes of elderly (≥75 years) and non-elderly (<75 years) groups. 2.1 Multimodality Therapy According to our institutional policy for the treatment of metastatic CRC with an unresectable primary tumor, all patients enrolled in the study received 4–5 months of initial chemotherapy. All patients were informed about the possibility of secondary multimodality therapy using cytoreductive surgery and/or radiofrequency ablation (RFA) before their initial chemotherapy. Cytoreductive therapy was defined as surgery and/or RFA aimed at reducing tumor volume. Whether to proceed with multimodality therapy was determined by the response to chemotherapy. Cytoreductive surgery and/or RFA were considered for patients with partial responses or stable disease after systemic chemotherapy. Multidisciplinary discussions during chemotherapy determined the nature and timing of cytoreductive therapy for each patient. Our institutional Ethics Committee approved the study, and written informed consent was obtained from all patients who entered the study. 2.2 Chemotherapy Approval for drugs takes much longer in Japan than in the west. Because Japanese national insurance did not allow treatment of CRC with oxaliplatin between 2000 and 2005, first-line chemotherapy for advanced CRC was 5-FU with or without irinotecan in the first five years of the study. For the remaining seven years, the study patients received the then first-line chemotherapy for advanced or recurrent CRC triple-drug chemotherapy, namely 5-fluorouracil (5-FU), folinic acid and oxaliplatin or irinotecan (FOLFOX or FOLFIRI), with or without bevacizumab or cetuximab. The molecular-targeted agents bevacizumab, cetuximab, and panitumumab were approved for use in 2007, 2008, and 2010, respectively. From 2007, bevacizumab with FOLFOX or FOLFIRI was used as first-line chemotherapy for advanced or recurrent CRC. From 2008, cetuximab with or without irinotecan was used as second- or third-line chemotherapy. From 2010, cetuximab or panitumumab with FOLFIRI or FOLFOX have been available in Japan as first-line chemotherapy in patients with wild-type KRAS. In patients with incomplete cytoreduction, chemotherapy was reintroduced depending on their performance status (PS). Patients with complete secondary cytoreduction received 5-FU-based adjuvant chemotherapy. Those with no extrahepatic metastases but with unresectable hepatic metastases underwent hepatic arterial infusion chemotherapy with 5-FU followed by secondary surgery (Kusunoki et al., 1997). Radiotherapy with concurrent 5-FU-based chemotherapy was used to improve the resectability of locally inoperable rectal cancer. 2.3 Study Variables Patients were categorized into two groups based on age: an elderly group, aged ≥ 75 years and a non-elderly group aged <75 years. The age of 75 years was selected to divide the sample because approximately 40% of cases occurring in CRC patients aged over 75 years and the incidence increases with advancing age (Köhne et al., 2008; Yang et al., 2004; Boyle et al., 2005; Christensen et al., 2009). The comorbidity status was assessed by the Charlson index (Birim et al., 2003), which is a partially modified score including neither cancer nor age (Charlson et al., 1987). No patients had clinical evidence of infection or inflammatory conditions such as obstructive colitis or obstructive jaundice at that time. Routine laboratory tests, including serum C-reactive protein (CRP) and albumin concentrations and tumor markers such as carcinoembryonic antigen (CEA) (cut-off value, 6 ng/mL) were performed on the day of the first medical examination. Serum CRP serum concentrations were measured by turbidimetric immunoassay using an N-Assay TIA CRP-S kit (Nittobo Medical, Tokyo, Japan). Because this CRP assay has a lower detection limit than assays used in other studies (0.2 mg/dL vs >0.5 mg/dL) (McMillan et al., 2003, Crozier et al., 2006), the cut-off value for abnormal serum CRP was set at 0.5 mg/dL. As previously described, the original Glasgow Prognostic Score (GPS) (Forrest et al., 2003) was modified by the present authors according to the best predictive values calculated by receiver operating characteristic analysis to 71 Cancer and Clinical Oncology www.ccsenet.org/cco Vol. 2, No. 2; 2013 create the instrument used in this study: the mGPS (Toiyama et al., 2011, Inoue et al., 2013). Briefly, patients with high CRP concentrations (>0.5 mg/dL) plus hypoalbuminemia (<3.5g/dL) were allocated a score of 2, patients with only one of these factors a score of 1, and patients with neither of these factors a score of 0. create the instrument used in this study: the mGPS (Toiyama et al., 2011, Inoue et al., 2013). Briefly, patients with high CRP concentrations (>0.5 mg/dL) plus hypoalbuminemia (<3.5g/dL) were allocated a score of 2, patients with only one of these factors a score of 1, and patients with neither of these factors a score of 0. JMP version 7 software (SAS Institute, Cary, NC, USA) was used to perform statistical analyses. Data are presented as the mean ± standard deviation. 2.3 Study Variables Contingency tables were analyzed using Fisher’s exact test or the χ2 test with Yates’ correction. Correlations between continuous and categorical variables were evaluated by the Mann-Whitney U test. Survival curves were constructed according to the Kaplan-Meier method and differences analyzed using the log-rank test. Each significant predictor identified was assessed by multivariate analysis using Cox’s proportional hazards model. A P value of < 0.05 was considered significant. 3.1 Relevant Patient Variables and Comorbidities We performed a retrospective review of 272 patients treated in our department for unresectable primary, synchronous metastatic, and metachronous metastatic or recurrent CRC. There were 163 men (60%) and 109 women (40%), with a mean age of 64 years (range 29–85 years). Of these, we analyzed 37 elderly (age ≥75 years: 79 ± 3 years; range 75–85 years) and 235 non-elderly (age <75 years: 60 ± 10 years; range, 29–74 years) patients over the 12 years of the study. Table 1 summarizes the background characteristics of the 272 patients by age group. We found no differences between groups in tumor characteristics such as extent or tumor. However, there were significant correlations between age and PS (P = 0.0122) and comorbidity (P = 0.0003). To assess comorbid condition, we also measured the Charlson comorbidity index (CCI). Although this index was not significantly different for elderly (mean 0.270) than for non-elderly patients (mean 0.149), hypertension (P =0.0077) and cardiovascular disease (P = 0.0324) were present significantly more frequently in elderly than in non-elderly patients. 72 able 1. Patient characteristics Variables Total (n=272) elderly (≥75) (n=37) non-elderly (<75) (n=235) p-value Gender 0.6721 Female 109 16 93 Male 163 21 142 PS 0.0122 0 112 8 104 1 113 21 92 2 31 8 23 3 14 0 14 4 2 0 2 Comorbidities All comorbidities 83 20(54%) 63(27%) 0.0008 Diabetes mellitus 17 2(5%) 15(6%) 0.8194 Cardiovascular disease 4 2(5%) 2(1%) 0.0324 Hypertension 52 13(35%) 39(17%) 0.0077 Plumonary disease 13 3(8%) 10(4%) 0.3072 Charlson index 0.1553 0 133 28 205 1 33 8 25 2 6 1 5 modified GPS 0.9201 0 157 22 135 1 81 10 71 2 34 5 29 Table 1. Patient characteristics 72 Cancer and Clinical Oncology Vol. 2, No. 2; 2013 Vol. 2, No. 3.1 Relevant Patient Variables and Comorbidities 2; 2013 www.ccsenet.org/cco Variables Total (n=272) elderly (≥75) (n=37) non-elderly (<75) (n=235) p-value Tumor state 0.8215 synchronous 159 21 138 metachronous 113 16 97 Tumor site 0.7905 Liver 78 12 66 Lung 44 5 39 Liver and lung 18 4 14 Local recurrence 28 3 25 Unresectable primary 22 3 19 Lymphnode 13 2 11 Dissemination 66 7 59 Others 2 1 1 Pathology 0.9748 Differentiated 243 33 210 Non differentiated 29 4 25 CEA 0.6373 <12ng/ml 120 15 105 >12ng/ml 152 22 130 Chemotherapy 0.0335 5FU-based 38 7(19%) 31(13%) Irinotecan-based 95 8(22%) 87(37%) Oxaliplatin-based 72 16(43%) 56(24%) Molecular agents 67 6(16%) 61(26%) Response Rate (measurable) 0.7349 CR 6 0 6 PR 90 10 80 SD 129 21 108 PD 33 5 28 Radiotherapy 0.2629 yes 47 4 43 no 225 33 192 Secondary surgery and/or RFA 0.3472 yes 92 10(27%) 82(35%) no 180 27 153 carcinoembryonic antigen; CR, complete response; mGPS, modified Glasgow Prognostic Score; PD, progressive disease; PR, part nse; PS, performance status; RFA, radiofrequency ablation; SD, stable disease. CEA, carcinoembryonic antigen; CR, complete response; mGPS, modified Glasgow Prognostic Score; PD, progressive disease; PR, partial response; PS, performance status; RFA, radiofrequency ablation; SD, stable disease. 3.2 Therapeutic Approach All patients received initial therapeutic chemotherapy; 92/272 (34%) also received chemotherapy as part of multimodality therapy following secondary cytoreductive surgery and/or RFA. A greater proportion of non-elderly patients than elderly patients received multimodality therapy, (82/235 [35%] and 10/37 [27%]), respectively; this difference was not significant. We also assessed the proportion of patients receiving each of four main chemotherapy regimens, the categories being 5-FU- based, irinotecan-based, oxaliplatin-based, and combinations including molecular agents. The main regimens of patients who received chemotherapy alone were significantly different: non-elderly patients most commonly received irinotecan-based or combinations including molecular agents, whereas elderly patients most often received oxaliplatin- based or 5-FU-based chemotherapy. As to the main initial regimen received by patients who subsequently underwent multimodality therapy, there 73 Vol. 2, No. 2; 2013 Cancer and Clinical Oncology www.ccsenet.org/cco were no significant differences between the two groups. were no significant differences between the two groups. 3.3 Cancer-specific Survival were no significant differences between the two group 3.3 Cancer-specific Survival 3.3 Cancer-specific Survival We assessed cancer-related mortality by Kaplan-Meier survival analysis. After a median follow up of 21 months, there were no significant differences in median survival time (MST) between elderly (32 months, 95% CI 24–39 months) and non-elderly patients (27 months, 95% CI 23–31 months). Multimodality therapy was associated with significantly better survival than chemotherapy alone in both elderly (P =0.0423) and non-elderly patients (P < 0.0001) (Fig. 1). Multimodality therapy resulted in longer overall survival than did chemotherapy alone for both elderly (median 53 months [95% CI 20–85 months] and 30 months [95% CI 20–40 months], respectively), and non-elderly patients (median 44 months [95% CI 32–55 months] and 20 months [95% CI 17–23 months, respectively). Cox univariate regression analyses identified different factors affecting cancer-specific survival in elderly than in non-elderly patients. In non-elderly patients, timing of metastasis (synchronous vs. metachronous), pathology (undifferentiated vs. differentiated), CEA concentration (≥ 12 vs. < 12 ng/mL), extent of dissemination, mGPS, PS, rate of response to chemotherapy, and multimodality therapy correlated significantly with cancer-specific survival (Table 2a). Conversely, in elderly patients there were no correlations between tumor characteristics and cancer-specific survival; however, patient characteristics, including comorbidity and therapeutic factors such as response rate, did correlate significantly with cancer-specific survival (Table 2b). Multivariate analysis using these characteristics showed that pathology (undifferentiated vs. differentiated), CEA concentration (> 12 vs. < 12 ng/mL), PS, response rate of chemotherapy, and multimodality therapy all had significant independent correlations with cancer specific survival time in non-elderly patients, whereas response rate was the only significant independent prognostic factor in elderly patients (Table 3). Figure 1. Multimodality therapy resulted in significantly better survival than did chemotherapy alone in both elderly (a) and non-elderly patients (b) Figure 1. Multimodality therapy resulted in significantly better survival than did chemotherapy alone in both elderly (a) and non-elderly patients (b) 74 74 Cancer and Clinical Oncology Vol. 2, No. 2; 2013 www.ccsenet.org/cco Table 2a. Univariate analysis in relation to cancer-specific mortality in non-elderly patients Variables p-value Odds Ratio 95%CI Gender (male vs. female) 0.1442 0.7825 0.563-1.087 PS (2-4 vs. 0-1) 0.0005 1.9685 1.342-2.890 Comorbidity (yes vs. no) 0.5496 1.119 0.775-1.613 Timing of metastasis (synchronous vs. metachronous) 0.0038 1.632 1.171-2.272 Pathology (undifferentiated vs. differentiated) 0.0017 2.173 1.337-3.533 Metastatic site (dissemination vs. others) 0.0008 1.845 1.289-2.646 CEA (>12 vs. <12ng/ml) <0.0001 2.086 1.492-2.916 mGPS (1-2 vs.0) 0.0002 1.861 1.347-2.572 Response rate (others vs. 3.4 Multimodality Therapy in Elderly Patients Having confirmed that multimodality therapy prolonged survival in both non-elderly and elderly patients, we evaluated the correlation between comorbidities and the use of multimodality therapy. The CCI index correlated significantly with the use of multimodality therapy in the overall group of 272 patients (P = 0.0463). Of the patients who underwent multimodality therapy, 85/92 (92%) were in the category CCI 0, 7/92 (8%) CCI 1 and none in CCI 2. In contrast, of the patients who received chemotherapy alone, 148/180 (82%) were in the category CCI 0, 26/180 (14%) in CCI 1 and 6/180 (4%) in CCI 2. This trend differed between elderly and non-elderly patients. All comorbidities and the CCI category were significantly correlated with contraindication to multimodality therapy in non-elderly patients; however, this was not so in elderly patients (P = 0.0234, P = 0.0166, respectively). In the latter group, only hypertension tended toward correlating with contraindications to multimodality therapy (P = 0.0513). To explorer useful predictors of indications for multimodality therapy in elderly patients, we also evaluated the prognostic significance of several clinical factors in 10 elderly patients who underwent multimodality therapy. In these elderly patients, the disease sites were initially unresectable primary, local recurrence, liver, bladder and peritoneal metastases. We performed reductive surgery in eight patients; three with liver metastases, two peritoneal metastases, one primary, one local recurrence, and one bladder involvement. We administered RFA for liver metastases to the other two patients. Kaplan-Meier survival analysis showed that patients with comorbidity (n = 4) had poorer survival than did those without comorbidity (n = 6) (3-year survival 83% vs. 25%, P = 0.0649) (Fig. 2). However, there were no significant differences in survival according to tumor characteristics such as timing of metastasis, pathology, and CEA concentration. Furthermore, we found no difference in survival between elderly and non-elderly patients who received chemotherapy alone. Interestingly, there were significant differences in survival according to mGPS (0 vs.1–2), with MST of 53 months (95% CI 20–85 months) for mGPS 0 (n = 8), and 12 months (95% CI 7–31 months) for mGPS1-2 (n = 2) (P = 0.0012) (Fig. 3). In contrast, there was no significant correlation between mGPS and survival in the elderly patients who received chemotherapy alone. Figure 2. Kaplan-Meier survival curves of elderly patients receiving multimodality therapy showing that those with comorbidity had inferior survival to without comorbidity (a). were no significant differences between the two group 3.3 Cancer-specific Survival CR, PR) 0.0015 1.736 1.235-2.445 Secondary surgery and/or RFA <0.0001 2.5641 1.786-3.676 PS, performance status; CEA, carcinoembryonic antigen; mGPS, modified Glasgow Prognostic Score Table 2b. Multivariate analysis in relation to cancer-specific mortality in non-elderly patients Variables p-value Odds Ratio 95%CI PS (2-4 vs. 0-1) 0.1291 1.328 0.867-2.033 Timing of metastasis (synchronous vs. metachronous) 0.6411 1.092 0.755-1.577 Pathology (undifferentiated vs. differentiated) 0.0077 2.005 1.202-3.347 Metastatic site (dissemination vs. others) 0.0323 1.502 1.035-2.179 CEA (>12 vs. <12ng/ml) 0.0005 1.862 1.311-2.644 mGPS (1-2 vs.0) 0.0003 1.982 1.367-2.874 Response rate (others vs. CR, PR) 0.0041 1.727 1.189-2.513 Secondary surgery and/or RFA <0.0001 2.247 1.516-3.333 Table 2b. Multivariate analysis in relation to cancer-specific mortality in non-elderly patients Table 3a. Univariate analysis in relation to cancer-specific mortality in elderly patients Variables p-value Odds Ratio 95%CI Gender (male vs. female) 0.744 1.179 0.439-3.175 PS (2-4 vs. 0-1) 0.521 0.709 0.248-2.024 Comorbidity (yes vs. no) 0.06 2.747 0.959-7.874 Hypertension (yes vs. no) 0.023 3.663 1.196-11.236 Timing of metastasis (synchronous vs. metachronous) 0.8676 0.924 0.363-2.350 Pathology (undifferentiated vs. differentiated) 0.7243 0.689 0.097-5.456 Metastatic site (dissemination vs. others) 0.539 0.629 0.143-2.672 CEA (>12 vs. <12ng/ml) 0.3232 1.658 0.608-4.522 mGPS (1-2 vs.0) 0.1315 2.197 0.790-6.114 Response rate (others vs. CR, PR) 0.0228 4.545 1.235-16.666 Secondary surgery and/or RFA 0.051 3.322 0.995-11.111 CEA, carcinoembryonic antigen; mGPS, modified Glasgow Prognostic Score; PS, performance status Table 3a. Univariate analysis in relation to cancer-specific mortality in elderly patients 75 Vol. 2, No. 2; 2013 Cancer and Clinical Oncology www.ccsenet.org/cco Table 3b. Multivariate analysis in relation to cancer-specific mortality in elderly patients Variables p-value Odds Ratio 95%CI Hypertension (yes vs. no) 0.9123 1.0235 0.673-1.558 Response rate (others vs. CR, PR) 0.0016 1.7631 1.231-2.445 4. Discussion Colorectal cancer is a major cause of cancer deaths in developed countries. Because the populations of these nations, including Japan, are rapidly aging, clinicians will be faced with treating many patients with advanced cancer, including those with metastatic or recurrent CRC. Recent advances in CRC treatment have resulted in prolongation of survival even in patients with advanced stages of disease. However, clinical decision-making regarding modern multimodality therapy for advanced CRC in elderly patients is complex, because of the lack of data regarding optimal chemotherapy regimens, timing of cytoreductive therapy and the use of multimodality therapy in these patients. In a systematic review of surgery for CRC in 34,194 elderly patients, researchers found the incidence of postoperative morbidity and mortality increased progressively with advancing age. However, although overall survival was less in elderly than non-elderly patients, age-related differences in cancer-specific survival were much less pronounced (Colorectal Cancer Collaborative Group., 2000). Because most of the definitive clinical trials have excluded subjects of advanced age or with a poor PS, there is still uncertainly regarding the optimal use of systemic chemotherapy in elderly patients with CRC. Many of the clinical trials that have included elderly patients have documented similar survival rates and toxicity profiles for elderly and younger patients (Köhne et al., 2008). Most researchers agree that age alone should not be a contraindication to the use of standard chemotherapy and that fit elderly patients can receive similar aggressive chemotherapy as younger patients; however, identification of the elderly patients who are most likely to benefit from chemotherapy warrants further investigation (Colorectal Cancer Collaborative Group., 2000). Indications for modern chemotherapy followed by cytoreductive therapy in elderly patients remain largely unknown. In addition, because of the potential for worsening comorbidities to cause poorer outcomes, the utility of metastasectomy, including hepatectomy and pulmonary resection, in elderly patients is also controversial. There have been recent reports of decreased survival and higher postoperative morbidity associated with this procedure in the elderly (Adam et al., 2010; Nagano et al., 2005; Endoh et al., 2013). In our study, although elderly patients were more likely to have a poor PS and comorbidities, there were no significant differences in cancer-specific survival time between elderly and non-elderly patients. Interestingly, factors affecting cancer-specific survival were different in the elderly and non-elderly groups. Treatment-related prognostic factors were common to both groups. 3.4 Multimodality Therapy in Elderly Patients For elderly patients who received chemotherapy alone, there was no comorbidity-related difference in survival (b) Figure 2. Kaplan-Meier survival curves of elderly patients receiving multimodality therapy showing that those with comorbidity had inferior survival to without comorbidity (a). For elderly patients who received chemotherapy alone, there was no comorbidity-related difference in survival (b) 76 Cancer and Clinical Oncology Vol. 2, No. 2; 2013 www.ccsenet.org/cco Figure 3. Relationship between mGPS (0, 1–2) and cancer-specific survival time in patients who received multimodality therapy (a) or chemotherapy alone (b) Figure 3. Relationship between mGPS (0, 1–2) and cancer-specific survival time in patients who received multimodality therapy (a) or chemotherapy alone (b) 4. Discussion However, elderly patients had various host-related prognostic factors such as comorbidities, whereas tumor-related prognostic factors were characteristic in the non-elderly group. Multivariate analysis revealed that various tumor-related (pathology, serum CEA concentration), host-related (PS) and treatment-related factors (response rate and multimodality therapy) all had significant independent correlations with survival in non-elderly patients, whereas only treatment-related factors (response rate) was an independent prognostic factor in elderly patients. Although multimodality therapy was an independent prognostic factor in only non-elderly patients, we found that the survival benefit of multimodality therapy in both elderly and non-elderly patients was comparable to that of chemotherapy alone. As a consequence of differences in the main chemotherapy regimen that each group received, elderly patients were less likely than non-elderly patients to receive multimodality therapy (27 vs. 35%). 77 Cancer and Clinical Oncology Vol. 2, No. 2; 2013 www.ccsenet.org/cco It seems likely that elderly patients were under-treated compared with non-elderly patients, especially those who received chemotherapy alone. Of the patients who received chemotherapy alone, elderly patients were more likely to receive oxaliplatin-based or 5-FU-based chemotherapy, probably because these types of chemotherapy are less aggressive and more readily tolerated than modern first line chemotherapy that includes molecular agents. These findings imply that under-treatment occurred because of physician preferences; however, cancer-specific survival of elderly and non-elderly patients who received chemotherapy alone was not significantly different. Elderly and non-elderly groups who went on to undergo multimodality therapy did not differ significantly in the initial main chemotherapy regimen they received. In other words, elderly patients who were able to undergo the same aggressive chemotherapy as non-elderly patients were more likely to continue on to multimodality therapy. In elderly patients, therapeutic decisions concerning palliative chemotherapy versus conversion chemotherapy prior to multimodality therapy must be made on an individual basis, however, the factors that determine the optimal therapeutic approach in elderly patients are not well known. In this regard, one important consideration is overall survival time after secondary cytoreductive therapy. In the current study, the survival of elderly patients with comorbidity was inferior to that of those without comorbidity, even after multimodality therapy. Our findings suggest that both tumor-related factors such as treatment markers and host-related factors may be more reliable prognostic indicators in elderly patients undergoing multimodality therapy than they are in non-elderly patients. 4. Discussion Because GPS is a well-known surrogate marker for response to treatment and can be used to predict tumor recurrence in a variety of cancers (McMillan et al., 2003; Crozier et al., 2006; Forrest et al., 2003; Proctor et al., 2011), and because we recently reported the usefulness of a modified GPS in patients undergoing multimodality therapy for advanced CRC (Toiyama et al., 2011; Inoue et al., 2013), we assessed its value in the current study. Our findings suggest that modified GPS is a potential prognostic marker for elderly patients receiving multimodality therapy. Limitations of our study included the small number of patients with huge difference considering its dimension (37 versus 235) and that it was a single-site study; consequently, the findings may not be applicable to all elderly patients with CRC, and more patients should be involved to explore useful predictors of indications for multimodality therapy in elderly patients. However, we were able to identify patterns associated with multimodality therapy in patients aged >75 years with CRC and to compare the outcomes of elderly and non-elderly patients. 5. Conclusion Chronological age alone should not be a contraindication to multimodality therapy of CRC in elderly patients. To improve survival in the elderly, selection of palliative chemotherapy versus or active multimodality therapy for patients is very important. Furthermore, appropriate selection criteria for multimodality therapy in elderly patients may include not only tumor characteristics but also host- or treatment-related factors such as comorbidities or surrogate markers, including modified GPS. Douillard, J. Y., Oliner, K. S., Siena, S., Tabernero, J., Burkes, R., Barugel, M., … Patterson, S. D. (2013). References Adam, R., Frilling, A., Elias, D., Laurent, C., Ramos, E., Capussotti, L., … de Haas, R. J. (2010). Liver resection of colorectal metastases in elderly patients. Br J Surg, 97, 366-376. http://dx.doi.org/10.1002/bjs.6889. Birim, O., Maat, A. P., Kappetein, A. P., van Meerbeeck, J. P., Damhuis, R. A., & Bogers, A. J. (2003). Validation of the Charlson comorbidity index in patients with operated primary non-small cell lung cancer. Eur J Cardiothorac Surg, 23, 30-34. http://dx.doi.org/10.1016/S1010-7940(02)00721-2 Boyle, P., & Ferlay, J. (2005). Cancer incidence and mortality in Europe. Ann Oncol, 16, 481-488. Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis, 40, 373-383. Christensen, K., Doblhammer, G., Rau. R., & Vaupel, J. W. (2009). Ageing populations: thechallenges ahead. Lancet, 374, 1196-1208. http://dx.doi.org/10.1016/S0140-6736(09)61460-4 Colorectal Cancer Collaborative Group. (2000). Surgery for colorectal cancer in elderly patients: a systematic review. Lancet, 356, 968-974. http://dx.doi.org/10.1016/S0140-6736(00)02713-6 Crozier, J. E., McKee, R. F., McArdle, C. S., Angerson, W. J., Anderson, J. H., Horgan, P. G., … McMillan, D. C. (2006). The presence of a systemic inflammatory response predicts poorer survival in patients receiving adjuvant 5-FU chemotherapy following potentially curative resection for colorectal cancer. Br J Cancer, 94, 1833-1836. http://dx.doi.org/10.1038/sj.bjc.6603185 Douillard, J. Y., Oliner, K. S., Siena, S., Tabernero, J., Burkes, R., Barugel, M., … Patterson, S. D. (2013). 78 Vol. 2, No. 2; 2013 Cancer and Clinical Oncology www.ccsenet.org/cco Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. N Engl J Med, 369, 1023-1034. http://dx.doi.org/10.1056/NEJMoa1305275 Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. N Engl J Med, 369, 1023-1034. http://dx.doi.org/10.1056/NEJMoa1305275 Endoh, H., Yamamoto, R., Satoh, Y., Kuwano, H., & Nishizawa, N. (2013). Risk analysis of pulmonary resection for elderly patients with lung cancer. Surg Today, 43, 514-520. http://dx.doi.org/10.1007/s00595-012-0401-8 Folprecht, G., Gruenberger, T., Bechstein, W. O., Raab, H. R., Lordick, F., Hartmann, J. T., … Köhne, C. H. (2010). Tumour response and secondary resectability of colorectal liver metastases following neoadjuvant chemotherapy with cetuximab: the CELIM randomized phase 2 trial. Lancet Oncol, 11, 38-47. http://dx.doi.org/10.1016/S1470-2045(09)70330-4 Forrest, L. M., McMillan, D. C., McArdle, C. S., Angerson, W. J., & Dunlop, D. J. (2003). Evaluation of cumulative prognostic scores based on the systemic inflammatory response in patients with inoperable non-small-cell lung cancer. Br J Cancer, 89, 1028-1030. http://dx.doi.org/10.1038/sj.bjc.6601242 Fuchs, C. References S., Marshall, J., Mitchell, E., Wierzbicki, R., Ganju, V., Jeffery, M., … Barrueco, J. (2007). Randomized, controlled trial of irinotecan plus infusional, bolus, or oral fluoropyrimidines in first-line treatment of metastatic colorectal cancer: results from the BICC-C study. J Clin Oncol, 25, 4779-4786. http://dx.doi.org/10.1200/JCO.2007.11.3357 Grosso, G., Biondi, A., Marventano, S., Mistretta, A., Calabrese, G., & Basile, F. (2012). Major postoperative complications and survival for colon cancer elderly patients. BMC Surg, 12(Suppl 1), S20. http://dx.doi.org/10.1186/1471-2482-12-S1-S20 Hurwitz, H., Fehrenbacher, L., Novotny, W., Cartwright, T., Hainsworth, J., Heim, W., … Kabbinavar, F. (2004). Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med, 350, 2335-2342. http://dx.doi.org/10.1056/NEJMoa032691 Inoue, Y., Iwata, T., Okugawa, Y., Kawamoto, A., Hiro, J., Toiyama, Y., … Kusunoki, M. (2013). Prognostic significance of a systemic inflammatory response in patients undergoing multimodality therapy for advanced colorectal cancer. Oncology, 84, 100-107. http://dx.doi.org/10.1159/000343822 Kopetz, S., Chang, G. J., Overman, M. J., Eng, C., Sargent, D. J., Larson, D. W., … McWilliams, R. R. (2009). Improved survival in metastatic colorectal cancer is associated with adoption of hepatic resection and improved chemotherapy. J Clin Oncol, 27, 3677-3683. http://dx.doi.org/10.1200/JCO.2008.20.5278 Kozloff, M., Yood, M. U., Berlin, J., Flynn, P. J., Kabbinavar, F. F., Purdie, D. M., … Grothey, A. (2009). Clinical outcomes associated with bevacizumab-containing treatment of metastatic colorectal cancer: the BRiTE observational cohort study. Oncologist, 14, 862-870. http://dx.doi.org/10.1634/theoncologist.2009-0071 Köhne, C. H., Folprecht, G., Goldberg, R. M., Mitry, E., & Rougier, P. (2008). Chemotherapy in elderly patients with colorectal cancer. Oncologist, 13, 390-402. http://dx.doi.org/10.1634/theoncologist.2007-0043 Kusunoki, M., Noda, M., Yanagi, H., Kotera, H., & Yamamura, T. (1997). Second-look hepatectomy after 5-FU arterial infusion in patients with primary unresectable hepatic colorectal metastases: molecular characteristics of the tumors. Int J Oncol, 10, 107-111. Lee, L., Jannapureddy, M., Albo, D., Awad, S. S., Farrow, B., Bellows, C. C., … Berger, D. H. (2007). Outcomes of Veterans Affairs patients older than age 80 after surgical procedures for colon malignancies. Am J Surg, 194, 646-651. http://dx.doi.org/10.1016/j.amjsurg.2007.08.003 McMillan, D.C., Canna, K., & McArdle, C.S. (2003). Systemic inflammatory response predicts survival following curative resection for colorectal cancer. Br J Surg, 90, 215-219. http://dx.doi.org/10.1002/bjs.4038 Nagano, Y., Nojiri, K., Matsuo, K., Tanaka, K., Togo, S., Ike, H., … Shimada, H. (2005). The impact of advanced age on hepatic resection of colorectal liver metastases. J Am Coll Surg, 201, 511-516. http://dx.doi.org/10.1016/j.jamcollsurg.2005.05.010 Proctor, M. J., Morrison, D. S., Talwar, D., Balmer, S. M., O'Reilly, D. Oncol, 2, 1891-1897. http://dx.doi.org/10.1200/JCO.2006.10.1220 Oncol, 2, 1891-1897. http://dx.doi.org/10.1200/JCO.2006.10.1220 Schiffmann, L., Ozcan, S., Schwarz, F., Lange, J., Prall, F., & Klar, E. (2008). Colorectal cancer in the elderly: surgical treatment and long-term survival. Int J Colorectal Dis, 23, 601-610. http://dx.doi.org/10.1007/s00384-008-0457-5 She, W. H., Poon, J. T., Fan, J. K., Lo, O. S., & Law, W. L. (2013). Outcome of laparoscopic colectomy for cancer in elderly patients. Surg Endosc, 27, 308-312. http://dx.doi.org/10.1007/s00464-012-2466-2 Toiyama, Y., Miki, C., Inoue, Y., Tanaka, K., Mohri, Y., & Kusunoki, M. (2011). Evaluation of inflammation-based prognostic score for the identification of patients requiring postoperative adjuvant chemotherapy for stage II colorectal cancer. Experimental and Therapeutic Medicine, 2, 95-101. Yang, L., Parkin, D. M., Li, L. D., Chen, Y. D., & Bray, F. (2004). Estimation and projection of the national profile of cancer mortality in China: 1991-2005. Br J Cancer, 90, 2157-2166. http://dx.doi.org/10.1038/sj.bjc.6601813 Van Cutsem, E., Köhne, C. H., Hitre, E., Zaluski, J., Chang Chien, C. R., Makhson, A., … Rougier, P. (2009). Cetuximab and chemotherapy as initial treatment for metastatic colorectal cancer. N Engl J Med, 360, 1408-1417. http://dx.doi.org/10.1056/NEJMoa0805019 van Leeuwen, B. L., Påhlman, L., Gunnarsson, U., Sjövall, A., & Martling, A. (2008). The effect of age and gender on outcome after treatment for colon carcinoma. A population-based study in the Uppsala and Stockholm region. Crit Rev Oncol Hematol, 67, 229-236. http://dx.doi.org/10.1016/j.critrevonc.2008.03.005 Wong, R., Cunningham, D., Barbachano, Y., Saffery, C., Valle, J., Hickish, T., … Chau, I. (2011). A multicentre study of capecitabine, oxaliplatin plus bevacizumab as perioperative treatment of patients with poor-risk colorectal liver-only metastases not selected for upfront resection. Ann Oncol, 22, 2042-2048. http://dx.doi.org/10.1093/annonc/mdq714 References S., Foulis, A. K., … McMillan, D. C. (2011). An inflammation-based prognostic score (mGPS) predicts cancer survival independent of tumour site: a Glasgow Inflammation Outcome Study. Br J Cancer, 104, 726-734. http://dx.doi.org/10.1038/sj.bjc.6606087 Sanoff, H. K, Bleiberg, H., & Goldberg, R. M. (2007). Managing older patients with colorectal cancer. J Clin 79 Vol. 2, No. 2; 2013 Cancer and Clinical Oncology www.ccsenet.org/cco This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). Copyrights Copyright for this article is retained by the author(s), with first publication rights granted to the journa Copyright for this article is retained by the author(s), with first publication rights granted to the journal. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). 80
https://openalex.org/W2998904818	https://europepmc.org/articles/pmc7022919?pdf=render	English	null	From Acidifiers to Intestinal Health Enhancers: How Organic Acids Can Improve Growth Efficiency of Pigs	Animals	2,020	cc-by	11,820	Received: 29 November 2019; Accepted: 9 January 2020; Published: 14 January 2020 Simple Summary: Organic acids have been used for a long time to support pig growth particularly at weaning, and more recently have become the number one alternative to growth promoters to improve the production eﬃciency of pigs. This article will review the antimicrobial properties of organic acids and elucidate the diﬀerent modes of action that organic acids can exert along the gastrointestinal tract of pigs. Moreover, it will be introduced the advantage of microencapsulation as a tool to deliver organic acids along the intestine and allow their positive eﬀects. Abstract: Organic acids have been used successfully in pig production as a cost-eﬀective performance-enhancing option and they continue to be the number one alternative to antibiotic growth promoters. The aim of this review is to provide the biological rationale behind organic acids use in pig production, focusing on their diﬀerent eﬀects along the gastrointestinal tract of pigs. Organic acids are reviewed for their antimicrobial properties and for their classic use as acidiﬁers, with particular attention to pH modulation and microﬂora control. Additional beneﬁcial eﬀects on intestinal health and general metabolism are presented and we explain the advantage of microencapsulation as a tool to deliver organic acids along the intestine. Keywords: organic acids; pig; acidiﬁers; metabolic eﬀect; microencapsulation Animals 2020, 10, 134; doi:10.3390/ani10010134 animals animals animals animals 1. Introduction Today the modern livestock industry requests increasingly high animal productivity and eﬃciency. In the past, high levels of production were obtained using pharmaceutical strategies such as antibiotic growth promoters. Health concerns have driven worldwide regulatory restrictions, limiting the use of antibiotics and favoring the diﬀusion of new types of production (such as antibiotic-free and no antibiotic ever) whose feasibility requires the highest standards of management, feeding, and nutrition. To support animal growth and health without drugs, diet has a pivotal role beyond the supply of nutrients, and high quality feed ingredients and feed additives are therefore needed. Among them, organic acids have been widely used over the last decades for their positive eﬀect on growth eﬃciency and they are considered a valid tool for pig production. Organic acids were reported as eﬀective growth enhancers throughout the production cycle of pigs, although there is a relatively large variation in responses due to various factors such as type and dose of organic acids used, supplementation duration, type of diet and buﬀering capacity, hygiene and welfare standards, health status, and age of the animals [1,2]. The magnitude of growth response was shown to be greater in weaning pigs than in older animals. For example, in a meta-analysis study, conducted by Tung and Pettigrew, the Animals 2020, 10, 134; doi:10.3390/ani10010134 www.mdpi.com/journal/animals www.mdpi.com/journal/animals 2 of 18 Animals 2020, 10, 134 improvements of growth rates were 12.25% and 6.03% for the ﬁrst 2 weeks or 4 weeks post-weaning respectively, while the enhancement was lower for growing (3.51%) or ﬁnishing (2.69%) pigs [2]. Even with a level of variability, all the literature available suggests that organic acids eﬀects on pig growth performance are consistent, thus supporting their use in practical conditions. From this starting point, the aim of the present work is to provide the biological rationale behind the use of organic acids in pig nutrition, focusing on the diﬀerent mode of action they can exert in the diet and then travelling along the gastro-intestinal tract of pigs. 2. Classiﬁcation of Organic Acids Organic acids can be classiﬁed in three main functional categories: short chain fatty acids ( medium chain fatty acids (MCFA), and tricarboxylic acids (TCA) [3]. SCFA are carboxylic acids with max 5 carbon atoms and they are produced in the lower intestine of animals by the microbial fermentation of indigestible sugars and amino acids. In particular, acetic, propionic, and butyric acid are produced in a physiological ratio (60:25:15), having a key role for intestinal mucosa trophism and general metabolism [4,5]. In swine and piglets, they improve intestinal morphology and have a beneﬁcial eﬀect on the intestinal barrier, decreasing intestinal inﬂammation [6,7]. Thanks to their liquid status, SCFA are mostly used as feed acidiﬁers, silage inoculants, and preservatives in animal nutrition. MCFA have aliphatic chains with 6 to 12 carbon atoms. These acids can be quickly incorporated in the membrane of phospholipids and they play a relevant role in the nutrition of young piglets, where they represent an important energy source [8,9]. Thanks to their high pKa MCFA have a higher antibacterial activity compared to SCFA in the hindgut [10]. TCA are metabolic intermediates of Krebs cycle, thus involved in energy metabolism. These acids improve gut morphology and barrier function, with a favorable inﬂuence on microbiota [11,12]. Other than these categories, there are a few organic acids like benzoic, sorbic, and lactic acid that are widely used in food and feed preservation thanks to their antifungal and antimould properties. When managing organic acids in animal nutrition there are several properties to consider, i.e., physical form, ﬂavoring properties like odor and taste, solubility in water and also the safety aspect. Table 1 shows the list of the most common organic acids and their properties, grouped by the aforementioned categories. Organic acids with the highest antimicrobial activity have a pKa between 3 and 5 [13]. It follows that there are diﬀerent organic acids that can act at diﬀerent levels on the basis of environmental pH: organic acids with higher pKa can be used with the aim of preserving food/feed, while, from a nutritional point of view, a lower pKa means that the acid acts in the stomach. 3 of 18 Animals 2020, 10, 134 Table 1. List of the most common acids and their properties. 2. Classiﬁcation of Organic Acids Category Acid Molecular Formula pKa Physical Form Odor and Taste Solubility Chemical Safety Short Chain Formic CH2O2 3.75 Colorless liquid Pungent odor Miscible in water, ether, acetone, ethyl acetate, methanol, ethanol Severe skin burns and eye damage Acetic C2H4O2 4.74 Colorless liquid Pungent vinegar-like odor Sour and burning taste Miscible in water, alcohol, glycerol, ether, carbon tetrachloride Flammable liquid and vapor, severe skin burns and eye damage Propionic C3H6O2 4.88 Colorless liquid oily Very pungent rancid odor Miscible in water, soluble in alcohol, ether, chloroform Severe skin burns and eye damage Butyric C4H8O2 4.82 Colorless liquid oily Rancid unpleasant odor, acrid taste, with a sweetish after taste Miscible in water, alcohol, ether Severe skin burns and eye damage Medium Chain Caproic (or hexanoic) C6H12O2 5.09 Colorless to light-yellow liquid oily Characteristic goat-like odor Soluble in ethanol and ether Toxic in contact with skin, severe skin burns and eye damage Caprylic (or octanoic) C8H16O2 4.89 Colorless to light-yellow liquid oily Characteristic goat-like odor Miscible in ethanol, chloroform, acetonitrile Soluble in alcohol, chloroform, ether, carbon disulﬁde petroleum ether and glacial acetic acid Severe skin burns and eye damage; harmful to aquatic life with long lasting eﬀects Capric (or decanoic) C10H20O2 4.90 White crystalline powder Characteristic goat-like odor Soluble in ethanol, alcohol, ether, chloroform, benzene and carbon disulﬁde Skin and eye irritation; harmful to aquatic life with long lasting eﬀects Lauric (or dodecanoic) C12H24O2 5.30 White ﬂakes Bay-like odor Miscible with benzene Very soluble in methanol and ethanol Soluble in acetone Skin irritation, serious eye damage and irritation TCA Citric C6H8O7 3.13 4.76 6.49 White or colorless crystalline powder Odorless Pleasant and sour taste Very soluble in water and ethanol Soluble in ether and ethyl acetate Skin irritation, serious eye damage and irritation α-ketoglutaric C5H6O5 2.47 4.68 White crystalline powder Odorless Soluble in water and alcohol Skin irritation, serious eye damage. May cause respiratory irritation Fumaric C4H4O4 3.02 4.76 White crystalline powder Odorless Fruit-like taste Soluble in ethanol and concentrated sulfuric acid Serious eye irritation Malic C4H6O5 3.40 5.10 Liquid or white crystalline powder Odorless Apple or tart taste Very soluble in methanol, ethanol, acetone, ether, and other polar solvents Harmful if swallowed; skin and eye irritation. May cause respiratory irritation Table 1. List of the most common acids and their properties. 4 of 18 Animals 2020, 10, 134 Table 1. Cont. 3. Antimicrobial Properties of Organic Acids Since many years, organic acids are aﬃrmed as powerful antimicrobials. Their activity against microorganisms is traditionally employed in food preservation, but today their role in animal nutrition has reached the same importance. Organic acids owe their success to a broad spectrum of applications, that are based on antibacterial, antiviral, antifungal, and antimold properties [14–16]. Not all organic acids have the same antimicrobial activity. Their action against microorganisms depends on the carbon chain length and degree of unsaturation, but overall the pKa of the acid inﬂuences its antimicrobial mechanism of action [17]. Indeed, every organic acid is distinguished by a pH value at which 50% of the acid is found in a dissociated form (pKa). In 1998 Russell and Diez-Gonzalez for the ﬁrst time proposed the anion model, according to which the inhibitory eﬀect of organic acids is highly related to their undissociated form [18]. Based upon the environmental pH and pKa values, organic acids in their undissociated form can diﬀuse across the bacterial cell membrane and dissociate inside the cell, releasing H+ ions and decreasing intracellular pH. To overcome the lowering of pH, microorganisms activate proton pumps consuming energy and, at the same time, the anion RCOO- is toxic to DNA replication, disrupting metabolic functions and increasing osmotic cell pressure [19–21]. The combination of these two actions inhibits bacterial replication and growth, leading to bacteriostatic or bactericidal eﬀects. The anion model is generally accepted as mode of action for all the organic acids, but the eﬃcacy of diﬀerent organic acids can vary mainly upon two factors: on one hand, the lipophilic nature of the acid inﬂuences the ability to pass through the microorganism wall; on the other hand, upon dissociation inside the cell, diﬀerent anions can have diﬀerent inhibitory mechanisms on cellular functions [22–24]. As already mentioned, there are many other variables that aﬀect the antimicrobial activity of organic acids besides the pKa. Polar groups, number of double bonds, molecular size, and solubility in non-polar solvents are the four principal chemical and physical features that can predict the inhibitory eﬀect of organic acids, as described by the mathematical model of Principal Components Analysis calculated by Hsiao and Siebert [25]. The spectrum of eﬃcacy of organic acids can vary depending on the nature of the target organism and, in particular, on the complexity and structure of its outer cell wall and/or membrane [26]. 4. Targeting the Feed and the Stomach In pig nutrition, organic acids are traditionally used as “acidiﬁers” with two targets, the feed and the stomach where the main aim is to reduce the pH. 2. Classiﬁcation of Organic Acids Category Acid Molecular Formula pKa Physical Form Odor and Taste Solubility Chemical Safety Others Sorbic C6H8O2 4.76 White crystalline powder or granules Odorless Acrid and sour taste Very soluble in ether. Soluble in ethanol Skin and eye irritation. May cause respiratory irritation Benzoic C7H6O2 4.19 Colorless crystalline powder Pungent odor Bitter taste Soluble in alcohol, ether, and benzene Skin irritation, eye damage; damage to organs through prolonged or repeated exposure Lactic C3H6O3 3.86 Colorless to yellow viscous liquid or crystals Odorless Acrid taste Soluble in water, ethanol and diethyl ether Skin irritation, serious eye damage 5 of 18 Animals 2020, 10, 134 3. Antimicrobial Properties of Organic Acids Gram positive bacteria (i.e., Clostridium perfringens, Enterococcus spp., Streptococcus spp.) are mainly susceptible to MCFA, while Gram-negative bacteria (i.e., Escherichia coli, Campylobacter jejuni, Salmonella spp.) are more sensitive to SCFA [27,28]. This can be explained by the lipophilic nature of MCFA that allows them to have a stronger antibacterial activity mainly against Gram-positive species, whereas the presence of lipopolysaccharide (LPS) in the Gram-negative cell wall confers resistance to these species [29]. Propionic acid and butyric acid are strong mold inhibitors, while acetic acid is commonly used as antifungal, also reducing aﬂatoxins production [30–32]. 4.2. Eﬀects in the Stomach At the gastric level, the primary goal of organic acids is the reduction of pH, key-point particularly for young, newly-weaned animals. At this age the gastro-intestinal tract is not fully developed and stomach pH tends to be high (often over 5) due to a combination of poor endogenous HCl secretion, lack of lactic acid from lactose fermentation, and intake of large meals at infrequent intervals [6]. High pH ultimately impairs pepsin activation and function, that is optimal in acidic environment at pH 2 to 3.5, therefore strongly reducing the eﬃciency of protein digestion [36]. There is also evidence that high pH also increases the rate of gastric emptying thus reducing the time for the feed to be digested in the stomach [37]. Beside these physiological aspects, gastric pH of weaning piglets is kept high also by the high buﬀering capacity of the diet, particularly determined by some ingredients such as milk products, vegetable proteins, and meat/ﬁsh meal, so that acidiﬁcation can help the transition from milk to solid feed. Indeed, to overcome all these issues related to high pH, the inclusion of organic acids is aimed at lowering stomach pH. However, data showed inconsistent results about gastric pH modulation: while some studies reported a signiﬁcant reduction, others found no eﬀects on gastric pH following dietary acidiﬁcation. In a meta-analysis collecting data from 22 studies using individual organic acids at high doses (1% to 2%), stomach pH was lower in 55% of the cases, higher in 36% and equal in 9% of the cases for acidiﬁed diets compared to control [2]. More recently, no relevant eﬀects on gastric pH were reported for diﬀerent combinations of organic acids, mainly fumaric acid-based blends, MCFA, or mixtures of organic acids and MCFA fed to weaning pigs [38–41]. In a study conducted by Zentek and coworkers, a fumaric and lactic acid mixture did not have any eﬀect, whereas a blend of caprylic and capric acid reduced pH in pylorus, but not in cardia and fundus of the stomach [42] These inconsistencies among diﬀerent studies may be due to the fact that gastric contents are heterogeneous, and that there are diﬀerent pH values in diﬀerent stomach regions. In addition, the time of sampling can have an eﬀect on pH values as gastric pH is subjected to ﬂuctuations over time after feeding [43]. 4.1. Eﬀects in the Feed Organic acids are included in the feed to prevent spoilage and improve the hygienic quality, with particular reference to undesirable molds and bacterial growth inhibition. For this purpose, in the European Union several organic acids such as formic and propionic acids, as well as others (lactic, citric, fumaric, and sorbic acid) and their salts (e.g., calcium formate, calcium propionate) are authorized as feed additives in the category named “feed preservatives” [33]. A great variety of molds inhibitors are available on the market, from single acids in liquid form to blends of diﬀerent organic acids and their salts on carriers, with diﬀerent applications. In general, buﬀered products such as acids salts are preferred over acids for their higher ease in handling and lower corrosiveness [34]. Beside the Animals 2020, 10, 134 6 of 18 antimold eﬀects, organic acids can act as antimicrobials primarily by reducing the pH of the feed, so altering the growth conditions for microorganisms, and secondly by a direct inhibition on the growth of speciﬁc bacteria. The eﬀect of organic acids on diet pH has been thoroughly reviewed by Tung and Pettigrew that, in a meta-analysis, showed an appreciable reduction of diet pH (from 5.95 to 4.71) due to the addition of acids [2]. Moreover, organic acids can help in reducing the growth of speciﬁc pathogenic bacteria (such as Salmonella spp.) in the feed and therefore the risk of microbial contamination for the animal, gaining importance from a food safety perspective. In this regard, several studies investigated the eﬀects of acids either as individual components or as mixtures of acids on feed materials artiﬁcially contaminated with Salmonella. The eﬃcacy of organic acids in reducing Salmonella counts was variable upon various factors such as the type of acid, the nature of acid (free or salt), the inclusion dose, the bacterial load, and the feed ingredients. Interestingly, the lowest and slowest reduction of Salmonella with acidiﬁcation was observed in soybean meal (47.8% of crude protein), suggesting a protective eﬀect on Salmonella cells for protein-rich matrices [35]. 4.2. Eﬀects in the Stomach Interestingly, formic acid fed at 1.8% to weaning pigs kept gastric pH at values below 3 for several time-points after feeding, in particular it prevented the physiological rise in pH observed in control animals after the meal. However, when considering the mean of all sampling times from 0.5 to 8.5 h post-feeding, the gastric pH value was not aﬀected by formic acid [44]. This result shows that formic acid at high doses can help in counteract the buﬀering capacity of the diet but, at the same time, it suggests that sampling time can often explain the diﬀerent outcome between diﬀerent studies. Although with inconsistent eﬀects on gastric pH, organic acids were shown to positively aﬀect nutrient digestibility especially fecal dry matter (+0.82%) and crude protein digestibility (+1.33%) with the response being variable upon acid type, acid dose, and diet type. Improvements were generally higher for diets formulated with plant and animal origin feed ingredients compared to 7 of 18 Animals 2020, 10, 134 only plant-based diets [2]. Regarding MCFA, caprylic and capric acids, both individually and in combination, increased total tract protein and ﬁber digestibility in weaning pigs [39]. Other studies did not observe a beneﬁcial eﬀect on nutrient digestibility (either ileal or total tract) by supplementing weaning diets with either fumaric, formic, or lactic acid individually or with mixtures of various organic acids [45–47]. It is noteworthy that fumaric acid, included to a diet with low buﬀering capacity, increased the ileal digestibility of crude protein, gross energy and the majority of amino acids, but these beneﬁcial eﬀects were reduced in case of a high buﬀering capacity diet [48]. Regarding growing pigs, positive eﬀects on the ileal digestibility of several amino acids have been reported for propionic and lactic acid, whereas commercial blends of organic acids and MCFA showed higher apparent total tract digestibility of gross energy, dry matter, organic matter, and crude ﬁber without aﬀecting amino acids ileal digestibility [49–51]. It is important to point out that these digestibility trials, both ileal or total tract, provide estimates of the amount of nutrients digested and absorbed within the GIT, therefore the positive impact of organic acids can be explained by a promotion of either gastric digestive function or intestinal absorption, two aspects not easily discernible analytically. 4.2. Eﬀects in the Stomach In addition, various organic acids (such as citric, formic, fumaric, and lactic acid) may improve the absorption and retention of some minerals including calcium, phosphorus, magnesium, and zinc, thus increasing their biological value and limiting their excretion. More precisely, citric acid seems to increase calcium and phosphorus absorption, likely by chelating calcium and making the phytate structure less stable and more accessible to phytase action [2,52,53]. Low pH in the stomach is beneﬁcial for proper enzymatic activities and protein digestion but, at the same time, can select acid tolerant undesirable microorganisms. Bearson et al. reviewed that enterobacteria (including S. typhimurium and E. coli) have evolved inducible acid survival strategies to face acid stress conditions characterized by low pH and high concentration of weak organic acids [54]. These systems include the activation of acid shock proteins and inducible decarboxylases that consume protons for decarboxylation reactions and exchange end-products for new substrates thanks to membrane transporters [55,56]. Furthermore, it has been shown that some SCFA (such as formic and acetic acid) may serve as environmental signals to trigger the expression of invasion genes in Salmonella [57,58]. Therefore, an early exposure of pathogens to acidic conditions in the stomach can prepare them to endure stress conditions more distally in the small intestine and intensify their pathogenic action. 5. Targeting the Intestine Beyond the stomach, organic acids are meant to work along the intestine where most of the microbial colonization/proliferation occur and their antimicrobial action is needed. In addition, organic acids can have a metabolic eﬀect serving as energy source for the intestinal mucosa and overall modulate general metabolism. 5.2. Modulation of Fermentation Patterns Organic acids have been studied not only for their antimicrobial activity against speciﬁc bacterial strains, but also for their ability to modulate intestinal fermentation patterns. Some studies from our research group evaluated the eﬀects of diﬀerent organic acids on microbial growth and ammonia production by pig cecal microﬂora, using an in vitro fermentation system with fresh cecal content as bacterial inoculum. Various organic acids were tested (among which SCFA, Krebs cycle acids, and preservatives) showing that diﬀerent types and concentrations of organic acids can either inhibit (e.g., sorbic, benzoic, formic acid) or enhance (e.g., lactic, citric, fumaric acid) the bacterial fermentation, at the same time reducing ammonia production (particularly for sorbic, benzoic, and alpha-ketoglutaric acid) [71]. Furthermore, in the same system, a blend of Krebs cycle acids (citric, fumaric, malic acid plus phosphoric acid) was eﬀective in reducing both ammonia and the concentration of iso-acids such as iso-butyric and iso-valeric, thus indicating a strong reduction of microbial proteolysis, at the same time reducing also cellulolytic activity as shown by lower acetic acid levels [72]. Partanen and Jalava developed a diﬀerent in vitro fermentation model using frozen ileal digesta plus fresh fecal samples as bacterial inoculum: among others, formic acid, potassium sorbate, and sodium benzoate were the most eﬀective in modulating bacterial fermentation, by reducing total volatile fatty acids (VFA), acetic acid, and propionic acid concentrations compared to control, although not aﬀecting ammonia levels [73]. It is noteworthy to point out that these in vitro studies are strictly dependent on the bacterial inoculum and its source (intestinal tract vs. fecal, fresh vs. frozen) that can likely explain some diﬀerent ﬁnal outcomes. Moreover, as additional drawback, the batch fermentation set-up diﬀers from in vivo conditions primarily because microbial end-products accumulate in the medium due to lacking physiologic digesta ﬂow and mucosa absorption. However, such systems can be useful to screen the eﬀects of potential feed additives, among which organic acids, on intestinal fermentation patterns or to assess the fermentability of feed ingredients before in vivo studies [74]. In vivo, the eﬀects of dietary organic acids on fermentation patterns along the intestine appeared to be more variable, as previously reviewed [6,52]. Formic acid was one of the most studied and in some cases doses in the range of 0.3% to 3% did not alter ammonia and VFA production along the GIT [75–77]. 5.1. Modulation of Microﬂora Based on their antimicrobial mode of action, organic acids are able to inhibit the growth of undesired pH-sensitive microorganisms like Enterobacteriaceae, while they do not aﬀect beneﬁcial lactic acid bacteria. In this view, organic acids can be an optimal tool to control the dysbiosis characterized by coliforms overgrowth and Lactobacilli depression, typical of weaning [59]. Many studies investigated the eﬀects of dietary acidiﬁers on microbial populations along the digestive tract of weaning piglets with dissimilar results, mainly dependent on the acid used (type and dose) and the intestinal tract analyzed. For example, Grecco et al. reported that fumaric acid at 0.8% reduced total coliforms and E. coli numbers in the cecum of weaning pigs, whereas in other trials no eﬀects on microﬂora all along the gastro-intestinal tract were observed with 1.5% fumaric acid [40,60]. Also citric acid at 1.5% did not show signiﬁcant eﬀects on microﬂora populations (total anaerobes, Lactobacilli, Clostridia, E. coli) or incidence of scouring after an E. coli challenge in weaning pigs [60,61]. However, in weaned pig challenged with both S. typhimurium and E. coli, 0.5% of citric acid decreased both Salmonella and E. coli 8 of 18 Animals 2020, 10, 134 counts in fecal samples, while increasing Lactobacilli number [62]. Formic acid included at doses ≥1%, either as pure acid or salts, was generally eﬀective in decreasing coliforms along the gastro-intestinal tract, being the microbial changes more evident in the small intestine than in the lower gut in some cases [44,63–66]. These inconsistent eﬀects along the GIT are likely related to the lack of suﬃcient amounts of undissociated acids after the stomach, ﬁnally impairing organic acids antimicrobial action. Indeed, in many cases organic acids supplemented with the diet were recovered at concentrations higher than unsupplemented control only in the stomach and in the proximal small intestine then disappeared in the distal small intestine and large intestine contents [42,60,67,68]. Regarding MCFA, caprylic acid, and to a lesser extent capric acid at 0.3%, allowed to reduce E. coli counts both in jejunum and cecum digesta of weaning pigs [69]. Blends of organic acids and MCFA showed variable results either reducing E. coli and increasing microbial diversity in the colon or not aﬀecting at all large intestine or fecal microﬂora [40,41,70]. 5.2. Modulation of Fermentation Patterns However, in other studies formic acid or its salts increased acetic acid and decreased lactic acid concentrations in both ileum and cecum-colon contents [44,68,78,79]. These ﬁndings can indicate a shift in the composition of intestinal ﬂora and a modulation of microbial fermentations with more nutrients (like glucose not fermented into lactic acid) or metabolites (like acetate) available to the host [79]. Supplementation of propionic acid-based products to weaning pigs’ diets showed inconsistent eﬀects, generally not aﬀecting pH, ammonia, SCFA, and microﬂora populations along GIT, but sometimes decreasing E. coli counts in the stomach and increasing Lactobacilli number in the duodenum compared to controls [77,80,81]. Concerning Krebs cycle acids, the addition of either citric or fumaric acid at 1.5% did not aﬀect pH, VFA concentrations, or microﬂora populations along the GIT of weaning pigs [60,61]. Zentek and 9 of 18 Animals 2020, 10, 134 colleagues investigated the eﬀects of combinations of organic acids (fumaric and lactic acid) and MCFA (capric and caprylic acid) fed to weaning piglets ﬁnding only minor changes on the gastro-intestinal ecology and microbial metabolites production. Nevertheless, in the colon digesta an interesting reduction of some E. coli pathogenicity factors was observed particularly for organic acids that may be relevant to support intestinal health at weaning [42]. More recently, diﬀerent commercial blends of organic acids including SCFA, MCFA, and acids salts were shown to increase the levels of acetic, propionic, and butyric acid produced by microbial fermentation of carbohydrates in the large intestine, while reducing E. coli counts [41]. In addition, other mixtures of organic acids (including MCFA) fed to growing pigs increased VFA content in the ileum and ﬁber fecal digestibility, indicating a positive stimulation on microﬂora that can utilize carbohydrates in the large intestine and produce SCFA [51]. Overall, organic acids have the potential to modulate the microﬂora populations and consequently microbial metabolites production along the GIT. The modulation of VFA can be of particular interest because of their pivotal role for the intestinal physiology and metabolism, as presented in the following paragraph. Nevertheless, these results appear to be often inconsistent and it is quite diﬃcult to compare the experimental studies because of the employment of diﬀerent organic acids (pure acids or salts, individually or in variable combinations) but also because of additional factors such as the intestinal tract analyzed and, among all, the diet composition. 5.2. Modulation of Fermentation Patterns Furthermore, organic acids frequently appear as not eﬀective, even when they are used at high doses, likely due to their variable ability to reach the intestinal sites at suﬃcient levels for early dissociation or absorption in the stomach. 5.3. Metabolic Eﬀect-Trophic Factors A metabolic role in the intestinal tract can be notably attributed to SCFA (mainly acetic, propionic, and butyric acid) produced by microbial fermentations and physiologically relevant for the large intestine, but also to TCA involved in Krebs cycle. 5.3.1. Short Chain Fatty Acids SCFA are end-products of microbial fermentations of carbohydrates occurring mostly in the large intestine of hindgut fermenter animals. In the large intestine of pigs, SCFA are usually found at a molar ratio of approximately 60:25:15 for acetic:propionic:butyric and they are rapidly absorbed by the colonic epithelium to play a pivotal role in the intestinal and general energy metabolism. Butyric acid is almost completely oxidized within the mucosa, serving as preferred energy source for the colonocytes, whereas propionic acid is collected by the liver where it is converted into glucose, and acetic acid is instead used by peripheral tissues [4,82]. The metabolic eﬀects of supplementing diets for pigs with SCFA have been variably investigated. Literature about propionic and acetic acid in pigs is quite scarce although a possible role for acetic acid in improving protein metabolism through a nitrogen-sparing eﬀect was suggested in a single study with growing pigs but then not conﬁrmed under practical conditions [83,84]. For butyric acid most of the attention was given to its impact on the intestinal mucosa, not only for its role as trophic factor but also for its multi-functional properties such as anti-inﬂammatory, anti-oxidant and protective for the epithelial integrity that can be useful to enhance intestinal health [85–87]. Addition of butyric acid to diets for weaning piglets has been reported to positively aﬀect the gut morphology, favor gastric mucosa development, reduce the inﬂammatory stress and improve immune status and intestinal barrier function [88–91]. However, the ﬁnal outcome in terms of growth performance for dietary butyric acids was not always consistent and highly dependent on several factors such as dose, supplementation duration, and age of piglets [92]. Other studies investigated the impact of butyric acid triglyceride, named tributyrin, as a tool to provide more butyric acid and enhance its eﬀects. Dietary tributyrin combined with lactitol, a fermentable source of butyrate, improved the intestinal trophism of weaning piglets, as indicated by longer villi, shorter crypts, higher polyamines concentration and increased intestinal enzymes activities [93–95]. Additionally, tributyrin was particularly eﬀective in supporting the intestinal development and growth of intrauterine growth restriction (IGR) piglets 10 of 18 Animals 2020, 10, 134 namely piglets born with lower body weight than normal neonates. Tributyrin not only improved intestinal morphology, barrier function, and digestive activities but also attenuated lipid metabolism dysfunctions usually associated with IGR piglets, thus suggesting a role for TB in modulating general metabolism [96,97]. 5.3.2. Krebs Cycle Acids TCA involved in Krebs cycle contribute to energy metabolism by deﬁnition. Some of them, such as citric and fumaric acid, are commonly used in pig nutrition as acidiﬁers and it is interesting to understand whether they can actually have a metabolic role at the intestinal level beyond their acidic properties. p p The intestinal absorption of citric and fumaric acid was studied using brush border membrane vesicles isolated from pig proximal jejunum: citric and fumaric acid uptake was shown to occur through a Na-dependent co-transporter that is common for other TCA [99]. As demonstrated in hamster small intestine, TCA following the uptake at the apical side are metabolized inside the enterocyte and then leave the basolateral side partly as such and partly as metabolites. The extent of metabolism inside the cell appeared to be 60%–70% of the absorbed acid, with a 30%–40% eﬀectively being transported to the serosa side [100]. More recently, studies with pigs showed that alpha-ketoglutaric acid (AKG) is better absorbed in the small intestine than in distal sites, then quickly metabolized in the enterocytes and liver so that half-life in blood is very short [101,102]. These results provide evidence that dietary Krebs cycle acids can be metabolized within the enterocytes, therefore directly inﬂuencing intestinal metabolic status. It has been proposed that the local trophic eﬀect on the small intestine mucosa can explain the positive eﬀects of fumaric acid in terms of improved digestive and absorptive functions [48]. Moreover, Kirchgessner and Roth suggested that ingested fumarate is metabolized in Krebs cycle as well as fumaric acid produced by metabolic routes and, additionally, pigs can utilize it as energy source with an eﬃciency close to that of glucose [103]. More recently, succinate was shown to improve intestinal morphometry (longer villi, shorter crypts) in growing pigs and to enhance the barrier functionality of pig intestinal cells in vitro, by increasing tight junction markers’ expression [12]. Beyond the intestine, the eﬀects of dietary supplementation with citric and fumaric acid have been studied mainly on intermediary liver metabolism in rats: the activity of Krebs cycle enzymes was generally not aﬀected but liver transaminases (glutamate dehydrogenase, glutamate oxaloacetate transaminase, and glutamate pyruvate transaminase) were increased following citric/fumaric acid supplementation, suggesting that their carbon skeleton could have been used to synthesize non-essential amino acids [84]. 5.3.1. Short Chain Fatty Acids In this line, recent studies by Murray and colleagues demonstrated that tributyrin fed to neonatal piglets can promote muscle growth by directly increasing the myogenic potential of muscle satellite cells, therefore appearing a promising nutritional tool for proper muscle development and pig growth [98]. p g g Overall, dietary SCFA and especially butyric acid can exert a metabolic eﬀect not limited to the intestinal site but also involving general metabolism, with positive consequences on health and growth. 6. Use of Protected vs. Free Organic Acids In the last decade, the use of organic acids “protected” using diﬀerent coating and microencapsulation techniques has become well established. The protection of organic acids is of great interest for the pig nutrition industry because it allows various technological advantages: it improves handling and safety, increases stability, reduces dustiness and corrosiveness problems, and prevents damages due to temperature/pressure during processing, as well as undesirable interactions with other ingredients [120]. Moreover, microencapsulation allows to increase palatability by masking unpleasant smell/taste and, even more importantly, to deliver the ingredients to speciﬁc target sites of action inside the animal, providing therefore also “biological” advantages. Studies from our research group demonstrated that lipid microencapsulation of organic acids (i.e., the inclusion in a lipid matrix) prevents the rapid disappearance of organic acids right after the stomach and, conversely, allows them to be slowly released along the small and large intestine of pigs, thanks to the action of intestinal lipases [121]. As a consequence of their targeted delivery, lipid encapsulated organic acids were eﬀective in exerting their antimicrobial action along the intestine, by reducing coliforms counts both in distal jejunum and cecum, whereas free organic acids did not have any eﬀect [121]. Additionally, comparing free and protected organic acids in weaned pigs, lipid microencapsulation allowed to reduce the eﬀective dose of organic acids included in the diet by 10 fold compared to the non-encapsulated organic acids, still maintaining the same performance results [122]. Currently, the feed additive industry oﬀers a wide range of organic acids-based products, protected or encapsulated in diﬀerent forms, containing either individual organic acids, mixtures of organic acids, or blends of organic acids coupled with other active compounds. Particularly interesting and eﬃcacious is the combination of organic acids with botanicals, as they can interact positively and exert antimicrobial properties in a synergistic way: botanical aromatic compounds, as pore-forming agents, can alter the bacterial membrane, eventually facilitating the organic acids entrance and therefore the antimicrobial action [3]. Some in vitro studies demonstrated the synergy between organic acids and pure botanicals, showing that their combination is more eﬀective than each compound alone against food-safety related bacteria such as S. typhimurium and C. jejuni [3,123]. Again, a combination of organic acids and botanicals in pigs was eﬀective in reducing S. typhimurium shedding throughout a production cycle from weaning to slaughter [124]. 5.3.2. Krebs Cycle Acids Alpha-ketoglutaric acid is very attractive for its multiple role in cell metabolism, especially as bridge between energy and protein metabolism, but also precursor for the synthesis of glutamate and glutamine [104]. The speciﬁc role of AKG at the intestinal level has been already thoroughly reviewed with particular reference to its positive eﬀects on intestinal morphology, anti-oxidative capacity, and absorptive and barrier function [105,106]. In weaning piglets challenged with E. coli LPS, dietary AKG supplementation improved both small intestine and liver morphology and function thus ameliorating the endotoxin-induced injury [107–109]. Moreover, the addition of AKG to a low-protein diet in pigs allowed to improve nitrogen, Ca and P metabolism and promote both protein and lipid metabolism in skeletal muscle (favoring intramuscular fat and monounsaturated fatty acids content), thus being a promising nutritional strategy to support pig productivity and supply high-quality pork [110–112]. Animals 2020, 10, 134 11 of 18 11 of 18 5.3.3. Others Also, sorbic and benzoic acid, commonly used as feed preservatives for their antimicrobial properties, can modulate general metabolism with diﬀerent modes of action. Sorbic acid is metabolized in the body like other fatty acids, being subject to beta-oxidation and serving as energy source. Studies with murine models showed that sorbic acid is rapidly absorbed and completely oxidized to H2O and CO2 [113,114]. In addition, sorbic acid added to weaning piglets increased growth performance through the modulation of lipid metabolism and the enhancement of insulin-like growth factor system, particularly relevant for the GIT development [115,116]. On the other hand, benzoic acid is almost completely excreted in urine as its glycine conjugate, hippuric acid, in pigs [117]. Benzoic acid has been reported to reduce urine pH and ammonia emissions from pig slurry, thus with positive consequences in terms of environmental impact of pig production in commercial settings [118,119]. 7. Conclusions Based on the literature review, the key points about organic acids in pig nutrition can be summarized as follows: • The positive eﬀects on growth performance are consistent and largely supported by the extensive use of organic acids in practical conditions. Instead, the impact in terms of reduction of gastric pH, improvement of protein digestibility, and modulation of microﬂora is more variable and generally achieved including organic acids in the diet at high doses (above 1%). • Most of the published studies are quite dated and involve the use of formic, fumaric, and citric acid reﬂecting their use as feed additives, well-established, over the years. More recently, combinations of organic acids and MCFA have been thoroughly tested and proposed on the market with positive results. • More than acidiﬁers, organic acids can have a metabolic role, by improving the intestinal mucosa trophism and modulating general metabolism, particularly for SCFA and Krebs cycle acids. Butyric acid-based products have a strong history of use, while alpha-ketoglutaric acid is very attractive for future applications. • Protecting organic acids with encapsulation techniques can provide both technological and biological advantages, target-delivering organic acids along the intestine where they can exert their beneﬁcial eﬀects. In conclusion, organic acids have multiple properties that can be useful to support health and growth of pigs and they continue to be pivotal feed additives in pig nutrition. In conclusion, organic acids have multiple properties that can be useful to support health and growth of pigs and they continue to be pivotal feed additives in pig nutrition. Author Contributions: All of the authors contributed in an equal way to the realization of this paper. All authors have read and agreed to the published version of the manuscript. Author Contributions: All of the authors contributed in an equal way to the realization of this paper. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Funding: This research received no external funding. Funding: This research received no external funding. Conﬂicts of Interest: Andrea Piva serves as professor at the University of Bologna and is a member of the board of directors of Vetagro S.p.A. 6. Use of Protected vs. Free Organic Acids Beside the promising application of organic acids and botanicals for foodborne pathogens control, selected combinations of organic acids such as citric and sorbic acid together with monoterpenes were shown to exert a direct eﬀect on the intestinal mucosa, by improving the barrier integrity of Caco-2 intestinal cells, without the mediation of the microﬂora. When these active ingredients were encapsulated in a lipid matrix and fed to weaning pigs, improved barrier functionality in jejunum and ileum and reduced local and systemic inﬂammatory pressure were observed and then translated in better growth performance [125]. 12 of 18 Animals 2020, 10, 134 Without claiming to be exhaustive, the above-mentioned examples suggest that microencapsulation can confer an additional value to organic acids and that the combination with other bioactive molecules such as botanicals can further enhance their antimicrobial action and allow a general intestinal health-promoting eﬀect. References 1. Mroz, Z.; Koopmans, S.J.; Bannink, A.; Partanen, K.; Krasucki, W.; Overland, M.; Radcliﬀe, S. Carboxylic acids as bioregulator and gut growth promoters in nonruminant. In Biology of Nutrition in Growing Animals; Mosentin, R., Zentek, J., Zebrowska, T., Eds.; Elsevier Limited: Amsterdam, The Netherlands, 2006; Volume 4, pp. 81–133. 1. Mroz, Z.; Koopmans, S.J.; Bannink, A.; Partanen, K.; Krasucki, W.; Overland, M.; Radcliﬀe, S. Carboxylic acids as bioregulator and gut growth promoters in nonruminant. In Biology of Nutrition in Growing Animals; Mosentin, R., Zentek, J., Zebrowska, T., Eds.; Elsevier Limited: Amsterdam, The Netherlands, 2006; Volume 4, pp. 81–133. pp Tung, C.M.; Pettigrew, J.E. Critical Review of Acidiﬁers; National Pork Board: Des Moines, IA, USA, 2006. g g f ﬁ 3. Grilli, E.; Piva, A. Organic acids and their role in reduce foodborne pathogens in food animals. In On-Farm Strategies to Control Foodborne Pathogens; Callaway, T.R., Edrington, T.S., Eds.; Nova Science Pub. Inc.: Hauppauge, NY, USA, 2012; pp. 183–210. 3. Grilli, E.; Piva, A. Organic acids and their role in reduce foodborne pathogens in food animals. In On-Farm Strategies to Control Foodborne Pathogens; Callaway, T.R., Edrington, T.S., Eds.; Nova Science Pub. Inc.: Hauppauge, NY, USA, 2012; pp. 183–210. pp g pp 4. Cummings, J.H. Short chain fatty acids in the human colon. Gut 1981, 22, 763–779. [CrossRef] pp g pp 4. Cummings, J.H. Short chain fatty acids in the human colon. Gut 1981, 22, 763–779. [CrossRef] 5. Mroz, Z. Organic acids as potential alternatives to antibiotic growth promoters for pigs. Adv. Pork Prod. 2005, 16, 169–182. 5. Mroz, Z. Organic acids as potential alternatives to antibiotic growth promoters for pigs. Adv. Pork Prod. 2005, 16, 169–182. 6. Suiryanrayna, M.V.; Ramana, J.V. A review of the eﬀects of dietary organic acids fed to swine. J. Anim. Sci. Biotechnol. 2015, 6, 45–56. [CrossRef] 6. Suiryanrayna, M.V.; Ramana, J.V. A review of the eﬀects of dietary organic acids fed to swine. J. Anim. Sci. Biotechnol. 2015, 6, 45–56. [CrossRef] 7. Diao, H.; Jiao, A.R.; Yu, B.; Mao, X.B.; Chen, D.W. Gastric infusion of short-chain fatty acids can improve intestinal barrier function in weaned piglets. Genes Nutr. 2019, 14, 4–29. [CrossRef] 8. Odle, J. New insights into the utilization of medium-chain triglycerides by the neonate: Observations from a piglet model. J. Nutr. 1997, 127, 1061–1067. [CrossRef] 13 of 18 13 of 18 Animals 2020, 10, 134 9. References Zentek, J.; Buchheit-Renko, S.; Ferrara, F.; Vahjen, W.; Van Kessel, A.G.; Pieper, R. Nutritional and physiological role of medium-chain triglycerides and medium-chain fatty acids in piglets. Anim. Health Res. Rev. 2011, 12, 83–93. [CrossRef] Gupta, R.C.; Srivastava, A.; Lall, R. Nutraceuticals in Veterinary Medicine; Springer International Publishing Cham, Switzerland, 2019. 11. Chen, J.; Kang, B.; Jiang, Q.; Han, M.; Zhao, Y.; Long, L.; Fu, C.; Yao, K. Alpha-Ketoglutarate in low-protein diets for growing pigs: Eﬀects on cecal microbial communities and parameters of microbial metabolism. Front. Microbiol. 2018, 9, 1057. [CrossRef] 12. Li, X.; Mao, M.; Zhang, Y.; Yu, K.; Zhu, W. Succinate modulates intestinal barrier function and inﬂammation response in pigs. Biomolecules 2019, 9, 486. [CrossRef] 13. Dibner, J.; Buttin, P. Use of organic acids as a model to study the impact of gut microﬂora on nutrition and metabolism. J. Appl. Poult. Res. 2002, 11, 453–463. [CrossRef] 14. Cochrane, R.; Dritz, S.; Woodworth, J.; Huss, A.; Stark, C.; Saensukjaroenphon, M.; DeRouchey, J.; Tokach, M.; Goodband, R.; Bai, J.; et al. Assessing the Eﬀects of Medium Chain Fatty Acids and Fat Sources on Porcine Epidemic Diarrhea Virus Viral RNA Stability and Infectivity. Kans. Agric. Exp. Stn. Res. Rep. 2016, 2, 1. [CrossRef] 15. Balti´c, B.; Starˇcevi´c, M.; Djordjevi´c, J.; Mrdovi´c, B.; Markovi´c, R. Importance of medium chain fatty acids in animal nutrition. In IOP Conference Series: Earth and Environmental Science; IOP Publishing: Bristol, UK, 2017; Volume 85, pp. 012048–012054. 16. Leyva Salas, M.; Mounier, J.; Valence, F.; Coton, M.; Thierry, A.; Coton, E. Antifungal Microbial Agents for Food Biopreservation—A Review. Microorganisms 2017, 5, 37. [CrossRef] 17. Huyghebaert, G.; Ducatelle, R.; Van Immerseel, F. An update on alternatives to antimicrobial growth promoters for broilers. Vet. J. 2011, 187, 182–188. [CrossRef] 18. Russell, J.B.; Diez-Gonzalez, F. The eﬀects of fermentation acids on bacterial growth. Adv. Microb. Physiol. 1998, 39, 205–234. 19. Holyoak, C.D.; Stratford, M.; McMullin, Z.; Cole, M.B.; Crimmins, K.; Brown, A.J.; Coote, P.J. Activity of the plasma membrane H (+)-ATPase and optimal glycolytic ﬂux are required for rapid adaptation and growth of Saccharomyces cerevisiae in the presence of the weak-acid preservative sorbic acid. Appl. Environ. Microbiol. 1996, 62, 3158–3164. [CrossRef] 20. Lambert, R.J.; Stratford, M. Weak-acid preservatives: Modelling microbial inhibition and response. J. Appl. Microbiol. 1999, 86, 157–164. [CrossRef] 21. Cetin-Karaca, H. Evaluation of Natural Antimicrobial Phenolic Compounds against Foodborne Pathogens. Master’s Thesis, University of Kentucky, Lexington, KY, USA, 2011. 22. References BMC Vet. Res. 2013, 9, 81–89. [CrossRef] 36. Kim, Y.Y.; Kil, D.Y.; Oh, H.K.; Han, I.K. Acidiﬁer as an alternative material to antibiotics in animal feed. Asian Australas J Anim Sci 2005 18 1048 1060 [CrossRef] g , , [ ] 36. Kim, Y.Y.; Kil, D.Y.; Oh, H.K.; Han, I.K. Acidiﬁer as an alternative material to antibiotics in animal feed. Asian-Australas. J. Anim. Sci. 2005, 18, 1048–1060. [CrossRef] 37. Mayer, E.A. The physiology of gastric storage and emptying. In Physiology of the Gastrointestinal Tract, 3rd ed.; Johnson, L.R., Alpers, D.H., Christensen, J., Jacobsen, E., Eds.; Lippencott Raven Press: New York, NY, 1994; Volume 1, pp. 929–976. 38. Gomes, F.E.; Fontes, D.O.; Vasconcellos, C.H.F.; Silva, F.C.O. Fumaric acid associated with lactic acid or calcium propionate for weaning piglet diets. Arq. Bras. Med. Veterinária Zootec. 2011, 63, 678–686. [CrossRef] 38. Gomes, F.E.; Fontes, D.O.; Vasconcellos, C.H.F.; Silva, F.C.O. Fumaric acid associated with lactic acid or calcium propionate for weaning piglet diets. Arq. Bras. Med. Veterinária Zootec. 2011, 63, 678–686. [CrossRef] 39. Hanczakowska, E.; Szewczyk, A.; Oko´n, K. Eﬀects of dietary caprylic and capric acids on piglet performance 39. Hanczakowska, E.; Szewczyk, A.; Oko´n, K. Eﬀects of dietary caprylic and capric acids on piglet performance and mucosal epithelium structure of the ileum. J. Anim. Feed Sci. 2011, 20, 545–554. [CrossRef] 40. Grecco, H.A.T.; Amorim, A.B.; Saleh, M.A.D.; Tse, M.L.P.; Telles, F.G.; Miassi, G.M.; Pimenta, G.M.; Berto, D.A. Evaluation of growth performance and gastro-intestinal parameters on the response of weaned piglets to dietary organic acids. An. Acad. Bras. Ciências 2018, 90, 401–414. [CrossRef] 41. Li, S.; Zheng, J.; Deng, K.; Chen, L.; Zhao, X.L.; Jiang, X.; Fang, Z.; Che, L.; Xu, S.; Feng, B.; et al. Supplementation with organic acids showing diﬀerent eﬀects on growth performance, gut morphology, and microbiota of weaned pigs fed with highly or less digestible diets. J. Anim. Sci. 2018, 96, 3302–3318. [CrossRef] 42. Zentek, J.; Ferrara, F.; Pieper, R.; Tedin, L.; Meyer, W.; Vahjen, W. Eﬀects of dietary combinations of organic acids and medium chain fatty acids on the gastrointestinal microbial ecology and bacterial metabolites in the digestive tract of weaning piglets. J. Anim. Sci. 2013, 91, 3200–3210. [CrossRef] g g p g 43. Clemens, E.T.; Stevens, C.E.; Southworth, M. Sites of organic acid production and pattern of digesta movement in the gastrointestinal tract of swine. J. Nutr. 1975, 105, 759–768. [CrossRef] 44. References Hirshﬁeld, I.; Terzulli, S.; O’Byrne, C. Weak organic acids: A panoply of eﬀects on bacteria. Sci. Prog. 2003, 86, 245–269. [CrossRef] 23. van Dam, H. Organic acids and their salts. Feed Mix 2006, 14, 28–31. 3. van Dam, H. Organic acids and their salts. Feed M 24. Ng, W.K.; Koh, C.B. The utilization and mode of action of organic acids in the feeds of cultured aquatic animals. Rev. Aquac. 2017, 9, 342–368. [CrossRef] 25. Hsiao, C.P.; Siebert, K.J. Modeling the inhibitory eﬀects of organic acids on bacteria. Int. J. Food Microbiol. 1999, 47, 189–201. [CrossRef] 26. Jay, J.M.; Loessner, M.J.; Golden, D.A. Food Protection with Chemicals, and by Biocontrol. In Modern Food Microbiology, 7th ed.; Heldman, D.R., Ed.; Springer Science & Business Media: New York, NY, USA, 2006; pp. 301–341. 27. Giovagnoni, G.; Tugnoli, B.; Piva, A.; Grilli, E. Organic acids and nature identical compounds can increase the activity of conventional antibiotics against Clostridium Perfringens and Enterococcus Cecorum in vitro. J. Appl. Poult. Res. 2019, 28, 1398–1407. [CrossRef] 28. Kovanda, L.; Zhang, W.; Wei, X.; Luo, J.; Wu, X.; Atwill, E.R.; Vaessen, S.; Li, X.; Liu, Y. In vitro antimicrobial activities of organic acids and their derivatives on several species of gram-negative and gram-positive bacteria. Molecules 2019, 24, 3770. [CrossRef] 29. Sheu, C.W.; Freese, E. Lipopolysaccharide layer protection of gram-negative bacteria against inhibition by long-chain fatty acids. J. Bacteriol. 1973, 115, 869–875. [CrossRef] 30. Higgins, C.; Brinkhaus, F. Eﬃcacy of several organic acids against molds. J. Appl. Poult. Res. 1999, 8, 480–487. [CrossRef] 14 of 18 14 of 18 Animals 2020, 10, 134 31. Pundir, R.; Jain, P. Screening for antifungal activity of commercially available chemical food preservatives. Int. J. Pharm. Sci. Rev. Res. 2010, 5, 25–27. 32. El-Kadi, S. Eﬀect of some organic acids on some fungal growth and their toxins production. Int. J. Adv. Biol. 2015, 2, 1–11. 33. European Union Register of Feed Additives. Available online: https://ec.europa.eu/food/safety/animal-feed/ feed-additives/eu-register_en (accessed on 20 November 2019). 34. FEFANA. Organic Acids in Animal Nutrition. 2014. Available online: http://fefana.org/publication/organic- acids-in-animal-nutrition/ (accessed on 20 November 2019). 35. Koyuncu, S.; Andersson, M.G.; Löfström, C.; Skandamis, P.N.; Gounadaki, A.; Zentek, J.; Häggblom, P. Organic acids for control of Salmonella in diﬀerent feed materials. BMC Vet. Res. 2013, 9, 81–89. [CrossRef] 35. Koyuncu, S.; Andersson, M.G.; Löfström, C.; Skandamis, P.N.; Gounadaki, A.; Zentek, J.; Häggblom, P. Organic acids for control of Salmonella in diﬀerent feed materials. References The eﬀects of citric acid on phytate-phosphorus utilization in young chicks and pigs. J. Anim. Sci. 2000, 78, 682–689. [CrossRef] 54. Bearson, S.; Bearson, B.; Foster, J.W. Acid stress responses in enterobacteria. FEMS Microbiol. Lett. 1997, 147, phytate-phosphorus utilization in young chicks and pigs. J. Anim. Sci. 2000, 78, 682–689. [CrossRef] 54. Bearson, S.; Bearson, B.; Foster, J.W. Acid stress responses in enterobacteria. FEMS Microbiol. Lett. 1997, 147, 173–180. [CrossRef] 54. Bearson, S.; Bearson, B.; Foster, J.W. Acid stress responses in enterobacteria. FEMS Microbiol. Lett. 1997, 147, 173–180. [CrossRef] 55. Lin, J.; Smith, M.P.; Chapin, K.C.; Baik, H.S.; Bennett, G.N.; Foster, J.W. Mechanisms of acid resistance in enterohemorrhagic Escherichia coli. Appl. Environ. Microbiol. 1996, 62, 3094–3100. [CrossRef] 56. Park, Y.K.; Bearson, B.; Bang, S.H.; Bang, I.S.; Foster, J.W. Internal pH crisis, lysine decarboxylase and the acid tolerance response of Salmonella typhimurium. Mol. Microbiol. 1996, 20, 605–611. [CrossRef] 57. Durant, J.A.; Corrier, D.E.; Ricke, S.C. Short-chain volatile fatty acids modulate the expression of the hilA and invF genes of Salmonella typhimurium. J. Food Prot. 2000, 63, 573–578. [CrossRef] 58. Huang, Y.; Suyemoto, M.; Garner, C.D.; Cicconi, K.M.; Altier, C. Formate acts as a diﬀusible signal to induce Salmonella invasion. J. Bacteriol. 2008, 190, 4233–4241. [CrossRef] 59. Barrow, P.A.; Fuller, R.; Newport, M.J. Changes in the microﬂora and physiology of the anterior intestinal tract of pigs weaned at 2 days with special reference to the pathogenesis of diarrhea. Infect. Immun. 1977, 18, 586–595. [CrossRef] 60. Risley, C.R.; Kornegay, E.T.; Lindemann, M.D.; Wood, C.M.; Eigel, W.N. Eﬀect of feeding organic acids on selected intestinal content measurements at varying times postweaning in pigs. J. Anim. Sci. 1992, 70, 196–206. [CrossRef] 61. Risley, C.R.; Kornegay, E.T.; Lindemann, M.D.; Wood, C.M.; Eigel, W.N. Eﬀect of feeding organic acids on gastrointestinal digesta measurements at various times postweaning in pigs challenged with enterotoxigenic Escherichia coli. Can. J. Anim. Sci. 1993, 73, 931–940. [CrossRef] 62. Ahmed, S.T.; Hwang, J.A.; Hoon, J.; Mun, H.S.; Yang, C.J. Comparison of single and blend acidiﬁers as alternative to antibiotics on growth performance, fecal microﬂora, and humoral immunity in weaned piglets. Asian-Australas. J. Anim. Sci. 2014, 27, 93–100. [CrossRef] [PubMed] 63. Canibe, N.; Steien, S.H.; Overland, M.; Jensen, B.B. Eﬀect of K-diformate in starter diets on acidity, microbiota, and the amount of organic acids in the digestive tract of piglets, and on gastric alterations. J. Anim. Sci. 2001, 79, 2123–2133. [CrossRef] [PubMed] 64. References Canibe, N.; Højberg, O.; Højsgaard, S.; Jensen, B.B. Feed physical form and formic acid addition to the feed aﬀect the gastrointestinal ecology and growth performance of growing pigs. J. Anim. Sci. 2005, 83, 1287–1302. [CrossRef] 45. Gabert, V.M.; Sauer, W.C. The eﬀect of fumaric acid and sodium fumarate supplementation to diets for weanling pigs on amino acid digestibility and volatile fatty acid concentrations in ileal digesta. Anim. Feed Sci. Technol. 1995, 53, 243–254. [CrossRef] 46. Kil, D.Y.; Piao, L.G.; Long, H.F.; Lim, J.S.; Yun, M.S.; Kong, C.S.; Ju, W.S.; Lee, H.B.; Kim, Y.Y. Eﬀects of organic or inorganic acid supplementation on growth performance, nutrient digestibility and white blood cell counts in weanling pigs. Asian-Australas. J. Anim. Sci. 2006, 19, 252–261. [CrossRef] g p g 7. Gerritsen, R.; van Dijk, A.J.; Rethy, K.; Bikker, P. The eﬀect of blends of organic acids on apparent fa digestibility in piglets. Livest. Sci. 2010, 134, 246–248. [CrossRef] 48. Blank, R.; Mosenthin, R.; Sauer, W.C.; Huang, S. Eﬀect of fumaric acid and dietary buﬀering capacity on ileal and fecal amino acid digestibilities in early-weaned pigs. J. Anim. Sci. 1999, 77, 2974–2984. [CrossRef] 49. Mosenthin, R.; Sauer, W.C.; Ahrens, F.; De Lange, C.F.M.; Bornholdt, U. Eﬀect of dietary supplements of propionic acid, siliceous earth or a combination of these on the energy, protein and amino acid digestibilities and concentration of microbial metabolites in the digestive tract of growing pigs. Anim. Feed Sci. Technol. 1992, 37, 245–255. [CrossRef] 50. Kemme, P.A.; Jongbloed, A.W.; Mroz, Z.; Mäkinen, M. Apparent ileal amino acid digestibility in pigs as aﬀected by phytate, microbial phytase, and lactic acid. J. Anim. Sci. 1995, 73, 173. 51. Li, M.; Long, S.; Wang, Q.; Zhang, L.; Hu, J.; Yang, J.; Cheng, Z.; Piao, X. Mixed organic acids improve nutrients digestibility, volatile fatty acids composition and intestinal microbiota in growing-ﬁnishing pigs fed high-ﬁber diet. Asian-Australas. J. Anim. Sci. 2019, 32, 856–864. [CrossRef] Animals 2020, 10, 134 15 of 18 15 of 18 52. Partanen, K.H.; Mroz, Z. Organic acids for performance enhancement in pig diets. Nutr. Res. Rev. 1999, 12, 117–145. [CrossRef] 53. Boling, S.D.; Webel, D.M.; Mavromichalis, I.; Parsons, C.M.; Baker, D.H. The eﬀects of citric acid on phytate-phosphorus utilization in young chicks and pigs. J. Anim. Sci. 2000, 78, 682–689. [CrossRef] 53. Boling, S.D.; Webel, D.M.; Mavromichalis, I.; Parsons, C.M.; Baker, D.H. References Feed Sci. Technol. 2005, 123–124, 445–462. [CrossRef] 75. Eckel, B.; Roth, F.X.; Kirchgessner, M.; Eidelsburger, U. Inﬂuence of formic acid on concentrations of ammonia and biogenic amines in the gastrointestinal tract, 4: Nutritive value of organic acids in piglet rearing. J. Anim. Physiol. Anim. Nutr. 1992, 67, 198–205. [CrossRef] y 76. Gabert, V.M.; Sauer, W.C.; Schmitz, M.; Ahrens, F.; Mosenthin, R. The eﬀect of formic acid and buﬀering capacity on the ileal digestibilities of amino acids and bacterial populations and metabolites in the small intestine of weanling pigs fed semipuriﬁed ﬁsh meal diets. Can. J. Anim. Sci. 1995, 75, 615–623. [CrossRef] 76. Gabert, V.M.; Sauer, W.C.; Schmitz, M.; Ahrens, F.; Mosenthin, R. The eﬀect of formic acid and buﬀering capacity on the ileal digestibilities of amino acids and bacterial populations and metabolites in the small intestine of weanling pigs fed semipuriﬁed ﬁsh meal diets. Can. J. Anim. Sci. 1995, 75, 615–623. [CrossRef] 77. Bolduan, V.G.; Jung, H.; Schneider, R.; Block, J.; Klenke, B. Inﬂuence of propionic-acid and formic-acid on capacity on the ileal digestibilities of amino acids and bacterial populations and metabolites in the small intestine of weanling pigs fed semipuriﬁed ﬁsh meal diets. Can. J. Anim. Sci. 1995, 75, 615–623. [CrossRef] 77. Bolduan, V.G.; Jung, H.; Schneider, R.; Block, J.; Klenke, B. Inﬂuence of propionic-acid and formic-acid on piglets. J. Anim. Physiol. Anim. Nutri. 1988, 59, 72–78. [CrossRef] g p g p 77. Bolduan, V.G.; Jung, H.; Schneider, R.; Block, J.; Klenke, B. Inﬂuence of propionic-acid and formic-acid on piglets. J. Anim. Physiol. Anim. Nutri. 1988, 59, 72–78. [CrossRef] 78. Roth, F.X.; Eckel, B.; Kirchgessner, M.; Eidelsburger, U. Inﬂuence of formic acid on pH, dry matter content, and concentrations of volatile fatty acids and lactic acid in the gastrointestinal tract, 3: Nutritive value of organic acids in piglet rearing. J. Anim. Physiol. Anim. Nutr. 1992, 67, 148–156. [CrossRef] 79. Partanen, K.; Jalava, T.; Valaja, J. Eﬀects of a dietary organic acid mixture and of dietary ﬁbre levels on ileal and faecal nutrient apparent digestibility, bacterial nitrogen ﬂow, microbial metabolite concentrations and rate of passage in the digestive tract of pigs. Animal 2007, 1, 389–401. [CrossRef] 80. Sutton, A.L.; Mathew, A.G.; Scheidt, A.B.; Patterson, J.A.; Kelly, D.T. Eﬀects of carbohydrate sources and organic acids on intestinal microﬂora and performance of the weanling pig. References Gedek, B.; Kirchgessner, M.; Eidelsburger, U.; Wiehler, S.; Bott, A.; Roth, F.X. Inﬂuence of formic acid on the microﬂora in diﬀerent segments of the gastrointestinal tract. J. Anim. Physiol. Anim. Nutr. 1992, 67, 206–214. [CrossRef] 65. Kirchgessner, M.; Gedek, B.; Wiehler, S.; Bott, A.; Eidelsburger, U.; Roth, F.X. Inﬂuence of formic acid, calcium formate and sodium hydrogen carbonate on the microﬂora in diﬀerent segments of the gastrointestinal tract. J. Anim. Physiol. Anim. Nutr. 1992, 68, 73–81. [CrossRef] 66. Franco, L.D.; Fondevila, M.; Lobera, M.B.; Castrillo, C. Eﬀect of combinations of organic acids in weaned pig diets on microbial species of digestive tract contents and their response on digestibility. J. Anim. Physiol. Anim. Nutr. 2005, 89, 88–93. [CrossRef] 67. Risley, C.R.; Kornegay, E.T.; Lindemann, M.D.; Weakland, S.M. Eﬀects of organic acids with and without a microbial culture on performance and gastrointestinal tract measurements of weanling pigs. Anim. Feed Sci. Technol. 1992, 35, 259–270. [CrossRef] 68. Partanen, K.; Jalava, T.; Valaja, J.; Perttila, S.; Siljander-Rasi, H.; Lindeberg, H. Eﬀect of dietary carbadox or formic acid and ﬁbre level on ileal and faecal nutrient digestibility and microbial metabolite concentrations in ileal digesta of the pig. Anim. Feed Sci. Technol. 2001, 93, 137–155. [CrossRef] 69. Hanczakowska, E.; ´Swi ˛atkiewicz, M.; Natonek-Wi´sniewska, M.; Oko´n, K. Medium chain fatty acids (MCFA) and/or probiotic Enterococcus faecium as a feed supplement for piglets. Livest. Sci. 2016, 192, 1–7. [CrossRef] 70. Walsh, M.C.; Sholly, D.M.; Hinson, R.B.; Trapp, S.A.; Sutton, A.L.; Radcliﬀe, J.S.; Smith, J.W., 2nd; Richert, B.T. Eﬀects of Acid LAC and Kem-Gest acid blends on growth performance and microbial shedding in weanling pigs. J. Anim. Sci. 2007, 85, 459–467. [CrossRef] 71. Biagi, G.; Piva, A. In vitro eﬀects of some organic acids on swine cecal microﬂora. Ital. J. Anim. Sci. 2007, 6, 361–374. [CrossRef] 72. Piva, A.; Casadei, G.; Biagi, G. An organic acid blend can modulate swine intestinal fermentation and reduce microbial proteolysis. Can. J. Anim. Sci. 2002, 82, 527–532. [CrossRef] 16 of 18 16 of 18 Animals 2020, 10, 134 73. Partanen, K.; Jalava, T. Eﬀects of some organic acids and salts on microbial fermentation in the digestive tract of piglets estimated using an in vitro gas production technique. Agric. Food Sci. 2005, 14, 311–324. [CrossRef] 74. Williams, B.A.; Bosch, M.W.; Boer, H.; Verstegen, M.W.A.; Tamminga, S. An in vitro batch culture method to assess potential fermentability of feed ingredients for monogastric diets. Anim. References Dietary Tributyrin Supplementation Attenuates Insulin Resistance and Abnormal Lipid Metabolism in Suckling Piglets with Intrauterine Growth Retardation. PLoS ONE 2015, 10, e0136848. [CrossRef] [PubMed] 96. He, J.; Dong, L.; Xu, W.; Bai, K.; Lu, C.; Wu, Y.; Huang, Q.; Zhang, L.; Wang, T. Dietary Tributyrin Supplementation Attenuates Insulin Resistance and Abnormal Lipid Metabolism in Suckling Piglets with Intrauterine Growth Retardation. PLoS ONE 2015, 10, e0136848. [CrossRef] [PubMed] 97. Dong, L.; Zhong, X.; He, J.; Zhang, L.; Bai, K.; Xu, W.; Wang, T.; Huang, X. Supplementation of tributyrin improves the growth and intestinal digestive and barrier functions in intrauterine growth-restricted piglets. Clin. Nutr. 2016, 35, 399–407. [CrossRef] [PubMed] 98. Murray, R.L.; Zhang, W.; Iwaniuk, M.; Grilli, E.; Stahl, C.H. Dietary tributyrin, an HDAC inhibitor, promotes muscle growth through enhanced terminal diﬀerentiation of satellite cells. Physiol. Rep. 2018, 6, e13706. [CrossRef] [PubMed] 99. Wolﬀram, S.; Hagemann, C.; Grenacher, B.; Scharrer, E. Characterization of the transport of tri- and dicarboxylates by pig intestinal brush-border membrane vesicles. Comp. Biochem. Physiol. Comp. Physiol. 1992, 101, 759–767. [CrossRef] 100. Browne, J.L.; Sanford, P.A.; Smyth, D.H. Transfer and metabolism of citrate, succinate, alpha-ketoglutarate and pyruvate by hamster small intestine. Proc. R. Soc. Lond. Ser. B Biol. Sci. 1978, 200, 117–135. 101. Buddington, R.K.; Pajor, A.; Buddington, K.K.; Pierzynowski, S. Absorption of alpha-ketoglutarate by the gastrointestinal tract of pigs. Comp. Biochem. Physiol. Part A Mol. Integr. Physiol. 2004, 138, 215–220. [CrossRef] 102. Dabek, M.; Kruszewska, D.; Filip, R.; Hotowy, A.; Pierzynowski, L.; Wojtasz-Pajak, A.; Szymanczyk, S.; Valverde Piedra, J.L.; Werpachowska, E.; Pierzynowski, S.G. Alpha-Ketoglutarate (AKG) absorption from pig intestine and plasma pharmacokinetics. J. Anim. Physiol. Anim. Nutr. 2005, 89, 419–426. [CrossRef] 103. Kirchgessner, M.; Roth, F.X. Fumaric acid as a feed additive in pig nutrition. Pig News Inf. 1982, 3, 259–264. 104. He, L.; Xu, Z.; Yao, K.; Wu, G.; Yin, Y.; Nyachoti, C.M.; Kim, S.W. The Physiological Basis and Nutritional Function of Alpha-ketoglutarate. Curr. Protein Pept. Sci. 2015, 16, 576–581. [CrossRef] 105. Hou, Y.; Wang, L.; Ding, B.; Liu, Y.; Zhu, H.; Liu, J.; Li, Y.; Kang, P.; Yin, Y.; Wu, G. Alpha-Ketoglutarate and intestinal function. Front. Biosci. 2011, 16, 1186–1196. [CrossRef] [PubMed] 106. Xiao, D.; Zeng, L.; Yao, K.; Kong, X.; Wu, G.; Yin, Y. The glutamine-alpha-ketoglutarate (AKG) metabolism and its nutritional implications. Amino Acids 2016, 48, 2067–2080. [CrossRef] [PubMed] 107. References In Proceedings of the 5th International Symposium on Digestible Physiology in Pigs, Wageningen, The Netherlands, 24–26 April 1991; Verstegen, M.W.A., Huisman, J., den Hartog, L.A., Eds.; Pudoc: Wageningen, The Netherlands, 1991; pp. 422–427. 81. Mathew, A.G.; Sutton, A.L.; Scheidt, A.; Bforsyth, D.M.; Patterson, J.A.; Kelly, D.T. Eﬀects of a propionic acid containing feed additive on performance and intestinal microbial fermentation of the weanling pig. In Proceedings of the 5th International Symposium on Digestible Physiology in Pigs, Wageningen, The Netherlands, 24–26 April 1991; Verstegen, M.W.A., Huisman, J., den Hartog, L.A., Eds.; Pudoc: Wageningen, The Netherlands, 1991; pp. 464–469. 82. Bergman, E.N. Energy contributions of volatile fatty acids from the gastrointestinal tract in various species. Physiol. Rev. 1990, 70, 567–590. [CrossRef] 83. Imoto, S.; Namioka, S. Nutritive value of acetate in growing pigs. J. Anim. Sci. 1983, 56, 858–866. [CrossRef] [PubMed] 84. Gabert, V.M.; Sauer, W.C. The eﬀects of supplementing diets for weanling pigs with organic acids. A review. J. Anim. Feed Sci. 1994, 3, 73–87. [CrossRef] 85. Hamer, H.M.; Jonkers, D.; Venema, K.; Vanhoutvin, S.; Troost, F.J.; Brummer, R.J. Review article: The role of butyrate on colonic function. Aliment. Pharmacol. Ther. 2008, 27, 104–119. [CrossRef] [PubMed] 86. Guilloteau, P.; Martin, L.; Eeckhaut, V.; Ducatelle, R.; Zabielski, R.; Van Immerseel, F. From the gut to the peripheral tissues: The multiple eﬀects of butyrate. Nutr. Res. Rev. 2010, 23, 366–384. [CrossRef] 87. Leonel, A.J.; Alvarez-Leite, J.I. Butyrate: Implications for intestinal function. Curr. Opin. Clin. Nutr. Metab. Care 2012, 15, 474–479. [CrossRef] 88. Kotunia, A.; Woli´nski, J.; Laubitz, D.; Jurkowska, M.; Romé, V.; Guilloteau, P.; Zabielski, R. Eﬀect of sodium butyrate on the small intestine development in neonatal piglets fed by artiﬁcial sow. J. Physiol. Pharmacol. 2004, 55, 59–68. 89. Mazzoni, M.; Le Gall, M.; De Filippi, S.; Minieri, L.; Trevisi, P.; Wolinski, J.; Lalatta-Costerbosa, G.; Lallès, J.P.; Guilloteau, P.; Bosi, P. Supplemental sodium butyrate stimulates diﬀerent gastric cells in weaned pigs. J. Nutr. 2008, 138, 1426–1431. [CrossRef] 90. Fang, C.L.; Sun, H.; Wu, J.; Niu, H.H.; Feng, J. Eﬀects of sodium butyrate on growth performance, haematological and immunological characteristics of weanling piglets. J. Anim. Physiol. Anim. Nutr. 2014, 98, 680–685. [CrossRef] 91. Grilli, E.; Tugnoli, B.; Foerster, C.J.; Piva, A. Butyrate modulates inﬂammatory cytokines and tight junctions components along the gut of weaned pigs. J. Anim. Sci. 2016, 94, 433–436. [CrossRef] 17 of 18 17 of 18 Animals 2020, 10, 134 92. References Lallès, J.P.; Bosi, P.; Janczyk, P.; Koopmans, S.J.; Torrallardona, D. Impact of bioactive substances on the gastrointestinal tract and performance of weaned piglets: A review. Animal 2009, 3, 1625–1643. [CrossRef] 92. Lallès, J.P.; Bosi, P.; Janczyk, P.; Koopmans, S.J.; Torrallardona, D. Impact of bioactive substances on the gastrointestinal tract and performance of weaned piglets: A review. Animal 2009, 3, 1625–1643. [CrossRef] , J ; , ; J y , ; p , J ; , p gastrointestinal tract and performance of weaned piglets: A review. Animal 2009, 3, 1625–1643. [CrossRef] 93. Piva, A.; Prandini, A.; Fiorentini, L.; Morlacchini, M.; Galvano, F.; Luchansky, J.B. Tributyrin and lactitol synergistically enhanced the trophic status of the intestinal mucosa and reduced histamine levels in the gut of nursery pigs. J. Anim. Sci. 2002, 80, 670–680. [CrossRef] 93. Piva, A.; Prandini, A.; Fiorentini, L.; Morlacchini, M.; Galvano, F.; Luchansky, J.B. Tributyrin and lactitol synergistically enhanced the trophic status of the intestinal mucosa and reduced histamine levels in the gut of nursery pigs. J. Anim. Sci. 2002, 80, 670–680. [CrossRef] 94. Hou, Y.Q.; Liu, Y.L.; Hu, J.; Shen, W.H. Eﬀects of lactitol and tributyrin on growth performance, small intestinal morphology and enzyme activity in weaned pigs. Asian-Australas. J. Anim. Sci. 2006, 19, 1470–1477. [CrossRef] 95. Piva, A.; Grilli, E.; Fabbri, L.; Pizzamiglio, V.; Gatta, P.P.; Galvano, F.; Bognanno, M.; Fiorentini, L.; Woli´nski, J.; Zabielski, R.; et al. Intestinal metabolism of weaned piglets fed a typical United States or European diet with or without supplementation of tributyrin and lactitol. J. Anim. Sci. 2008, 86, 2952–2961. [CrossRef] [PubMed] 96. He, J.; Dong, L.; Xu, W.; Bai, K.; Lu, C.; Wu, Y.; Huang, Q.; Zhang, L.; Wang, T. Dietary Tributyrin Supplementation Attenuates Insulin Resistance and Abnormal Lipid Metabolism in Suckling Piglets with 95. Piva, A.; Grilli, E.; Fabbri, L.; Pizzamiglio, V.; Gatta, P.P.; Galvano, F.; Bognanno, M.; Fiorentini, L.; Woli´nski, J.; Zabielski, R.; et al. Intestinal metabolism of weaned piglets fed a typical United States or European diet with or without supplementation of tributyrin and lactitol. J. Anim. Sci. 2008, 86, 2952–2961. [CrossRef] [PubMed] Zabielski, R.; et al. Intestinal metabolism of weaned piglets fed a typical United States or European diet with or without supplementation of tributyrin and lactitol. J. Anim. Sci. 2008, 86, 2952–2961. [CrossRef] [PubMed] 96. He, J.; Dong, L.; Xu, W.; Bai, K.; Lu, C.; Wu, Y.; Huang, Q.; Zhang, L.; Wang, T. References He, L.; Zhou, X.; Huang, N.; Li, H.; Cui, Z.; Tian, J.; Jiang, Q.; Liu, S.; Wu, J.; Li, T.; et al. Administration of alpha-ketoglutarate improves epithelial restitution under stress injury in early-weaning piglets. Oncotarget 2017, 8, 91965–91978. [CrossRef] [PubMed] 108. Hou, Y.; Wang, L.; Ding, B.; Liu, Y.; Zhu, H.; Liu, J.; Li, Y.; Wu, X.; Yin, Y.; Wu, G. Dietary alpha-ketoglutarate supplementation ameliorates intestinal injury in lipopolysaccharide-challenged piglets. Amino Acids 2010, 39, 555–564. [CrossRef] 109. Wang, L.; Hou, Y.; Yi, D.; Li, Y.; Ding, B.; Zhu, H.; Liu, J.; Xiao, H.; Wu, G. Dietary supplementation with glutamate precursor α-ketoglutarate attenuates lipopolysaccharide-induced liver injury in young pigs. Amino Acids 2015, 47, 1309–1318. [CrossRef] 110. Chen, J.S.; Wu, F.; Yang, H.S.; Li, F.N.; Jiang, Q.; Liu, S.J.; Kang, B.J.; Li, S.; Adebowale, T.O.; Huang, N.; et al. Growth performance, nitrogen balance, and metabolism of calcium and phosphorus in growing pigs fed diets supplemented with alpha-ketoglutarate. Anim. Feed Sci. Technol. 2017, 226, 21–28. [CrossRef] 111. Chen, J.; Su, W.; Kang, B.; Jiang, Q.; Zhao, Y.; Fu, C.; Yao, K. Supplementation with α-ketoglutarate to a low-protein diet enhances amino acid synthesis in tissues and improves protein metabolism in the skeletal muscle of growing pigs. Amino Acids 2018, 50, 1525–1537. [CrossRef] 18 of 18 18 of 18 Animals 2020, 10, 134 112. Chen, J.; Zhang, H.; Gao, H.; Kang, B.; Chen, F.; Li, Y.; Fu, C.; Yao, K. Eﬀects of Dietary Supplementation of Alpha-Ketoglutarate in a Low-Protein Diet on Fatty Acid Composition and Lipid Metabolism Related Gene Expression in Muscles of Growing Pigs. Animals 2019, 9, 838–849. [CrossRef] 113. Walker, R. Toxicology of sorbic acid and sorbates. Food Addit. Contam. 1990, 7, 671–676. [CrossRef] [P r, R. Toxicology of sorbic acid and sorbates. Food Addit. Contam. 1990, 7, 671–676. [CrossRef] [PubMed] o, G. On the metabolism of sorbic acid in the mouse. Acta. Chem. Scand. 1964, 18, 1373–1378. [CrossRef] Westoo, G. On the metabolism of sorbic acid in the mouse. Acta. Chem. Scand. 1964, 18, 1373–1378. [CrossRef 115. Simmen, F.A.; Badinga, L.; Green, M.L.; Kwak, I.; Song, S.; Simmen, R.C. The porcine insulin-like growth factor system: At the interface of nutrition, growth and reproduction. J. Nutr. 1998, 128, 315S–320S. [CrossRef] [PubMed] 116. Luo, Z.F.; Fang, X.L.; Shu, G.; Wang, S.B.; Zhu, X.T.; Gao, P.; Chen, L.L.; Chen, C.Y.; Xi, Q.Y.; Zhang, Y.L.; et al. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). References Sorbic acid improves growth performance and regulates insulin-like growth factor system gene expression in swine. J. Anim. Sci. 2011, 89, 2356–2364. [CrossRef] [PubMed] 117. Bridges, J.W.; French, M.R.; Smith, R.L.; Williams, R.T. The fate of benzoic acid in various species. Biochem. J. 1970, 118, 47–51. [CrossRef] [PubMed] 118. Kristensen, N.B.; Nørgaard, J.V.; Wamberg, S.; Engbaek, M.; Fernández, J.A.; Zacho, H.D.; Poulsen, H.D. Absorption and metabolism of benzoic acid in growing pigs. J. Anim. Sci. 2009, 87, 2815–2822. [CrossRef] 119. Eriksen, J.; Adamsen, A.P.; Nørgaard, J.V.; Poulsen, H.D.; Jensen, B.B.; Petersen, S.O. Emissions of sulfur-containing odorants, ammonia, and methane from pig slurry: Eﬀects of dietary methionine and benzoic acid. J. Environ. Qual. 2010, 39, 1097–1107. [CrossRef] 120. Temiz, U.; Ozturk, E. Encapsulation Methods and Use in Animal Nutrition. Selcuk J. Agric. Food Sci. 2018, 32, 624–631. [CrossRef] 121. Piva, A.; Pizzamiglio, V.; Morlacchini, M.; Tedeschi, M.; Piva, G. Lipid microencapsulation allows slow release of organic acids and natural identical ﬂavors along the swine intestine. J. Anim. Sci. 2007, 85, 486–493. [CrossRef] 122. Piva, A.; Grilli, E.; Fabbri, L.; Pizzamiglio, V.; Campani, I. Free versus microencapsulated organic acids in medicated or not medicated diet for piglets. Livest. Sci. 2007, 108, 214–217. [CrossRef] 123. Grilli, E.; Vitari, F.; Domeneghini, C.; Palmonari, A.; Tosi, G.; Fantinati, P.; Massi, P.; Piva, A. Development of a feed additive to reduce caecal Campylobacter jejuni in broilers at slaughter age: From in vitro to in vivo, a proof of concept. J. Appl. Microbiol. 2013, 114, 308–317. [CrossRef] [PubMed] 124. Grilli, E.; Foresti, F.; Tugnoli, B.; Fustini, M.; Zanoni, M.G.; Pasquali, P.; Callaway, T.R.; Piva, A.; Alborali, G.L. Microencapsulated sorbic acid and pure botanicals aﬀect Salmonella Typhimurium shedding in pigs: A close-up look from weaning to slaughter in controlled and ﬁeld conditions. Foodborne Pathog. Dis. 2015, 12, 813–819. [CrossRef] [PubMed] 125. Grilli, E.; Tugnoli, B.; Passey, J.L.; Stahl, C.H.; Piva, A.; Moeser, A.J. Impact of dietary organic acids and botanicals on intestinal integrity and inﬂammation in weaned pigs. BMC Vet. Res. 2015, 11, 96–105. [CrossRef] [PubMed] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W2047946045	https://dash.harvard.edu/bitstream/1/41542991/1/63166%20Journal%20of%20Virology-2011-Yan-7483.full.pdf	English	null	Virion Structure of Baboon Reovirus, a Fusogenic Orthoreovirus That Lacks an Adhesion Fiber	Journal of virology	2,011	cc-by	13,923	Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#LAA Permanent link http://nrs.harvard.edu/urn-3:HUL.InstRepos:41542991 Citation Yan, X., K. N. Parent, R. P. Goodman, J. Tang, J. Shou, M. L. Nibert, R. Duncan, and T. S. Baker. 2011. “Virion Structure of Baboon Reovirus, a Fusogenic Orthoreovirus That Lacks an Adhesion Fiber.” Journal of Virology 85 (15): 7483–95. doi:10.1128/JVI.00729-11. Received 11 April 2011/Accepted 10 May 2011 on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from on October 13, 201 http://jvi.asm.org/ Downloaded from Baboon reovirus (BRV) is a member of the fusogenic subgroup of orthoreoviruses. Unlike most other members of its genus, BRV lacks S-segment coding sequences for the outer ﬁber protein that binds to cell surface receptors. It shares this lack with aquareoviruses, which constitute a related genus and are also fusogenic. We used electron cryomicroscopy and three-dimensional image reconstruction to determine the BRV virion structure at 9.0-Å resolution. The results show that BRV lacks a protruding ﬁber at its icosahedral 5-fold axes or elsewhere. The results also show that BRV is like nonfusogenic mammalian and fusogenic avian orthoreoviruses in having 150 copies of the core clamp protein, not 120 as in aquareoviruses. On the other hand, there are no hub-and-spoke complexes attributable to the outer shell protein in the P2 and P3 solvent channels of BRV, which makes BRV like fusogenic avian orthoreoviruses and aquareoviruses but unlike nonfusogenic mammalian orthoreoviruses. The outermost “ﬂap” domains of the BRV core turret protein appear capable of conformational variability within the virion, a trait previously unseen among other ortho- and aquareoviruses. New cDNA sequence determinations for the BRV L1 and M2 genome segments, encoding the core turret and outer shell proteins, were helpful for interpreting the structural features of those proteins. Based on these ﬁndings, we conclude that the evolution of ortho- and aquareoviruses has included a series of discrete gains or losses of particular components, several of which cross taxonomic boundaries. Gain or loss of adhesion ﬁbers is one of several common themes in double-stranded RNA virus evolution. on October 13, 2019 by guest http://jvi.asm.org/ rom that additional comparative studies are likely to provide im- portant new insights into not only their evolution but also their structural and functional properties. Baboon reovirus (BRV) is the prototype strain of species Baboon orthoreovirus, one of ﬁve species formally recognized to date in the genus Orthoreovirus (subfamily Spinareovirinae, family Reoviridae) of double-stranded RNA (dsRNA) viruses with 10-segment genomes and two-layer, icosahedral capsids. Viruses in the other four species—Mammalian orthoreovirus, Avian orthoreovirus, Nelson Bay orthoreovirus, and Reptilian orthoreovirus—infect mammals, birds, or reptiles (16, 17). Broome virus (BroV), a recently described Australian isolate from an encephalitic fruit bat, is divergent from Nelson Bay virus (NBV) and related strains of bat-borne orthoreoviruses and represents a new, though not yet formally recognized, orthoreovirus species (50). * Corresponding author. Mailing address for Timothy S. Baker: De- partment of Chemistry and Biochemistry, University of California— San Diego, La Jolla, CA 92093. Phone: (858) 534-5845. Fax: (858) 534-5846. E-mail: tsb@ucsd.edu. Mailing address for Roy Duncan: Department of Microbiology and Immunology, Dalhousie University, Halifax NS B3H4R2, Canada. Phone: (902) 494-6770. Fax: (902) 494-5125. E-mail: roy.duncan@dal.ca. † M.L.N., R.D., and T.S.B. are co-senior authors. Published ahead of print on 18 May 2011. Virion Structure of Baboon Reovirus, a Fusogenic Orthoreovirus That Lacks an Adhesion Fiber Xiaodong Yan,1 Kristin N. Parent,1 Russell P. Goodman,2,3 Jinghua Tang,1 J Max L. Nibert,2† Roy Duncan,4†* and Timothy S. Baker1,5†* on October 13, 2019 by gue http://jvi.asm.org/ Downloaded from Received 11 April 2011/Accepted 10 May 2011 Share Your Story The Harvard community has made this article openly available. Please share how this access benefits you. Submit a story . The Harvard community has made this article openly available. Please share how this access benefits you. Submit a story . Accessibility Vol. 85, No. 15 JOURNAL OF VIROLOGY, Aug. 2011, p. 7483–7495 0022-538X/11/$12.00 doi:10.1128/JVI.00729-11 Copyright © 2011, American Society for Microbiology. All Rights Reserved. JOURNAL OF VIROLOGY, Aug. 2011, p. 7483–7495 0022-538X/11/$12.00 doi:10.1128/JVI.00729-11 Copyright © 2011, American Society for Microbio / /$ / Copyright © 2011, American Society for Microbiology. All Rights Reserved. Virion Structure of Baboon Reovirus, a Fusogenic Orthoreovirus That Lacks an Adhesion Fiber Xiaodong Yan,1 Kristin N. Parent,1 Russell P. Goodman,2,3 Jinghua Tang,1 Jingyun Shou,4 Max L. Nibert,2† Roy Duncan,4†* and Timothy S. Baker1,5†* Department of Chemistry and Biochemistry1 and Division of Biological Sciences,5 University of California—San Diego, La Jolla, California 92093; Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, Massachusetts 021152; Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts 021153; and Department of Microbiology and Immunology, Dalhousie University, Halifax NS B3H4R2, Canada4 Received 11 April 2011/Accepted 10 May 2011 In GCRV and SBRV virions, in contrast, the tops of the core turrets (formed by VP1) adopt a more open arrangement (10, 32), presumably reﬂecting that the aquareovirus turrets need not be closed since there is no adhesion ﬁber to be anchored at each icosa- hedral 5-fold (5f) axis. reovirus (GCRV) (from species Aquareovirus C) have been determined by transmission electron cryomicroscopy (cryo- TEM) and three-dimensional (3D) image reconstruction (10, 32) and show substantial similarities to those of ARV and MRV virions (15, 58). One notable exception is the absence of an adhesion ﬁber in the aquareovirus cryo-reconstructions (10, 32). In ARV and MRV virions, the base of the adhesion ﬁber is anchored to the ﬂattened top of the pentameric core turret protein (ARV C and MRV 2) (15, 58). In GCRV and SBRV virions, in contrast, the tops of the core turrets (formed by VP1) adopt a more open arrangement (10, 32), presumably reﬂecting that the aquareovirus turrets need not be closed since there is no adhesion ﬁber to be anchored at each icosa- hedral 5-fold (5f) axis. rus (RRV) strains (16, 17), but prototypical mammalian reo- virus (MRV) strains are nonfusogenic. The only fusogenic mammalian orthoreoviruses in addition to BRV are the bat- borne NBV-like viruses (11, 21) and the divergent bat-borne isolate BroV (50). Notably, several of the NBV-like viruses have been recently associated with human cases of high fever and acute respiratory illness in eastern Asia (21). rus (RRV) strains (16, 17), but prototypical mammalian reo- virus (MRV) strains are nonfusogenic. The only fusogenic mammalian orthoreoviruses in addition to BRV are the bat- borne NBV-like viruses (11, 21) and the divergent bat-borne isolate BroV (50). Notably, several of the NBV-like viruses have been recently associated with human cases of high fever and acute respiratory illness in eastern Asia (21). p y ( ) Full-length cDNA sequences have been reported to date for the four small (S) genome segments of BRV, and their en- coded, -class proteins show only low levels of sequence iden- tity with homologous proteins of other representative or- thoreoviruses (35%), consistent with the evolutionary divergence of BRV and its assignment to a separate species (13, 16, 17). Received 11 April 2011/Accepted 10 May 2011 Related 11-segment dsRNA viruses isolated from ﬁsh deﬁne a distinct phylogenetic cluster and are grouped in the separate genus Aquareovirus (1, 19). Interest- ingly, a 10-segment ﬁsh reovirus (piscine reovirus, or PRV) was recently described and proposed to represent the prototype strain of another new orthoreovirus, not aquareovirus, species (38). The diversity of viruses from these two genera suggests BRV was ﬁrst described in 1995 (18), having been isolated in 1993 to 1994 from a single colony of Papio cynocephalus baboons at the Southwest Foundation for Biomedical Re- search in San Antonio, TX (24). It was recovered from the brains of ﬁve juvenile animals that showed clinical signs of progressive meningoencephalomyelitis as well as histopatho- logical evidence for that disease. A causative role for BRV was established by intracranial inoculations of two young baboons, both of which developed the disease, although the source of the virus and the mode of transmission among the baboons remain unknown. The severity of disease in these primates nevertheless raises concern that BRV may represent a group of viruses circulating in nature that has the potential to emerge as a serious human pathogen. In fact, partial sequence evidence for another virus closely related to BRV has been reported, based on samples derived from skunk brain tissue collected in 1974 by the California Department of Public Health (52). In cultured cells, BRV is fusogenic, inducing formation of multinucleated syncytia by cell-cell fusion (18). This unusual activity for a nonenveloped virus is mediated by a virally en- coded nonstructural protein, called the FAST (fusion-associ- ated small transmembrane) protein (46), which is expressed on the surface of infected cells but is not present in mature virions (12, 13). Possession of a FAST protein and associated fusoge- nicity are shared by avian reovirus (ARV) and reptilian reovi- 7483 7484 YAN ET AL. J. VIROL. YAN ET AL. 7484 TABLE 1. Virion proteins of orthoreoviruses BRV, MRV, and ARV and aquareovirus GCRV . V o p ote s o o t o eov uses V, V, a d V a d aqua eov us GC V Protein namea No. Received 11 April 2011/Accepted 10 May 2011 The virion structures of striped bass reovirus (SBRV) (from species Aquareovirus A) and grass carp Received 11 April 2011/Accepted 10 May 2011 One distinctive aspect of the BRV proteins is that they do not include a homolog of the outer ﬁber protein found in most other orthoreoviruses (e.g., ARV C or MRV 1) (12, 13, 16) (Table 1). Each of the orthoreovirus ﬁber proteins, which bind to cell surface receptors (3, 39), is distin- guished by heptad repeats of hydrophobic residues in the N- terminal half of the deduced amino acid sequence, consistent with formation of a long, -helical coiled coil (4, 17, 47). In most other orthoreoviruses, the outer ﬁber protein is encoded on the same bi- or tricistronic S genome segment as the FAST protein and/or a poorly conserved nonstructural protein of unclear function (4, 17, 47). The S4 genome segment that encodes the FAST protein of BRV is also bicistronic, but the second encoded protein, p16, shows no sequence similarity to the outer ﬁber protein. Moreover, p16 is a nonstructural pro- tein (12) and is therefore not a functional analog of the ﬁber protein. Which protein in BRV virions then plays the primary role in adhesion to cell surface receptors remains undeﬁned. ( ) Given the several notable characteristics of BRV, as well as its status as the prototype strain of a distinct, fusogenic mam- malian orthoreovirus species, we sought to determine its virion structure. In particular, based on the structures of other ortho- and aquareoviruses, we hypothesized that the virions of BRV might differ from those of other orthoreoviruses by not only lacking an adhesion ﬁber but also having more open core turrets, like those in aquareovirus virions. To address these hypotheses, we determined the structure of BRV virions by cryo-TEM and 3D image reconstruction to an estimated res- olution of 9.0 Å. Sequences of the BRV L1 and M2 genome segments, respectively, encoding the core turret protein (BRV A) and the outer shell protein (BRV ), were also newly determined to aid in interpreting the structure. Results indi- cate that the evolution of ortho- and aquareoviruses has in- cluded a series of discrete gains or losses of particular compo- nents, several of which cross taxonomic boundaries. Even though their genomes comprise 11 dsRNA segments, members of the genus Aquareovirus exhibit clear homologies to orthoreoviruses in all but a few of their proteins, both struc- tural and nonstructural (1, 19, 23) (Table 1). Also, like many orthoreoviruses as noted above, aquareoviruses are fusogenic in cultured cells (43). Received 11 April 2011/Accepted 10 May 2011 of copies per virionb Protein homolog Protein size (aa) BRV MRV ARV GCRV BRVc MRVd ARVe GCRVf Core RdRp 12 B 3 B VP2 1,261 1,267 1,259 1,274 Core NTPase 24 A 2 A VP5 738 736 732 728 Core shell 120 C 1 A VP3 1,231 1,275 1,293 1,214 Core clamp 150/120 A 2 A VP6 413 418 416 412 Core turret 60 A 2 C VP1 1,284 1,289 1,285 1,299 Outer shell 600 B 1 B VP4 676 708 676 648 Outer clamp 600 B 3 B VP7 396 365 367 276 Outer ﬁber 36/0 1 C 455 326 Protein size (aa) p ( ) b Values are for MRV virions, which are probably also correct for ARV and BRV except that BRV lacks an outer ﬁber protein. The values with approximate signs () are estimates from best available data for MRV. These values are probably also correct for aquareovirus virions, except that aquareoviruses contain only 120 core clamp subunits and also lack an outer ﬁber protein. on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from on October 13, 201 http://jvi.asm.org/ Downloaded from c GenBank accession numbers are (top to bottom) HQ847903, HQ847904, HQ847905, HQ847906, HQ847907, AAC18124, and AAC18128. d Values are for MRV-T3D. GenBank accession numbers are (top to bottom) AF129822, J03488, M31058, AF461684, M20161, X01161, L19776, and K02739, in respective order. p e Values are for ARV-176. GenBank accession numbers are (top to bottom) ACH72474, ACH72476, ACH72478, AAT52025, and AF218358. f k i b ( b ) 511 511 511 1 1 on October 13, 2019 by guest http://jvi.asm.org/ ed from and AF218358. f GenBank accession numbers are (top to bottom) AF260511S3, AF260511S2, AF260511S1, AF403391, AF403392, AF403394, and AAM92742. ion numbers are (top to bottom) AF260511S3, AF260511S2, AF260511S1, AF403391, AF403392, AF403394, and AAM92742. f GenBank accession numbers are (top to bottom) AF260511S3, AF260511S2, AF260511S1, AF403391, AF403392, AF403394, and AAM92742. reovirus (GCRV) (from species Aquareovirus C) have been determined by transmission electron cryomicroscopy (cryo- TEM) and three-dimensional (3D) image reconstruction (10, 32) and show substantial similarities to those of ARV and MRV virions (15, 58). One notable exception is the absence of an adhesion ﬁber in the aquareovirus cryo-reconstructions (10, 32). In ARV and MRV virions, the base of the adhesion ﬁber is anchored to the ﬂattened top of the pentameric core turret protein (ARV C and MRV 2) (15, 58). MATERIALS AND METHODS Cell debris was removed from the pooled cell lysates (50 ml) by centrifugation at 10,000 g for 30 min at 4°C, and virus particles were concentrated from the supernatant by centrifuga- tion at 100,000 g for 60 min at 4°C through a 20% sucrose cushion. The virus pellet was resuspended in TMN buffer, and virus particles were puriﬁed on linear CsCl density gradients as previously described (58). Isolated virus particles were dialyzed against TMN buffer and stored at 4°C until analysis. For obtaining genomic dsRNA, puriﬁed BRV virions were disrupted by incubation for 30 min at 37°C in the presence of 1% SDS, and the dsRNA was then isolated by phenol-chloroform extraction and ethanol precipitation. Resuspended dsRNA was subjected to centrifugation through a CHROMA SPIN 1000 column (Clon- tech) to remove small oligonucleotides as a ﬁnal step in puriﬁcation. on October 13, 2019 by guest http://jvi.asm.org/ aded from on October 13, 2019 by guest jvi.asm.org/ Cryo-TEM. Small (3.5-l) aliquots of puriﬁed BRV particles (5 1012 particles/ml) were vitriﬁed and examined using standard procedures (2). Brieﬂy, this involved applying samples to holey carbon grids that were glow discharged for 15 s in an Emitech K350 evaporation unit. Grids were blotted with What- man ﬁlter paper for 5 s, plunged into liquid ethane, and transferred into a precooled, FEI Polara multispecimen holder, which maintained the specimen at liquid nitrogen temperature. Micrographs were recorded on Kodak SO-163 electron-image ﬁlm at both 300 and 200 keV in an FEI Polara microscope under minimal-dose conditions (ranging from 25 to 35 electrons/Å2) at a nominal magniﬁcation of 39,000 (Table 2). Sequence alignments. Pairwise sequence alignments were performed using EMBOSS Stretcher, with default settings except for use of the Blosum 30 matrix, as implemented at http://www.ebi.ac.uk/Tools/psa/emboss_stretcher/. Multiple sequence alignments were performed using the Clustal W2 algorithm as imple- mented at http://guidance.tau.ac.il/in combination with the GUIDANCE algo- rithm to calculate guide tree-based alignment conﬁdence scores (40). Conﬁdence scores from 100 bootstraps for the core turret and outer shell protein alignments were 97.4% and 97.2%, respectively, compared with only 45.5% from a parallel analysis of the outer clamp protein. The program PROMALS3D as implemented at http://prodata.swmed.edu/promals3d/promals3d.php and the program MUSCLE, version 3.7, and T-Coffee, version 6.85, as implemented at http://phylogeny.fr/ were used as alternatives for multiple sequence alignments to conﬁrm results. MATERIALS AND METHODS Final density maps used for visualization, analysis, and interpretation were sharpened to enhance high-reso- lution features by imposing an inverse temperature factor (1/400 Å2) (22). During the course of the image-processing procedures, we sometimes obtained cryo-reconstructions of the BRV virion that were missing the core turret ﬂap densities. Upon investigating this phenomenon, we determined that it varied from grid to grid. The presence or absence of ﬂaps in the resulting reconstruction was found to be independent of the virion preparation (three independent preparations were tested at different times, and each produced a mixture of reconstructions with or without ﬂaps in a grid-speciﬁc manner) and also inde- pendent of the three types of support ﬁlm on which the particles were frozen (home-made continuous, home-made holey, or Quantifoil carbon grids). In ad- dition, the presence or absence of ﬂaps in the resulting reconstruction seemed to reﬂect an all-or-none phenomenon, whereby all particles processed from any particular grid either contributed ﬂap densities or did not. The images were therefore sorted into two groups to give rise to the cryo-reconstructions of the BRV virion with or without core turret ﬂaps as described below. During the course of the image-processing procedures, we sometimes obtained cryo-reconstructions of the BRV virion that were missing the core turret ﬂap densities. Upon investigating this phenomenon, we determined that it varied from grid to grid. The presence or absence of ﬂaps in the resulting reconstruction was found to be independent of the virion preparation (three independent preparations were tested at different times, and each produced a mixture of reconstructions with or without ﬂaps in a grid-speciﬁc manner) and also inde- pendent of the three types of support ﬁlm on which the particles were frozen (home-made continuous, home-made holey, or Quantifoil carbon grids). In ad- dition, the presence or absence of ﬂaps in the resulting reconstruction seemed to reﬂect an all-or-none phenomenon, whereby all particles processed from any particular grid either contributed ﬂap densities or did not. The images were therefore sorted into two groups to give rise to the cryo-reconstructions of the BRV virion with or without core turret ﬂaps as described below. Fitting MRV crystal structures. In preparation for model-ﬁtting experiments, the BRV 3D density maps were scaled to the same size as the MRV virion structure from Zhang et al. (57) by using the program UCSF Chimera (http: //www.cgl.ucsf.edu/chimera). MATERIALS AND METHODS GenBank accession numbers for published sequences used in these analyses were AF378003 (MRV type 1 Lang [T1L]), AF378005 (MRV type 2 Jones [T2J]), EU707937 (ARV-138), EU707938 (ARV-176), GQ258978 (BroV), GU994014 (PRV), AF260511S1 (GCRV), and EF589098 (American grass carp reovirus [AGCRV]) for alignment with BRV core turret protein A and AF490617 (MRV-T1L), M19355 (MRV-T2J), AY50052 (ARV-138), AY50053 (ARV-176), GQ258981 (BroV), GU994016 (PRV), AF403392 (GCRV), and EF589103 (AGCRV) for alignment with BRV outer shell protein B. ober 13, 2019 by guest 3D image reconstruction. Micrographs were digitized either on a Nikon Super CoolScan 8000 scanner at 6.35-m intervals (representing 1.63-Å pixels at the specimen) or on a Zeiss SCAI microdensitometer at 7.0-m intervals (1.80-Å pixels) (Table 2). Only micrographs that exhibited negligible astigmatism and specimen drift or charging were selected for further processing. Estimation of the defocus level of each micrograph was performed using ctfﬁnd3 (30). RobEM (http://cryoem.ucsd.edu/programs/robem_usersguide_v4.01.htm) was used to ex- tract individual BRV virion images from the digitized micrographs (Table 2). We used the random model computation method to create an initial, low-resolution model of BRV (53) and then used the AUTO3DEM automated 3D reconstruc- tion package (version 4.01.8) to determine and reﬁne the origin and orientation parameters for each particle (54). The estimated resolution limits of all recon- structed density maps were computed based on a conservative Fourier shell correlation threshold value of 0.5 (FSC0.5) (51). Final density maps used for visualization, analysis, and interpretation were sharpened to enhance high-reso- lution features by imposing an inverse temperature factor (1/400 Å2) (22). 3D image reconstruction. Micrographs were digitized either on a Nikon Super CoolScan 8000 scanner at 6.35-m intervals (representing 1.63-Å pixels at the specimen) or on a Zeiss SCAI microdensitometer at 7.0-m intervals (1.80-Å pixels) (Table 2). Only micrographs that exhibited negligible astigmatism and specimen drift or charging were selected for further processing. Estimation of the defocus level of each micrograph was performed using ctfﬁnd3 (30). RobEM (http://cryoem.ucsd.edu/programs/robem_usersguide_v4.01.htm) was used to ex- tract individual BRV virion images from the digitized micrographs (Table 2). We used the random model computation method to create an initial, low-resolution model of BRV (53) and then used the AUTO3DEM automated 3D reconstruc- tion package (version 4.01.8) to determine and reﬁne the origin and orientation parameters for each particle (54). The estimated resolution limits of all recon- structed density maps were computed based on a conservative Fourier shell correlation threshold value of 0.5 (FSC0.5) (51). MATERIALS AND METHODS d Nikon-scanned data were interpolated to 1.80-Å pixels so that they could be on October 13, 2019 by gue http://jvi.asm.org/ Downloaded from on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from on October 13, 201 http://jvi.asm.org/ Downloaded from First-strand cDNA synthesis, PCR ampliﬁcation, and sequencing. To synthe- size viral cDNA, puriﬁed dsRNA was ﬁrst denatured in 30 mM methylmercury hydroxide for 10 min at room temperature, followed by addition of a poly(A) tail to both 3 ends of each dsRNA molecule using a Saccharomyces cerevisiae poly(A) polymerase (Epicentre Technologies). First-strand cDNA synthesis was next conducted using SuperScript III reverse transcriptase (Invitrogen) and an oligo(dT)20 primer with the 5 20-nucleotide (nt) extension GCGATAGTCTC CGTTCAGGA, followed by digestion of cRNA with RNase H (New England BioLabs). PCR was then performed at low annealing temperatures (50°C) with iProof DNA polymerase (Bio-Rad) for 35 cycles of ampliﬁcation, using a primer to the anchor sequence along with one of 28 internal primers based on reported skunk orthoreovirus partial sequences (GenBank accession numbers EU789369, EU789370, and EU789372) (52). Each ampliﬁed sample was then sequenced directly using the same internal primer that was used for ampliﬁcation, and sequences from the targeted BRV genome segments (L1 and M2) were identi- ﬁed by BLAST homology (http://blast.ncbi.nlm.nih.gov/) with previously re- ported sequences of other orthoreoviruses. These internal sequence toe holds were then used to design new primers for PCR amplifying the remaining portions of BRV L1 and M2, followed again by direct sequencing. In this manner, every nucleotide of the BRV L1 and M2 genome segments was read in both forward and reverse directions, except for about 50 nt at either end of each segment, which were read with high conﬁdence in a single (outward) direction. All primers were obtained from Integrated DNA Technologies. formation was extensive, but monolayers remained intact. Culture medium was removed, and cells were disrupted in TMN buffer (25 mM Tris, pH 7.5, 5 mM MgCl2, 150 mM NaCl) supplemented with 50 g/ml RNase A, 1% Triton X-100, and protease inhibitors (200 nM aprotinin, 1 M leupeptin, and 1 M pepstatin). The inclusion of RNase and Triton was required to remove the nonstructural protein NS, which tends to associate with BRV virions. MATERIALS AND METHODS Virion and dsRNA puriﬁcations. Vero cells were cultured in roller bottles at 37°C in medium 199 with Hank’s salts supplemented with 5% fetal bovine serum (FBS). Subconﬂuent monolayers (2 108 cells per bottle; 10 bottles) were infected with BRV at a multiplicity of infection of 3 to 5 PFU per cell, and infected cultures were incubated at 37°C for 21 h. At that time, syncytium VOL. 85, 2011 VIRION STRUCTURE OF BABOON REOVIRUS 7485 7485 TABLE 2. Parameters for cryoTEM and 3D image reconstruction of BRV virions Samplea No. of micrographs by instrument typeb Pixel size (Å) Defocus range (m) No. of boxed particle images/no. of boxed particle images used for 3D reconstruction Resolution (Å)c Polara Zeiss Nikon 200 keV 300 keV BRV 0 91 48 19 1.80d 0.25–5.09 2669/2138 9.0 BRV ﬂap 30 146 0 176 1.63 0.41–5.38 3778/2993 9.2 a BRV, baboon reovirus with turret ﬂaps; BRV ﬂap, baboon reovirus without turret ﬂaps. b No. of micrographs obtained or scanned with each indicated instrument. Zeiss, Zeiss SCAI microdensitometer (7.0-m step size); Nikon, Nikon Super CoolScan 8000 scanner (6.35-m step size). c Resolution of ﬁnal 3D reconstruction as estimated by FSC0.5 criterion (51). d Nikon-scanned data were interpolated to 1.80-Å pixels so that they could be merged with Zeiss-scanned data. TABLE 2. Parameters for cryoTEM and 3D image reconstruction of BRV virions TABLE 2. Parameters for cryoTEM and 3D image reconstruction of BRV virions Samplea No. of micrographs by instrument typeb Pixel size (Å) Defocus range (m) No. of boxed particle images/no. of boxed particle images used for 3D reconstruction Resolution (Å)c Polara Zeiss Nikon 200 keV 300 keV BRV 0 91 48 19 1.80d 0.25–5.09 2669/2138 9.0 BRV ﬂap 30 146 0 176 1.63 0.41–5.38 3778/2993 9.2 a BRV b b i ith t t ﬂ BRV b b i ith t t t ﬂ No. of micrographs by instrument typeb a BRV, baboon reovirus with turret ﬂaps; BRV ﬂap, baboon reovirus without turret ﬂaps. b No. of micrographs obtained or scanned with each indicated instrument. Zeiss, Zeiss SCA 8000 scanner (6.35-m step size). ( p ) c Resolution of ﬁnal 3D reconstruction as estimated by FSC0.5 criterion (51). d Nikon-scanned data were interpolated to 1.80-Å pixels so that they could be merged with Zeiss-scanned data. c Resolution of ﬁnal 3D reconstruction as estimated by FSC0.5 criterion (51). RESULTS Electron cryomicrographs of puriﬁed BRV virions. BRV virions were puriﬁed from infected Vero cells by differential and density gradient centrifugation. Denaturing gel electro- phoresis of the puriﬁed virions showed the expected pattern of , , and structural proteins (Fig. 1A) as well as the expected pattern of L, M, and S dsRNA genome segments (Fig. 1B) (16). Transmission electron cryomicrographs were recorded from unstained, vitriﬁed samples of the puriﬁed virions and showed ﬁelds of regularly sized, intact particles with diameters near 850 Å (Fig. 1C). As evident in the ﬁgure, the projected proﬁles of individual particles appear largely round though some are more angular. There are no evident structures pro- jecting far out from the particle surfaces. Striated surface fea- tures viewed in proﬁle in some areas are reminiscent of regu- larly spaced subunits. Overall, the images are very similar to those of other orthoreoviruses (15, 58). on October 13, 2019 by guest /jvi.asm.org/ ( ) Cryo-reconstruction of BRV virions. Using current process- ing procedures (49), we obtained a cryo-reconstruction at 9.0-Å resolution from 2,138 particle images of the BRV viri- ons. A density-coded equatorial section of the cryo-reconstruc- tion highlights the approximately round proﬁle of the virion’s outer surface, with ﬂattened areas, or “divots,” around the icosahedral 5f axes (Fig. 2, top left). These divots are explained by absence of the short projections, each 40 Å high, that regularly stud the remainder of the surface. The outermost radius approximates 425 Å to the tops of these projections near the icosahedral 2-fold (2f) and 3-fold (3f) axes and 355 Å near the icosahedral 5f axes. A radially color-coded surface view of the BRV virion in stereo (Fig. 2, bottom) clariﬁes these surface features and is discussed in detail below. The inner- most radius of the BRV capsid layers approximates 250 Å near the icosahedral 2f and 3f axes and is less near the 5f axes though its limit is hard to discern at those positions (Fig. 2, top left). Outer and inner capsids are separated by large solvent cavities at many positions in the equatorial section. Around the icosahedral 2f and 3f axes, these cavities appear smaller be- cause “nodules” reach up by about 25 Å at regular intervals from the top of the inner capsid to contact the bottom of the outer capsid. Around each 5f axis, the solvent cavity is larger and heart shaped in proﬁle. MATERIALS AND METHODS Puriﬁed BRV virions. (A) Protein gel lane of puriﬁed viri- ons. Positions of the structural proteins are labeled at right. Size marker positions (in kDa) are indicated at left. (B) dsRNA gel lane of puriﬁed virions. Positions of the 10 genome segments are labeled at right. Size marker positions (in bp) are indicated at left. (C) Trans- mission electron cryomicrograph of unstained, vitriﬁed BRV virions, with scale bar as indicated. on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from MATERIALS AND METHODS The entire asymmetric unit containing one or more copies of all ﬁve major structural proteins of MRV (two core shell 1 subunits, two and a half core clamp 2 subunits, one core turret 2 subunit, 10 outer shell 1 subunits, and 10 outer clamp 3 subunits) was then ﬁtted into each BRV map without any manual adjustments. The resulting, overall ﬁt among all of these structural components was extremely good, as described below. Images of the MRV 2 and 1 subunits from this ﬁt were then extracted for depictions in ﬁgures (see Fig. 6 and 7). Phylogenetic analysis. Phylogenetic analysis (see Fig. 8) was performed using programs implemented at http://phylogeny.fr/, with default settings unless oth- erwise indicated. Multiple-sequence alignment was performed with the program YAN ET AL. J. VIROL. 7486 center-to-center distance of about 30 Å and attributable to genomic dsRNA, are evident in the particle interior (Fig. 2, top left). The RNA rings are interrupted at the icosahedral 5f axes FIG. 1. Puriﬁed BRV virions. (A) Protein gel lane of puriﬁed viri- ons. Positions of the structural proteins are labeled at right. Size marker positions (in kDa) are indicated at left. (B) dsRNA gel lane of puriﬁed virions. Positions of the 10 genome segments are labeled at right. Size marker positions (in bp) are indicated at left. (C) Trans- mission electron cryomicrograph of unstained, vitriﬁed BRV virions, with scale bar as indicated. MUSCLE, version 3.7, without subsequent curation. Maximum-likelihood anal- ysis was then performed with the program PhyML, version 3.0, using the chi- square parametric method of the approximate-likelihood ratio test to generate conﬁdence scores. The tree was drawn with the program TreeDyn, version 198.3, and reﬁned for publication with the program FigTree, version 1.3.1. GenBank accession numbers for published core clamp protein sequences used in this analysis were AAA47239 (MRV-T1L), AAA47248 (MRV-T2J), AAA47261 (MRV type 3 Dearing [MRV-T3D]), AAC18124 (BRV), AAY54282 (BroV), AAC18123 (NBV), AAR13234 (Pulau reovirus), ABM67658 (Melaka reovirus), AAC18122 (ARV-138), AAC18121 (ARV-176), ABK96865 (ARV Muscovy duck S12), ABY57290 (psittacine reovirus), AAM93413 (SBRV), AAM92740 (GCRV), and ABV01047 (AGCRV). Nucleotide sequence accession numbers. GenBank accession numbers for newly determined BRV sequences reported here are HQ847903 for L1/A and HQ847907 for M2/B. GenBank accession numbers for additional new BRV sequences to be reported separately are HQ847904 for L2/B, HQ847905 for L3/C, HQ847906 for M1/A, and HQ847908 for M3/NS. on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from FIG. 1. RESULTS Punctate and linear features visible within most capsid regions appear likely to represent second- ary-structure elements including -helices and -sheets. Two or three regularly spaced rings of density, with an average FIG. 1. Puriﬁed BRV virions. (A) Protein gel lane of puriﬁed viri- ons. Positions of the structural proteins are labeled at right. Size marker positions (in kDa) are indicated at left. (B) dsRNA gel lane of puriﬁed virions. Positions of the 10 genome segments are labeled at right. Size marker positions (in bp) are indicated at left. (C) Trans- mission electron cryomicrograph of unstained, vitriﬁed BRV virions, with scale bar as indicated. center-to-center distance of about 30 Å and attributable to genomic dsRNA, are evident in the particle interior (Fig. 2, top left). The RNA rings are interrupted at the icosahedral 5f axes by what appear to be limited regions of protein density that project inward from the bottom of the inner capsid, as also seen in MRV and attributable to the viral transcriptase com- plex (14, 59). Overall similarities of these BRV features to those of the MRV virion (Fig. 2, top right) are obvious. center-to-center distance of about 30 Å and attributable to genomic dsRNA, are evident in the particle interior (Fig. 2, top left). The RNA rings are interrupted at the icosahedral 5f axes by what appear to be limited regions of protein density that project inward from the bottom of the inner capsid, as also seen in MRV and attributable to the viral transcriptase com- plex (14, 59). Overall similarities of these BRV features to those of the MRV virion (Fig. 2, top right) are obvious. BRV L1 and M2 sequences. To assist in further interpreting features of the BRV virion, we determined the full-length cDNA sequences of the genome segments encoding the core turret and outer shell proteins. The full-length cDNA sequence has been previously reported for the outer clamp protein B (16) but not for these other two proteins that are expected to cover the remaining bulk of the BRV surface. Based on the lengths of the cDNA sequences recently determined for the other four L and M genome segments of BRV (R. P. Good- BRV L1 and M2 sequences. To assist in further interpreting features of the BRV virion, we determined the full-length cDNA sequences of the genome segments encoding the core turret and outer shell proteins. RESULTS The full-length cDNA sequence has been previously reported for the outer clamp protein B (16) but not for these other two proteins that are expected to cover the remaining bulk of the BRV surface. Based on the lengths of the cDNA sequences recently determined for the other four L and M genome segments of BRV (R. P. Good- VIRION STRUCTURE OF BABOON REOVIRUS 7487 VOL. 85, 2011 FIG. 2. Cryo-reconstruction of the BRV virion at 9.0-Å resolution. (Top) Density-coded, equatorial 2f-axis sections are shown for BRV and MRV. Highest and lowest densities are depicted in black and white, respectively. The MRV map is that from Zhang et al. (57) recalculated at 9.0 Å. Icosahedral symmetry axes (2f, 3f, and 5f) are labeled for one quadrant of MRV. A difference in the relative erectness of the outer clamp subunits of BRV and MRV is highlighted by black arrows. RNA ring densities in the two viruses are highlighted by white arrows. (Bottom) Stereo surface representation of the BRV cryo-reconstruction viewed along an icosahedral 2f axis with radial color cueing applied to highlight features at different radii (from blue outside to red inside). The scale bar applies to the entire ﬁgure. p j g on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from on October 13, 2019 by guest http://jvi.asm.org/ aded from FIG. 2. Cryo-reconstruction of the BRV virion at 9.0-Å resolution. (Top) Density-coded, equatorial 2f-axis sections are shown for BRV and MRV. Highest and lowest densities are depicted in black and white, respectively. The MRV map is that from Zhang et al. (57) recalculated at 9.0 Å. Icosahedral symmetry axes (2f, 3f, and 5f) are labeled for one quadrant of MRV. A difference in the relative erectness of the outer clamp subunits of BRV and MRV is highlighted by black arrows. RNA ring densities in the two viruses are highlighted by white arrows. (Bottom) Stereo surface representation of the BRV cryo-reconstruction viewed along an icosahedral 2f axis with radial color cueing applied to highlight features at different radii (from blue outside to red inside). The scale bar applies to the entire ﬁgure. ober 13, 2019 by guest man., J. Shou., R. Duncan, and M. L. Nibert, unpublished data), the segments encoding the core turret and outer shell proteins were found to be L1 and M2, respectively, and the encoded BRV proteins were therefore designated A and B (Table 1). RESULTS The number at the right end of each line indicates the position of the last amino acid in the sequence. Identities (), strong similarities (:), and weaker similarities (.) in the aligned sequences are indicated at bottom. Other annotations are for the N-terminal myristoylation site (myr), the conserved myristoylation consensus sequence (ﬁlled circles), the putative transmembrane -strands (solid lines at top) described in the text, the autolytic cleavage site (auto, arrow), and the disulﬁde-bonding Cys residue of MRVs (ds). on October 13, 201 http://jvi.asm.org/ Downloaded from on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from FIG. 3. Multiple-sequence alignment of ortho- and aquareovirus outer shell proteins including the newly determined sequence from BRV (bold). N-terminal (top) and C-terminal (bottom) portions of the alignment are shown. The alignment was generated using the program Clustal W2 as implemented at http://guidance.tau.ac.il/, using sequences as identiﬁed in Materials and Methods. The number at the right end of each line indicates the position of the last amino acid in the sequence. Identities (), strong similarities (:), and weaker similarities (.) in the aligned sequences are indicated at bottom. Other annotations are for the N-terminal myristoylation site (myr), the conserved myristoylation consensus sequence (ﬁlled circles), the putative transmembrane -strands (solid lines at top) described in the text, the autolytic cleavage site (auto, arrow), and the disulﬁde-bonding Cys residue of MRVs (ds). on October 13, 2019 by guest http://jvi.asm.org/ rom the extreme N-terminal myristoylation motif (1-MGNASS-6 in MRV-T3D; 1-MGAGSS-6 in BRV) (35), a -hairpin motif implicated in membrane interaction by the 1N peptide (11-I NVTGDGNVFKPSAETSSTAVPSLSLSPG-39 in MRV-T3D [predicted strands underlined]; 11-FEVYGDGNVFSPDAR QATSAVPQLDLNIG-39 in BRV) (56), and the autocleavage motif spanning the 1N/1C junction [42-(N/P)GG-45 in both MRV-T3D and BRV; cleavage between Asn42 and Pro43 as indicated by a slash] (34). not conserved in BRV or in other ortho- or aquareoviruses. A few long regions notable for variability and poor scores in the alignments include BRV aa 693 to 722 (MRV-T3D; aa 696 to 725), which contributes to a small domain that bridges the methyltransferase domains and probably serves as a structural support for them as seen in MRV 2 (44), and BRV aa 1056 to 1104 (MRV-T3D aa 1060 to 1113), which contributes to the ﬁrst (N-terminal-most) of the three C-terminal ﬂap domains as seen in MRV 2 (44). ober 13, 2019 by guest y ] ( ) 3D structure of the BRV virion. RESULTS A stereo surface view of the BRV virion (Fig. 2, bottom) reveals many characteristic fea- tures such as the radially directed P2 and P3 solvent channels that perforate the outer capsid, the small projections attribut- able to outer clamp protein B that cover much of the virion surface, and the extent to which the ﬂattened top of the pen- tameric core turret is inset (i.e., having divots) relative to other regions of the surface. These and other surface details are quite similar to those of MRV (15, 57, 59) though closer inspection reveals three notable differences, as follows. (i) An open P1 channel is present at the icosahedral 5f axes of BRV, through which underlying portions of the core turret and core shell proteins are visible from outside the virion (Fig. 4, top). This ﬁnding is indicative of the fact that BRV lacks an adhe- sion ﬁber, the base of which obscures this channel in MRV and ARV virions by inserting into the top of the pentameric core turret at this position (15). (ii) The hub-and-spoke complexes in the P2 and P3 channels of MRV (57) are missing from those channels in BRV, allowing underlying portions of the core clamp and core shell proteins to be visible again from outside the virion (Fig. 4, bottom). BRV is like ARV in this regard (58). The absence of these complexes in BRV is consistent with new sequence evidence above, namely, that the BRV outer ( ) (ii) M2/B. The M2 genome segment is 2,143 bp long, and its plus strand contains one long ORF spanning nt positions 24 to 2054. The encoded B protein comprises 676 aa, with a calculated molecular mass of 74 kDa. In pairwise sequence alignments, it shows clear homology to the outer shell protein of representative MRV and ARV strains (1 and B; 59 to 64% similarity and 33 to 39% identity) as well as to the outer shell protein of aquareoviruses (VP4; 47 to 48% similarity and 20 to 22% identity). This protein is also known to mediate cellular membrane penetration during cell entry by MRV and presumably other ortho- and aquareoviruses as well (6, 25). In multiple sequence alignments, the BRV B protein aligns very well with the homologous proteins of other ortho- and aquareoviruses. RESULTS Both are discussed below. core turret protein of ortho- and aquareoviruses is known to comprise multiple domains (10, 44), and the end-to-end align- ments indicate that the homologous BRV protein contains all of these domains, including the C-terminal ﬂap domains (see below) that are missing from the core turret proteins of certain, more divergent members of the subfamily Spinareovirinae (9, 31). Speciﬁc residues of apparent functional importance for the substrate binding or enzymatic activities are conserved in the RNA 5 guanylyltransferase domain (e.g., Lys171, Lys190, His223, and His232 in MRV-T3D; Lys172, Lys191, His221, and His230 in BRV), in the RNA methyltransferase-1 domain (e.g., Ser518, His521, Tyr552, Asp553, Asp561, Asp577, and Val578 in MRV-T3D; Ser515, His518, Tyr548, Asp549, Asp557, Asp573, Val574, and 570-VYCDVDQV-577 in BRV), and in the RNA methyltransferase-2 domain (e.g., Asn812, Asp827, Gly829, Asp850, Asp871, Tyr872, and Ser891 in MRV-T3D; Asn807, Asp822, Gly824, Asp845, Asp866, Tyr867, and Ser886 in BRV) (5, 26, 28, 41, 44). Two potential integrin-binding motifs conserved in MRV 2 as noted by Maginnis et al. (29) (882-RGD-884 and 1083-KGE-1085) are ( ) (i) L1/A. The L1 genome segment is 3,903 bp long, and its plus strand contains one long open reading frame (ORF) span- ning nucleotide positions 18 to 3872. The encoded A protein comprises 1,284 amino acids (aa), with a calculated molecular mass of 147 kDa. In pairwise sequence alignments, it exhibits homology to the core turret protein of representative MRV and ARV strains (2 and C; 51 to 53% similarity and 26 to 29% identity) as well as to the core turret protein of aquareo- viruses (VP1; 49 to 50% similarity and 23 to 25% identity). This protein is also known to mediate three of the four enzy- matic reactions during cotranscriptional 5 capping of or- thoreovirus plus-strand RNAs (26, 28, 44). In multiple se- quence alignments, the BRV A protein aligns well—end to end and with only occasional, small internal gaps—with the homologous protein of other ortho- and aquareoviruses. The J. VIROL. J. VIROL. 7488 YAN ET AL. YAN ET AL. FIG. 3. Multiple-sequence alignment of ortho- and aquareovirus outer shell proteins including the newly determined sequence from BRV (bold). N-terminal (top) and C-terminal (bottom) portions of the alignment are shown. The alignment was generated using the program Clustal W2 as implemented at http://guidance.tau.ac.il/, using sequences as identiﬁed in Materials and Methods. RESULTS The alignment includes not only occasional, small, internal gaps scattered throughout the sequences but also an 30-aa C-terminal extension present in MRVs and PRV but missing from BRV, BroV, ARVs, and aquareoviruses (Fig. 3). In the MRV virion, this C-terminal extension is known to form the hub-and-spoke complexes visible by cryo-EM within the P2 and P3 channels of the MRV outer capsid (57), structures that are missing from ARV and aquareovirus virions (10, 58). On the other hand, sequences near the N terminus are among the best conserved in the alignment (Fig. 3), including VIRION STRUCTURE OF BABOON REOVIRUS VOL. 85, 2011 7489 FIG. 4. Close-up surface views of BRV and MRV virion cryo-reconstructions. The map and color coding of the BRV virion (left) are the same as used as in Fig. 2, but the map is shown in close-up view to highlight the P1 and P2 (top) and P3 (bottom) channel areas (arrows). Matching regions of the MRV virion surface are shown on the right for comparison to BRV. The MRV map is the same as that for Fig. 2, with radial color cueing applied to match that for BRV. on October 13, 2019 by g http://jvi.asm.org/ Downloaded from on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from on October 13, 2019 by guest http://jvi.asm.org/ ded from on October 13, 2019 by guest vi.asm.org/ FIG. 4. Close-up surface views of BRV and MRV virion cryo-reconstructions. The map and color coding of the BRV virion (left) are the same as used as in Fig. 2, but the map is shown in close-up view to highlight the P1 and P2 (top) and P3 (bottom) channel areas (arrows). Matching regions of the MRV virion surface are shown on the right for comparison to BRV. The MRV map is the same as that for Fig. 2, with radial color cueing applied to match that for BRV. ober 13, 2019 by guest positions (10, 32). As one examines more carefully for speciﬁc high-density features at particular radii in the BRV and MRV capsids, which likely represent individual -helices and -sheets, one again sees a number of similarities between BRV and MRV (Fig. 5, right columns), suggesting that many indi- vidual components of their protein folds are conserved. RESULTS For example, again at radius 275 Å, there is a curving helix in the guanylyltransferase domain at the base of the core turret pro- tein on the perimeter of this domain facing the P1 channel, which is present in both BRV and MRV. In addition, at radius 329 Å in both BRV and MRV, intersecting, straight helices adopt a “backward-L” pattern in each outer shell protein sub- unit, such that the perimeter of each outer shell trimer is rimmed by these features. Both of these exempliﬁed helical features are indeed present in X-ray crystal structures of the MRV core turret and outer shell proteins (25, 44) (Fig. 6). On the other hand, one can also ﬁnd many differences in high- density features between BRV and MRV at essentially all capsid radii. Such differences are consistent with the fact that all ﬁve major BRV and MRV structural proteins demonstrate 41% sequence identity in pairwise alignments: core shell, 41%; core turret, 29%; core clamp, 29%; outer shell, 33%; and outer clamp, 18%) (16; also this study and R. P. Goodman, unpublished). shell protein is missing the C-terminal sequences that form these complexes in MRV (Fig. 3). (iii) There are differences in the surface topography of the outer clamp proteins of BRV and MRV, that is, differences in the positions and sizes of various “bumps” on the protein surfaces. In addition, as per- haps best seen in the equatorial sections of the two virions (Fig. 2, top), the outer clamp subunits of BRV are “less erect” than those of MRV. For example, the outer clamp subunits that ﬂank the icosahedral 2f axis are less parallel to that axis in BRV than in MRV. Such differences are consistent with the fact that the outer clamp proteins of BRV and MRV share the lowest level of sequence identity in pairwise alignments (20%) among all of the structural proteins of these viruses (16; also this study and R. P. Goodman et al., unpublished), especially in the sequences that have been shown to form the upper lobe of the MRV outer clamp protein 3 (16, 25, 37). Density-coded radial sectioning of the cryo-reconstructions provided a means to compare the more internal features of the BRV and MRV capsids. Overall features are similar between BRV and MRV at most radii (Fig. 5, left columns). RESULTS One notable observation is that sections through the core clamp protein, such as at 275 Å (Fig. 5), reveal that both BRV and MRV have 150 of these subunits per virion, which is also like ARV (58) but unlike aquareoviruses, which have only 120 per virion because none are present at the 30 icosahedral 2f-axis 7490 YAN ET AL. J. VIROL. 7490 YAN ET AL. FIG. 5. Selected radial sections of BRV and MRV virion cryo-reconstructions. Density-coded radial sections (each 2.3 Å thick) are shown for BRV and MRV (left columns as labeled for each virus). The particle radius at which each section was generated is indicated at left. Corner brackets frame the region in each section that is shown in close-up view in one of the right columns. The proteins for which densities are visible in each close-up view are indicated. The MRV map is the same as that for Fig. 2. Filled ellipse, triangle, and pentagon symbols demarcate the positions of 2f, 3f, and 5f axes in each close-up panel. FIG. 5. Selected radial sections of BRV and MRV virion cryo-reconstructions. Density-coded radial sections (each 2.3 Å thick) are s BRV and MRV (left columns as labeled for each virus). The particle radius at which each section was generated is indicated at left. Corner frame the region in each section that is shown in close-up view in one of the right columns. The proteins for which densities are visibl close-up view are indicated. The MRV map is the same as that for Fig. 2. Filled ellipse, triangle, and pentagon symbols demarcate the of 2f, 3f, and 5f axes in each close-up panel. on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from on October 13, 2019 by guest http://jvi.asm.org/ aded from on October 13, 2019 by guest jvi.asm.org/ ober 13, 2019 by guest FIG. 5. Selected radial sections of BRV and MRV virion cryo-reconstructions. Density-coded radial sections (each 2.3 Å thick) are shown for BRV and MRV (left columns as labeled for each virus). The particle radius at which each section was generated is indicated at left. Corner brackets frame the region in each section that is shown in close-up view in one of the right columns. The proteins for which densities are visible in each close-up view are indicated. The MRV map is the same as that for Fig. 2. RESULTS Filled ellipse, triangle, and pentagon symbols demarcate the positions of 2f, 3f, and 5f axes in each close-up panel. the correspondence between individual secondary-structure el- ements within the outer clamp protein subunits of MRV and BRV is poorer, especially within their outer lobes, consistent with the greater sequence divergence of these proteins. Fitting MRV crystal structures into the BRV virion cryo- reconstruction. To assist with comparisons, we also ﬁtted the atomic model of an asymmetric unit containing the ﬁve major capsid proteins of the MRV virion (37, 44, 57) into the BRV cryo-reconstruction. The results corroborated that the capsid positions and overall domain organizations of these proteins are largely the same in the two viruses. For instance, the series of domains in the core turret protein subunit and the trimeric organization of the outer shell protein are readily seen to be shared between MRV and BRV (Fig. 6). For these two pro- teins, as well as the core shell and core clamp proteins, the goodness of ﬁt extends even to many individual secondary- structure elements within each subunit. As examples, several -helices in the MRV core turret and outer shell proteins are seen to ﬁt very well into corresponding rod-like densities in the BRV cryo-reconstruction (Fig. 6), including the backward-L pattern of intersecting helices that rim the 1 trimer, as also highlighted in the radial sections (Fig. 5). On the other hand, g q g p Cryo-reconstruction and 3D structure of BRV virions miss- ing the A ﬂap domains. During the course of this study, we encountered many frozen grids in which the BRV particles, when separately subjected to 3D image reconstruction (3,798 particles; 9.2-Å resolution), were found to be missing the C- terminal ﬂap domains of the core turret protein (Fig. 7A and B). Difference maps between these “ﬂapless” particles and those with ﬂaps showed the missing ﬂap domains to be the only substantive difference (Fig. 7C). The phenomenon was ob- served with three different preparations of puriﬁed BRV viri- ons and moreover was conﬁned to only certain grids; in other words, from the same virion preparation, the particles from some grids did not show ﬂaps in the subsequent reconstruction, whereas the particles from other grids did. SDS-PAGE of one VIRION STRUCTURE OF BABOON REOVIRUS 7491 VOL. 85, 2011 FIG. 6. Atomic model of the MRV asymmetric unit ﬁtted into the BRV cryo-reconstruction. RESULTS Excerpted portions are shown in stereo to illustrate goodness of ﬁt for the core turret protein (top) and the outer shell protein (bottom). For the core turret protein, the MRV 2 model is shown in magenta for the three ﬂap domains and in green for the other domains of one subunit. For the outer shell protein, the MRV 1 model is shown in a different color (red, yellow, or cyan) for the three subunits within one trimer. on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from on October 13, 2019 by guest http://jvi.asm.org/ aded from FIG. 6. Atomic model of the MRV asymmetric unit ﬁtted into the BRV cryo-reconstruction. Excerpted portions are shown in stereo to illustrate goodness of ﬁt for the core turret protein (top) and the outer shell protein (bottom). For the core turret protein, the MRV 2 model is shown in magenta for the three ﬂap domains and in green for the other domains of one subunit. For the outer shell protein, the MRV 1 model is shown in a different color (red, yellow, or cyan) for the three subunits within one trimer. FIG. 6. Atomic model of the MRV asymmetric unit ﬁtted into the BRV cryo-reconstruction. Excerpted portions are shown in stereo to illustrate goodness of ﬁt for the core turret protein (top) and the outer shell protein (bottom). For the core turret protein, the MRV 2 model is shown in magenta for the three ﬂap domains and in green for the other domains of one subunit. For the outer shell protein, the MRV 1 model is shown in a different color (red, yellow, or cyan) for the three subunits within one trimer. cussion), has been observed in association with structural tran- sitions between virions and cores during MRV particle disassembly and assembly (7, 15) but not within MRV virions per se. The capacity for such variable arrangement in BRV virions may reﬂect the absence of an outer ﬁber protein, which inserts at the icosahedral 5f axes in MRV and ARV virions (15, 58) and thereby stabilizes the ﬂaps of those viruses in the “down” position. Why ﬂap rearrangement occurred irregularly, and only on certain grids, with the BRV virions is difﬁcult to answer but presumably reﬂects that the potential for move- ment of these domains was actuated only under certain speciﬁc conditions. RESULTS such virion preparation, after the ﬂaps were found to be miss- ing from the cryo-reconstruction, showed that the proteins appeared to remain intact, suggesting that the ﬂap had not been removed (e.g., proteolytically) during storage. Instead, the ﬂaps were either removed or exhibited irregular mobility only after being placed on those grids. y g p g By examining the ﬁtted MRV core turret protein structure within the context of the ﬂapless BRV particles, we were able to discern that the cryo-reconstruction of these particles is missing all three of the C-terminal domains (designated A, B, and C) that constitute the ﬂap (MRV 2 aa 1024 to 1289; BRV A aa 1020 to 1284 according to our alignment) of each BRV core turret subunit (Fig. 7D and E). Based on these observa- tions, we conclude that the ﬂap domains of the BRV core turret protein are capable of assuming variable positions within the virion, probably consequent to “swiveling” about a hinge near amino acid position 1020 in BRV A, within a region of sequence that forms an extended chain between the methyl- transferase-2 domain and the A (i.e., most N-terminal) ﬂap domain, as seen in MRV 2 (44). A similar type of ﬂap rear- rangement, though involving a different hinge region (see Dis- DISCUSSION (D and E) Close-up view down the icosahedral 5f axis of a single core turret from the cryo-reconstruction of ﬂapless BRV virions. In panel D, a pentamer of the atomic model of the full-length MRV core turret protein 2 is ﬁtted. The ﬂap domains are shown in magenta, and the other domains are shown in green. In panel E, a pentamer of the atomic model of the MRV core turret protein 2 minus its ﬂap domains is ﬁtted, with the remaining domains shown in green. The scale bar applies to panels D and E. per virion and the tops of which are also well exposed on the BRV virion surface. The homologous MRV protein 2 has been suggested to play a role in coreceptor (1 integrin) bind- ing and uptake of virions from the cell surface (29); however, proposed integrin-binding sites in the MRV core turret protein are not conserved in the BRV protein. the cryo-reconstructions of ARV and MRV virions (15, 58) are indeed missing from the cryo-reconstruction of the BRV virion shown here, resulting in an open P1 solvent channel at the icosahedral 5f axis. Thus, although it was previously conceiv- able that an outer ﬁber protein might be encoded by one of the L or M genome segments of BRV that have yet to be se- quenced, the current results suggest that this is unlikely and that BRV does indeed lack an adhesion ﬁber anchored at the icosahedral 5f axes or elsewhere. Completion of the BRV genome sequences in this and a pending report (R. P. Good- man et al., unpublished) conﬁrm that the BRV L and M seg- ments encode only the expected - and -class homologs of ARV and MRV proteins and thus lack the coding capacity for an outer ﬁber protein (Table 1). p Aquareoviruses also lack an outer ﬁber protein and more- over have an altered conformation of their core turret protein (1, 10, 19, 32). The conformation of the aquareovirus turret protein differs in that the ﬂap domains have pivoted around a hinge point following the ﬁrst (A) domain so that the second and third (B and C) domains project to higher radii and further away from the icosahedral 5f axes (10), thereby opening a larger P1 channel than do the homologous, down-positioned ﬂap B and C domains in MRV, ARV, and BRV virions. DISCUSSION The region of the missing ﬂap domain densities of one core turret is indicated by an arrow. (B) Stereo surface representation viewed down an icosahedral 2f axis, comparable to the stereo view of the ﬂap-containing BRV virion shown at the bottom of Fig. 2 with the same radial color map employed. The scale bar applies to panels A, B, and C. (C) Difference map obtained by subtracting the cryo-reconstruction of the BRV virion lacking core turret ﬂaps from that of the BRV virion containing core turret ﬂaps. The left side shows a density-coded, equatorial 2f-axis section of the difference map. The right side shows a radial section at 345 Å through the level of the ﬂap domains in the difference map, as indicated by the dashed half circle on the left side. (D and E) Close-up view down the icosahedral 5f axis of a single core turret from the cryo-reconstruction of ﬂapless BRV virions. In panel D, a pentamer of the atomic model of the full-length MRV core turret protein 2 is ﬁtted. The ﬂap domains are shown in magenta, and the other domains are shown in green. In panel E, a pentamer of the atomic model of the MRV core turret protein 2 minus its ﬂap domains is ﬁtted, with the remaining domains shown in green. The scale bar applies to panels D and E. FIG. 7. Cryo-reconstruction at 9.2-Å resolution of ﬂapless BRV virions. (A) Density-coded, equatorial 2f-axis section, comparable to the sections shown at top of Fig. 2. The region of the missing ﬂap domain densities of one core turret is indicated by an arrow. (B) Stereo surface representation viewed down an icosahedral 2f axis, comparable to the stereo view of the ﬂap-containing BRV virion shown at the bottom of Fig. 2 with the same radial color map employed. The scale bar applies to panels A, B, and C. (C) Difference map obtained by subtracting the cryo-reconstruction of the BRV virion lacking core turret ﬂaps from that of the BRV virion containing core turret ﬂaps. The left side shows a density-coded, equatorial 2f-axis section of the difference map. The right side shows a radial section at 345 Å through the level of the ﬂap domains in the difference map, as indicated by the dashed half circle on the left side. DISCUSSION Because the S genome segments of BRV lack coding capac- ity for a homolog of the outer ﬁber protein encoded by most other orthoreoviruses (13, 16), BRV has been thought likely to contain only seven structural proteins, like aquareoviruses (1, 19) but unlike ARV and MRV strains, which have eight (Table 1). Limited densities corresponding to the adhesion ﬁbers in J. VIROL. 7492 YAN ET AL. 7492 YAN ET AL. YAN ET AL. 7492 FIG. 7. Cryo-reconstruction at 9.2-Å resolution of ﬂapless BRV virions. (A) Density-coded, equatorial 2f-axis section, comparable to the sections shown at top of Fig. 2. The region of the missing ﬂap domain densities of one core turret is indicated by an arrow. (B) Stereo surface representation viewed down an icosahedral 2f axis, comparable to the stereo view of the ﬂap-containing BRV virion shown at the bottom of Fig. 2 with the same radial color map employed. The scale bar applies to panels A, B, and C. (C) Difference map obtained by subtracting the cryo-reconstruction of the BRV virion lacking core turret ﬂaps from that of the BRV virion containing core turret ﬂaps. The left side shows a density-coded, equatorial 2f-axis section of the difference map. The right side shows a radial section at 345 Å through the level of the ﬂap domains in the difference map, as indicated by the dashed half circle on the left side. (D and E) Close-up view down the icosahedral 5f axis of a single core turret from the cryo-reconstruction of ﬂapless BRV virions. In panel D, a pentamer of the atomic model of the full-length MRV core turret protein 2 is ﬁtted. The ﬂap domains are shown in magenta, and the other domains are shown in green. In panel E, a pentamer of the atomic model of the MRV core turret protein 2 minus its ﬂap domains is ﬁtted, with the remaining domains shown in green. The scale bar applies to panels D and E. 7492 YAN ET AL. J. VIROL. FIG. 7. Cryo-reconstruction at 9.2-Å resolution of ﬂapless BRV virions. (A) Density-coded, equatorial 2f-axis section, comparable to the on October 13, 2019 by guest http://jvi.asm.org/ aded from on October 13, 2019 by guest jvi.asm.org/ FIG. 7. Cryo-reconstruction at 9.2-Å resolution of ﬂapless BRV virions. (A) Density-coded, equatorial 2f-axis section, comparable to the sections shown at top of Fig. 2. VIRION STRUCTURE OF BABOON REOVIRUS A reciprocal possibility is that the open P2 and P3 chan- nels, as well as the open P1 channels, in BRV and others may facilitate or otherwise inﬂuence conformational changes in the outer capsid that are associated with cell entry (6, 15, 25). Like BRV, ARV, and aquareoviruses, BroV also lacks a C-terminal extension on its outer shell protein, but, interestingly, PRV contains an extension, though not the disulﬁde-bonding Cys residue found in MRVs (Fig. 3). g p In both MRV and aquareovirus particles, pivoting of the core turret ﬂap domains occurs during the transitions between virions and cores that accompany particle disassembly and assembly (7, 10, 15, 32). In both cases, the hinge point for these rearrangements appears to be in an extended region of struc- ture between the A and B ﬂap domains in each subunit, such that the A domain remains largely in the same position in virions and cores. The new results for BRV virions, however, point to rearrangements also involving the A domain, with the main hinge region being between the methyltransferase-2 do- main and the ﬂap A domain. The ﬂap rearrangements in BRV virions are therefore more comparable to those that occur in MRV cores upon heating to 52°C, in which the A ﬂap domain of each core turret subunit is also involved (27). The results for BRV virions are also distinct in that all three ﬂap domains of each subunit appear to assume variable positions such that they are missing from the cryo-reconstruction of the ﬂapless particles, whereas in MRV and aquareovirus particles, though pivoted, the B and C ﬂap domains of each subunit assume regular positions and remain well visualized in the reconstruc- tions of both virions and cores (7, 10, 15, 32). ober 13, 2019 by guest ( g ) The present cryo-reconstruction of BRV highlights struc- tural differences in the ortho- and aquareoviruses that cross taxonomic boundaries. For instance, while BRV resembles ARV and MRV in containing 150 copies of the core clamp protein, rather than 120 copies as in aquareoviruses, BRV resembles aquareoviruses in its lack of an adhesion ﬁber and dissimilarities in the arrangement or stability of its core turret structure compared to ARV and MRV (10, 15, 32, 58). VIRION STRUCTURE OF BABOON REOVIRUS VOL. 85, 2011 VIRION STRUCTURE OF BABOON REOVIRUS 7493 FIG. 8. Phylogenetic tree and evolution schematic. A maximum-likelihood phylogenetic tree is shown for the core clamp proteins of the indicated ortho- and aquareovirus species or tentative species. Species names are abbreviated as follows: MRV, Mammalian orthoreovirus; BRV, Baboon orthoreovirus; NBV, Nelson Bay orthoreovirus; ARV, Avian orthoreovirus; AqRVA, Aquareovirus A; AqRVC, Aquareovirus C; and AqRVG, Aquareovirus G. For MRV, NBV, and ARV, three representative isolates were used in generating the tree. Tentative new species are represented by Broome virus (BroV) and psittacine reovirus (PsRV) isolates. Branch labels are conﬁdence scores (%). The tree is rooted on the aquareovirus clade. The scale bar indicates horizontal branch lengths in units of substitutions per amino acid position in the multiple-sequence alignment. Highlighted properties of each species for which one or more representative virion structures has been determined (MRV, BRV, ARV, and AqRVA and C) are shown at right. Suggested evolutionary events to explain differences in the highlighted features are labeled on appropriate branches. on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from on October 13, 201 http://jvi.asm.org/ Downloaded from FIG. 8. Phylogenetic tree and evolution schematic. A maximum-likelihood phylogenetic tree is shown for the core clamp proteins of the indicated ortho- and aquareovirus species or tentative species. Species names are abbreviated as follows: MRV, Mammalian orthoreovirus; BRV, Baboon orthoreovirus; NBV, Nelson Bay orthoreovirus; ARV, Avian orthoreovirus; AqRVA, Aquareovirus A; AqRVC, Aquareovirus C; and AqRVG, Aquareovirus G. For MRV, NBV, and ARV, three representative isolates were used in generating the tree. Tentative new species are represented by Broome virus (BroV) and psittacine reovirus (PsRV) isolates. Branch labels are conﬁdence scores (%). The tree is rooted on the aquareovirus clade. The scale bar indicates horizontal branch lengths in units of substitutions per amino acid position in the multiple-sequence alignment. Highlighted properties of each species for which one or more representative virion structures has been determined (MRV, BRV, ARV, and AqRVA and C) are shown at right. Suggested evolutionary events to explain differences in the highlighted features are labeled on appropriate branches. on October 13, 2019 by guest http://jvi.asm.org/ ded from turret features with BRV and aquareoviruses and to compare the strategies for adhesion to, and uptake by, cells of orthoreo- viruses lacking the ﬁber protein. formation of intersubunit disulﬁde bonds within this region (36). DISCUSSION Which proteins function in aquareovirus adhesion to cell surface re- ceptors remains unknown. However, it is intriguing that both aquareovirus and BRV virions display distinct differences in the arrangement or structural plasticity of their core turrets relative to the ﬁber-containing MRV and ARV virions. It will be interesting to determine whether BroV, another orthoreo- virus that lacks an outer ﬁber protein (50), shares similar core The absence of an adhesion ﬁber raises the question of which other BRV protein might take the place of the ﬁber in binding to cell surface receptors. One strong candidate is the outer clamp protein B, which is present in 600 copies per virion and the upper lobes of which are the most dominant feature on the BRV virion surface. Another strong candidate to participate in adhesion to cell surface receptors is the pen- tameric core turret protein A, which is present in 60 copies VIRION STRUCTURE OF BABOON REOVIRUS Sequence of reovirus haemagglutinin predicts a coiled-coil structure. Nature 315:421–423. 5. Breun, L. A., et al. 2001. Mammalian reovirus L2 gene and 2 core spike protein sequences and whole-genome comparisons of reoviruses type 1 Lang, type 2 Jones, and type 3 Dearing. Virology 287:333–348. 6. Chandran, K., D. L. Farsetta, and M. L. Nibert. 2002. Strategy for nonen- veloped virus entry: a hydrophobic conformer of the reovirus membrane penetration protein 1 mediates membrane disruption. J. Virol. 76:9920– 9933. 7. Chandran, K., et al. 1999. In vitro recoating of reovirus cores with baculo- virus-expressed outer-capsid proteins 1 and 3. J. Virol. 73:3941–3950. y p We have recently described the structure of a provisional member of family Totiviridae—penaeid shrimp infectious myo- necrosis virus (IMNV)—whose virions are adorned at the ico- sahedral 5f axes by surface ﬁbers, making it unlike other toti- viruses (48). Here, we reciprocally describe a member of family Reoviridae, subfamily Spinareovirinae, genus Orthoreovirus— BRV—whose virions lack surface ﬁbers, making it unlike most other orthoreoviruses. These and other examples suggest that gain or loss of adorning ﬁbers is one of several common themes in dsRNA virus evolution. The outer ﬁber protein VP4 of rotaviruses, as one other example from the family Reoviridae (subfamily Sedoreovirinae, genus Rotavirus), shows little if any sequence similarity to the outer ﬁber protein of orthoreo- viruses and instead anchors its base into the P2 channels of the rotavirus outer capsid (45, 55). This malleability seems likely to reﬂect the ease of accommodating a surface ﬁber that simply anchors atop an established capsid architecture. Perhaps the bigger obstacle is how to incorporate ﬁber-encoding sequences into the viral genome so as to ensure their proper expression. Interestingly, in fusogenic orthoreoviruses, the ﬁber protein is commonly encoded on the same bi- or tricistronic S genome segment as the FAST protein (17, 47). The evolution of these multicistronic segments was presumably facilitated by the fact that neither the FAST protein nor the ﬁber protein needs to be expressed at high levels (there are no more than 36 copies of the ﬁber protein per orthoreovirus virion), allowing expression of these proteins to be mediated by atypical and possibly inef- ﬁcient translation initiation mechanisms (42). VIRION STRUCTURE OF BABOON REOVIRUS Pressures driving the gain or loss of ﬁber proteins and FAST proteins, for that matter, are likely based in the ever-continuing battles between viruses and their hosts, relating to such elements as innate and adaptive immune responses, tropism, virulence, transmission, and host range. 8. Chappell, J. D., A. E. Prota, T. S. Dermody, and T. Stehle. 2002. Crystal structure of reovirus attachment protein 1 reveals evolutionary relationship to adenovirus ﬁber. EMBO J. 21:1–11. on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from on October 13, 201 http://jvi.asm.org/ Downloaded from 9. Cheng, L., et al. 2011. Atomic model of a cypovirus built from cryo-EM structure provides insight into the mechanism of mRNA capping. Proc. Natl. Acad. Sci. U. S. A. 108:1373–1378. 10. Cheng, L., et al. 2010. Backbone model of an aquareovirus virion by cryo- electron microscopy and bioinformatics. J. Mol. Biol. 397:852–863. on October 13, 2019 by guest http://jvi.asm.org/ aded from 11. Chua, K. B., et al. 2007. A previously unknown reovirus of bat origin is associated with an acute respiratory disease in humans. Proc. Natl. Acad. Sci. U. S. A. 104:11424–11429. 12. Dawe, S., J. Boutilier, and R. Duncan. 2002. Identiﬁcation and character- ization of a baboon reovirus-speciﬁc nonstructural protein encoded by the bicistronic S4 genome segment. Virology 304:44–52. 13. Dawe, S., and R. Duncan. 2002. The S4 genome segment of baboon reovirus is bicistronic and encodes a novel fusion-associated small transmembrane protein. J. Virol. 76:2131–2140. 14. Dryden, K. A., et al. 1998. Internal structures containing transcriptase- related proteins in top component particles of mammalian orthoreovirus. Virology 245:33–46. 15. Dryden, K. A., et al. 1993. Early steps in reovirus infection are associated with dramatic changes in supramolecular structure and protein conforma- tion: analysis of virions and subviral particles by cryoelectron microscopy and image reconstruction. J. Cell Biol. 122:1023–1041. 16. Duncan, R. 1999. Extensive sequence divergence and phylogenetic relation- ships between the fusogenic and nonfusogenic orthoreoviruses: a species proposal. Virology 260:316–328. 17. Duncan, R., J. Corcoran, J. Shou, and D. Stoltz. 2004. Reptilian reovirus: a new fusogenic orthoreovirus species. Virology 319:131–140. 18. Duncan, R., F. A. Murphy, and R. R. Mirkovic. 1995. Characterization of a novel syncytium-inducing baboon reovirus. Virology 212:752–756. 19. Fang, Q., et al. 2000. Sequence of genome segments 1, 2, and 3 of the grass carp reovirus (Genus Aquareovirus, family Reoviridae). Biochem. Biophys. Res. Commun. 274:762–766. ober 13, 2019 by guest 20. Furlong, D. B., M. L. VIRION STRUCTURE OF BABOON REOVIRUS Some more divergent members of the subfamily Spinareovirinae, cy- poviruses and oryzaviruses, whose core turret proteins wholly lack the ﬂap domains, also contain only 120 copies of the core clamp protein (9, 31). In combination, this absence of a cor- relation between the presence of ﬁbers and corresponding changes in the core turret with the number of copies of the core clamp protein implies that these structural features are not strictly linked in evolutionary terms. The signiﬁcance of the various differences for particle function or assembly, however, has yet to be deciphered. ( ) BRV virions are additionally seen to be like ARV and aquareovirus virions in lacking the hub-and-spoke complexes that reside within the P2 and P3 solvent channels of the MRV outer capsid. In MRV virions, these unique structures are formed by a 33-aa C-terminal extension of the outer shell protein (25, 57), which is missing from the homologous pro- teins of BRV, ARV, and aquareoviruses. The signiﬁcance of these P2/P3 structures in MRV is not fully known, but they appear to play a role in stabilizing the outer capsid, including The taxon-crossing differences in structural features and coding capacities among ortho- and aquareoviruses have led us to ponder the detailed steps in their evolution. A series of discrete transitions that could account for gain or loss of par- ticular features is diagrammed in Fig. 8. It is important to recognize, however, that because ortho- and aquareoviruses YAN ET AL. J. VIROL. 7494 have segmented genomes, evolution can proceed by both drift (random point substitutions) and shift (transfer of an entire genome segment), analogous to inﬂuenza virus antigenic vari- ation. Moreover, horizontal transfer of genetic elements be- tween different viruses may well have occurred during the complex evolutionary pathway of the nonenveloped viruses, as suggested by the previously described relationships between orthoreovirus and adenovirus ﬁber proteins (8, 20, 33). This diagram is therefore clearly speculative and intended solely to help frame certain questions or predictions that may be ad- dressed by future experiments. 2. Baker, T. S., N. H. Olson, and S. D. Fuller. 1999. Adding the third dimension to virus life cycles: three-dimensional reconstruction of icosahedral viruses from cryo-electron micrographs. Microbiol. Mol. Biol. Rev. 63:862–922. from cryo-electron micrographs. Microbiol. Mol. Biol. Rev. 6 3. Barton, E. S., et al. 2001. Junction adhesion molecule is a receptor for reovirus. Cell 104:441–451. 4. Bassel-Duby, R., et al. 1985. VIRION STRUCTURE OF BABOON REOVIRUS Nibert, and B. N. Fields. 1988. Sigma 1 protein of mammalian reoviruses extends from the surfaces of viral particles. J. Virol. 62:246–256. 21. Gard, G., and R. W. Compans. 1970. Structure and cytopathic effects of Nelson Bay virus. J. Virol. 6:100–106. 22. Havelka, W. A., R. Henderson, and D. Oesterhelt. 1995. Three-dimensional structure of halorhodopsin at 7 Å resolution. J. Mol. Biol. 247:726–738. 23. Kim, J., Y. Tao, K. M. Reinisch, S. C. Harrison, and M. L. Nibert. 2004. Orthoreovirus and Aquareovirus core proteins: conserved enzymatic sur- faces, but not protein-protein interfaces. Virus Res. 101:15–28. 24. Leland, M. M., G. B. Hubbard, H. T. Sentmore III, K. F. Soike, and J. K. Hilliard. 2000. Outbreak of Orthoreovirus-induced meningoencephalomyeli- tis in baboons. Comp. Med. 50:199–205. ACKNOWLEDGMENTS 25. Liemann, S., K. Chandran, T. S. Baker, M. L. Nibert, and S. C. Harrison. 2002. Structure of the reovirus membrane-penetration protein, 1, in a complex with is protector protein, 3. Cell 108:283–295. We thank Xing Zhang for assistance with MRV maps, Alex Rusnak for exploratory TEM work on BRV, Kelly Dryden for help with the Zeiss SCAI microdensitometer at The Scripps Research Institute, and Kevin Chiang and Debbie Kim for assistance with scanning images. 26. Luongo, C. L., C. M. Contreras, D. L. Farsetta, and M. L. Nibert. 1998. Binding site for S-adenosyl-L-methionine in a central region of mammalian reovirus 2 protein. Evidence for activities in mRNA cap methylation. J. Biol. Chem. 273:23773–23780. g g g This work was supported in part by NIH grant R01 AI46440 (to M.L.N.), Canadian Institutes of Health Research grants 13723 and 100584 and Natural Sciences and Engineering (NSERC) grant 0183745 (to R.D.), NIH fellowship F32 AI078624 (to K.N.P.), NIH grants R37 GM-033050 and 1S10 RR-020016, and support from Uni- versity of California—San Diego and the Agouron Foundation to establish and equip cryo-TEM facilities at USCD (to T.S.B.). 27. Luongo, C. L., et al. 1997. Localization of a C-terminal region of 2 protein in reovirus cores. J. Virol. 71:8035–8040. 28. Luongo, C. L., K. M. Reinisch, S. C. Harrison, and M. L. Nibert. 2000. Identiﬁcation of the guanylyltransferase region and active site in reovirus mRNA capping protein 2. J. Biol. Chem. 275:2804–2810. pp g p 29. Maginnis, M. S., et al. 2006. Beta1 integrin mediates internalization of mammalian reovirus. J. Virol. 80:2760–2770. REFERENCES 30. Mindell, J. A., and N. Grigorieff. 2003. Accurate determination of local defocus and specimen tilt in electron microscopy. J. Struct. Biol. 142:334– 347. 1. Attoui, H., et al. 2002. Common evolutionary origin of aquareoviruses and orthoreoviruses revealed by genome characterization of golden shiner reo- virus, grass carp reovirus, striped bass reovirus and golden ide reovirus (genus Aquareovirus, family Reoviridae). J. Gen. Virol. 83:1941–1951. 1. Attoui, H., et al. 2002. Common evolutionary origin of aquareoviruses and orthoreoviruses revealed by genome characterization of golden shiner reo- virus, grass carp reovirus, striped bass reovirus and golden ide reovirus (genus Aquareovirus, family Reoviridae). J. Gen. Virol. 83:1941–1951. 31. Miyazaki, N., et al. 2008. Structural evolution of reoviridae revealed by oryzavirus in acquiring the second capsid shell. J. Virol. 82:11344–11353. VOL. 85, 2011 7495 VIRION STRUCTURE OF BABOON REOVIRUS 32. Nason, E. L., S. K. Samal, and B. V. V. Prasad. 2000. Trypsin-induced structural transformation in aquareovirus. J. Virol. 74:6546–6555. Harrison. 2011. Atomic model of an infectious rotavirus particle. EMBO J. 30:408–416. Harrison. 2011. Atomic model of an infectious rotavirus particle. EMBO J. 30:408–416. 46. Shmulevitz, M., and R. Duncan. 2000. A new class of fusion-associated small transmembrane (FAST) proteins encoded by the non-enveloped fusogenic reoviruses. EMBO J. 19:902–912. 33. Nibert, M. L., T. S. Dermody, and B. N. Fields. 1990. Structure of the reovirus cell-attachment protein: a model for the domain organization of 1. J. Virol. 64:2976–2989. 47. Shmulevitz, M., et al. 2002. Sequential partially overlapping gene arrange- ment in the tricistronic S1 genome segments of avian reovirus and Nelson Bay reovirus: implications for translation initiation. J. Virol. 76:609–618. 34. Nibert, M. L., A. L. Odegard, M. A. Agosto, K. Chandran, and L. A. Schiff. 2005. Putative autocleavage of reovirus 1 protein in concert with outer- capsid disassembly and activation for membrane permeabilization. J. Mol. Biol. 345:461–474. 48. Tang, J., et al. 2008. Infectious myonecrosis virus has a totivirus-like, 120- subunit capsid, but with ﬁber complexes at the ﬁvefold axes. Proc. Natl. Acad. Sci. U. S. A. 105:17526–17531. 35. Nibert, M. L., L. A. Schiff, and B. N. Fields. 1991. Mammalian reoviruses contain a myristoylated structural protein. J. Virol. 65:1960–1967. 49. Tang, J., et al. 2010. Backbone trace of partitivirus capsid protein from electron cryomicroscopy and homology modeling. Biophys. J. 99:685–694. 36. Odegard, A. L., K. Chandran, S. Liemann, S. C. Harrison, and M. L. Nibert. 2003. REFERENCES Disulﬁde bonding among 1 trimers in mammalian reovirus outer capsid: a late and reversible step in virion morphogenesis. J. Virol. 77:5389– 5400. 50. Thalmann, C. M., et al. 2010. Broome virus, a new fusogenic Orthoreovirus species isolated from an Australian fruit bat. Virology 402:26–40. 51. van Heel, M., and M. Schatz. 2005. Fourier shell correlation threshold criteria. J. Struct. Biol. 151:250–262. 37. Olland, A. M., J. Jane´-Valbuena, L. A. Schiff, M. L. Nibert, and S. C. Harrison. 2001. Structure of the reovirus outer capsid and dsRNA-binding protein 3 at 1.8 Å resolution. EMBO J. 20:979–989. 52. Victoria, J. G., A. Kapoor, K. Dupuis, D. P. Schnurr, and E. L. Delwart. 2008. Rapid identiﬁcation of known and new RNA viruses from animal tissues. PLoS Pathog. 4:e1000163. on October 13, 2019 by guest http://jvi.asm.org/ Downloaded from on October 13, 201 http://jvi.asm.org/ Downloaded from 38. Palacios, G., et al. 2010. Heart and skeletal muscle inﬂammation of farmed salmon is associated with infection with a novel reovirus. PLoS One 5:e11487. g 53. Yan, X., K. A. Dryden, J. Tang, and T. S. Baker. 2007. Ab initio random model method facilitates 3D reconstruction of icosahedral particles. J. Struct. Biol. 157:211–225. 39. Paul, R. W., A. H. Choi, and P. W. K. Lee. 1989. The -anomeric form of sialic acid is the minimal receptor determinant recognized by reovirus. Vi- rology 172:382–385. 54. Yan, X., R. S. Sinkovits, and T. S. Baker. 2007. AUTO3DEM—-an auto- mated and high throughput program for image reconstruction of icosahedral particles. J. Struct. Biol. 157:73–82. on October 13, 2019 by guest http://jvi.asm.org/ ded from 40. Penn, O., E. Privman, G. Landan, D. Graur, and T. Pupko. 2010. An alignment conﬁdence score capturing robustness to guide-tree uncertainty. Mol. Biol. Evol. 27:1759–1767. 55. Yeager, M., K. A. Dryden, N. H. Olson, H. B. Greenberg, and T. S. Baker. 1990. Three-dimensional structure of rhesus rotavirus by cryoelectron mi- croscopy and image reconstruction. J. Cell Biol. 110:2133–2144. 41. Qiu, T., and C. L. Luongo. 2003. Identiﬁcation of two histidines necessary for reovirus mRNA guanylyltransferase activity. Virology 316:313–324. 56. Zhang, L., et al. 2009. Requirements for the formation of membrane pores by the reovirus myristoylated 1N peptide. J. Virol. 83:7004–7014. 5 2005 f i id i 1 l d 42. Racine, T., and R. Duncan. 2010. Facilitated leaky scanning and atypical ribosome shunting direct downstream translation initiation on the tricis- tronic S1 mRNA of avian reovirus. REFERENCES Nucleic Acids Res. 38:7260–7272. by the reovirus myristoylated 1N peptide. J. Virol. 83:7004–7014 57. Zhang, X., et al. 2005. Features of reovirus outer capsid protein 1 revealed by electron cryomicroscopy and image reconstruction of the virion at 7.0 Å resolution. Structure 13:1545–1557. 43. Racine, T., et al. 2009. Aquareovirus effects syncytiogenesis by using a novel member of the FAST protein family translated from a noncanonical trans- lation start site. J. Virol. 83:5951–5955. 58. Zhang, X., et al. 2005. Structure of avian orthoreovirus virion by electron cryomicroscopy and image reconstruction. Virology 343:25–35. 44. Reinisch, K. M., M. L. Nibert, and S. C. Harrison. 2000. Structure of the reovirus core at 3.6 Å resolution. Nature 404:960–967. y py g gy 59. Zhang, X., S. B. Walker, P. R. Chipman, M. L. Nibert, and T. S. Baker. 2003. Reovirus polymerase 3 localized by cryo-electron microscopy of virions at a resolution of 7.6 Å. Nat. Struct. Biol. 10:1011–1018. 45. Settembre, E. C., J. Z. Chen, P. R. Dormitzer, N. Grigorieff, and S. C.
https://openalex.org/W2885139158	https://www.epj-conferences.org/10.1051/epjconf/201818202117/pdf	English	null	Cosmic matter in the laboratory - the Compressed Baryonic Matter experiment at FAIR	EPJ web of conferences	2,018	cc-by	5,170	https://doi.org/10.1051/epjconf/201818202117 https://doi.org/10.1051/epjconf/201818202117 EPJ Web of Conferences 182, 02117 (2018) ICNFP 2017 EPJ Web of Conferences will be set by the p © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). ae-mail: p.senger@gsi.de Cosmic matter in the laboratory - the Compressed Baryonic Mat- ter experiment at FAIR Peter Senger1,a 1GSI Helmholtzzentrum für Schwerionenforschung GmbH, Planckstr. 1, 64291 Darmstadt, Germany Peter Senger1,a 1GSI Helmholtzzentrum für Schwerionenforschung GmbH, Planckstr. 1, 64291 Darmstadt, Germany 1GSI Helmholtzzentrum für Schwerionenforschung GmbH, Planckstr. 1, 64291 Darmstadt, Germany Abstract. The Compressed Baryonic Matter (CBM) experiment will be one of the major scientiﬁc pillars of the future Facility for Antiproton and Ion Research (FAIR) in Darm- stadt. The goal of the CBM research program is to explore the QCD phase diagram in the region of high baryon densities using high-energy nucleus-nucleus collisions. This in- cludes the study of the equation-of-state of nuclear matter at neutron star core densities, and the search for the deconﬁnement and chiral phase transitions. The CBM detector is designed to measure rare diagnostic probes such as hadrons including multi-strange (anti-) hyperons, lepton pairs, and charmed particles with unprecedented precision and statistics. Most of these particles will be studied for the ﬁrst time in the FAIR energy range. In order to achieve the required precision, the measurements will be performed at very high reaction rates of 1 to 10 MHz. This requires very fast and radiation-hard detectors, a novel data read-out and analysis concept based on free streaming front-end electronics, and a high-performance computing cluster for online event selection. The physics program and the status of the proposed CBM experiment will be discussed. 1 Introduction The experimental and theoretical investigation of the properties of elementary matter under extreme conditions is a major topic of international fundamental research. Experiments with high-energy heavy-ion collisions are performed worldwide at the large accelerator facilities, in order to create and to study extremely hot and dense matter in the laboratory. In heavy-ion collisions at very high beam energies, as provided by the Relativistic Heavy Ion Collider at BNL in USA or by the Large Hadron Collider at CERN in Switzerland, elementary matter at extremely high temperatures is created, more than hundred-thousand times hotter than the core of our sun. Under such conditions, a plasma is created which consists of quarks and gluons. This plasma is a mixture of elementary matter and antimatter, similar to the primordial soup in early universe about a microsecond after the big bang. In heavy-ion collisions lower bombarding energies, as available at the SPS at CERN and at the future Facility for Antiproton and Ion Research (FAIR) at Darmstadt, nuclear matter is strongly compressed. The matter density in the collision zone of two heavy nuclei exceeds more than ﬁve times the density of an atomic nucleus. In nature, such densities are expected to exist in the core of a neutron star. The common goal of these experiments is to explore the various forms and phases of strongly- interacting matter. Similar to water, which exhibits diﬀerent states separated by phase transitions, https://doi.org/10.1051/epjconf/201818202117 EPJ Web of Conferences 182, 02117 (2018) ICNFP 2017 it is expected that the degrees of freedom of nuclear matter change with temperature and density, thereby creating novel forms of elementary matter. The fundamental theory of strong interaction, Quantum Chromo Dynamics (QCD), predicts a smooth crossover transition from hadronic matter to the quark-gluon plasma at a pseudo-critical temperature of about 160 MeV for vanishing net- baryon density [1–3]. However, QCD still fails to make predictions about the properties of matter at high net-baryon densities. Our conception of compressed nuclear matter is based on eﬀective- model calculations which predict structures in the QCD phase diagram at high net-baryon densities, like a critical endpoint followed by a ﬁrst-order phase transition, or new forms of exotic matter [4–6]. Figure 1 illustrates the conjectured phases of nuclear matter together with the locations of cosmic matter in a diagram of temperature versus net-baryon density. 1 Introduction it is expected that the degrees of freedom of nuclear matter change with temperature and density, thereby creating novel forms of elementary matter. The fundamental theory of strong interaction, Quantum Chromo Dynamics (QCD), predicts a smooth crossover transition from hadronic matter to the quark-gluon plasma at a pseudo-critical temperature of about 160 MeV for vanishing net- baryon density [1–3]. However, QCD still fails to make predictions about the properties of matter at high net-baryon densities. Our conception of compressed nuclear matter is based on eﬀective- model calculations which predict structures in the QCD phase diagram at high net-baryon densities, like a critical endpoint followed by a ﬁrst-order phase transition, or new forms of exotic matter [4–6]. Figure 1 illustrates the conjectured phases of nuclear matter together with the locations of cosmic matter in a diagram of temperature versus net-baryon density. Figure 1. Sketch of the phase diagram of strongly interacting matter with conjectured structures and locations of cosmic matter. In the early universe, the temperature was very high, but the net-baryon density was zero. In contrast, the net-baryon density in neutron star cores is very high, about 5 times the density of atomic nuclei, but the temperature is very low. The matter in neutron star mergers is also highly compressed but much hotter than in neutron stars, similar to the matter created in a heavy-ion collision at FAIR energies. Figure 1. Sketch of the phase diagram of strongly interacting matter with conjectured structures and locations of cosmic matter. In the early universe, the temperature was very high, but the net-baryon density was zero. In contrast, the net-baryon density in neutron star cores is very high, about 5 times the density of atomic nuclei, but the temperature is very low. The matter in neutron star mergers is also highly compressed but much hotter than in neutron stars, similar to the matter created in a heavy-ion collision at FAIR energies. Experimentally, only little is known about the QCD phase diagram. The comparison of the mea- sured yields of particles and antiparticles to the results of thermal model calculations provides a so called freeze-out temperature as function of baryon chemical potential [7, 8]. This temperature is measured at a late stage of the collision, when the produced particles cease to interact, and the density has dropped well below saturation density. 1 Introduction However, for LHC and top RHIC energies, the freeze-out temperature coincides with the pseudo-critical temperature predicted by QCD, indicating that freeze- out and hadronization might happen simultaneously. The mission of heavy-ion experiments at lower beam energies includes the search for the landmarks in the QCD phase diagram, such as the critical point and a ﬁrst order phase transition. Another fundamental ingredient for our understanding of nu- 2 EPJ Web of Conferences 182, 02117 (2018) ICNFP 2017 https://doi.org/10.1051/epjconf/201818202117 clear matter is its equation-of-state, which describes the matter properties as function of pressure and density, and, determines the evolution of supernova explosions, the size and radius of neutron stars, and the dynamics of neutron star mergers. clear matter is its equation-of-state, which describes the matter properties as function of pressure and density, and, determines the evolution of supernova explosions, the size and radius of neutron stars, and the dynamics of neutron star mergers. Moreover, heavy-ion collision experiments oﬀer the unique opportunity to experimentally address an-other fundamental question: why is the mass of protons and neutrons about 50 times larger than the sum of the masses of their building blocks, the up and down quarks? The mass of the nucleons, and, hence, the mass of the visible universe, is not explained by the Higgs mechanism. According to our present knowledge, the kinetic energy of quarks and gluons essentially contributes to the nucleon mass. If nucleons dissolve in hot and /or dense matter, they should not only lose their identity, but also their mass. The discovery of such a Chiral phase transition would be a breakthrough in our understanding of the origin of the mass of the visible universe. 2 The CBM Physics Program at SIS100 FAIR will provide heavy-ion beam energies from 2A - 11A GeV for Au or U nuclei, up to 14A GeV for N=Z nuclei, and up to 29 GeV for protons with the SIS100 synchrotron. Already in central Au+Au collisions at 5 A GeV the nuclear ﬁreball will be compressed - according to transport model and hydro calculations [9] - to more than 6 times saturation density ρ0, and at 10 A GeV even a density above 8 ρ0 is reached as illustrated in ﬁgure 2. At such densities, the nucleons will start to melt and to dissolve into their constituents. The calculations predict that the dense ﬁreball spends a relatively long time within the phase coexistence region or even beyond. Figure 2. Evolution of the central net baryon density ρ(t) as function of elapsed time as calculated by diﬀerent transport models and by a 3-ﬂuid hydrodynamics code for central Au+Au collision at 5 A GeV (left panel) and at 10 A GeV (right panel)[9]. Figure 2. Evolution of the central net baryon density ρ(t) as function of elapsed time as calculated by diﬀerent transport models and by a 3-ﬂuid hydrodynamics code for central Au+Au collision at 5 A GeV (left panel) and at 10 A GeV (right panel)[9]. Further indication, that a phase transition might occur at densities reached at SIS100 beam ener- gies, comes from a non-local 3-ﬂavor Nambu Jona-Lasinio model calculation of a neutron star, which predicts the development of a mixed phase of hadrons and quarks above densities of about 5 ρ0, and the transition to pure quark matter above 8 ρ0 [10]. The results of this calculation, which is able to reproduce a 2 solar mass neutron star, is depicted in ﬁgure 3. 3 https://doi.org/10.1051/epjconf/201818202117 EPJ Web of Conferences 182, 02117 (2018) EPJ Web of Conferences 182, 02117 (2018) ICNFP 2017 Figure 3. Particle population in a neutron star calculated with a Nambu Jona-Lasinio (n3NJL) model with repulsive vector interactions. The model is able to describe neutron stars with 2 solar masses and radii between 12 and 13 km [10]. Figure 3. Particle population in a neutron star calculated with a Nambu Jona-Lasinio (n3NJL) model with repulsive vector interactions. The model is able to describe neutron stars with 2 solar masses and radii between 12 and 13 km [10]. 2 The CBM Physics Program at SIS100 In conclusion, the beam energies available at SIS100 appear to be especially well suited for gen- erating signals of the phase transition, and, therefore, oﬀer the opportunity to address fundamental scientiﬁc questions as mentioned above. The focus of the CBM experiment at FAIR is to study mes- sengers from the dense ﬁreball such as multiple strange hyperons, lepton pairs, and hadrons containing charm quarks in order to ﬁnd answers to the questions raised above. A survey of the theoretical con- cepts and the experimental programs devoted to the exploration of the QCD phase diagram with focus on high baryon densities is given in the CBM Physics Book [11]. The CBM research program at SIS100 includes the physics cases and observables as discussed in the following. 2.1 The equation of state of nuclear matter at high baryon densities According to transport models, multi-strange (anti-)hyperons are produced in sequential collisions involving kaons and Lambda hyperons, and, therefore, are sensitive to the density in the ﬁreball. This sensitivity is expected to increase towards lower beam energies close to or even below the pro- duction threshold. The excitation functions of multi-strange hyperons (Ξ−(dss) and Ω−(sss)) and anti-hyperons (Ξ+( ¯d¯s¯s) and Ω+ (¯s¯s¯s)) in A+A collisions with diﬀerent A values at SIS100 beam en- ergies are very promising observables which will shed light on the matter equation of state at neutron star core densities. Another promising observable is the collective ﬂow of identiﬁed particles which is driven by the pressure gradient inside the ﬁreball, and hence, is expected to be sensitive to the com- pressibility of nuclear matter. Up to now, only the proton ﬂow excitation function has been measured at the AGS [12], and was used to extract the compressibility of nuclear matter [13]. 4 4 4 https://doi.org/10.1051/epjconf/201818202117 EPJ Web of Conferences 182, 02117 (2018) ICNFP 2017 2.3 In-medium modiﬁcations of hadrons, onset of chiral symmetry restoration Lepton pairs will be measured over a wide range of invariant masses, starting from low-mass vector mesons including their Dalitz decays up to charmonium. The precise measurement of lepton pairs at low invariant masses will allow to analyze modiﬁcations of vector meson properties in dense baryonic matter. This observable is expected to be sensitive to chiral symmetry restoration [20]. The thermal radiation at intermediate invariant dilepton masses includes a broadened in-medium ρ meson, radia- tion from the QGP, and dileptons from multi-pion annihilation. The latter contribution reﬂects ρ −a1 chiral mixing, and, therefore, provides a direct link to chiral symmetry restoration. 2.2 Phase Transitions, phase coexistence, critical point The experimental observation, that in ultra-relativistic heavy-ion collisions multi-strange hyperons including Ω−and Ω+ are in chemical equilibrium like all the other produced particles, was taken as strong indication that the system had undergone a transition from a partonic phase to the hadronic ﬁnal state, with the equilibration being driven by multi-body collisions in the high particle density regime near the phase boundary [14]. Agreement of the hyperon yield with thermal model calculations was found also at 40A GeV in Pb+Pb collisions at the SPS [15]. In Ar + KCl collisions at an energy of 1.76A GeV, however, the measured yield of Ξ−hyperons exceeds the thermal model prediction by about a factor of 20, indicating that Ξ−hyperons are far oﬀchemical equilibrium [16]. High precision measurements of excitation functions of multi-strange hyperons in A+A collision at SIS100 energies will allow to study the degree of equilibration of the ﬁreball, and, hence, open the possibility to ﬁnd a signal for the onset of deconﬁnement in QCD matter at high net-baryon densities. 2 The slope of the dilepton invariant mass spectrum between 1 and about 2.5 GeV/c2 reﬂects the av- erage temperature of the ﬁreball [17]. The precise measurement of the spectral slope as a function of beam energy opens the unique possibility to measure the caloric curve, which would be the ﬁrst direct experimental signature for phase coexistence in high-density nuclear matter. This measurement would also provide indications for the onset of deconﬁnement and the location of the critical endpoint. Another direct experimental proof for a ﬁrst order phase transition would be the discovery of phase coexistence by observing an enhanced production of composite particles or multi-particle correlations caused by the spinodal ampliﬁcation of density ﬂuctuations [18]. In classical binary liquids, density ﬂuctuations grow when approaching the critical point, leading to critical opalescence. A similar phenomenon is predicted in heavy-ion collisions, namely event-by- event ﬂuctuations of conserved quantities such as baryon number, strangeness and electrical charge, which can be related to thermodynamical susceptibilities. Lattice QCD calculations suggest, that higher moments of these distributions are sensitive to vicinity of the critical point. Recent mea- surements by the STAR collaboration provide evidence for an increase of higher-order ﬂuctuations towards lower beam energies [19]. Corresponding data in the FAIR energy range are required in order to search for a peak of these ﬂuctuations, as expected for a critical behavior. 2.5 Hypernuclei and strange objects Thermal models predict a maximum in the production probability of light hypernuclei in heavy-ion collisions at SIS100 beam energies [22, 23]. The CBM experiment is well suited to measure hydro- gen and helium hypernuclei, including light double-lambda hypernuclei. The discovery of (double-) lambda hypernuclei and the determination of their lifetimes will provide valuable information on the hyperon-nucleon and hyperon-hyperon interactions, which are essential ingredients for the un- derstanding of the nuclear matter equation-of-state at high densities, and, hence, of the structure of neuron stars [24]. 2.4 Charm With CBM at SIS100, charm production will be studied for the ﬁrst time at beam energies close to production threshold. At these energies, the formation time of charmonium is small compared to the lifetime of the reaction system. CBM is thus uniquely suited to study the interactions between fully formed J/ψ and the dense medium with appropriate counting statistics and systematics. Systematic measurements of charmonium in p+A collisions with varying target mass number A at proton energies up to 29 GeV will shed light on the charmonium interaction with cold nuclear matter, and constitute 5 5 5 EPJ Web of Conferences 182, 02117 (2018) ICNFP 2017 https://doi.org/10.1051/epjconf/201818202117 an important baseline for measurements in heavy-ion collisions. Moreover, the simultaneous mea- surement of open charm will give access to the basically unknown charm production cross section at or near the kinematic threshold. According to a recent UrQMD calculation the subthreshold charm production in central Au+Au collision increased dramatically when considering secondary processes like N∗→Λc + D and N∗→N + J/ψ [21]. 3 Experimental challenge Figure 4. Interaction rates achieved by existing and planned heavy-ion experiments as function of beam energy (taken from [30]). Figure 4. Interaction rates achieved by existing and planned heavy-ion experiments as function of beam energy (taken from [30]). 3 Experimental challenge Most of the diagnostic probes discussed above are rarely produced at beam energies where the high- est baryon densities will be created. Therefore, the key feature of existing and future experiments is rate capability in order to measure these observables with unprecedented precision. The exper- imental challenge is to combine a large-acceptance and fast detector setup with a high-speed data read-out system and a high luminosity beam. Several experiments concentrate on the exploration of the QCD phase diagram at in the region of high baryon densities. Explorative heavy-ion experiments at AGS in Brookhaven [12] and CERN-SPS [25] mainly measured abundantly produced hadrons. The NA61/SHINE experiment at CERN-SPS continues to scan the QCD phase diagram light and medium size beams [26]. The STAR collaboration at RHIC plans for a second beam energy scan to improve the statistical signiﬁcance of the data taken in the ﬁrst series of measurements [27]. In order to go down in collision energy and up in baryon-chemical potential, the STAR detector can be operated in a ﬁxed target mode. The HADES collaboration investigates the production of dileptons and (strange) hadrons in collisions of beams of pions, protons, and nuclei with nuclear targets at GSI SIS18 [28]. At the Joint Institute for Nuclear Research (JINR) in Dubna, the collider facility NICA together with a multi-purpose detector is planned at JINR [29]. As an intermediate step, a ﬁxed target experiment called "Baryonic Matter at Nuclotron" (BM@N) is being installed at the Nuclotron to study heavy-ion collisions at gold-beam energies up to about 4.5A GeV. The luminosity limitations of these existing and future facilities constrain the research programs to the investigation of bulk observables, and pre- vent high precision measurements of rare diagnostic probes. In contrast, the Compressed Baryonic Matter (CBM) experiment at the Facility for Antiproton and Ion Research (FAIR) is designed to run at extremely high interaction rates (up to 10 MHz). This feature is the key requirement for the mea- surement of multi-diﬀerential observables including rare probes with unprecedented precision. The rate capabilities of existing and planned heavy-ion experiments are plotted in ﬁg. 4 as function of beam energy. The comparison clearly demonstrates the uniqueness of the CBM setup in terms of rate capability. 6 6 6 https://doi.org/10.1051/epjconf/201818202117 EPJ Web of Conferences 182, 02117 (2018) ICNFP 2017 Figure 4. Interaction rates achieved by existing and planned heavy-ion experiments as function of beam energy (taken from [30]). 4 The Compressed Baryonic Matter (CBM) Experiment The research program on dense QCD matter at FAIR will be performed by the experiments CBM and HADES which are shown in ﬁgure 5. The two setups will be operated alternatively. The HADES de- tector, with its large polar angle acceptance ranging from 18 to 85 degrees, is well suited for reference measurements with proton beams and heavy-ion collision systems with moderate particle multiplici- ties, such as Ni+Ni or Ag+Ag at low SIS100 energies. With the HADES detector, electron-positron pairs and hadrons including multi-strange hyperons can be reconstructed. The CBM setup is a ﬁxed target experiment covering polar emission angles between 2.5 and 25 degrees. The detector system has been designed as a multipurpose device which will be capable to measure hadrons, electrons and muons in elementary nucleon and heavy-ion collisions over the full FAIR beam energy range [11]. For particle identiﬁcation several detectors are needed. The masses of hadrons which traverse the full setup (protons, pions, kaons) are determined by the measurement of their momentum and velocity. The particle momentum is derived from the curved trajectories in a magnetic ﬁeld. The particle trajectories are measured by the Silicon Tracking System (STS) based on double-sided silicon micro-strip sensors arranged in eight stations inside a dipole magnet. In order to reduce multiple scattering, the sensors are read out via low-mass micro cables by free-streaming ASICs which are mounted on frontend boards outside the active detector area. The superconducting dipole magnet has a gap between the pole shoes of 1.44 m height and a ﬁeld integral of 1 Tm. The particle time-of-ﬂight is measured by a wall of Multi-Gap Resistive Plate Chambers (MRPC) with low resistivity glass electrodes providing a time resolution of 50 ps at rates up to 25 kHz/cm2. The MRPC wall has an active area of 100 m2 and is located about 7 m downstream the target. Both centrality and reaction plane angle of the collision will be measured with the Project Spectator Detector, which is a segmented hadronic calorimeter located about 10 me downstream the target. 7 https://doi.org/10.1051/epjconf/201818202117 EPJ Web of Conferences 182, 02117 (2018) ICNFP 2017 Short-lived hadrons (e.g. the hyperons Λ, Ξ, Ω) and hypernuclei decaying inside the magnetic ﬁeld are identiﬁed by their invariant mass which is reconstructed via the measured trajectories and the time-of-ﬂight of their decay products. In this case, the decay vertex has to be determined accurately by the STS. 4 The Compressed Baryonic Matter (CBM) Experiment Even hyperons with one neutral daughter (e.g. Σ+ →p + π0 ) can be identiﬁed via the missing mass method. In order to identify open charm hadrons (Λc, D mesons) with decay lengths between cτ = 60 µm and 312 µm, the Micro-Vertex-Detector (MVD) is required, which consists of four layers of silicon monolithic active pixel sensors, located between 5 cm and 20 cm downstream the target in the vacuum of the beam pipe. The identiﬁcation of lepton pairs requires additional detector systems. Electrons and positrons up to momenta of 6 GeV/c can be separated from pions and heavier hadrons by a Ring Imaging Cherenkov (RICH) detector. The RICH consists of a large volume CO2 radiator, where the high energetic electrons travel faster than light and produce Cherenkov photons, which are measured by a UV photon detector based on multi-anode photomultipliers. Additional pion suppression is achieved - in particular for electrons with momenta above 2 GeV/c - by a four layer Transition Radiation Detector which can be operated at particle densities up to 100 kHz/cm2. Alternative to electrons, also muons will measured by a Muon Chamber (MuCh) system which consists of Gas Electron Multiplier (GEM) detector triplets sandwiched between hadron absorber plates made of graphite and iron. For muon measurements, the RICH detector will be replaced by the MuCh, and the TRD serves as tracking detector after the last hadron absorber. Figure 5. The HADES detector (left) and the CBM experimental setup (right) with Ring Imaging Cherenkov detector in measuring position, and the muon detection system in parking position. Figure 5. The HADES detector (left) and the CBM experimental setup (right) with Ring Imaging Cherenkov detector in measuring position, and the muon detection system in parking position. In order to perform high-precision multi-diﬀerential measurements of rare probes the experiment should run at event rates of up to 10 MHz for several months per year. To ﬁlter out weakly decaying particles like multi-strange hyperons, hypernuclei or D mesons, no simple trigger signal can be gen- erated. Instead, the full events have to be reconstructed, and the decay topology has to be identiﬁed 8 8 EPJ Web of Conferences 182, 02117 (2018) EPJ Web of Conferences 182, 02117 (2018) ICNFP 2017 https://doi.org/10.1051/epjconf/201818202117 online by fast algorithms running on a high-performance computing farm hosted by the GSI GreenIT cube. 4 The Compressed Baryonic Matter (CBM) Experiment To utilize the maximum rates, the data acquisition is based on self-triggered front-end electron- ics, which provides a time stamp for each signal. The online track reconstruction algorithms combine detector hits to tracks based on the space and time information of the hits ("4-D" reconstruction). Figure 6 depicts the tracks of charged particles produced in a simulated central collision of two gold nuclei at a beam energy of 10A GeV as reconstructed by the MVD and the STS. Figure 6. Simulation of a central collision of two gold nuclei at a beam energy of 10A GeV. The trajectories of about 700 charged particles including the decay products of short-lived particles will be measured with the CBM silicon tracking detectors MVD and STS. Figure 6. Simulation of a central collision of two gold nuclei at a beam energy of 10A GeV. The trajectories of about 700 charged particles including the decay products of short-lived particles will be measured with the CBM silicon tracking detectors MVD and STS. The development of the experimental components is well in progress. The Technical Design Reports on the Superconducting Dipole Magnet, on the Silicon Tracking System, on the Ring Imaging Cherenkov Detectors, on the Projectile Spectator Detector, on the Time-of-Flight detector, and on the Muon Chamber system have been approved by FAIR. The TDRs on Data Acquisition and First Level Event Selection, on the Micro-Vertex-Detector, and on the Transition Radiation Detector will be submitted in 2017. Until the start of FAIR, the nuclear matter research program at GSI will be pursued with the HADES experiment at SIS18. At the SIS100 accelerator, the HADES detector can be used to perform di-electron and hadron reference measurements in collision systems with moderate particle multiplic- ities, such as proton-proton, proton-nucleus and nucleus-nucleus collisions with light nuclei. 5 Summary The Compressed Baryonic Matter (CBM) experiment will be one of the major scientiﬁc pillars of the future Facility for Antiproton and Ion Research (FAIR) in Darmstadt. The goal of the CBM research program is to explore the QCD phase diagram in the region of high baryon densities using high-energy nucleus-nucleus collisions. This includes the study of the equation-of-state of nuclear matter at neutron star core densities, and the search for the deconﬁnement and chiral phase transitions. The CBM detector is designed to measure rare diagnostic probes such as multi-strange hyperons, charmed particles and vector mesons decaying into lepton pairs with unprecedented precision and 9 9 9 EPJ Web of Conferences 182, 02117 (2018) ICNFP 2017 https://doi.org/10.1051/epjconf/201818202117 statistics. Most of these particles will be studied for the ﬁrst time in the FAIR energy range. In order to achieve the required precision, the measurements will be performed at reaction rates between 100 kHz and 10 MHz. This requires very fast and radiation hard detectors, and a novel data read-out and analysis concept based on free streaming front-end electronics and a high-performance computing cluster for online event selection. The use of the most modern detector and computer technology is the prerequisite for a substantial discovery potential of heavy-ion collision experiments at FAIR energies. statistics. Most of these particles will be studied for the ﬁrst time in the FAIR energy range. In order to achieve the required precision, the measurements will be performed at reaction rates between 100 kHz and 10 MHz. This requires very fast and radiation hard detectors, and a novel data read-out and analysis concept based on free streaming front-end electronics and a high-performance computing cluster for online event selection. The use of the most modern detector and computer technology is the prerequisite for a substantial discovery potential of heavy-ion collision experiments at FAIR energies. Acknowledgments The design and development of the CBM experiment is performed by the CBM Collaboration which actually consists of more than 470 persons from 56 institutions and 11 countries. The CBM project is supported by the German Ministry of Education and Research, the Helmholtz Association, and national funds of the CBM member institutions. References [7] F. Becattini et al., Phys. Rev. Lett. 111 (2013) 082302 [8] A. Andronic et al., arXiv:1611.01347 [nucl-th] [8] A. Andronic et al., arXiv:1611.01347 [nucl-th] [9] I.C. Arsene et al., Phys. Rev. C75 (2007) 24902 [9] I.C. Arsene et al., Phys. Rev. C75 (2007) 24902 [10] M. Orsaria et al., Phys. Rev. C89 (2014)015806 [10] M. Orsaria et al., Phys. Rev. C89 (2014)015806 [10] M. Orsaria et al., Phys. Rev. C89 (2014)015806 [11] The CBM Physics Book, B. Friman et al.(eds): Series: Lecture Notes in Physics, Vol. 814 (2011) 11] The CBM Physics Book, B. Friman et al.(eds): Series: Lecture Notes in Physics, Vol. 8 [11] The CBM Physics Book, B. Friman et al.(eds): Series: Lectur [11] The CBM Physics Book, B. Friman et al.(eds): Series: Lecture Notes in Physics, Vol. 814 (2011) [12] C. Pinkenburg et al., Phys. Rev. Lett. 83 (1999) 1295 [12] C. Pinkenburg et al., Phys. Rev. Lett. 83 (1999) 1295 [13] P. Danielewicz et al., Science 298 (2002) 1592 [13] P. Danielewicz et al., Science 298 (2002) 1592 [14] P. Braun-Munzinger, J. Stachel and C. Wetterich, Phys.Lett. B 596 (2004) 61 14] P. Braun-Munzinger, J. Stachel and C. Wetterich, Phys.Lett. B 596 (2004) 61 [15] A. Andronic, P. Braun-Munzinger and J. Stachel, Acta Phys. Polon. B 40 (20 15] A. Andronic, P. Braun-Munzinger and J. Stachel, Acta Phys. Polon. B 40 (2009) 1005 [16] G. Agakishiev et al. (HADES collaboration),arXiv:1512.07070v1 16] G. Agakishiev et al. (HADES collaboration),arXiv:1512.07070v1 [17] R. Rapp and H. van Hees, Phys. Lett. B 753 (2016) 586 [17] R. Rapp and H. van Hees, Phys. Lett. B 753 (2016) 586 [18] J. Steinheimer and J. Randrup, Phys. Rev. Lett. 109 (2012) 212301 [18] J. Steinheimer and J. Randrup, Phys. Rev. Lett. 109 (2012) 212301 [19] L. Adamczyk et al., Phys. Rev. Lett. 113 (2014) 092301 [19] L. Adamczyk et al., Phys. Rev. Lett. 113 (2014) 092301 [20] P. Hohler and R. Rapp, Phys. Lett. B731 (2014) 103-109 [20] P. Hohler and R. Rapp, Phys. Lett. B731 (2014) 103-109 21] J. Steinheimer, A. Botvina, M. Bleicher, arXiv:1605.03439v1 [21] J. Steinheimer, A. Botvina, M. Bleicher, arXiv:1605.03439v1 [22] A. Andronic et al., Phys. Lett. B697 (2011) 203 [23] H. Stöcker et al., Nucl. Phys. A 827 (2009) 624c [24] A. S. Botvina et al., Phys. Lett. B 742 (2014) 7 [25] C. Alt et al., Phys. Rev. C77 (2008)024903 [26] A. Laszlo, Pos(CPOD07)054 (2007) [28] G. Agakishiev et al., (HADES Collaboration), Eur. Phys. J. 41 (2009) 243 [30] T. Ablyazimov et al., [CBM Collaboration], Eur. Phys. J. A 53 (2017) 60 [29] D. Blaschke et al. (eds), Eur. Phys. J. A, 52 8 (2016) 267 [26] A. Laszlo, Pos(CPOD07)054 (2007) [27] https://drupal.star.bnl.gov/STAR/starnotes/public/sn0598 10 EPJ Web of Conferences 182, 02117 (2018) ICNFP 2017 https://doi.org/10.1051/epjconf/201818202117 11
https://openalex.org/W3020926124	https://www.duo.uio.no/bitstream/10852/72613/7/euclid.ba.1545296448.pdf	English	null	Rejoinder for the discussion of the paper "A novel algorithmic approach to Bayesian Logic Regression"	arXiv (Cornell University)	2,020	cc-by	32,951	∗Department of mathematics, University of Oslo, aliaksah@math.uio.no †Department of mathematics, University of Oslo, geirs@math.uio.no ‡Department of Medical Statistics (CEMSIIS), Medical University of Vienna, ﬂorian.frommlet@meduniwien.ac.at §Norwegian Computing Center, aliaksandr.hubin@nr.no c⃝2020 International Society for Bayesian Analysis https://doi.org/10.1214/18-BA1141 A Novel Algorithmic Approach to Bayesian Logic Regression (with Discussion) Aliaksandr Hubin∗,§, Geir Storvik†, and Florian Frommlet‡ Abstract. Logic regression was developed more than a decade ago as a tool to construct predictors from Boolean combinations of binary covariates. It has been mainly used to model epistatic eﬀects in genetic association studies, which is very appealing due to the intuitive interpretation of logic expressions to describe the in- teraction between genetic variations. Nevertheless logic regression has (partly due to computational challenges) remained less well known than other approaches to epistatic association mapping. Here we will adapt an advanced evolutionary al- gorithm called GMJMCMC (Genetically modiﬁed Mode Jumping Markov Chain Monte Carlo) to perform Bayesian model selection in the space of logic regres- sion models. After describing the algorithmic details of GMJMCMC we perform a comprehensive simulation study that illustrates its performance given logic re- gression terms of various complexity. Speciﬁcally GMJMCMC is shown to be able to identify three-way and even four-way interactions with relatively large power, a level of complexity which has not been achieved by previous implementations of logic regression. We apply GMJMCMC to reanalyze QTL (quantitative trait locus) mapping data for Recombinant Inbred Lines in Arabidopsis thaliana and from a backcross population in Drosophila where we identify several interesting epistatic eﬀects. The method is implemented in an R package which is available on github. Keywords: logic regression, Bayesian model averaging, mode jumping Monte Carlo Markov Chain, genetic algorithm, QTL mapping. 15, Number 1, pp. 263–333 15, Number 1, pp. 263–333 Bayesian Analysis (2020) A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression based on Cockerham’s coding to detect interactions illustrated the advantages of logic regression to detect epistasic eﬀects in QTL mapping (Malina et al., 2014). Given the potential of logic regression to detect interpretable interaction eﬀects in a regression setting it is rather surprising that it has not yet become wider addressed in applica- tions. based on Cockerham’s coding to detect interactions illustrated the advantages of logic regression to detect epistasic eﬀects in QTL mapping (Malina et al., 2014). Given the potential of logic regression to detect interpretable interaction eﬀects in a regression setting it is rather surprising that it has not yet become wider addressed in applica- tions. Originally logic regression was introduced together with likelihood based model se- lection, where simulated annealing served as a strategy to obtain one “best” model (see Ruczinski et al., 2003, for details). However, assuming that there is one “best” model disregards the problem of model uncertainty. Whilst this approach works well in simulation studies, it seems to be quite an unrealistic assumption in real world appli- cations, where there often is no “true” model. Hence Bayesian model averaging, which implicitly takes into account model uncertainty, becomes important. Bayesian versions of logic regression combined with model exploration include Monte Carlo logic regres- sion (MCLR) (Kooperberg and Ruczinski, 2005) and the full Bayesian version of logic regression (FBLR) by Fritsch (2006). Both MCLR and FBLR use Markov Chain Monte Carlo (MCMC) algorithms for searching through the space of models and parameters. Inference is then based on a large number of models instead of just one model as in the original version of logic regression. MCLR utilizes a geometric prior on the size of the model (deﬁned through the number of logic terms and their complexity). All mod- els of the same size get the same prior probability while larger models implicitly are penalized. Regression parameters are marginalized out, signiﬁcantly simplifying com- putational complexity. In contrast FBLR is performed on a joint space of parameters and models. FBLR uses multivariate normal priors for regression parameters, while model size is furnished with a slightly diﬀerent prior serving similar purposes as the MCLR prior. In case of a large number of binary covariates these MCMC based meth- ods might require extremely long Markov chains to guarantee convergence which can make them infeasible in practice. 1 Introduction Logic regression (not to be confused with logistic regression) was developed as a general tool to obtain predictive models based on Boolean combinations of binary covariates (Ruczinski et al., 2003). Its primary application area is epistatic association mapping as pioneered by Ruczinski et al. (2004) and Kooperberg and Ruczinski (2005) although already early on the method was also used in other areas (Keles et al., 2004; Janes et al., 2005). Important contributions to the development of logic regression were later made by the group of Katja Ickstadt (Fritsch, 2006; Schwender and Ickstadt, 2008), which also provided a comparison of diﬀerent implementations of logic regression (Fritsch and Ickstadt, 2007). Schwender and Ruczinski (2010) gave a brief introduction with various applications and potential extensions of logic regression. Recently a systematic com- parison of the performance of logic regression and a more classical regression approach 264 A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression Additionally both of them utilize simple Metropolis- Hastings settings which, together with the fact that the search space is often multimodal, increases the probability that they are stuck in local extrema for a signiﬁcant amount of time. In this paper we propose a new approach for Bayesian logic regression including model uncertainty. We introduce a novel prior for the topology of logic regression models which is slightly simpler to compute than the one used by MCLR and which still shows excellent properties in terms of controlling false discoveries. We consider two diﬀerent priors for regression coeﬃcients: Jeﬀreys’ prior and the robust g-priors as a state of the art choice for priors of regression coeﬃcients in variable selection problems. For Jeﬀreys’ prior computing the marginal likelihoods can be performed with the Laplace approximation as in BIC (Bayesian information criterion) and similar model selection criteria. For the robust g-prior the marginal likelihood is eﬃciently computed using the integrated Laplace approximation (Li and Clyde, 2018). The main contribution of this paper is the proposed search algorithm, named GMJMCMC, which provides a better search strategy for exploring the model space than previous approaches. GMJMCMC combines genetic algorithm ideas with the mode jumping Markov Chain Monte Carlo (MJMCMC) algorithm (Tjelmeland and Hegstad, 2001; Hubin and Storvik, 2018) in order to be able to jump between local modes in the 265 A. Hubin, G. Storvik, and F. Frommlet model space. After formally introducing logic regression and describing the GMJMCMC algorithm in detail we will present results from a comprehensive simulation study. The performance of GMJMCMC is compared with MCLR and FBLR in case of logistic models (binary responses) and additionally analyzed for linear models (quantitative responses). Models of diﬀerent complexities are studied which allows us to illustrate the potential of GMJMCMC to detect higher order interactions. Finally we apply our logic regression approach to perform QTL mapping using two publicly available data sets. The ﬁrst study is concerned with the hypocotyledonous stem length in Arabidopsis thaliana using Recombinant Inbred Line (RIL) data (Balasubramanian et al., 2009), the second one considering various traits from backcross data of Drosophila Simulans and Drosophila Mauritana is presented in the web supplement (Hubin et al., 2018b). The method is implemented as an R package which is freely available on GitHub at http:// aliaksah.github.io/EMJMCMC2016/, where one can also ﬁnd examples of further logic regression applications. We will study logic regression in the context of the generalized linear model (GLM, see McCullagh and Nelder (1989)) of the form Y ∼ f (y \| μ(X); φ) , (1) Y ∼ f (y \| μ(X); φ) , 2.1 Logic regression The method of logic regression (Ruczinski et al., 2003) was speciﬁcally designed for the situation where covariates are binary and predictors are deﬁned as logic expressions operating on these binary variables. Logic regression can be applied in the context of the generalized linear model (GLM) as demonstrated in Malina et al. (2014). It can also be easily expanded to the domain of generalized linear mixed models (GLMM), but to keep our presentation as simple as possible we will focus here on generalized linear regression models. Consider a response variable Y ∈R, together with m binary covariates X1, X2, . . . , Xm. Our primary example will be genetic association studies where, depending on the context, each binary covariate, Xj, j ∈{1, 2, . . . , m}, can have a diﬀerent interpretation. In QTL mapping with backcross design or recombinant inbred lines Xj simply codes the two possible genetic variants. In case of intercross design or in outbred populations diﬀerent Xj will be used to code dominant and recessive eﬀects (see for example Malina et al., 2014). We will adapt the usual convention that a value 1 corresponds to logical TRUE and a value 0 to logical FALSE where the immediate interpretation in our ex- amples is that a speciﬁc marker is associated with a trait or not. Each combination of the binary variables Xj with the logical operators ∧(AND), ∨(OR) and Xc (NOT X), is called a logic expression (for example L = (X1 ∧X2) ∨Xc 3). Following the nomen- clature of Kooperberg and Ruczinski (2005) we will refer to logic expressions as trees, whereas the primary variables contained in each tree are called leaves. The set of leaves of a tree L will be denoted by v(L), that is for the speciﬁed example above we have v(L) = {X1, X2, X3}. We will study logic regression in the context of the generalized linear model (GLM, see McCullagh and Nelder (1989)) of the form Y ∼ f (y \| μ(X); φ) , Y ∼ f (y \| μ(X); φ) , (1) 266 A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression h (μ(X)) = α + q j=1 γjβjLj, (2) (2) where f denotes the parametric distribution of Y belonging to the exponential family with mean μ(X) and dispersion parameter φ. 2.1 Logic regression The function h is an appropriate link function, α and βj, j ∈{1, . . . , q} are unknown regression parameters, and γj is the indicator variable which speciﬁes whether the tree Lj is included in the model. For the sake of simplicity we abbreviate by μ(X) the complex dependence of the mean μ on X via the logic expressions Lj according to (2). Our primary examples are linear regression for quantitative responses and logistic regression for dichotomous responses but the implementation of our approach works for any generalized linear model. We will restrict ourselves to trees with no more than Cmax leaves. Consequently the total number of trees q will be ﬁnite. The considered models are restricted to include no more than kmax trees. The vector of binary random variables M = (γ1, . . . , γq) fully characterizes a model in terms of which logical expressions are included. Here we go along with the usual convention in the context of variable selection that ‘model’ refers to the set of regressors and does not take into account the speciﬁc values of the non-zero regression coeﬃcients. Bayesian model speciﬁcation For a fully Bayesian approach one needs prior speciﬁcations for the model topology characterized by the index vector M as well as for the coeﬃcients α and βj belonging to a speciﬁc model M. This is a common approach in Bayesian model selection, used for example in Clyde et al. (2011) or Hubin and Storvik (2018). We start with deﬁning the prior for M by p(M) ∝I (\|M\| ≤kmax) q j=1 ρ(γj). (3) (3) Here \|M\| = q j=1 γj is the number of logical trees included in the model and kmax is the maximum number of trees allowed per model. The factors ρ(γj) are introduced to give smaller prior probabilities to more complex trees. Speciﬁcally we consider ρ(γj) = aγjc(Lj) (4) (4) with 0 < a < 1 and c(Lj) ≥0 being a non-decreasing measure for the complexity of the corresponding logical trees. In case of γj = 0 it holds that ρ(γj) = 1 and thus the prior probability for model M only consists of the product of ρ(γj) for all trees included in the model. It follows that if M and M ′ are two vectors only diﬀering in one component, say γ′ j = 1 and γj = 0, then p(M ′) p(M) = ac(Lj) < 1 showing that larger models are penalized more. This result easily generalizes to the comparison of more diﬀerent models and provides the basic intuition behind the chosen prior. showing that larger models are penalized more. This result easily generalizes to the comparison of more diﬀerent models and provides the basic intuition behind the chosen prior. 267 A. Hubin, G. Storvik, and F. Frommlet The prior choice implies a distribution for the model size \|M\| which can be inter- preted as a multiple-testing penalty (Scott and Berger, 2008). For kmax = q and a constant complexity value on all trees, \|M\| follows a binomial distribution. With vary- ing complexity measures, \|M\| follows the Poisson binomial distribution (Wang, 1993) which is a unimodal distribution with E[\|M\|] = q j=1 pj and Var[\|M\|] = q j=1 pj(1−pj) where pj = ac(Lj)/(1 + ac(Lj)). A truncated version of this distribution is obtained for kmax < q. The choices of a and the complexity measure c(Lj) are crucial for the quality of the model prior. Let N(s) be the total number of trees having s leaves. Bayesian model speciﬁcation Choosing a = e−1 and c(Lj) = log N(sj) as long as the number of leaves is not larger than Cmax results for γj = 1 in ac(Lj) = 1 N(sj) , sj ≤Cmax . Therefore the multiplicative contribution of a speciﬁc tree of size s to the model prior will be indirectly proportional to the total number of trees N(s) having s leaves as long as s ≤Cmax. Given that N(s) is rapidly growing with the tree size s this choice gives smaller prior probabilities for larger trees. The resulting penalty closely resembles the Bonferroni correction in multiple testing as discussed for example by Bogdan et al. (2008) in the context of modiﬁcations of the BIC. The number N(s) will in practice be diﬃcult to compute. To compute a rough approximation of N(s) we ignore logic expressions including the same variable multiple times. Then there are m s possibilities to select variables. Each variable can undergo logic negation giving s binary choices and furthermore there are s−1 logic symbols (∨, ∧) to be chosen resulting in 22s−1 diﬀerent expressions. However, due to De Morgan’s law half of the expressions provide identical logic regression models. This gives N(s) ≈ m s 22s−2. (5) (5) Using this approximation, for a model of size k = \|M\| the full model prior is of the form P(M) ∝I (k ≤kmax) k r=1 I (sjr ≤Cmax) m sjr 22sjr −2 , (6) (6) where Jn(α, β) is the observed information. To obtain model posterior probabilities one needs to evaluate the marginal likelihood of the model P(Y \| M) by integrating over all parameters of the model which is often a fairly diﬃcult task. The greatest advantage of Jeﬀreys’ prior is that this integral can be approximated simple and accurate through the Laplace approximation. In case of the Gaussian model choosing Jeﬀreys’ prior (8) for the coeﬃcients and the simple prior (7) for the variance term yields that the Laplace approximation becomes exact (Raftery et al., 1997) and gives a marginal likelihood of the simple form P(Y \| M) ∝P(Y \| M, ˆθ) n \|M\| 2 , (9) (9) where ˆθ refers to the maximum likelihood estimates of all parameters involved. On the log scale this exactly corresponds to the BIC model selection criterion (Schwarz, 1978) when using a uniform model prior. In case of logistic regression the marginal likelihood under Jeﬀreys’ prior becomes approximately (9) with an error of order O(n−1) (Tierney and Kadane, 1986; Claeskens and Hjort, 2008). Barber et al. (2016) also describe that Laplace approximations of the marginal likelihood yield very accurate results and can be trusted in Bayesian model selection problems. Although there are many situations in which selection based on BIC like criteria works well, within the Bayesian literature using Jeﬀreys’ prior for model selection has been widely criticized for not being consistent once the true model coincides with the null model (all γj = 0, Bayarri et al., 2012). A large number of alternative priors have been studied, see for example Li and Clyde (2018) who give a comprehensive review on the state of the art of g-priors. In a recent paper Bayarri et al. (2012) gave theoretical arguments in case of the linear model recommending the robust g-prior, which is consistent in all situations and yields errors diminishing signiﬁcantly faster than other prior choices. Thus we will introduce the robust g-prior as an alternative to Jeﬀreys’ prior. Our description of robust g-priors follows Li and Clyde (2018) who consider an improper constant prior for the intercept, P(α) ∝1, and a mixture g-prior for the regression coeﬃcients βj, j ∈{j1, . . . , j\|M\|} of the form P(β \| g) ∼N\|M\| 0, g · φJn(β)−1 . where j1, . . . , jk refer to the k trees of model M. where j1, . . . , jk refer to the k trees of model M. We will next discuss priors for the parameters given a speciﬁc model M. The GLM formulation (1) includes a dispersion parameter φ, which for example in case of the linear model is connected with the variance term σ2 for the underlying normal distribution. If a GLM has a dispersion parameter then for the sake of simplicity we will adapt the commonly used improper prior (Li and Clyde, 2018; Bayarri et al., 2012) π(φ) =φ−1. (7) (7) If a GLM does not include a dispersion parameter (like logistic regression) then one simply sets φ = 1. 268 A Novel Algorithmic Approach to Bayesian Logic Regression Concerning the intercept α and the regression coeﬃcients βj, where j ∈{j1, . . . , j\|M\|} correspond to the non-zero coeﬃcients of model M, we will consider two diﬀerent types of priors, simple Jeﬀreys’ priors and robust g-priors. Jeﬀreys’ prior (Jeﬀreys, 1946, 1961; Gelman et al., 2013) assumes for the parameters of the model an improper prior distri- bution of the form πα(α)πβ(β) =\|Jn(α, β)\| 1 2 , (8) (8) here Jn(α, β) is the observed information. A. Hubin, G. Storvik, and F. Frommlet 269 This family of mixtures of g-priors includes a large number of priors discussed in the literature, see Li and Clyde (2018) for more details. The recommended robust g-prior is a particular case with the following choice of parameters: a = 1, b = 2, r = 1.5, s = 0, v = n + 1 \|M\| + 1, κ = 1. Under this prior speciﬁcation precise integrated Laplace approximations of the marginal likelihood for GLM are given by Li and Clyde (2018), whilst exact values are available for Gaussian models (Li and Clyde, 2018; Bayarri et al., 2012). where Jn(α, β) is the observed information. (10) (10) Here Jn(β) is the subblock of the full observed information matrix Jn(α, β) related to β and g itself is assumed to be distributed according to the so called truncated Compound Conﬂuence Hypergeometric (tCCH) prior Here Jn(β) is the subblock of the full observed information matrix Jn(α, β) related to β and g itself is assumed to be distributed according to the so called truncated Compound Conﬂuence Hypergeometric (tCCH) prior P 1 1 + g ∼tCCH a 2, b 2, r, s 2, v, κ . (11) (11) A. Hubin, G. Storvik, and F. Frommlet 2.2 Computing posterior probabilities Based on model posterior probabilities one can easily obtain an estimate of the posterior probability for a logic expression Lj to be included in a model (also referred 270 A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression to as the marginal inclusion probability) by to as the marginal inclusion probability) by ˜P(Lj \| Y ) = M∈Ω∗:γj=1 ˜P(M \| Y ).1 (14) (14) Inference on trees can then be performed by means of selecting those trees with a poste- rior probability being larger than some threshold probability πC. In case of exploratory studies where the main aim is to discover many potentially interesting features to be explored in further studies it can be reasonable to use low threshold values on ˜P(Lj \| Y ). High threshold values can be used if false discoveries need to be avoided. In general the threshold can be speciﬁed through a decision theoretic framework, including the aim of controlling false discovery rates, see (Wakeﬁeld, 2007). Inference on trees can then be performed by means of selecting those trees with a poste- rior probability being larger than some threshold probability πC. In case of exploratory studies where the main aim is to discover many potentially interesting features to be explored in further studies it can be reasonable to use low threshold values on ˜P(Lj \| Y ). High threshold values can be used if false discoveries need to be avoided. In general the threshold can be speciﬁed through a decision theoretic framework, including the aim of controlling false discovery rates, see (Wakeﬁeld, 2007). A threshold of 0.5 corresponds to the median probability model of Barbieri et al. (2004) which under certain circumstances has greater predictive power than the most probable model. However, one of the criteria for the median probability model to be optimal in the linear Gaussian case, the graphical model structure criterion, will not always be valid in cases where one makes restrictions on the number of trees that can be included. The graphical model structure criterion requires that the median probability model results in a legal model. 2.2 Computing posterior probabilities Given prior probabilities for any logic regression model M the model posterior proba- bility can be computed according to Bayes formula as P(M \| Y ) = P(Y \| M)P(M) M ′∈Ω P(Y \| M ′)P(M ′) , (12) (12) where P(Y \| M) denotes the integrated (or marginal) likelihood for model M and Ω is the set of all models in the model space. The sum in the denominator involves a huge number of terms and it is impossible to compute all of them. Classical MCMC based approaches (like MCLR and FBLR) overcome this problem by estimating model poste- riors with the relative frequency with which a speciﬁc model M occurs in the Markov chain. In case of an ultrahigh-dimensional model space (like in case of logic regression) this is computationally extremely challenging and might require chain lengths which are prohibitive for practical applications. An alternative approach makes use of the fact that most of the summands in the denominator of (12) will be so small that they can be neglected. Considering a subset Ω∗⊆Ω containing the most important models we can therefore approximate (12) by P(M \| Y ) ≈˜P(M \| Y ) = P(Y \| M)P(M) M ′∈Ω∗P(Y \| M ′)P(M ′) . (13) (13) To obtain good estimates we have to search in the model space for those models that contribute signiﬁcantly to the sum in the denominator, that is for those models with large posterior probabilities or equivalently with large values of P(Y \| M)P(M). In Frommlet et al. (2012) speciﬁc memetic algorithms were developed to perform the model search for linear regression. Here we will rely upon the GMJMCMC algorithm, which is described in the next section. For now we assume that some method for computing the marginal likelihood P(Y \| M) is available. The details of such computation depend on the prior speciﬁcations of the parameters of a particular model and are given for the examples in the experimental sections. 2.2 Computing posterior probabilities Consider the case with three covariates x1, x2, x3 but with kmax = 2 and the posterior probabilities for models γ = (1, 1, 0), γ = (1, 0, 1) and γ = (0, 1, 1) each equal to 1/3. Then all marginal inclusion probabilities are 2/3 and the median probability model includes all variables which then has a model size larger than kmax. The median probability model can however still be a useful model to consider even in cases where the optimality results do not apply. 1Here by P(Lj \| Y ) we mean P(γj = 1 \| Y ). 2.3 The GMJMCMC algorithm To ﬁx ideas consider ﬁrst a variable selection problem with q potential covariates to enter a model. Recall that γj needs to be 1 if the j-th variable is to be included into the model and 0 otherwise. A model M is thus speciﬁed by the vector γ = (γ1, . . . , γq) and the general model space Ω is of size 2q. If this discrete model space is multimodal in terms of model posterior probabilities then simple MCMC algorithms typically run into problems by staying for too long in the vicinity of local maxima. Recently, the mode jumping MCMC procedure (MJMCMC) was proposed by Hubin and Storvik (2018) to overcome this issue in a model selection setting. MJMCMC is a proper MCMC algorithm equipped with the possibility to jump between diﬀerent modes within the discrete model space. The key to the success of MJMCMC is the generation of good proposals of models which are not too close to the current state. This is achieved by ﬁrst making a large jump (changing many model components) and then performing local optimization within the discrete model space to obtain a proposal model. Within a Metropolis-Hastings setting a valid acceptance probability is then constructed using symmetric backward kernels, which guarantees that the resulting Markov chain is ergodic and has the desired limiting distribution (Tjelmeland and Hegstad, 2001; Hubin and Storvik, 2018). 271 A. Hubin, G. Storvik, and F. Frommlet A. Hubin, G. Storvik, and F. Frommlet The MJMCMC algorithm requires that all of the covariates deﬁning the model space are known in advance and are all considered at each iteration of the algorithm. In case of logic regression the covariates are trees and a major problem in this setting is that it is quite diﬃcult to fully specify the space Ω. In fact it is even diﬃcult to specify q, the total number of feasible trees. To solve this problem we present an adaptive algorithm called Genetically Modiﬁed MJMCMC (GMJMCMC), where MJMCMC is embedded in the iterative setting of a genetic algorithm. In each iteration only a given set S of trees (of ﬁxed size d) is considered. Each S then induces a separate search space for MJMCMC. In the language of genetic algorithms S is the population, which dynamically evolves to allow MJMCMC exploring diﬀerent reasonable parts of the unfeasibly large total search space. 2.3 The GMJMCMC algorithm To be more speciﬁc, we consider diﬀerent populations S1, S2, . . . where each St is a set of d trees. For each given population a ﬁxed number of MJMCMC steps is per- formed. Since the MJMCMC algorithm is speciﬁed in full detail in Hubin and Storvik (2018), we will concentrate here on describing the evolutionary dynamics yielding sub- sequent populations St. Utilization of the approximation (13) in combination with exact or approximated marginal likelihoods allows us to compute posterior probabilities for all models in Ω∗which have been visited at least once by the algorithm. Consequently we do not need a proper MCMC (an algorithm with convergence towards the target distribution) which is needed if model posterior probabilities are estimated by the rel- ative frequency of how often a model has been visited. In principle it is possible to construct a proper MCMC algorithm which aims at simulating from extended models of the form P(M, S \| Y ) having P(M \| Y ) as a stationary distribution. This version of the algorithm is considered in (Hubin et al., 2018a) where the main idea is to perform both forward and backward swaps between populations in order to obtain a reversible Markov chain. The algorithm is initialized by ﬁrst running MJMCMC for a given number of iter- ations Ninit on the set of all binary covariates X1, . . . , Xm as potential regressors, but not including any interactions. The ﬁrst d1 < d members of population S1 are then deﬁned to be the d1 covariates with largest marginal inclusion probability. In our cur- rent implementation we select the d1 leaves which have marginal posterior probabilities (estimated from the ﬁrst Ninit iterations) larger than ρmin, thus d1 is not pre-speciﬁed but is obtained in a data driven way. For later reference we denote this set of d1 leaves by S0. The remaining d −d1 members of S1 are obtained by forming logic expressions from the leaves of S0 where trees are generated randomly by means of the crossover operation described below. In practice one ﬁrst has to choose some kmax which will depend on the expected number of trees to enter the model in the problem one studies. The choice of d can then be guided by the results of Theorem 1 given below. A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression Let Dt be the set of trees to be deleted from St. Then \|Dt\| replacement trees must be generated instead. Each replacement tree is generated randomly by a crossover op- erator with probability Pc and by a mutation operator with probability Pm = 1 −Pc. A reduction operator is applied if mutation or crossover gives a tree larger than the maximal tree size Cmax. Crossover: Two parent trees are selected from St with probabilities proportional to the approximated marginal inclusion probabilities of trees in St. Then each one of the parents is inverted with probability Pnot by the logical not c operator, before they are combined with a ∧operator with probability Pand and with a ∨operator otherwise. Hence the crossover operator gives trees of the form Lj1 ∧Lj2 or Lj1 ∨Lj2 where either Lji or Lc ji is in St for i = 1, 2. Mutation: One parent tree is selected from St with probability proportional to the approximated marginal inclusion probabilities of trees in St, whilst the other parent tree is selected uniformly from the set of m −d1 leaves which did not make it into the initial population S0. Then just like for the crossover operator each of the parents is inverted with probability Pnot by the logical not c operator, before they are combined with a ∧operator with probability Pand and with a ∨operator otherwise. The mutation operator gives trees of the form Lj1 ∧X or Lj1 ∨X where either Lj1 or Lc j1 is in St and X or Xc is in D0. Reduction: A new tree is generated from a tree by deleting a subset of leaves, where each leave has a probability of ρdel to be deleted. The pruning of the tree is performed in a natural way meaning that the ‘closest’ logical operators of the deleted leaves are also deleted. If the deleted leave is not on the boundaries of the original tree the oper- ation is resulting in obtaining two separated subtrees. The resulting subtrees are then combined in a tree with a ∧operator with probability Pand or with a ∨operator oth- erwise. For all three operators it holds that if the newly generated tree is already present in St then it is not considered for St+1 but rather a new replacement tree is proposed instead. 2.3 The GMJMCMC algorithm After S1 has been initialized MJMCMC is performed for a ﬁxed number of iterations Nexpl before the next population S2 is generated. This process is iterated for Tmax populations St, t ∈{1, . . . , Tmax}. The d1 input trees from the initialization procedure remain in all populations St throughout our search. Other trees from the population St with low marginal inclusion probabilities (below a threshold ρmin) will be substituted by trees which are generated by crossover, mutation and reduction operators to be described in more detail below. 272 Algorithm 1 GMJMCMC. 1: Run the MJMCMC algorithm for Ninit iterations on X1, . . . , Xm and deﬁne S0 as the set of d1 variables among them with the largest estimated marginal inclusion probabilities. 1: Run the MJMCMC algorithm for Ninit iterations on X1, . . . , Xm and deﬁne S0 as the set of d1 variables among them with the largest estimated marginal inclusion probabilities. p 2: Generate d −d1 trees by randomly selecting crossover operations of elements from S0 and add those trees to the set S0 to obtain S1. 3: Run the MJMCMC algorithm within search space S1. f d 3: Run the MJMCMC algorithm within search space S1. 4: for t = 2, . . . , Tmax do 5: Delete trees within St−1\S0 which have estimated inclusion probabilities less than ρmin. ρ 6: Add new trees which are generated by crossover, mutation or reduction operato until the having again a set of size d, which becomes St. 6: Add new trees which are generated by crossover, mutation or reduction operators until the having again a set of size d, which becomes St. g g 7: Run the MJMCMC algorithm within search space St. 8: end for models. Since S0 is generated in the ﬁrst step and never changed, we will consider it to be ﬁxed. Deﬁne MSt to be the last model visited by the MJMCMC algorithm on search space St. Then the construction of St+1 only depends on (St, MSt, X) while MSt+1 only depends on St+1. Therefore {(St, MSt, X)} is a Markov chain. Assume now S and S′ are two populations diﬀering in one component with L ∈S, L′ ∈S′, L ̸= L′. Deﬁne Lsub to be any tree that is a subtree of both L and L′ (where a subtree is deﬁned as a tree which can be obtained by reduction) and Ssub to be the search space where L is substituted with Lsub in S. Then it is possible to move from S to Ssub in l steps using ﬁrst mutations and crossovers to grow a tree L∗of size larger than Cmax, which can undergo reduction (note that although only trees that have low enough estimated marginal inclusion probabilities can be deleted, there will always be a positive probability that marginal inclusion probabilities are estimated to be smaller than the threshold ρmin) to get to Lsub. A Novel Algorithmic Approach to Bayesian Logic Regression The pseudo-code Algorithm 1 describes the full GMJMCMC algorithm. For each iteration t the initial model for the next MJMCMC run is constructed by randomly selecting trees from St with probability Pinit. For the ﬁnal population STmax, MJMCMC is run until Mfin unique models are visited (within STmax). Mfin should be suﬃciently large to obtain good MJMCMC based approximations of the posterior parameters of interest based on the ﬁnal search space STmax. The following result is concerned with consistency of probability estimates of GMJMCMC when the number of iterations increases. Theorem 1. Assume Ω∗is the set of models visited through the GMJMCMC algorithm where d−d1 ≥kmax. Assume further the marginal likelihoods are calculated without er- rors. Then the model estimates based on (13) will converge to the true model probabilities as the number of iterations Tmax goes to ∞. Proof. Note that the approximation (13) will provide the exact answer if Ω∗= Ω. It is therefore enough to show that the algorithm in the limit will have visited all possible A. Hubin, G. Storvik, and F. Frommlet 273 Algorithm 1 GMJMCMC. Further, assuming the diﬀerence in size between Lsub and L′ is r, a move from Ssub to S′ can be performed by r steps of mutations or crossovers. Two search spaces which diﬀer in s trees can be reached by s combinations of the moves described above. Since also any model within a search space can be visited, the Markov chain {(St, MSt, X)} is irreducible. Since the state space for this Markov chain is ﬁnite, it is also recurrent, and there exists a stationary distribution with positive probabilities on every model. Thereby, all states, including all possible models of maximum size d, will eventually be visited. When d1 > 0, some restrictions on the possible search spaces are introduced. How- ever, when d −d1 ≥kmax, any model of maximum size kmax will eventually be vis- ited. Remark 1. If d −d1 < kmax, then every model of size up to d −d1 plus some of the larger models will eventually be visited, although the model space will get some additional constraints. In practice it is more important that d −d1 ≥k∗, where k∗ is the size of the true model. Unfortunately neither k∗nor d1 are known in advance, 274 A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression and one has to make reasonable choices of kmax and d depending on the problem one analyses. and one has to make reasonable choices of kmax and d depending on the problem one analyses. and one has to make reasonable choices of kmax and d depending on the problem one analyses. and one has to make reasonable choices of kmax and d depending on the problem one analyses. Remark 2. The result of Theorem 1 relies on exact calculation of the marginal likeli- hood P(Y \| M). Apart from the linear model, the calculation of P(Y \| M) is typically based on an approximation, giving similar approximations to the model probabilities. How precise these approximations are will depend on the type of method used. The current implementation includes Laplace approximations, integrated Laplace approx- imations, and integrated nested Laplace approximations. In principle other methods based on MCMC outputs (Chib, 1995; Chib and Jeliazkov, 2001) could be incorporated relatively easily resulting however in longer runtimes. Parallelization Due to our interest in exploring as many unique high quality models as possible and doing it as fast as possible, running multiple parallel chains is likely to be computation- ally beneﬁcial compared to running one long chain. The process can be embarrassingly parallelized into B chains using several CPUs (Central processing units), GPUs (graph- ics processing units) or clusters. If one is mainly interested in model probabilities, then Equation (13) can be directly applied with Ω∗now being the set of unique models visited within all runs. However, we suggest a more memory eﬃcient approach. If some statistic Δ is of interest, one can utilize the following posterior estimates based on weighted sums over individual runs: ˜P(Δ \| Y ) = B b=1 wb ˜Pb(Δ \| Y ). (15) (15) Here wb is a set of weights which will be speciﬁed below and ˜Pb(Δ \| Y ) are the posteriors obtained with formula (4) from run b of GMJMCMC. Due to the irreducibility of the GMJMCMC procedure it holds that for b wb = 1 we obtain limTmax→∞˜P(Δ \| Y ) = P(Δ \| Y ) where Tmax is the number of iterations within each run. Thus for any set of normalized weights the approximation ˜P(Δ \| Y ) converges to the true posterior probability P(Δ \| Y ). Therefore in principle any normalized set of weights wb would work, like for example wb = 1 B . However, uniform weights have the disadvantage to potentially give too much weight to posterior estimates from chains that have not quite converged. In the following heuristic improvement wb is chosen to be proportional to the posterior mass detected by run b, wb = M ′∈Ω∗ b P(Y \| M ′)P(M ′) B b=1 M ′∈Ω∗ b P(Y \| M ′)P(M ′) . This choice indirectly penalizes chains that cover smaller portions of the model space. When estimating posterior probabilities using these weights we only need, for each run, to store the following quantities: ˜Pb(Δ \| Y ) for all statistics Δ of interest and sb = M ′∈Ω∗ b P(Y \| M ′)P(M ′) as a ‘suﬃcient’ statistic of the run. There is no further need of data transfer between processes. 275 A. Hubin, G. Storvik, and F. 2The Abel cluster node (http://www.uio.no/english/services/it/research/hpc/abel/) with 16 dual Intel E5-2670 (Sandy Bridge, 2.6 GHz.) CPUs and 64 GB RAM under 64 bit CentOS-6 is a shared resource for research computing. Parallelization Frommlet Alternatively (as mentioned above) one might use (4) directly to approximate P(Δ \| Y ) based on the totality Ω∗of unique models explored through all of the parallel chains. This procedure might give in some cases slightly better precision than the weighted sum approach (15), but it is still only asymptotically unbiased. Moreover keeping track of all models visited by all chains requires signiﬁcantly more storage in the quick memory and RAM and requires signiﬁcantly more data transfers across the processes. Consequently this approach is not part of the current implementation of GMJMCMC. The consistency result of Theorem 1 also holds in case of the suggested embarrassing parallelization. Moreover it holds that even when the number of iterations per chain is ﬁnite that letting the numbers of chains B go to inﬁnity yields consistency of the posterior estimates as shown in Theorem A.1 in the web supplement. The main practical consequence is that running more chains in parallel allows for having a smaller number of iterations within each thread. Choice of algorithmic parameters Apart from the number of parallel chains, the GMJMCMC algorithm relies upon the choice of a number of tuning parameters which were described above. Section A of the web supplement presents the values that were used in the following simulation study and in real data analysis. Binary responses The responses of the ﬁrst three scenarios were sampled as modes of Bernoulli random variables with individual success probability π speciﬁed according to S.1 : logit(π) = −0.7 + L1 + L2 + L3, S.2 : logit(π) = −0.45 + 0.6 L1 + 0.6 L2 + 0.6 L3, S.3 : logit(π) = 0.4 −5 L1 + 9 L2 −9 L3, where the corresponding logic expressions are provided in Table 1. The ﬁrst two scenarios with models including only two-way interactions were copied from Fritsch (2006) except that we deliberately did not specify the trees in lexicographical order. The reason for this is that for some procedures (like stepwise search) it might be an algorithmic advantage if the eﬀects are speciﬁed in a particular order. The second scenario is slightly more challenging than the ﬁrst one due to the smaller eﬀect sizes. The third scenario is more demanding with a model including three-way and four-way interactions. As mentioned above the corresponding regression coeﬃcients were chosen rather large to make sure that these higher order trees can be detected for the given sample size. In practice when interested in smaller eﬀects one would need larger sample sizes. For the binary response scenarios GMJMCMC was compared with FBLR (Fritsch, 2006) and MCLR (Kooperberg and Ruczinski, 2005), where GMJMCMC was run with Jeﬀreys’ prior as well as with the robust g-prior. For GMJMCMC the default setting of the maximal number of leaves per tree is Cmax = 5. For Scenarios 1 and 2 we additionally report the results for Cmax = 2, which were the values used in the original study of Fritsch (2006) and which we also used here for MCLR and FBLR. For Scenario 3 we set Cmax = 5 for all three approaches. The maximal number of trees per model was set to kmax = 10 for GMJMCMC and FBLR whereas for MCLR it is only possible to specify a maximum of kmax = 5. This is apparently due to the complexity of prior computations in MCLR. Apart from the speciﬁcation of Cmax and kmax we used for all 3 algorithms their default priors. In all scenarios we used d = 15 for the population size in GMJMCMC. A Novel Algorithmic Approach to Bayesian Logic Regression would be necessary in practice to detect higher order interactions with smaller eﬀect sizes. would be necessary in practice to detect higher order interactions with smaller eﬀect sizes. 3.1 Simulation study The GMJMCMC algorithm was evaluated in a simulation study divided into two parts. The ﬁrst part considered three scenarios (numbered 1–3) with binary responses and the second part three scenarios (4–6) with quantitative responses. For each scenario we generated N = 100 datasets according to a regression model described by Equations (1) and (2) with n = 1000 observations and p = 50 binary covariates. The covariates were assumed to be independent and were simulated for each simulation run as Xj ∼ Bernoulli(0.3) for j ∈{1, . . . , 50} in the ﬁrst two scenarios and as Xj ∼Bernoulli(0.5) for j ∈{1, . . . , 50} in the last four scenarios. All computations were performed on the Abel cluster.2 For Scenarios 3, 5 and 6 the eﬀect sizes (βj’s) for higher order interactions might seem unrealistically large compared to real applications. To obtain more realistic scenarios with moderate eﬀect sizes and still suﬃcient power to detect larger trees one would have to increase the sample sizes. However, this would be quite challenging computationally for a simulation study. In the section on sensitivity analysis additional simulations for Scenario 5 illustrate which eﬀect sizes are needed with a sample size of n = 1000 for GMJMCMC to detect trees of diﬀerent size. Furthermore, we demonstrate that increasing the sample size by a factor 10 and reducing the eﬀect sizes by a factor 1/ √ 10 yields approximately the same power. This relationship indicates which sample sizes 276 A Novel Algorithmic Approach to Bayesian Logic Regression Binary responses GMJMCMC was run until up to 1.6×106 models were visited in the ﬁrst two scenar- ios and up to 2.7×106 models were visited for the third scenario (divided approximately equally on 32 parallel runs). The length of the Markov chains for FBLR and MCLR were chosen to be 2 × 106 for the ﬁrst two scenarios and 3 × 106 for the third scenario. By default a tree is classiﬁed as detected if the (estimated) marginal inclusion prob- ability is larger than 0.5. This corresponds to the median probability model of Barbieri et al. (2004). To evaluate the performance of the diﬀerent algorithms we estimated the following metrics: Individual power – the power to detect a particular tree from the data generating model; Overall power – the average power over all true trees; Individual power – the power to detect a particular tree from the data generating model; Individual power – the power to detect a particular tree from the data generating model; Overall power – the average power over all true trees; Overall power – the average power over all true trees; 277 277 A. Hubin, G. Storvik, and F. Frommlet FP – the expected number of false positive trees; FDR – the false discovery rate of trees; WL – the total number of wrongly detected leaves. Further computational details are given in Section B 1 of the web supplement Further computational details are given in Section B.1 of the web supplement. FBLR MCLR GMJMCMC Scenario 1 Jef. R. A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression A summary of the results for the ﬁrst three simulation scenarios is provided in Table 1. In all three scenarios, MCLR performed better than FBLR, even when taking into account the positively biased summary statistics of MCLR (see Section B.1 in the web supplement). On the other hand, GMJMCMC clearly outperformed MCLR and FBLR both in terms of power and in terms of controlling the number of false positives, where using Jeﬀreys’ prior gave slightly better results than using the robust g-prior. In the ﬁrst two scenarios GMJMCMC with Jeﬀreys’ prior worked almost perfectly both for Cmax = 5 and Cmax = 2. In the few instances where it did not detect the true tree it reported instead the two corresponding main eﬀects. Note however that in case of Cmax = 5 there were several instances where GMJMCMC detected Lc i ∧Lc j with (1 ≤i < j ≤3), which according to De Morgan’s law is equivalent to Li + Lj and was therefore counted as true positive both for Li and Lj. GMJMCMC with the robust g-prior had a few more instances where pairs of singletons were reported instead of the correct two-way interaction, especially when Cmax = 5 was used. FBLR and MCLR were also good at detecting the true leaves in these simple scenarios, but GMJMCMC was much better in terms of identifying the exact logical expressions. The third scenario is more complex than the previous ones but nevertheless GMJMCMC with Jeﬀreys’ prior performed almost perfectly. GMJMCMC with the robust g-prior had more diﬃculties to correctly identify the three-way and four-way interaction. Both FBLR and MCLR had severe problems to detect the true logic ex- pressions and they also reported a considerable number of wrongly detected leaves. For a more in depth discussion of these simulation results we refer to Section B.1 of the web supplement. Binary responses g L1 = Xc 1 ∧X4 0.30 ≤0.67 0.99 (0.97) 1.00 (0.98) L2 = X5 ∧X9 0.42 ≤0.61 0.99 (1.00) 0.96 (0.95) L3 = X11 ∧X8 0.33 ≤0.59 0.95 (0.91) 0.53 (0.77) Overall Power 0.35 ≤0.62 0.98 (0.96) 0.84 (0.90) FP 3.88 ≥2.70 0.08 (0.25) 1.01 (0.63) FDR 0.77 ≥0.06 0.03 (0.06) 0.25 (0.16) WL 1 0 0 (0) 0 (0) Scenario 2 L1 = Xc 1 ∧X4 0.32 ≤0.66 0.98 (0.97) 0.98 (0.97) L2 = X5 ∧X9 0.40 ≤0.67 0.99 (0.99) 0.94 (0.96) L3 = X11 ∧X8 0.37 ≤0.60 0.96 (0.86) 0.54 (0.76) Overall Power 0.36 ≤0.64 0.98 (0.94) 0.82 (0.90) FP 3.83 ≥2.58 0.10 (0.38) 1.08 (0.66) FDR 0.75 ≥0.06 0.03 (0.09) 0.27 (0.16) WL 1 1 0 (0) 0 (0) Scenario 3 L1 = X2 ∧X9 0.93 ≤0.93 1.00 1.00 L2 = X7 ∧X12 ∧X20 0.04 ≤0.67 0.91 0.56 L3 = X4∧X10∧X17∧X30 0.00 ≤0.19 1.00 0.56 Overall Power 0.32 ≤0.60 0.97 0.71 FP 6.40 ≥2.98 0.15 1.74 FDR 0.54 ≥0.06 0.04 0.39 WL 90 72 1 0 Further computational details are given in Section B.1 of the web supplement. Table 1: Results for the three simulation scenarios for binary responses. Power for in- dividual trees, overall power, expected number of false positives (FP) and FDR are compared between FBLR, MCLR and GMJMCMC using either Jeﬀreys’ prior (Jef.) or the robust g-prior (R.g.). For GMJMCMC the default Cmax = 5 is used. For the ﬁrst two scenarios we also present results for Cmax = 2 (inside parentheses) corresponding to the parameters used by MCLR and FBLR. All algorithms were tuned to use ap- proximately the same computational resources. In case of MCLR we can only provide upper bounds for the power and lower bounds for FP. We also report the total number of wrongly detected leaves (WL) over all simulation runs. 278 Continuous responses Responses were simulated according to a Gaussian distribution with error variance σ2 = 1 and the following three models for the expectation: S.4 : E(Y ) = 1+1.43 L1 + 0.89 L2 + 0.7 L3, S.5 : E(Y ) = 1+1.5 L1 + 3.5 L2 + 9 L3 + 7 L4, S.6 : E(Y ) = 1+1.5 L1 + 1.5 L2 + 6.6 L3 + 3.5 L4 +9 L5 + 7 L6 + 7 L7 + 7 L8. S.4 : E(Y ) = 1+1.43 L1 + 0.89 L2 + 0.7 L3, S.5 : E(Y ) = 1+1.5 L1 + 3.5 L2 + 9 L3 + 7 L4, S.6 : E(Y ) = 1+1.5 L1 + 1.5 L2 + 6.6 L3 + 3.5 L4 +9 L5 + 7 L6 + 7 L7 + 7 L8. The logic expressions used in the three diﬀerent scenarios are provided in Table 2. Scenario 4 is similar to the ﬁrst two scenarios for binary responses and contains only two-way interactions. The models of the last two scenarios both include trees of size 1 to 4, where Scenario 5 has one tree of each size. Scenario 6 is the most complex one with two trees of each size, resulting in a model with 20 leaves in total. For scenarios with Gaussian observations we were only able to study the performance of GMJMCMC since the other approaches cannot handle continuous responses (MCLR has an implementation but that did not work properly). For these scenarios the settings of GMJMCMC were adapted to the increasing complexity of the model. We used kmax = 279 A. Hubin, G. Storvik, and F. A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression Table 2 summarizes the results and further details are provided in Section B.2 of the web supplement. Scenario 4 illustrates that given a suﬃciently large sample size GMJMCMC can reliably detect two-way interactions with eﬀect sizes smaller than one standard deviation. Both Jeﬀreys’ prior and the robust g-prior worked almost perfectly in terms of power. In this simple scenario even the type I error was almost perfectly controlled with false discovery rates equal to 0.005 for Jeﬀreys’ prior and 0 for the robust g-prior. Interestingly the only false discovery over all 100 simulation runs was of the form X1 ∧X4 ∨X8 ∧X11 and is equal to L3 ∨L2. One might argue to which extent such a combination of trees should actually be counted as a false positive, a question which is further elaborated in Section B.2 of the web supplement and in the Discussion section. The remaining two scenarios are way more complex due to the higher order inter- action terms involved. In Scenario 5 the power to detect any of the four trees was very large, with only slightly smaller power for the four-way interaction. The robust g-prior had only a rather small advantage compared with Jeﬀreys’ prior both in terms of power (overall 97% against 96%) and in terms of type I error (FDR of 4% against 6%). For both priors the majority of false positive results were connected to detecting subtrees of true trees and in all simulation runs there were only 2 wrongly detected leaves for Jeﬀreys’ prior and 5 wrongly detected leaves for the robust g-prior. For the last scenario we again observed large power for all true trees up to order three. For the ﬁnal two expressions L7 and L8 of order for the results became slightly more ambiguous with power estimated to 0.32 and 0.21, respectively, for Jeﬀreys’ prior and 0.45 and 0.16 for the robust g-prior. However, among the false positive detections we very often found the expressions X11∧X13, X19∧X50 as well as X11∧X13∧X19∧X50. In fact in 72 simulation runs for Jeﬀreys’ prior and 69 simulation runs for the robust g-prior all of these three expressions were detected. According to the logic equivalence L8 = X11 ∧X13 + X19 ∧X50 −X11 ∧X13 ∧X19 ∧X50 one might actually consider these ﬁndings as true positives. A Novel Algorithmic Approach to Bayesian Logic Regression The numbers in parentheses in Table 2 were based on taking such similarities into account, resulting in much higher power. Among the remaining false positive detections more than two thirds were sub- trees of true trees or trees with misspeciﬁed logical operators but consisting of leaves corresponding to a true tree. Thus again the vast majority of false detections points towards true epistatic eﬀects where the exact logic expression was not identiﬁed. Inter- estingly like in Scenario 5 GMJMCMC with the robust g-prior detected again a larger number of wrong leaves than with Jeﬀreys’ prior. Continuous responses Frommlet 279 Scenario 4 Jeﬀreys Robust g L1 = X5 ∧X9 1.00 1.00 L2 = X8 ∧X11 0.99 1.00 L3 = X1 ∧X4 0.97 0.98 Overall Power 0.99 0.99 FP 0.01 0.00 FDR 0.005 0.00 WL 0 0 Scenario 5 Jeﬀreys Robust g L1 = X37 1.00 1.00 L2 = X2 ∧X9 1.00 0.99 L3 = X7 ∧X12 ∧X20 0.96 1.00 L4 = X4 ∧X10 ∧X17 ∧X30 0.89 0.90 Overall Power 0.96 0.97 FP 0.37 0.28 FDR 0.06 0.04 WL 2 5 Scenario 6 Jeﬀreys Robust g L1 = X7 0.95 0.99 L2 = X8 0.98 0.99 L3 = X2 ∧X9 0.98 0.99 L4 = X18 ∧X21 0.96 0.95 L5 = X1 ∧X3 ∧X27 1.00 1.00 L6 = X12 ∧X20 ∧X37 0.95 0.96 L7 = X4 ∧X10 ∧X17 ∧X30 0.32 0.45 L8 = X11 ∧X13 ∨X19 ∧X50 0.21 (0.93) 0.16 (0.85) Overall Power 0.79 (0.88) 0.81 (0.90) FP 4.28 (2.05) 4.24 (1.96) FDR 0.38 (0.19) 0.36 (0.16) WL 3 7 Table 2: Results for the three simulation scenarios for linear regression. Power for in- dividual trees, overall power, expected number of false positives (FP), FDR and the total number of wrongly detected leaves (WL) are given for parallel GMJMCMC. The four estimates in parentheses for Scenario 6 refer to results obtained when counting an equivalent logic expression of L8 as true positive as explained in the text. 10, 10 and 20, and d = 15, 20 and 40, respectively, for the three scenarios thus allowing for models larger than twice the size of the data generating model and populations at least twice the size of the number of correct leaves involved. Furthermore, the total number of models visited by GMJMCMC before it stopped was increased to 3.5 × 106 for Scenario 6. Cmax is set to 5 for all three of these scenarios. Otherwise all parameters of GMJMCMC were set as described for the binary responses. 280 A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression 4. A misspeciﬁed leave within L4, where the misspeciﬁed leave is substituted by a correlated leave with the correlation r varying from 0.1 to 1. In cases 2, 3 and 4 the relevant parameters were increased uniformly in 10 steps, in all cases kmax was set to 20. For computational reasons the sensitivity analysis was performed only using 10 simulation runs for each parameter value, both for Jeﬀreys’ prior and for the robust g-prior. This number of repetitions is not suﬃcient to give high resolution estimates of the power but it is enough to illustrate the general dependence on each of the considered parameters. The ﬁrst two plots of Figure 1 illustrate how the power to detect L4 changes when either the regression coeﬃcient β4 or the sample size n are varied. With a sample size of 1000 the power seems to deteriorate only for eﬀect sizes smaller than 4, whereas for the large eﬀect size of Scenario 5 a sample size of n = 600 still seems to provide reasonable power to detect L4. The ﬁrst plots of Figures S1–S3 of the web supplement show that for the lower order trees a sample size of n = 1000 is suﬃcient to obtain reasonable power for much smaller eﬀect sizes. Notably the three-way interaction L3 can be detected with large power already for β3 = 1 which is of the same order as the standard deviation of the error term. To reach suﬃcient power to detect four-way interactions with smaller regression coeﬃcients one would have to increase the sample size. For many statistical models there is the notion that when decreasing the eﬀect size by a factor 1/K one would roughly have to increase the sample size by a factor K2 to end up with the same power. Figure S4 from the web supplement indicates that this relationship also holds for the logic regression approach and together with the results from the ﬁrst plot of Figure 1 one can induce that a sample size of n > 10000 is needed to have suﬃcient power to detect four-way interactions with regression coeﬃcients which are of the order of the error standard deviation. The third plot of Figure 1 is concerned with the inﬂuence of the population size d from the GMJMCMC algorithm on the power to detect L4. Sensitivity analysis We performed sensitivity analysis for the power to detect trees in Scenario 5 based on ˜P(Lj\|Y ) > 0.5 for j ∈{1, . . . , 4}. Figure 1 presents the results for the four-way interaction L4. Results for the trees with fewer leaves are provided in Figures S1–S3 of the web supplement. Speciﬁcally we wanted to study how the power is eﬀected by the following factors: A. Hubin, G. Storvik, and F. Frommlet 281 Figure 1: Dependence of power to detect L4 for Jeﬀreys’ prior (red) and the robust g- prior (blue) when varying diﬀerent parameters as speciﬁed above each plot. Parameters which are not explicitly varied are kept ﬁxed at the levels from the original Scenario 5, except for the ﬁrst plot where all four coeﬃcients β1 . . . , β4 are simultaneously varied by multiplying with the same factor. Figure 1: Dependence of power to detect L4 for Jeﬀreys’ prior (red) and the robust g- prior (blue) when varying diﬀerent parameters as speciﬁed above each plot. Parameters which are not explicitly varied are kept ﬁxed at the levels from the original Scenario 5, except for the ﬁrst plot where all four coeﬃcients β1 . . . , β4 are simultaneously varied by multiplying with the same factor. 1. A change in the corresponding coeﬃcients βj, where all coeﬃcients are varied simultaneously by multiplying them with a factor K ∈{0.05, 0.1, 0.2, . . . , 1} and all other parameters are kept constant. 2. A change in the sample size n, where the sample size n is varied from 100 to 1000 and all other parameters are kept constant. 3. A change in the population size d, where the population size d is varied from 15 to 150 and all other parameters are kept constant. 282 3Data obtained from the second to ﬁfth column of the ﬁle http://journals.plos.org/plosone/article/ ﬁle?type=supplementary&id=info:doi/10.1371/journal.pone.0004318.s002. A Novel Algorithmic Approach to Bayesian Logic Regression Corresponding plots for the trees of lower size, for which the power is almost always equal to one, are provided in the web-supplement. As one can see for both priors power to detect L4 grows gradually from 0 to 1 when d changes from 15 to 45. For values of d > 30 the power remains stable at 1. This illustrates the statement of Theorem 1, according to which one requires d−d1 ≥kmax to have an irreducible algorithm in the restricted space of logic regression models. In these simulations we have kmax = 20 and d1 = 10. Hence according to Theorem 1 a population size d ≥30 is suﬃcient for asymptotic irreducibility of the GMJMCMC algorithm. For d −d1 < kmax irreducibility is no longer guaranteed and hence we cannot expect the approximations of the model posteriors to be precise in all cases, speciﬁcally when the model size of the data generating model is larger than d−d1. The ﬁnal plot of Figure 1 considers the eﬀect of misspeciﬁcation of one leave. This setting is motivated by genetic association studies, where it often happens that not a causal SNP (single nucleotide polymorphism) itself is genotyped but rather a strongly correlated tag SNP. As long as the correlation of the misspeciﬁed leave to the original leave is larger than 0.5 there appears to be no dramatic loss of power which indicates that a certain amount of model misspeciﬁcation can be tolerated by our method. 283 A. Hubin, G. Storvik, and F. Frommlet Phenotype Chr Marker expression ˜P(L \| Y ) Signif. Blue Light 4 X44606688 0.767 * Blue Light 5 X44607250 0.335 Blue Light 2 X21607656 0.309 ** Blue Light 4∧2 X44606688∧X44606810 0.203 * Red Light 2 MSAT2.36 0.441 Red Light 2 PHYB 0.353 Red Light 2∧1 PHYBc∧X44606541 0.112 * Red Light 2 X21607013 0.092 * Far Red Light 4 MSAT4.37 0.302 Far Red Light 4 NGA1107 0.302 White Light 5 X44606159 0.632 * White Light 1 X21607165 0.427 Table 3: Potential additive and epistatic QTL for hypocytol length under diﬀerent light conditions for Arabidopsis thaliana. Recombinant inbreed line data set taken from Bal- asubramanian et al. (2009). The last column shows the level of conﬁdence with * corresponding to ˜P(L \| Y ) > 0.5, to ˜P(L \| Y ) > 0.3 and ∗to ˜P(L \| Y ) > 0.05. 3.2 Analysis of Arabidopsis data According to our simulation results there is no large diﬀerence in the performance of GMJMCMC between using Jeﬀreys’ prior or the robust g-prior. On the other hand the clear computational advantage of Jeﬀreys’ prior seems to justify to omit the robust g-prior for analyzing real data. Hence in this section we only use Jeﬀreys’ prior for GMJMCMC. Furthermore we used kmax = 15 and d = 25 which allows for way more complex models than we would expect to see. Balasubramanian et al. (2009) mapped several diﬀerent quantitative traits in Ara- bidopsis thaliana using an advanced intercross-recombinant inbred line (RIL). Their data is publicly available as supporting information of their PLOS ONE article (Bala- subramanian et al., 2009) which also gives all the details of the breeding scheme and the measurement of the diﬀerent traits. We consider here only the hypocytol length in mm under diﬀerent light conditions.3 Genotype data is available for 220 markers distributed over the 5 chromosomes of Arabidopsis thaliana with 61, 39, 43, 31 and 46 markers, re- spectively. Balasubramanian et al. (2009) had genotyped 224 markers but we dismissed 4 markers which had identical genotypes with other markers. The amount of missing genotype data is relatively small with a genotype rate of 93.9% and most importantly the data contains only homozygotes (AA:49.6% vs. BB:50.4%). This means that the RIL population contains no heterozygote markers and logic regression can be directly applied using the genotype data as Boolean variables. Missing data were imputed using the R package R-QTL (http://www.rqtl.org/). The imputed data was then analyzed with our algorithm GMJMCMC to detect po- tential epistatic eﬀects and the results are summarized in Table 3. Under blue light Bala- 284 A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression subramanian et al. (2009) reported 4 potential QTL’s, the strongest one on chromosome 4 in the regions of marker X44606688 and three further fairly weak QTL on chromo- somes 2, 3 and 5. Our analysis based on logic regression conﬁrmed X44606688 and also detected those markers on chromosomes 2 and 5, though with a posterior probability slightly below 0.5. There was also some indication of a two-way interaction between the strong QTL on chromosome 4 and the QTL on chromosome 2. subramanian et al. 3.2 Analysis of Arabidopsis data (2009) reported 4 potential QTL’s, the strongest one on chromosome 4 in the regions of marker X44606688 and three further fairly weak QTL on chromo- somes 2, 3 and 5. Our analysis based on logic regression conﬁrmed X44606688 and also detected those markers on chromosomes 2 and 5, though with a posterior probability slightly below 0.5. There was also some indication of a two-way interaction between the strong QTL on chromosome 4 and the QTL on chromosome 2. Under red light the original interval mapping analysis reported the region of MSAT2.36 as a strong QTL on chromosome 2 and x44607889 as a weaker QTL on chro- mosome 1. Our logic regression analysis distributes the marker posterior weights on three diﬀerent markers on chromosome 2 which are all in the neighborhood of MSAT2.36. Ad- ditionally there is some rather small posterior probability for an epistatic eﬀect between this region and a marker on chromosome 1 which is rather close to x44607889. Finally both for Far Red Light and for White Light our analysis essentially yielded the same results as the interval mapping analysis, when observing that under the ﬁrst condition the posterior probability was again almost equally distributed between the neighboring markers MSAT4.37 and NGA1107. In summary the sample size in this data set might be slightly too small to detect epistatic eﬀects, although under the ﬁrst two light conditions there was at least some indication for a two-way interaction. We have analyzed a second data set concerned with QTL mapping for Drosophila where we compare logic regression with a more traditional approach to modeling epis- tasis. Further details and results are presented in Section D of the web supplement. 4 Discussion The success of GMJMCMC relies upon resolving the local extrema issue, which is mainly achieved by combining the following two ideas. First, when iterating through a ﬁxed search space S, GMJMCMC utilizes the MJMCMC algorithm (Hubin and Storvik, 2018) which was speciﬁcally constructed to explore multi-modal regression spaces eﬃ- ciently. Second, the evolution of the search spaces is governed within the framework of a genetic algorithm where a population consists of a ﬁnite number of trees forming the current search space. The population is updated by discarding trees with low estimated marginal posterior probability and generating new trees with a probability depending on the approximations of marginal inclusion probabilities from the current search space. The aim of the genetic algorithm is to converge towards a population which includes the most important trees. Finally the performance of GMJMCMC is additionally boosted by running it in parallel with diﬀerent starting points. Irreducibility of the proposals both for search spaces and for models within the search spaces guarantees that asymp- totically the whole model space will be explored by GMJMCMC and global extrema will at some point be reached under some weak regularity conditions. Clearly the genetic algorithm used to update search spaces results in a Markov chain of model spaces. One important question in the context of logic regression is concerned with how to deﬁne true positive and false positive detections in simulations. We adapted a rather strict point of view which might be called an ‘exact tree approach’: Only those de- tected logic expressions which were logically equivalent with trees from the data gener- ating model were counted as true positives. While this seems to be a natural deﬁnition there are certain pitfalls and ambiguities that occur in logic regressions which might speak against this strict deﬁnition. Apart from the more obvious logic equivalences ac- cording to Boolean algebra, for example due to De Morgan’s laws or the distributive law, there can be slightly more hidden logic identities in logic regression. For exam- ple the expressions (X1 ∨X2) −X1 and X2 −(X1 ∧X2) give identical models. We have seen a less trivial example including four-way interactions in Scenario 6 of our simulation study, where the data generating tree L8 is equivalent to the expression X11 ∧X13 + X19 ∧X50 −X11 ∧X13 ∧X19 ∧X50 consisting of three trees. 4 Discussion We have introduced GMJMCMC as a novel algorithm to perform Bayesian logic regres- sion and compared it with the two existing methods MCLR (Kooperberg and Ruczinski, 2005) and FBLR (Fritsch, 2006). The main advantage of GMJMCMC is that it is de- signed to identify more complex logic expressions than its predecessors. Our approach diﬀers both in terms of prior assumptions and in algorithmic details. Concerning the prior of regression coeﬃcients we compared the simple Jeﬀreys’ prior with the robust g-prior. Jeﬀreys’ prior in combination with the Laplace approximation coincides with a BIC-like approximation of the marginal likelihood, which was also used by MCLR. The robust g-prior has some very appealing theoretical properties for the linear model. However, in our simulation study it gave only slightly better results than Jeﬀreys’ prior for the linear model and in case of logistic regression actually performed worse in terms of power to detect the trees of the data generating logic regression model. With respect to the model topology we chose a prior which is rather similar to the one suggested by Fritsch (2006) for FBLR, but instead of using a truncated geometric prior for the number of leaves of a tree we suggest a prior which penalizes the complexity of a tree indirectly proportionally to the total number of trees of a given size. The motivation behind this prior is to control the number of false positive detections of trees in a similar way to how the Bonferroni correction works in multiple testing. GMJMCMC has the capacity to explore a much larger model search space than MCLR and FBLR because it manages to eﬃciently resolve the issue of not getting 285 A. Hubin, G. Storvik, and F. Frommlet stuck in local extrema, a problem that both MCLR and FBLR have in common. In logic regression the marginal posterior probability function is typically multi-modal in the space of models, with a large number of extrema which are often rather sparsely located. Additionally, the search space for logic regression is extremely large, where even computing the total number of models is a sophisticated task. As discussed in more detail in Hubin and Storvik (2018), in such a setting simple MCMC algorithms often get stuck in local extrema, which signiﬁcantly slows down their performance and convergence might only be reached after run times which are infeasible in practice. A Novel Algorithmic Approach to Bayesian Logic Regression Given estimates of posterior model probabilities, other prediction procedures such as the median probability model (Barbieri et al., 2004) or the posterior weighted median (Clarke et al., 2013) can also easily be applied. where Δ might be for example the predictor of unobserved data based on a speciﬁc set of covariates. Given estimates of posterior model probabilities, other prediction procedures such as the median probability model (Barbieri et al., 2004) or the posterior weighted median (Clarke et al., 2013) can also easily be applied. Supplementary Material Supplementary Material for: A novel algorithmic approach to Bayesian Logic Regression (DOI: 10.1214/18-BA1141SUPP; .pdf). https://github.com/aliaksah/EMJMCMC2016/ tree/master/supplementaries/Bayesian%20Logic%20Regression. A Novel Algorithmic Approach to Bayesian Logic Regression A Novel Algorithmic Approach to Bayesian Logic Regression reported two possible summaries of the simulation results, one based strictly on the exact tree approach and the other one counting simultaneous detections of X11∧X13, X19∧X50 and X11 ∧X13 ∧X19 ∧X50 also as true positives. This was slightly ad hoc and we believe that good reporting of logic regression results is an area which needs further research. The output of MCLR takes a step in that direction, where only the leaves of trees are reported and if a tree has been detected then also all its subtrees are reported. However, in our opinion MCLR throws away too much information. We believe that several diﬀerent layers of reporting might be more desirable, for example the exact tree approach, the MCLR approach and then something in between which does not reduce trees completely to their set of leaves. We have started to think more systematically in that direction and leave this topic open for another publication. reported two possible summaries of the simulation results, one based strictly on the exact tree approach and the other one counting simultaneous detections of X11∧X13, X19∧X50 and X11 ∧X13 ∧X19 ∧X50 also as true positives. This was slightly ad hoc and we believe that good reporting of logic regression results is an area which needs further research. The output of MCLR takes a step in that direction, where only the leaves of trees are reported and if a tree has been detected then also all its subtrees are reported. However, in our opinion MCLR throws away too much information. We believe that several diﬀerent layers of reporting might be more desirable, for example the exact tree approach, the MCLR approach and then something in between which does not reduce trees completely to their set of leaves. We have started to think more systematically in that direction and leave this topic open for another publication. This paper has had a focus on model selection and selection of features of interest. The method is however directly applicable to prediction as well. One can approximate the posterior probability of some parameter/variable Δ via model averaging by ˜P(Δ \| Y ) = M∈Ω∗ P(Δ \| M, Y ) ˜P(M \| Y ), where Δ might be for example the predictor of unobserved data based on a speciﬁc set of covariates. 4 Discussion Furthermore, diﬀerent logic expressions can be highly correlated even when they are not exactly identical. Especially the results from the most complex Scenario 6 impose the question whether the exact tree approach is slightly too strict to deﬁne false positives. Subtrees of true trees give valuable information even if they are not describing the exact interaction. Often combinations of several subtrees and trees with misspeciﬁed logical operators can give expressions which are very close to the correct interaction term. For Scenario 6 we 286 A Novel Algorithmic Approach to Bayesian Logic Regression References Balasubramanian, S., Schwartz, C., Singh, A., Warthmann, N., Kim, M., Maloof, J., Loudet, O., Trainer, G., Dabi, T., Borevitz, J., Chory, J., and Weigel, D. (2009). “QTL mapping in new Arabidopsis thaliana advanced intercross-recombinant inbred lines.” PLoS One, 4(2). 265, 283 Barber, R. F., Drton, M., and Tan, K. M. (2016). Laplace Approximation in High- Dimensional Bayesian Regression, 15–36. Cham: Springer International Publishing. MR3616262. 268 Barbieri, M. M., Berger, J. O., et al. (2004). “Optimal predictive model selection.” The annals of statistics, 32(3): 870–897. MR2065192. doi: https://doi.org/10.1214/ 009053604000000238. 270, 276, 286 Bayarri, M. J., Berger, J. O., Forte, A., Garc´ıa-Donato, G., et al. (2012). “Criteria for Bayesian model choice with application to variable selection.” The Annals of statis- 287 A. Hubin, G. Storvik, and F. Frommlet tics, 40(3): 1550–1577. MR3015035. doi: https://doi.org/10.1214/12-AOS1013. 267, 268, 269 Bogdan, M., Ghosh, J. K., and Tokdar, S. T. (2008). “A comparison of the Simes- Benjamini-Hochberg procedure with some Bayesian rules for multiple testing.” IMS Collections, Vol. 1, Beyond Parametrics in Interdisciplinary Research: Fetschrift in Honor of Professor Pranab K. Sen, edited by N. Balakrishnan, Edsel Pe˜na and Mervyn J. Silvapulle, 211–230. MR2462208. doi: https://doi.org/10.1214/ 193940307000000158. 267 Chib, S. (1995). “Marginal likelihood from the Gibbs output.” Journal of the American Statistical Association, 90(432): 1313–1321. MR1379473. 274 Chib, S. and Jeliazkov, I. (2001). “Marginal likelihood from the Metropolis– Hastings output.” Journal of the American Statistical Association, 96(453): 270–281. MR1952737. doi: https://doi.org/10.1198/016214501750332848. 274 Claeskens, G. and Hjort, N. L. (2008). Model Selection and Model Averaging. Cam- bridge Series in Statistical and Probabilistic Mathematics. Cambridge University Press. MR2431297. doi: https://doi.org/10.1017/CBO9780511790485. 268 Clarke, J. L., Clarke, B., Yu, C.-W., et al. (2013). “Prediction in M-complete Prob- lems with Limited Sample Size.” Bayesian Analysis, 8(3): 647–690. MR3102229. doi: https://doi.org/10.1214/13-BA826. 286 Clyde, M. A., Ghosh, J., and Littman, M. L. (2011). “Bayesian adaptive sam- pling for variable selection and model averaging.” Journal of Computational and Graphical Statistics, 20(1): 80–101. MR2816539. doi: https://doi.org/10.1198/ jcgs.2010.09049. 266 Fritsch, A. (2006). “A Full Bayesian Version of Logic regression for SNP Data.” Ph.D. thesis, Diploma Thesis. 263, 264, 276, 284 Fritsch, A. and Ickstadt, K. (2007). “Comparing Logic Regression Based Methods for Identifying SNP Interactions.” Springer Berlin / Heidelberg, Lecture Notes in Computer Science, 4414: 90–103. MR2291281. doi: https://doi.org/10.1080/ 09332480.2006.10722798. 263 Frommlet, F., Ljubic, I., Arnardottir, H., and Bogdan, M. (2012). References “QTL Mapping Using a Memetic Algorithm with modiﬁcations of BIC as ﬁtness function.” Statis- tical Applications in Genetics and Molecular Biology, 11(4): Article 2. MR2944873. doi: https://doi.org/10.1515/1544-6115.1793. 269 Gelman, A., Stern, H. S., Carlin, J. B., Dunson, D. B., Vehtari, A., and Rubin, D. B. (2013). Bayesian data analysis. Chapman and Hall/CRC. MR3235677. 268 Hubin, A. and Storvik, G. (2018). “Mode jumping MCMC for Bayesian variable selection in GLMM .” Computational Statistics and Data Analysis. MR3820324. doi: https://doi.org/10.1016/j.csda.2018.05.020. 264, 266, 270, 271, 285 Hubin, A., Storvik, G., and Frommlet, F. (2018a). “Deep Bayesian regression models.” arXiv preprint arXiv:1806.02160. Submitted for publication. 271 288 A Novel Algorithmic Approach to Bayesian Logic Regression Hubin, A., Storvik, G., and Frommlet, F. (2018b). “Supplementary Material for: A novel algorithmic approach to Bayesian Logic Regression.” Bayesian Analysis. doi: https://doi.org/10.1214/18-BA1141SUPP. 265 Janes, H., Pepe, M., Kooperberg, C., and Newcomb, P. (2005). “Identifying target pop- ulations for screening or not screening using logic regression.” Statistics in Medicine, 24: 1321–1338. MR2134561. doi: https://doi.org/10.1002/sim.2021. 263 Jeﬀreys, H. (1946). “An invariant form for the prior probability in estimation prob- lems.” Proceedings of the Royal Society of London. Series A, 186(1007): 453–461. MR0017504. doi: https://doi.org/10.1098/rspa.1946.0056. 268 Jeﬀreys, H. (1961). Theory of probability. Oxford University Press, London. 268 Keles, S., van der Laan, M., and Vulpe, C. (2004). “Regulatory motif ﬁnding by logic regression.” Bioinformatics, 20: 2799–2811. 263 Kooperberg, C. and Ruczinski, I. (2005). “Identifying Interacting SNPs Using Monte Carlo Logic Regression.” Genetic Epidemiology, 28: 157–170. 263, 264, 265, 276, 284 Li, Y. and Clyde, M. A. (2018). “Mixtures of g-priors in generalized linear models.” Journal of the American Statistical Association, (just-accepted). 264, 267, 268, 269 Malina, M., Ickstadt, K., Schwender, H., Posch, M., and Bogdan, M. (2014). “Detection of epistatic eﬀects with logic regression and a classical linear regression model.” Sta- tistical Applications in Genetics and Molecular Biology, 13(1): 83–104. MR3159119. doi: https://doi.org/10.1515/sagmb-2013-0028. 264, 265 McCullagh, P. and Nelder, J. (1989). Generalized Linear Models. 2nd Edi- tion. Chapman and Hall, London. MR3223057. doi: https://doi.org/10.1007/ 978-1-4899-3242-6. 265 Raftery, A. E., Madigan, D., and Hoeting, J. A. (1997). “Bayesian model averaging for linear regression models.” Journal of the American Statistical Association, 92(437): 179–191. MR1436107. doi: https://doi.org/10.2307/2291462. 268 Ruczinski, I., Kooperberg, C., and LeBlanc, M. (2003). “Logic regression.” Journal of Computational and Graphical Statistics, 12(3): 474–511. MR2002632. doi: https://doi.org/10.1198/1061860032238. 263, 264, 265 Ruczinski, I., Kooperberg, C., and LeBlanc, M. (2004). “Exploring Interactions in High- Dimensional Genomic Data: An Overview of Logic Regression, with Applications.” Journal of Multivariate Analysis, 90: 178–195. MR2086341. doi: https://doi.org/ 10.1016/j.jmva.2004.02.010. 263 Schwarz, G. (1978). “Estimating the dimension of a model.” The Annals of Statistics, 6: 461–464. MR0468014. 268 Schwender, H. and Ickstadt, K. (2008). “Identiﬁcation of SNP interactions using logic regression.” Biostatistics, 9: 187–198. 263 289 A. Hubin, G. Storvik, and F. Frommlet A. Hubin, G. Storvik, and F. Frommlet Schwender, H. and Ruczinski, I. (2010). “Logic Regression and Its Extensions.” Ad- vances in Genetics, 72: 25–45. 263 Schwender, H. and Ruczinski, I. (2010). “Logic Regression and Its Extensions.” Ad- vances in Genetics, 72: 25–45. 263 Scott, J. G. A Novel Algorithmic Approach to Bayesian Logic Regression and Berger, J. O. (2008). “Bayes and empirical-Bayes multiplicity ad- justment in the variable-selection problem.” Annalls of Statistics, 38(5): 2587–2619. MR2722450. doi: https://doi.org/10.1214/10-AOS792. 266 Tierney, L. and Kadane, J. B. (1986). “Accurate Approximations for Posterior Moments and Marginal Densities.” Journal of the American statistical association, 81(393): 82– 86. MR0830567. 268 Tjelmeland, H. and Hegstad, B. K. (2001). “Mode jumping proposals in MCMC.” Scan- dinavian Journal of Statistics, 28(1): 205–223. MR1844357. doi: https://doi.org/ 10.1111/1467-9469.00232. 264, 270 Wakeﬁeld, J. (2007). “A Bayesian measure of the probability of false discovery in genetic epidemiology studies.” The American Journal of Human Genetics, 81(2): 208–227. 270 Wang, Y. H. (1993). “On the number of successes in independent trials.” Statistica Sinica, 295–312. MR1243388. 267 Invited Discussion Ingo Ruczinski∗, Charles Kooperberg†, and Michael LeBlanc‡ The logic regression project started some 20 years ago as part of the doctoral disserta- tion of Ingo Ruczinski. The initial motivation was indeed to develop and implement a method speciﬁcally to detect epistatic interactions in genetic studies measuring single nucleotide polymorphisms (SNPs). However, it soon occurred to us that logic regression could also be useful for many other data types and settings, particularly in medical studies where often many binary data are collected. In its ﬁnal version, the disserta- tion contained two applications of logic regression: 1) a genetic association study using SNP data from the Genetic Analysis Workshop (GAW), and 2) a medical study to infer which brain regions aﬀected by infarcts inﬂuence the cognitive state of patients. Logic regression was quickly adopted by the community after the ﬁrst open source soft- ware release, with many applications analyzing data with predictors other than SNPs. Moreover, various groups also developed new methodology extending the original logic regression framework (including Bayesian versions of logic regression), which is partic- ularly rewarding for its creators. The algorithm introduced here by Hubin, Storvik and Frommlet (HSF hereafter) is an improvement for Bayesian model selection in the space of logic regression models. We would like to congratulate the authors and say “thank you” for their contribution to the ﬁeld, and would like to oﬀer a few additional thoughts and perspectives. We completely agree with the authors that the search algorithm in the original Bayesian version of logic regression (Kooperberg and Ruczinski, 2005) had room for improvement. We also experienced that the Markov Chain Monte Carlo (MCMC) algo- rithm can get stuck in a particular part of the model space, and we are not surprised that the method and implementation put forth by HSF based on “mode jumping”, the Genetically modiﬁed Mode Jumping Markov Chain Monte Carlo (GMJMCMC) algo- rithm, performs better in this regard. Our implementation of MCMC was in essence a modiﬁcation of the simulated annealing algorithm we developed to maximize the (fre- quentist) likelihood function. Great care in setting the parameters for the annealing scheme is required, since the algorithm for the search of the global optimum can easily get trapped in local extrema as well. The main contribution by HSF is a greatly im- proved search strategy to explore an extremely “ragged” likelihood landscape. Acknowledgments The ﬁrst two authors gratefully acknowledge the ﬁnancial support of the CELS project at the University of Oslo, http://www.mn.uio.no/math/english/research/groups/cels/index.html. 290 Invited Discussion , , g ‡Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, mleblanc@fhcrc.org ∗Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health, ingo@jhu.edu †Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, clk@fhcrc.org ∗Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health, ingo@jhu.edu †Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, clk@fhcrc.org ‡Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, mleblanc@fhcrc.org Invited Discussion We think that a non-Bayesian equivalent of GMJMCMC can also be used for maximizing such a likelihood, similar to simulated annealing. We discuss this in more detail further below. In the abstract, HSF motivate the GMJMCMC algorithm by stating that logic re- gression “has been mainly used to model epistatic eﬀects in genetic association studies, which is very appealing due to the intuitive interpretation of logic expressions to describe the interaction between genetic variations. Nevertheless logic regression has (partly due 291 I. Ruczinski, C. Kooperberg, and M. LeBlanc to computational challenges) remained less well known than other approaches to epistatic association mapping.” We believe the strong focus on epistasis in the abstract is not necessary, and it might actually deter practitioners who do not work with genetic data. The Boolean combinations of binary variables used in logic regression indeed lend them- selves to genetic analyses, be it as recorded mutations, markers in breeding studies, or bi-allelic SNPs recorded as two binary variables in dominant and recessive coding. In our initial eﬀorts, we used data from genetic candidate studies, which at the time recorded between a few dozen and a few hundred markers. The early 2000s, when the logic re- gression methodology was published and the software released (Kooperberg et al., 2001; Ruczinski et al., 2003), saw the emergence of commercial SNP genotyping arrays (Bum- garner, 2013), typing tens of thousands of markers at a time. Trying to infer higher-order epistatic interactions (with logic regression or otherwise) in data generated from even these ﬁrst generation arrays is futile, given the size of the search space to explore these interactions and the ensuing multiple comparisons problem. Even when only pairwise interactions are considered, 100,000 markers yield 4.5 billion possible SNP-SNP interac- tions! We note that this is also reﬂected in the data HSF present: the simulation study comprises of 50 markers, the Arabidopsis data have 220 markers, and the Drosophila data in the supplementary materials have 45 markers. Logic regression and similar algo- rithms are suitable for data of this dimensionality, but certainly not for modern genome scans with millions of markers typed or sequenced. But this is not a limitation of the GMJMCMC algorithm – logic regression is still being used in many settings, particularly in the medical literature (see for example the Introduction of Tietz et al. Invited Discussion Jeﬀreys’ prior proposed by HSF for generalized linear models leads to an objective function that in essence corresponds to a penalized likelihood. Thus, when adopting the framework of HSF, one could dramatically save on CPU time by only running one annealing chain using the above described posterior probability terms as the objective function, without the need for cross-validation or permutation tests for model size selection. We note that our soft- ware allows for the speciﬁcation of one’s own objective function, as described in the software manual (https://CRAN.R-project.org/package=LogicReg). In addition to numerous options to deﬁne complexity for logic regression models, we concur with the authors that it is also not clear-cut how the performance of algorithms to detect Boolean interactions should be evaluated (HSF, p. 23). In their simulation study, HSF classify a tree as detected if the marginal inclusion probability is estimated to be at least 50%, and report various metrics (the power to detect a particular tree from the data generating model, the average power over all true trees, the expected number of false positive trees, the false discovery rate of trees, and the total number of wrongly detected leaves) to evaluate the performance of the algorithm. Since “Jeﬀreys’ prior for model selection has been widely criticized for not being consistent once the true model coincides with the null model” (HSF, p. 6) the authors also evaluate the GMJMCMC algorithm using the robust g-prior. While no dramatic diﬀerences are observed in their simulation based on non-null models, it appears that Jeﬀreys’ prior performs a bit better for the logistic models than the robust g-prior according to the above mentioned metrics. So which one to choose in practice? For the analysis of the Arabidopsis data the authors argue that “the clear computational advantage of Jeﬀreys’ prior seems to justify to omit the robust g-prior for analyzing real data” (HSF, p. 21). We suggest a simple two-step procedure for the practitioner that circumvents the need to make this decision. The original logic regression framework oﬀers an easily executed permutation test to determine whether there is any signal in the data (the “null model test” in Ruczinski et al., 2003), which answers the questions whether the assumption of a non-null model is correct. If there is a signal, simply proceed with the GMJMCMC algorithm using Jeﬀreys’ prior. Invited Discussion (2019) for a number of recent examples), and these applications will also beneﬁt from advancements such as GMJMCMC. The GMJMCMC algorithm puts a prior on model size (equations (3) and (6) in HSF) and therefore depends on a deﬁnition of model complexity. This is not a trivial issue, since the predictors in logic regression models are Boolean combinations of binary covariates, and the number of parameters in the model is the same regardless how com- plex the Boolean terms are. In addition, equivalent Boolean terms can have diﬀerent expressions and therefore also a diﬀerent number of binary predictors, for example the Boolean expressions X1 ∧(X2 ∨X3) and (X1 ∧X2) ∨(X1 ∧X3). HSF write the model prior (equation (3)) as a product of terms, introduced to give smaller probabilities to more complex trees (subject to the total number of trees not exceeding the number of trees allowed). These terms are chosen so the multiplicative contribution of a logic tree of a given size is inversely proportional to the number of possible trees having the same number of leaves. To estimate this number and to deal with the thorny issue of tree complexity, HSF propose to ignore Boolean terms that include the same binary covariate multiple times. This is a very reasonable proposal we believe, and allows for a straightforward estimate of the number of possible trees to be incorporated in the model prior. These terms are also chosen so larger models are penalized more (i.e. the prior probability of a model is always larger than the prior probability of any other model it is nested in). Using Jeﬀreys’ prior for the regression parameters, the authors highlight that the model posterior probabilities can be calculated using the Laplace approximation, and discuss the relationship with the Bayesian Information Criterion. We would like to add that this approach also allows for an alternative model selec- tion strategy in the original logic regression approach as introduced in Ruczinski et al. 292 Invited Discussion Invited Discussion (2003), where we try to obtain a model that best explains the observed data. Logic regression uses simulated annealing to ﬁnd optimal models (according to the objective function used) for a variety of possible model sizes, and then employs cross-validation or permutation tests to select the suitable size for the model. Invited Discussion For linear models as the ones presented in this manuscript one could of course also regress these variables out, and use the residuals as dependent variables to search for the Boolean expressions of the binary predictors, but that strategy is not possible for generalized linear models with non- linear link functions, such as the logistic model in the simulation study. Another practical question is how the relevant parameters in the GMJMCMC al- gorithm should be chosen to obtain dependable results. Clearly, this depends on the problem at hand – more predictors require a longer run, but also a more complex data structure (i.e. higher order interactions) demands a longer search. The authors use the default tuning parameters of the implementation of the underlying Mode Jumping Markov Chain Monte Carlo (MJMCMC) algorithm in all simulations and data analyses presented, but use a range of values for the parameters related to the genetic algorithm of the GMJMCMC algorithm (HSF, Supplementary Table A.1). Do the authors have some general guidance how to choose these? A simulated annealing approach like the one implemented in logic regression – in theory – converges to the optimal solution as long as the chain is aperiodic and irreducible (van Laarhoven and Aarts, 1987). We do not have inﬁnite CPU time in practice however, so rely on some observable metrics to guide the annealing algorithm in logic regression. We implemented the search as a se- quence of Metropolis-Hastings algorithms by keeping the temperature ﬁxed for a chain, and then gradually decreasing the temperature to generate a sequence of limiting distri- butions converging to the optimum. In our implementation we suggest to monitor the acceptance probabilities of the proposed moves in each of the chains: these probabilities have to be essentially 100% early on at higher temperatures when almost every move has to be accepted, and slowly have to converge to 0% as only moves that improve the score should be accepted for very low temperatures (and once the optimum has been reached, the acceptance probability for any move at that temperature should be zero in essence). Further, these types of Metropolis-Hastings based simulated annealing approaches also undergo a phase transition (van Laarhoven and Aarts, 1987), and the variance of the scores visited in a chain should be constant before dropping to zero. Invited Discussion It was a bit surprising to us that the authors simulated independent binary predictors to assess the performance of the algorithms. In real data we commonly see dependent random variables (e.g. genetic markers can be highly correlated due to linkage dise- quilibrium), and Bayesian approaches are particularly suitable to address the ensuing model uncertainty as the notion of one “best” model is very questionable due to the correlation structure between the binary variables (the sensitivity analyses presented in HSF Figure 1 and the supplementary materials speak to that to some degree). We also wonder if in the here presented simulation study, especially for the models with large eﬀect sizes, the original logic regression approach as introduced in Ruczinski et al. (2003) might have been a more suitable approach than for example Markov Chain logic regression (MCLR)? Due to the independence of the predictors and the large eﬀect sizes used, it would not be surprising to us if the original annealing based approach would 293 I. Ruczinski, C. Kooperberg, and M. LeBlanc consistently detect the underlying interactions in the simulation study. As mentioned above and discussed by HSF (p. 23), it could also be debated whether any algorithm employed really needs to detect the exact Boolean trees, or simply harnesses the power to explore binary interactions to detect the leaves involved in these Boolean trees (this, we argue, would be the case for example in genetic association studies). Thus, in ad- dition to the total number of wrongly detected leaves (i.e. the speciﬁcity), we think the number of correctly detected leaves (i.e. the sensitivity) could also be of interest (and if all leaves are consistently detected due to the large eﬀect sizes, a maybe more challenging simulation could be considered). To end, we have a few more technical questions for the authors. In many biomedical applications we want to adjust for some predictors (additively) in the model, such as the age and body mass index of subjects, or some principal components to correct for genetic heterogeneity in association studies. Is this easily accommodated in the implementation of the GMJMCMC algorithm, specifying a prior for the corresponding parameters similar to the one for the intercept? Invited Discussion Are there similar metrics for the GMJMCMC algorithm that could be considered to guide the selection of the critical parameters such as Mfin and Tmax, and therefore the resulting chain length? Invited Discussion 294 Invited Discussion ∗Department of Mathematics, University of Wroclaw, Plac Grunwaldzki 2/4, 50-384 Wroclaw, Poland, malgorzata.bogdan@uwr.edu.pl †Faculty of Mathematics, Informatics and Mechanics, University of Warsaw, Banacha 2, 02-097 Warsaw, Poland, B.Miasojedow@mimuw.edu.pl ‡Department of Statistics, Lund University, Box 743, 220 07 Lund, Sweden, jonas.wallin@stat.lu.se Invited Discussion Malgorzata Bogdan∗,‡, Blazej Miasojedow†, and Jonas Wallin‡ First of all we would like to congratulate the authors for a very interesting and important article. Logic regression model introduced in Ruczinski (2000); Ruczinski et al. (2003, 2004) is a Generalized Linear Model (GLM) where individual predictors take form of logic expressions dependent on binary explanatory variables. This model arises naturally in the context of identifying epistatic eﬀects in genetic studies. Following Bateson and Mendel (1909), biological epistasis is usually understood as a phenomenon in which “a variant or allele at one locus [. . . ] prevents the variant or allele at another locus from manifesting its eﬀect” (see Cordell, 2002), or more generally as a situation when the eﬀect of one allele can only be observed when a second allele is also present. Such epistatic eﬀects can be naturally expressed using logic expressions of the binary variables dependent on the genotypes of genetic markers. While each logic expression can be also represented in the form of the regular linear model, this usually requires many main eﬀects and lower interaction terms. For example, a single “logic interaction” involving four variables (x1 ∨x2) ∧(x3 ∨x4) in classical representation takes the form x1x3 + x1x4 + x2x3 + x2x4 −x1x3x4 −x2x3x4 −x1x2x3 −x1x2x4 + x1x2x3x4 , (0.1) and its natural interpretation is lost in the large number of classical interaction terms. Moreover, the possible causal inﬂuence of this “logic interaction” is practically impos- sible to recover by the regular linear model, where the regression coeﬃcients by each component of (0.1) are estimated separately. Logic regression seems to be particularly useful for the analysis of outbred popula- tions (like humans), where the number of genetic variants is often much larger than in controlled populations (like e.g. domesticated animals or experimental crosses). Also, it can be applied in a much wider context, like e.g. for the natural representation of the joint inﬂuence of general qualitative variables or for the model selection for discrete multicolored graphical models, like the Potts model, in the spirit of (Miasojedow and Rejchel, 2018; Banerjee et al., 2008; H¨oﬂing and Tibshirani, 2009; Ravikumar et al., 2010). In case of multicolored graphical models logic expressions can naturally describe dependence between nodes of the graph. References Bumgarner, R. (2013). “Overview of DNA microarrays: types, applications, and their future.” Current Protocols in Molecular Biology, Chapter 22: Unit 22.1. 291 Kooperberg, C. and Ruczinski, I. (2005). “Identifying interacting SNPs using Monte Carlo logic regression.” Genetic Epidemiology, 28: 157–170. 290 Kooperberg, C., Ruczinski, I., LeBlanc, M., and Hsu, L. (2001). “Sequence analysis using logic regression.” Genetic Epidemiology, 21 Suppl 1: S626–S631. MR2002632. doi: https://doi.org/10.1198/1061860032238. 291 Ruczinski, I., Kooperberg, C., and LeBlanc, M. (2003). “Logic regression.” Journal of Computational and Graphical Statistics, 12(3): 475–511. MR2002632. doi: https:// doi.org/10.1198/1061860032238. 291, 292 Tietz, T., Selinski, S., Golka, K., Hengstler, J. G., Gripp, S., Ickstadt, K., Ruczinski, I., and Schwender, H. (2019). “Identiﬁcation of interactions of binary variables associated with survival time using survivalFS.” Archives of Toxicology, 93: 585–602. 291 van Laarhoven, P. J. and Aarts, E. H. (1987). Simulated Annealing: Theory and Appli- cations. Kluwer Academic Publishers. MR0904050. doi: https://doi.org/10.1007/ 978-94-015-7744-1. 293 295 M. Bogdan, B. Miasojedow, and J. Wallin Invited Discussion Application of logic regression in real life problems requires solving complex compu- tational and statistical issues, resulting from the large number of possible logic expres- sion models and the possibility of writing a single logic expression in many equivalent tautological forms. For example, the logicFS program of Schwender and Ickstadt (2008) uses simulated annealing (Kirkpatrick et al., 1983) to maximize the likelihood function 296 Invited Discussion over all logic regression models with a given number of leaves. After selecting the “best” model, each logic expression is transformed into a disjunctive normal form (DNF) i.e., OR combination of AND combinations (i.e. prime implicants or logic interactions). The importance of individual interactions is estimated by repeating the whole procedure using many bootstrap samples from the original data and taking into account both the frequency with which a given interaction appears in bootstrap replications as well as its contribution to a total model likelihood. The disadvantage of the importance measures proposed in Schwender and Ickstadt (2008) is that their values depend on the size of the data set and there exist no natural thresholds which would allow to separate important interactions from false predictors. However, these importance measures can be used for ranking the potential interactions. Concerning model selection strategies, Malina et al. (2014) use logicFS importance measures to build a GLM model including a moderate number of most important inter- actions. Then the “statistically signiﬁcant” interactions are selected using the backward elimination procedure based on the multiplicity adjusted p-values. The multiplicity ad- justment takes into account that the number of interactions in the space searched by logicFS increases with the interaction complexity. In Hubin et al. (2020) the issue of identifying important logic interactions is ad- dressed within a Bayesian framework, where the importance of a given logic expression is measured by the sum of posterior probabilities of GLM models which contain this expression as one of predictors. The algorithm in Hubin et al. (2020) calculates the posterior probability for a GLM model M by an approximation to the Bayes rule. The marginal likelihood of the data given M is calculated using the analytical formulas or Laplace approximations for Jeﬀreys’ or robust g-priors. This allows to avoid computa- tional burden of Markov Chain Monte Carlo (MCMC) search over the space of model parameters. Similarly as in Bogdan et al. (2004); Baierl et al. Invited Discussion (2006), the prior for each model M depends on its complexity and is selected in such a way that the prior ex- pected numbers of logic expressions of diﬀerent lengths are approximately the same and do not depend on the number of predictors m. Since the number of complex interactions increases with m at a higher rate than the number of simple interactions, this eﬀectively introduces the additional penalty on the model complexity, which depends on m. The arguments presented in Bogdan et al. (2008b,a) illustrate that this penalty is related to the Bonferroni-type correction for multiplicity, similar to the multiple testing correction used in Malina et al. (2014). To calculate the posterior probability of a model M the authors use the Bayes rule P(M\|Y ) = P(Y \|M)P(M) Ω P(Y \|M)P(M) , (0.2) (0.2) where Ω contains all possible logic regression models. Since it is not possible to visit all these models, the main computational challenge relies on designing a search algorithm which can visit most of the likely models, thus well approximating the denominator of (0.2). Similar problem appears also when ﬁtting regular regression models and in Frommlet et al. (2012a) it was approached by the application of the genetic algorithm supplied with the “local” research in the neighborhood of promising models. In Hu- bin et al. (2020) the authors propose an iterative algorithm, where in each iteration 297 M. Bogdan, B. Miasojedow, and J. Wallin some new predictors are formed using the speciﬁcally designed crossover, mutation and reduction operators on the selected set of logic expressions and then apply the Mode Jumping MCMC (MJMCMC) of Hubin and Storvik (2018) to search the space of GLM models based on these predictors. some new predictors are formed using the speciﬁcally designed crossover, mutation and reduction operators on the selected set of logic expressions and then apply the Mode Jumping MCMC (MJMCMC) of Hubin and Storvik (2018) to search the space of GLM models based on these predictors. While we believe that the article of Hubin et al. (2020) is an interesting and impor- tant contribution to the research on the logic regression, we are rather reserved with respect to the proposed algorithm. In Section 2.3 of Hubin et al. (2020) it is mentioned that a proper MCMC algorithm is not needed if the main purpose is to visit many highly probable models. Invited Discussion We agree with the authors and believe that the reversibility of MJMCMC is actually not desired, since it creates unnecessary loops and increases the time of visiting many distinct models. In our opinion a better performance could be obtained by constructing an irreducible and well mixing algorithm of walking over the space of GLM models. In the recent years non-reversible MCMC algorithms received large attention (see e.g. Bouchard- Cˆot´e et al., 2018; Bierkens et al., 2019) due to the fact that non-reversible chains are able to explore the state space much faster than the reversible algorithms. For example, let us consider a uniform distribution on the set 0, 1, . . . , N. In this case the standard reversible MCMC algorithm reduces to a random walk. Hence, after n steps the expected number of explored states is proportional to √n and the number of moves to explore the whole space is proportional to N 2. Instead, we could construct a simple, non-reversible algorithm; i.e. we remember the direction of the previous move and go in the same direction until we hit 0 or N, where the direction is reversed. Then we can explore the whole space in at most 2N steps. In case of the problem discussed in Hubin et al. (2020), the construction of the non-reversible MCMC algorithm would be rather simple, since the convergence to the stationary measure is not needed. The only requirement is that the algorithm is irreducible and aperiodic. One solution here would be to deﬁne the global and local moves and accept the new state with probability (π(y)/π(x))α with some α > 0. The parameter α would control the permissible deviations of the posterior with respect to its maximum. Another solution could rely on storing the visited states in a priority queue, with priority proportional to the posterior probability. Then the elements from the queue could be modiﬁed by some kernel and placed back to the queue. Such an approach would allow us to explore the space starting from the more promising candidates. Also, this method could be easily parallelized without the need of post processing. Further, we are concerned with the lack of treatment for tautologies. It seems to us that this might lead to the dilution of the posterior probability among many tautological representations of a given interaction and the loss of power of identiﬁcation of this interaction. Invited Discussion While at the ﬁnal stage of the algorithm this problem can be solved by post-processing of the output, it is not clear what is the impact of this dilution on elimination of interesting interactions at the earlier stages of the algorithm. It is also important to observe that the number of tautological representations increases with the interaction complexity. Thus, if one merges all tautologies to a single logical expression in a post-processing step, the total prior probability assigned to this unique expression eﬀectively increases with its length and counterbalances the eﬀect of the multiplicity correcting priors suggested by the authors. 298 Invited Discussion The authors estimate the posterior probabilities of diﬀerent models using (0.2). It is not clear to us why the sum in the denominator of (0.2) contains only Mfin = 10000 models based on d trees from the ﬁnal stage of the algorithm. Why not use the information from the earlier stages? Further, in some of the reported simulation examples the authors use d = 15. Thus the ﬁnal search is performed only over d = 215 = 32768 models, which could be easily looked at without application of the MCMC algorithm. Also, a huge random reduction of the ﬁnal model space leads to substantially diﬀerent results for diﬀerent parallel runs of the algorithm. Therefore the authors aggregate results from diﬀerent runs using a weighting scheme speciﬁed in (15) of their paper. In our opinion it seems more reasonable to estimate the posterior probabilities of diﬀerent models simply by including all models visited in diﬀerent runs in denominator of (0.2). Also, as we mentioned above, it seems to us that the priority queues would allow for some synchronization between diﬀerent runs and more eﬃcient search through the model space. Concerning implementation issues – we observed that the denominator of (0.2) cal- culated by the currently implemented algorithm includes only the models accepted by MJMCMC. Taking into account that the acceptance rate is usually below 0.1, storing all the models proposed rather than only accepted would give a better estimate of the denominator of (0.2). Further, it seems to us that in the current implementation the denominator of (0.2) increases every time the model is accepted by MJMCMC, with- out checking if this model already appeared in the sum. Invited Discussion However, the detailed analysis of the hidden duplication problems would require a more careful analysis of the code, which is rather diﬃcult due to its structure. The authors conclude that there is almost no diﬀerence between the results when the Jeﬀreys’ or the robust g-prior is used when calculating the model marginal likelihood. However, it seems important to note that the simulations justifying this claim were performed using rather simple GLM models with independent predictors. Actually, it seems that many of the solutions proposed by the authors are speciﬁcally designed for this case. For example, consider the case when a given predictor is strongly correlated with other explanatory variables. Then the posterior probability of a “true” model including this predictor will be diluted between “neighboring” models and this predictor might easily miss the threshold for inclusion in the subsequent populations. As noted by the authors, the dilution of posterior probabilities actually occurred in the real data from the experimental recombinant inbred line, where the neighboring markers are rather strongly correlated. We simulated similar spatially correlated data and had a substantial diﬃculty with identifying a simple two-way logic interaction. Actually, the dilution issue seems to be even more problematic for interactions than for the main eﬀects since the number of correlated interaction terms is substantially higher than the number of respective correlated markers. Also, one of important features of the algorithm is the initial selection of d1 important binary variables, which stay as the single trees in the spaces Si in all iterations of the algorithm. The initial space S1 is formed by including logic expressions dependent only on these predictors. Other variables can enter the search space only during the 299 M. Bogdan, B. Miasojedow, and J. Wallin mutation, which occurs with a relatively low probability. Thus, the selection of these initial predictors eﬀectively reduces the search space. This approach again seems to be very well suited for the situation when explanatory variables are independent but might lead to missing important predictors otherwise. mutation, which occurs with a relatively low probability. Thus, the selection of these initial predictors eﬀectively reduces the search space. This approach again seems to be very well suited for the situation when explanatory variables are independent but might lead to missing important predictors otherwise. Another interesting property, worth studying, is the scaling of the algorithm with respect to the number of explanatory variables m. Invited Discussion This number seems to hinder the speed of MJMCMC only at the ﬁrst step, where d1 important main eﬀects are selected. However, the magnitude of m probably strongly inﬂuences the power of identifying logic interactions. Since the number of possible logic interactions increases rapidly with m, the prior probability for each of them quickly diminishes, which results in decrease of posterior probabilities. To summarize: it appears to us that the usefulness of the proposed algorithm and the GLM logic regression model is rather restricted to the case when predictors are roughly independent and n ≫m. This is however still of a great value in genetic studies, where the raw data are often pre-processed and only relatively few candidate genetic markers are used for building more sophisticated predictive models. If such markers are suﬃciently distant, they are almost not correlated. The candidate markers are usually selected using the prior biological knowledge. Since logic interactions usually have strong main eﬀects components, the candidate markers could also be selected using the classical Genome Wide Association Studies (see e.g. Frommlet et al., 2012b or Brzyski et al., 2017). Invited Discussion Invited Discussion Bogdan, M., Ghosh, J., and ˙Zak-Szatkowska, M. (2008b). “Selecting explanatory vari- ables with the modiﬁed version of Bayesian Information Criterion.” Quality and Re- liability Engineering International, 24: 627–641. 296 Bogdan, M., Ghosh, J., and ˙Zak-Szatkowska, M. (2008b). “Selecting explanatory vari- ables with the modiﬁed version of Bayesian Information Criterion.” Quality and Re- liability Engineering International, 24: 627–641. 296 Bouchard-Cˆot´e, A., Vollmer, S. J., and Doucet, A. (2018). “The Bouncy Particle Sam- pler: A Nonreversible Rejection-Free Markov Chain Monte Carlo Method.” Journal of the American Statistical Association, 113(522): 855–867. MR3832232. doi: https:// doi.org/10.1080/01621459.2017.1294075. 297 Brzyski, D., Peterson, C., Sobczyk, P., Cand`es, E., Bogdan, M., and Sabatti, C. (2017). “Controlling the rate of GWAS false discoveries.” Genetics, 205: 61–75. 299 Cordell, H. (2002). “Epistasis: what it means, what it doesn’t mean, and statistical methods to detect it in humans.” Human Molecular Genetics, 11: 2463–2468. 295 Frommlet, F., Ljubic, I., Arnardottir, H., and Bogdan, M. (2012a). “QTL Mapping Using a Memetic Algorithm with Modiﬁcations of BIC as Fitness Function.” Sta- tistical Applications in Genetics and Molecular Biology, 11: Art. 2. MR2944873. doi: https://doi.org/10.1515/1544-6115.1793. 296 Frommlet, F., Ruhaltinger, F., Twarog, P., and Bogdan, M. (2012b). “Modiﬁed versions of Bayesian Information Criterion for genome-wide association studies.” Computa- tional Statistics and Data Analysis, 56(5): 1038–1051. MR2897552. doi: https:// doi.org/10.1016/j.csda.2011.05.005. 299 H¨oﬂing, H. and Tibshirani, R. (2009). “Estimation of sparse binary pairwise Markov networks using pseudolikelihoods.” Journal of Machine Learning Research, 10: 883– 906. MR2505138. 295 Hubin, A. and Storvik, G. (2018). “Mode jumping MCMC for Bayesian variable selection in GLMM.” Computational Statistics and Data Analysis, 127: 281–297. MR3820324. doi: https://doi.org/10.1016/j.csda.2018.05.020. 297 Hubin, A., Storvik, G., and Frommlet, F. (2020). “A Novel Algorithmic Approach to Bayesian Logic Regression.” Bayesian Analysis. 296, 297 Kirkpatrick, S., Gelatt, C., and Vecchi, M. (1983). “Optimization by simulated anneal- ing.” Science, 220: 671–680. MR0702485. doi: https://doi.org/10.1126/science. 220.4598.671. 295 Malina, M., Ickstadt, K., Schwender, H., Posch, M., and Bogdan, M. (2014). “Detec- tion of epistatic eﬀects with logic regression and a classical linear regression model.” Statistical Applications in Genetics and Molecular Biology, 13: 83–104. MR3159119. doi: https://doi.org/10.1515/sagmb-2013-0028. 296 Miasojedow, B. and Rejchel, W. (2018). “Sparse Estimation in Ising Model via Penalized Monte Carlo Methods.” Journal of Machine Learning Research, 19(75): 1–26. URL http://jmlr.org/papers/v19/16-554.html MR3899777. 295 Ravikumar, P., Wainwright, M., and Laﬀerty, J. (2010). “High-dimensional Ising model selection using l1-regularized logistic regression.” The Annals of Statistics, 38: 1287– 1319. MR2662343. doi: https://doi.org/10.1214/09-AOS691. References Baierl, A., Bogdan, M., Frommlet, F., and Futschik, A. (2006). “On locating multiple interacting quantitative trait loci in intercross designs.” Genetics, 173: 1693–1703. 296 Banerjee, O., El Ghaoui, L., and d’Aspremont, A. (2008). “Model selection through sparse maximum likelihood estimation for multivariate Gaussian or binary data.” Journal of Machine Learning Research, 9: 485–516. MR2417243. 295 Bateson, W. and Mendel, G. (1909). Mendel’s principles of heredity. Cambridge Uni- versity Press: New York, G.P. Putnam’s Sons. 295 Bierkens, J., Fearnhead, P., and Roberts, G. (2019). “The Zig-Zag process and super- eﬃcient sampling for Bayesian analysis of big data.” Annals of Statistics, 47: 1288– 1320. MR3911113. doi: https://doi.org/10.1214/18-AOS1715. 297 Bogdan, M., Ghosh, J., and Doerge, R. W. (2004). “Modifying the Schwarz Bayesian Information Criterion to locate multiple interacting quantitative trait loci.” Genetics, 167: 989–999. 296 Bogdan, M., Frommlet, F., Biecek, P., Cheng, R., Ghosh, J., and Doerge, R. W. (2008a). “Extending the Modiﬁed Bayesian Information Criterion (mBIC) to dense markers and multiple interval mapping.” Biometrics, 64: 1162–1169. MR2522264. doi: https://doi.org/10.1111/j.1541-0420.2008.00989.x. 296 300 Invited Discussion Holger Schwender∗and Katja Ickstadt† Holger Schwender∗and Katja Ickstadt† We congratulate Hubin, Storvik, and Frommlet to this nice paper providing an approach to (Bayesian) logic regression that substantially diﬀers from the existing procedure for ﬁtting logic regression models. Their valuable work contributes a new view on how to generate such models, and, therefore, on how to model and identify interactions, in particular, in genetic association studies. Moreover, their work enhances the use of logic regression due to the algorithmic improvement by their GMJMCMC (Genetically modiﬁed Mode Jumping Markov Chain Monte Carlo) procedure. Since their methods diﬀer from the common approaches to logic regression, we will brieﬂy discuss in the following the original approach to logic regression that to some extent also underlies the other Bayesian logic regression procedures MCLR (Monte Carlo Logic Regression) and FBLR (Fully Bayesian Logic Regression) considered in the paper. We will also present another procedure called GPAS (Genetic Programming for Association Studies) that uses similar operators as GMJMCMC to move through the search space consisting of all possible models. Finally, we will mention how procedures based on logic regression can be employed to not just identify interactions associated with the outcome of interest, but also to rank these interactions by their importance and to guide statements on their relevance and signiﬁcance. ∗Mathematical Institute, Heinrich Heine University, D¨usseldorf, Germany, holger.schwender@hhu.de †Faculty of Statistics, TU Dortmund University, Dortmund, Germany, ickstadt@statistik.tu-dortmund.de Invited Discussion 295 M. Bogdan, B. Miasojedow, and J. Wallin 301 Ruczinski, I. (2000). “Logic Regression and Statistical Issues Related to the Protein Folding Problem.” Dissertation, Department of Statistics, University of Washington, Seattle, WA. MR2716929. 295 Ruczinski, I., Kooperberg, C., and LeBlanc, M. (2003). “Logic Regression.” Journal of Computational and Graphical Statistics, 12: 475–511. MR2002632. doi: https:// doi.org/10.1198/1061860032238. 295 Ruczinski, I., Kooperberg, C., and LeBlanc, M. (2004). “Exploring Interactions in High- Dimensional Genomic Data: An Overview of Logic Regression, with Applications.” Journal of Multivariate Analysis, 90: 178–195. MR2086341. doi: https://doi.org/ 10.1016/j.jmva.2004.02.010. 295 Schwender, H. and Ickstadt, K. (2008). “Identiﬁcation of snp interactions using logic regression.” Biostatistics, 9: 187–198. 295, 296 Invited Discussion 302 1 Fitting Logic Regression Models The starting point of the original logic regression developed by Ruczinski et al. (2003) is a regression model containing one logic expression consisting of one binary/logic variable as predictor. This model is then modiﬁed by either adding a new logic expression consisting of one binary variable as predictor to the model or by modifying the logic expression (or later, one of the logic expressions) already in the model by either changing one of the logic variables or logic operators in the logic expression or by adding or removing a logic variable to/from the model. This step is repeated until a score function assessing the ﬁt of the logic regression model converges. E.g., when considering a binary response, the binomial deviance serves as score. In both logic regression and GMJMCMC, tree-based structures are used to repre- sent the logic expressions in the regression models or populations, respectively, and to modify these logic expressions to move through the space of all possible models. While GMJMCMC borrows ideas from genetic algorithms for the modiﬁcation of the logic expressions, other, partly related moves directly embedded in the framework of logic trees are employed by logic regression for this purpose. In the nomenclature of logic 303 H. Schwender and K. Ickstadt regression, replacing a logic variable or operator is called “Alternating a leaf” or “Al- ternating an operator”, respectively. A logic variable can be added to a logic expression by “Splitting a leaf” or “Growing a branch” depending on the position in the logic tree at which this change should be made. Logic variables can also be removed from the models by the countermoves “Delete the leaf” and “Prune a branch” to the adding moves. Thus, a logic tree can be modiﬁed at any level of this tree and not just by adding a new level to the hierarchy of the tree. This is in contrast to CART (Classiﬁcation And Regression Trees, Breiman et al., 1984), the arguably most well-known tree-based classiﬁcation and regression method, in which the trees have a hierarchical structure. Because of their non-hierarchical structure, logic trees do not only provide a much more concise representation of logic expressions compared to CART trees, but also a ﬂexible framework to search for a logic regression model that best explains the considered response variable. 2 Genetic Programming for Association Studies Another procedure more closely related to GMJMCMC than logic regression is GPAS (Genetic Programming for Association Studies) proposed by Nunkesser et al. (2007). As implied by its name, GPAS employs genetic programming instead of genetic algorithms as search procedure. As in GMJMCMC, logic expressions are represented by trees and crossover, mutation as well as reduction operators are used to generate in each iteration of the search algorithm a new population of logic expressions that are then evaluated to remove dispensable logic expressions from the population. In GPAS, each logic expression is generated directly in disjunctive normal form, i.e. as OR-combination of AND-combinations of the logic variables, since the AND- combinations can (at least in a statistical sense) be interpreted as the interactions contained in the logic expression. Besides a crossover operator similar to the one of GMJMCMC, a logic variable that is part of a logic expression in the population of the current iteration of GPAS can be replaced by another logic variable. Also a new logic variable can be added either to an AND-combination or to the OR-combination as a start of a new AND-combination. As in logic regression, countermoves to these two operations are also part of the move set of GPAS. In each iteration of GPAS, two logic expressions are randomly selected to generate a new logic expression by performing a randomly chosen crossover operation on these two expressions. Moreover, ﬁve logic expressions – one expression for each of the described mutation operations – are chosen at random to apply a random modiﬁcation to each of these expressions. In this way, a new population consisting of all logic expressions from the population of the current iteration and six new logic expressions is generated. While in GMJMCMC a ﬁxed number of MJMCMC (Mode Jumping Markov Chain Monte Carlo) iterations is performed for each population of the genetic algorithm to compute marginal inclusion probabilities and to remove logic expressions with a marginal inclusion probability below some threshold from the population, GPAS em- ploys multi-objective optimization and domination selection in which logic expressions dominated by other logic expressions are removed from the population. Thus, multiple criteria are used to evaluate the performance of a logic expression. 1 Fitting Logic Regression Models Depending on the search algorithm, all choices in all steps of logic regression are either made randomly (when the stochastic search algorithm simulated annealing is employed) or by selecting the modiﬁcation of the currently considered logic regression model that leads to the largest improvement of the score (when a greedy search is used). As mentioned by Hubin et al., the standard search procedure in logic regression is simulated annealing. This stochastic search algorithm is based on Markov chains, where the probability of accepting a proposed new model is governed by a parameter called temperature that decreases the acceptance probability during the run of simulated annealing. As a result, many models are visited at the beginning of the search, but towards the end of the search it gets more and more unlikely that a modiﬁcation to the logic regression model gets accepted when the proposed model has a worse score than the current model. In LogicReg, the R package in which the original logic regression is implemented (Kooperberg and Ruczinski, 2019), also a greedy search is implemented that has the drawback that it is not capable to escape from a local minimum. However, this greedy search works also well when it is put into an ensemble framework such as bagging (Breiman, 1996; see also Section 3 of this discussion). As mentioned by Hubin et al., in particular in genetic association studies, there usually does not exist the one and only explanation for a response variable such as the disease status, but many competing models that ﬁt the data almost equally well. Hubin et al., therefore, correctly argue that generating just a single best model – as the original logic regression does – ignores the problem of model uncertainty. For this reason, they disregard the original logic regression in their following discussion and consider only Bayesian versions of logic regression enabling Bayesian model averaging by generating and considering many competing logic regression models. It might be, at ﬁrst sight, a drawback of the original logic regression that it generates just one model. However, as we will discuss in Section 3, it is straightforward to formulate an ensemble framework that can be employed to ﬁt a large number of logic regression models, and thus, to analyze uncertainty in the models. 304 Invited Discussion 3 Assessing the Importance of Identiﬁed Interactions After identifying potentially interesting interactions, i.e. interactions that potentially inﬂuence the outcome of interest and are important for a correct prediction of the outcome, with search procedures such as GMJMCMC, GPAS, or logic regression, the importance of these interactions for correctly predicting the outcome should be mea- sured to diﬀer between interactions associated with this outcome and interactions found almost at random. Such a measure can then also be used to rank the interactions by their importance, which is often of particular interest. Hubin et al. consider for this purpose the marginal inclusion probability ˜P Lj \| Y , i.e. the estimated probability for a logic expression Lj to be included in a model. This probability is, however, determined on the same data on which the models are built, which usually results in a positively biased estimate of the importance of Lj. It would, thus, be preferable to quantify the importance of a particular logic expression or inter- action based on new/independent data. One approach for such a quantiﬁcation on independent data that was originally de- veloped by Breiman (2001) for measuring the importance of variables in Random Forests is (in a modiﬁed version) also used in logicFS (logic Feature Selection; Schwender and Ickstadt, 2008). In logicFS, the original logic regression is applied to B bootstrap sam- ples drawn from the considered data set, resulting in B logic regression models. The logic expressions in each of these models are, afterwards, transformed into disjunctive normal forms to identify the interactions composing these models by the conjunctions/AND- combinations in these disjunctive normal forms. The importance of each of the inter- actions is then quantiﬁed by considering the respective out-of-bag observations, i.e. the observations that do not belong to the respective bootstrap sample and were, therefore, not used in the ﬁtting of the respective model. For the quantiﬁcation of the importances, the predictive power of each of the B logic regression models is determined, e.g., by the number of correctly out-of-bag observations. Afterwards, the interaction for which the importance should be computed is removed from all the logic regression models, and again, the predictive power of each of the now reduced models is determined. 3 Assessing the Importance of Identiﬁed Interactions The mean diﬀerence between the predictive powers of the models before and after an interaction has been removed from them can then be employed to quantify the improvement in the prediction due to this interaction and thus as a measure of the importance of this inter- action for a good prediction of the outcome. Using a permutation test, the importances of the interactions can also be tested (Schwender et al., 2011). Since logicFS also results in several logic regression models, a comparison of the performance of GMJMCMC and logicFS as well as their importance measures might be valuable. 2 Genetic Programming for Association Studies If this expression shows a worse value for one of these criteria than another logic expression in the popu- lation and not a better value for all the other criteria, then this expression is dominated by the other logic expression, and hence, removed from the population. Since GPAS has been developed for the association analysis of case-control data, the criteria considered in GPAS are the rate of correctly classiﬁed observations, the number of correctly pre- dicted controls (in simulations, it has turned out to be beneﬁcial to use the controls in two objectives), and the length of the logic expression. Because of the similarities in the operators considered in GPAS and GMJMCMC and the diﬀerences in the selection processes, it would be valuable to compare these two procedures, which might perhaps even enhance GMJMCMC. Furthermore, as indicated in FBLR, the representation of logic expressions in their disjunctive normal form could also be employed to formulate the full model prior P(M) for GMJMCMC. 305 H. Schwender and K. Ickstadt 4 Eﬀect Sizes in Genetic Association Studies It is much appreciated that Hubin et al. consider the same scenarios as Fritsch (2006). In the other scenarios, they, however, consider eﬀect sizes, in particular, for interactions of higher order from which they admit that they “might seem unrealistically large 306 Invited Discussion Invited Discussion compared to real applications.” In our opinion, this is a large understatement. E.g., in Scenario S.3, odds ratios of exp(5) ≈148 and exp(9) ≈8103 are, even for epidemiological risk factors such as smoking, implausibly vast. In genetic association studies, individual noteworthy SNPs (Single Nucleotide Poly- morphisms) seldomly show an odds ratio larger than 1.5 (see, e.g., Golka et al., 2011). Interactions of SNPs might have a substantially higher impact on diseases, but the eﬀect sizes do by far not reach the eﬀect sizes considered in the simulation study. E.g., Selinski et al. (2017) identiﬁed a combination of four SNPs showing an odds ratio of 2.59 in a subgroup of urinary bladder cancer patients (where the odds ratios of the individual SNPs ranged between 1.1 and 1.3). Even though the sensitivity analysis of Hubin et al. gives some insight in the per- formance of GMJMCMC for a range of eﬀect sizes and sample sizes, the eﬀect sizes in the simulated analysis in our opinion could have been chosen a bit more realistically. Considering these unrealistically large eﬀect sizes, however, does not really diminish this nice, well-thought-out paper, and in particular, not the proposed method. It is a very welcome contribution to the association analysis of interactions in genetic association studies and opens the ﬁeld for further research in this direction. Selinski, S., Blaszkewicz, M., Ickstadt, K., Gerullis, H., Otto, T., Roth, E., Volkert, F., Ovsiannikov, D., Moormann, O., Banﬁ, G., Nyirady, P., Vermeulen, S. H., Garcia- Closas, M., Figueroa, J. D., Johnson, A., Karagas, M. R., Kogevinas, M., Malats, N., Schwenn, M., Silverman, D. T., Koutros, S., Rothman, N., Kiemeney, L. A., Hengstler, J. G., and Golka, K. (2017). “Identiﬁcation and replication of the interplay of four genetic high-risk variants for urinary bladder cancer.” Carcinogenesis, 38: 1167–1179. 306 Schwender, H., Ruczinski, I., and Ickstadt, K. (2011). “Testing SNPs and sets of SNPs for importance in association studies.” Biostatistics, 12: 18–32. 305 ∗CEREMADE, Univerist´e Paris Dauphine, clarte@ceremade.dauphine.fr †CEREMADE, Univerist´e Paris Dauphine, xian@ceremade.dauphine.fr Contributed Discussion Gr´egoire Clart´e∗and Christian P. Robert† Gr´egoire Clart´e∗and Christian P. Robert† While logic regression is not to be confused with logistic regression, the distinction may be proved more delicate than stated. For one thing, as the central object of interest is a generalised linear model (or rather a family of such models) based on a vector of binary covariates, it covers the special case of logistic regression. For another thing, it does not very clearly slit from a standard generalised linear model—or generalised analysis of variance model—when all covariates are dummy variables. Culling the number of total covariates (trees) away from the exponential of exponential number of possible covariates deﬁned by logical combinations appears to be a signiﬁcant component of the approach but this selection of potential (sub-)models remains obscure. If this primary selection is to be data-dependent, there could be a connection with variable length Markov chain models (B¨uhlmann and Wyner, 1999). Cont Contributed Discussion Gr´egoire Clart´e∗and Christian P. Robert† 308 Contributed Discussion References Breiman, L. (1996). “Bagging Predictors.” Machine Learning, 26: 123–140. 303 Breiman, L. (2001). “Random Forests.” Machine Learning, 45: 5–32. MR3874153. 305 Breiman, L., Friedman, J. H., Olshen, R. A., and Stone, C. J. (1984). Classiﬁcation and Regression Trees. Wadsworth, Belmont, CA. MR0726392. 303 Fritsch, A. (2006). “A Full Bayesian Version of Logic Regression for SNP Data.” Master’s thesis, Faculty of Statistics, TU Dortmund University. 305 Golka, K., Selinski, S., Lehmann, M. L., Blaszkewicz, M., Marchan, R., Ickstadt, K., Schwender, H., Bolt, H. M., and Hengstler, J. G. (2011). “Genetic variants in urinary bladder cancer: collective power of “wimp SNPs”.” Archives of Toxicology, 85: 539– 554. 306 Kooperberg, C. and Ruczinski, I. (2019). LogicReg: Logic Regression. R package version 1.6.1. URL https://CRAN.R-project.org/package=LogicReg. 303 Nunkesser, R., Bernholt, T., Schwender, H., Ickstadt, K., and Wegener, I. (2007). “De- tecting High-Order Interactions of Single Nucleotide Polymorphisms Using Genetic Programming.” Bioinformatics, 23: 3280–3288. 304 Ruczinski, I., Kooperberg, C., and LeBlanc, M. (2003). “Logic Regression.” Journal of Computational and Graphical Statistics, 12: 475–511. MR2002632. doi: https:// doi.org/10.1198/1061860032238. 302 Schwender, H. and Ickstadt, K. (2008). “Identiﬁcation of SNP Interactions Using Logic Regression.” Biostatistics, 9: 187–198. 305 H. Schwender and K. Ickstadt 307 Schwender, H., Ruczinski, I., and Ickstadt, K. (2011). “Testing SNPs and sets of SNPs for importance in association studies.” Biostatistics, 12: 18–32. 305 1 Prior Issues With respect to the prior modeling adopted in the paper, it mostly relies on a rather standard decomposition in variable indicators—to signify whether or not some trees are included in the regression (and hence the model)—. The prior modelling on these indicators is purely a complexity penalisation in that it is only function of the number of active trees, hence not accounting for a possible speciﬁcity of some covariates, as for instance when dealing with imbalanced binary covariates (many more 1’s than 0’s, say). “. . . using Jeﬀreys’ prior for model selection has been widely criticized for not being con- sistent once the true model coincides with the null model.” A central issue with the prior modelling adopted in the paper is its loose handling of improper priors. It is well-known that the use of improper priors is debatable for model choice settings and hence that they should be best avoided altogether, to wit the Lindley-Jeﬀreys paradox (Lindley, 1957; DeGroot, 1973, 1982). Let us ﬁrst recall that Jeﬀreys (1939) distinguishes between estimation and testing reference priors (Bayarri and Garcia-Donato, 2007; Robert et al., 2009). Not only does the paper adopt the notion of a same, improper, prior on the GLM scale parameter, which is a position advocated in some part of the Bayesian literature (Berger et al., 1998), but it also seems to be using an improper prior on each set of model parameters (further undiﬀerentiated between models). Because the priors operate on diﬀerent (sub)sets of parameters, we wonder whether or not this jeopardises the later discourse on the posterior probabilities of the diﬀerent models, since such probabilities are not meaningful from a probabilistic viewpoint. Such a prior construct indeed implies there is no joint distribution and no marginal density. In some cases, it may even be that p(y\|M) becomes inﬁnite. Referring G. Clart´e and C. P. Robert 309 to a “simple Jeﬀreys”’ prior in this setting is therefore anything but simple as Jeﬀreys (1939) himself shied away from using improper priors on the parameter of interest. We therefore ﬁnd it surprising that this fundamental and well-known diﬃculty with improper priors in hypothesis testing is not even alluded to in the paper, the above quote being a much milder criticism, core setting thus seems to be ﬂawed. 1 Prior Issues Now, the numerical comparisons run in the paper between Jeﬀreys’ prior and a regular g-prior exhibit close numerical proximity and we wonder at the reason if the Bayes factor is deﬁned up to an arbitrary constant. Could it be that the culling and selection processes end up having a similar number of covariates and hence ignore the overall impact of the prior? Or is it rather a consequence of recoursing to a Laplace approximation of the marginal likelihood since it completely escapes the problem lack of deﬁnition of the said marginal? Contributed Discussion Contributed Discussion Contributed Discussion GMJMCMC starts from this idea to develop a more complex algorithm in which the previous exploration technique is used inside a subset of models which is then updated. We however wonder whether or not the algorithm is not wasting time overall by exploring some parts of an already explored section of the space. More generally, it seems to us that the genetic layer in the algorithm has solely been added to constrain the exploration to smaller spaces, hence are wondering of the eﬃciency gain brought by this addition. This method may in the end suﬀer from several ﬂaws. First, it does not provide the theoretical security of (asymptotic) unbiasedness that is attained with MCMC method. However, it could be of interest to study the variance of such estimators, as Markov chains with poor mixing properties can have huge variance in a multi-modal context. For example, assuming the function f is primarily supported by points outside of Ω∗, it is clear that the estimation is ineﬃcient; however, an MCMC algorithm will similarly be ineﬃcient in the sense that low probability states will also be underexplored, leading to a massive variance estimator. In our opinion, the main issue is to ensure that Ω∗is well- chosen, that is to say, that it contains the right amount of points. Several parameters account for this in the algorithm, choosing these parameters may be a tough calibration problem—especially the choice of the cutoﬀparameter ρmin, even though it could be chosen as a “quantile of posterior probability” or be adaptive. 2 Algorithmic Aspects Methinks the proposed strategy is fruitful in a discrete space; we agree with the au- thors that contrary to Metropolis-Hastings-like methods, it does not involve repeated computations of the same quantity (which can be expensive, especially when involv- ing marginal likelihoods). However, even a limited number of computations of these marginal likelihoods may constitute a real challenge, while the solutions mentioned in the paper are not necessarily the most eﬃcient (Geyer, 1993; Gutmann and Hyv¨arinen, 2012). “. . . we do not need a proper MCMC (an algorithm with convergence towards the target distribution) which is needed if model posterior probabilities are estimated by the relative frequency of how often a model has been visited.” While we have not read the referred article on MJMCMC in detail, a ﬁrst comment is that the name itself is somewhat unsuitable, as indeed the algorithm does not sample from a distribution but only explores its surface. There is no proper sampling part in the algorithm, as quantities are computed over Ω∗with integrals of the form x∈Ω∗ f(x)π(x)/Z∗, x∈Ω∗ f(x)π(x)/Z∗, where π the target and Z∗= x∈Ω∗π(x) is an approximation of the normalizing constant. Finding such a set Ω∗is the main goal of the method developed in the paper. “. . . hereby, all states, including all possible models of maximum sized, will eventually be visited.” In a self-avoiding mode, keeping track of all the previous states visited by the chain ensures that those states will never be visited again. As we are in a discrete setting, this implies that once a mode has been visited the algorithm is constrained to eventually visit another mode, even if the potential between the modes is almost zero. The set Ω∗is then built sequentially, removing states with too low posterior probability to add more interesting states which neighbours have been recently visited. 310 Robert, C., Chopin, N., and Rousseau, J. (2009). “Theory of Probability revisited (with discussion).” Statistical Science, 24(2): 141–172 and 191–194. MR2655841. doi: https://doi.org/10.1214/09-STS284. 308 Lindley, D. (1957). “A statistical paradox.” Biometrika, 44: 187–192. MR0087273. doi: https://doi.org/10.1093/biomet/44.1-2.179. 308 Jeﬀreys, H. (1939). Theory of Probability. Oxford: The Clarendon Press, ﬁrst edition. MR0187257. 308, 309 ∗Norwegian Computing Center, aliaksandr.hubin@nr.no †Department of Mathematics, University of Oslo, geirs@math.uio.no ‡Department of Medical Statistics (CEMSIIS), Medical University of Vienna, ﬂorian.frommlet@meduniwien.ac.at References Bayarri, M. and Garcia-Donato, G. (2007). “Extending conventional priors for test- ing general hypotheses in linear models.” Biometrika, 94: 135–152. MR2367828. doi: https://doi.org/10.1093/biomet/asm014. 308 Berger, J., Pericchi, L., and Varshavsky, J. (1998). “Bayes factors and marginal distri- butions in invariant situations.” Sankhya A, 60: 307–321. MR1718789. 308 B¨uhlmann, P. and Wyner, A. J. (1999). “Variable length Markov chains.” The An- nals of Statistics, 27(2): 480–513. MR1714720. doi: https://doi.org/10.1214/aos/ 1018031204. 308 DeGroot, M. (1973). “Doing what comes naturally: Interpreting a tail area as a posterior probability or as a likelihood ratio.” Journal of the American Statistical Association, 68: 966–969. MR0362639. 308 DeGroot, M. (1982). “Discussion of Shafer’s ‘Lindley’s paradox’.” Journal of the Amer- ican Statistical Association, 378: 337–339. MR0664677. 308 Geyer, C. (1993). “Estimating normalizing constants and reweighting mixtures in Markov chain Monte Carlo.” Technical Report 568, School of Statistics, Univ. of Minnesota. MR1341319. doi: https://doi.org/10.2307/1390763. 309 Gutmann, M. U. and Hyv¨arinen, A. (2012). “Noise-Contrastive Estimation of Unnor- malized Statistical Models, with Applications to Natural Image Statistics.” Journal of Machine Learning Research, 13(1): 307–361. MR2913702. 309 G. Clart´e and C. P. Robert 311 Jeﬀreys, H. (1939). Theory of Probability. Oxford: The Clarendon Press, ﬁrst edition. MR0187257. 308, 309 Lindley, D. (1957). “A statistical paradox.” Biometrika, 44: 187–192. MR0087273. doi: https://doi.org/10.1093/biomet/44.1-2.179. 308 Robert, C., Chopin, N., and Rousseau, J. (2009). “Theory of Probability revisited (with discussion).” Statistical Science, 24(2): 141–172 and 191–194. MR2655841. doi: https://doi.org/10.1214/09-STS284. 308 Rejoinder 312 Rejoinder Aliaksandr Hubin∗, Geir Storvik†, and Florian Frommlet‡ 2 Applications of Bayesian logic regression We very much appreciate that Ruczinski et al. (2020) have pointed out important ap- plications of logic regression outside of genetics. Our emphasis on genetic applications was not meant to indicate limitations of the usefulness of logic regression in other areas but rather reﬂects our own previous research interests. Also, the applications mentioned by Bogdan et al. (2020) in the context of multicolored graphical models sound quite interesting. We are however more sceptical whether logic regression, in whichever form, will ever be applicable directly to association studies of outbred populations, where the number of genetic variants is much larger than in controlled populations. For large numbers of binary covariates, already the number of pairwise logic expressions becomes prohibitively large to apply logic regression, both in terms of algorithmic feasibility and in terms of having suﬃcient power while controlling type I error. Realistic applications of logic regression (with the aim of identifying true logic expressions) will most likely be restricted to applications with a few hundred binary covariates unless technologic ad- vances allow one day to eﬃciently resolve the NP hard combinatorial problem of model search. However, one might consider bagging and boosting to obtain scalable versions of logic regression for prediction. In this rejoinder, we also discuss extensions of Bayesian logic regression, allowing for non-binary predictors and latent Gaussian variables to be included into the model. This could further extend the applications of Bayesian logic regression methodology to such ﬁelds as epidemiology, spatio-temporal statistics, environmetrics and econometrics. For example, in Hubin et al. (2020c), a model with latent Gaussian processes (where a subset of predictors are binary) was used for the analysis of DNA methylation. The paper discusses the potential of using logic expressions of the binary predictors as a direction for further research. With the extensions of BLR provided in this rejoinder, it would become feasible to perform logic regression in the settings of Hubin et al. (2020c). Both Ruczinski et al. (2020) and Bogdan et al. (2020) commented on the lack of correlated regressors in our simulation studies. This was mainly due to the fact that for the sake of comparison we wanted to use the scenarios from Fritsch (2006). For the more complex scenarios, we simply extended these scenarios by adding logic expres- sions of a higher order. 1 Introduction We have structured the rejoinder according to diﬀerent topics while trying to address all 313 A. Hubin, G. Storvik, and F. Frommlet the points raised by the discussants. We also provide several interesting extensions of the model. Finally, we give a brief tutorial on the relevant part of our R-package EMJMCMC http://aliaksah.github.io/EMJMCMC2016/ dealing with BLR. This should facilitate the practical application of the methodology developed in Hubin et al. (2020a). 1 Introduction We would like to begin this rejoinder with expressing our sincere gratitude to all of the discussants for their interesting and thought-provoking comments and remarks. We also feel heartily thankful to the editorial board of Bayesian Analysis for giving us the opportunity to publish our paper entitled “A novel algorithmic approach to Bayesian logic regression” (Hubin et al., 2020a) as a discussion article. Logic regres- sion is a tool to model non-linear relationships between binary covariates and some response variable by constructing predictors as Boolean combinations. The number of possible logic expressions grows exponentially with the number of binary variables in- volved, making the model search signiﬁcantly harder with the increasing complexity of Boolean combinations. Due to Boolean equivalence, it is in fact almost impossible to specify the full model space a priori even for a relatively small number of covari- ates. Our primary goal is to identify those logic expressions which are associated with the response variable. To this end, we want to estimate posterior probabilities of logic expressions within the framework of generalized linear models. The major contributions of our paper are two-fold: Firstly, we have introduced novel model priors for Bayesian logic regression (BLR), which yield good power to detect important logic expressions while controlling the number of false positive discoveries. Secondly, we have introduced a novel genetically modiﬁed mode jumping Markov chain Monte Carlo (GMJMCMC) algorithm to eﬃciently explore the space of logic regression models. The main idea of GMJMCMC is to embed the mode jumping Markov chain Monte Carlo (MJMCMC) algorithm (Hubin and Storvik, 2018) into the iterative setting of a genetic algorithm. Populations for the genetic algorithm consist of relatively small sets of logic expressions. Any such subset forms a well deﬁned model space which allows to run MJMCMC. The population is then regularly updated in such a way that the algo- rithm is guaranteed to be irreducible in the model space of all logic regression models. This is required for asymptotic unbiasedness of the estimated posterior probabilities, as we will discuss in more detail below. Although GMJMCMC is not a proper MCMC algorithm (in the sense that its stationary distribution does not coincide with the target distribution of interest), renormalized estimates of the posterior probabilities are readily available. The discussants have pointed out several interesting extensions and open problems. 2 Applications of Bayesian logic regression In our sensitivity analysis, though, we considered one scenario with correlations (but only for one true leaf), where reasonably good results were ob- tained when the correlation of a mis-speciﬁed leaf r was being varied from 0.1 to 1. However, speciﬁcally to respond to the remarks from Bogdan et al. (2020), who hypoth- esised that our approach would only work under independence, we will provide some additional simulation results, where we consider regressors with diﬀerent correlation structures. 314 Rejoinder Figure 1: Correlation structure of the simulated covariates with a general correlation structure (left) and from QTL back-cross (right). 314 Rejoinder Rejoinder Figure 1: Correlation structure of the simulated covariates with a general correlation structure (left) and from QTL back-cross (right). Response variables Y are simulated from the data generating model of Scenario 5 from Hubin et al. (2020a), where Y ∼N(μ, 1), with 2.1 Simulation study with correlated regressors Frommlet Figure 2: Genetic map of markers on ﬁve chromosomes of diﬀerent length (given in centiMorgan). For the second scenario of our simulation study, these marker positions are used to simulate genotype data from a back-cross design. The closer markers on the same chromosome are lying the stronger will be the correlation of the corresponding genotype data. Figure 2: Genetic map of markers on ﬁve chromosomes of diﬀerent length (given in centiMorgan). For the second scenario of our simulation study, these marker positions are used to simulate genotype data from a back-cross design. The closer markers on the same chromosome are lying the stronger will be the correlation of the corresponding genotype data. The exact deﬁnition of the trees L1–L4 is given in Table 1 below and is equivalent to the deﬁnition in Table 2 of our original article. For the QTL mapping scenario, the responses were generated for each simulation replicate after randomly permuting the order of the genetic markers. In this way, we considered diﬀerent patterns of correlations between the leaves of the data generating model. For both correlation structures, we generated N = 100 datasets with n = 1000 observations. Every data set was analysed with the Jeﬀreys’ prior and with the robust g-prior using GMJMCMC with the same tuning parameters as in Scenario 5 of the original article. We appreciate the comment of Schwender and Ickstadt (2020) that for the higher order interactions L3 and L4 the eﬀect sizes are unrealistically large. However, as il- lustrated by our sensitivity analysis, if one wants to have suﬃcient power do detect more complex logic expressions with realistic eﬀect sizes then one will need a much larger sample size. This would be potentially feasible for real data analysis (by means of simply collecting more observations) but not for a simulation study with hundreds of simulation runs. In any case, our goal here is to show that correlated regressors are not an impediment for our approach. Table 1 summarizes the results of our simulations with correlated regressors. For the ﬁrst scenario with the general structure, correlations are ranging between 0 and 0.5 in absolute values. Comparing the results with Table 2 from the original manuscript which was based on independent regressors the diﬀerences are relatively small. 2.1 Simulation study with correlated regressors In this study, we simulate the data using p = 50 regressors with two diﬀerent types of correlation structure: The ﬁrst one is rather general and uses fairly weak correlations, whereas the second one is typical for QTL mapping and gives very strong correlations. For the ﬁrst scenario, we consider covariates which are marginally distributed according to Xj ∼Bernoulli(0.5), j ∈{1, . . . , 50}. The correlation matrix is obtained using the approach from Joe (2006), which allows to generate positive deﬁnite matrices where all pairwise correlations are i.i.d. from a Beta distribution B(a, a) linearly transformed to the interval (-1, 1). The parameter of the Beta distribution equals a = alphad+(p−2)/2, where alphad > 0 can be chosen. In our case, for p = 50 and alphad = 5/2 it holds that the pairwise correlation lies between −0.2 and 0.2 with probability 0.85 and between −0.3 and 0.3 with probability 0.97. Correlations with an absolute value larger than 0.4 are extremely unlikely. Multivariate binary random variables Xj, j ∈{1, . . . , p} with such correlation structures are then simulated by thresholding normal distributions as described by Leisch et al. (1998). A typical correlation structure of covariates generated by this approach is shown in the left panel of Figure 1. The second scenario is based on the classical back-cross design for QTL mapping. We used the R/QTL package (Broman et al., 2003) to generate a map of 5 chromo- somes of diﬀerent lengths ranging from 100 cM to 40 cM with 10 equidistant markers per chromosome, see Figure 2. For experimental populations, there is a direct relation- ship between the genetic distance between markers on the same chromosome and their correlation as described in any textbook on QTL mapping (Chen, 2016). The corre- sponding correlation structure from simulated genotypes of n = 1000 individuals from a back-cross design is illustrated by the heatmap in the right panel of Figure 1. One can see that the correlations between markers on the same chromosome are very strong, getting close to 0.9 for neighbouring markers. Response variables Y are simulated from the data generating model of Scenario 5 from Hubin et al. (2020a), where Y ∼N(μ, 1), with (1) μ = 1 + 1.5L1 + 3.5L2 + 9L3 + 7L4. (1) μ = 1 + 1.5L1 + 3.5L2 + 9L3 + 7L4. 315 A. Hubin, G. Storvik, and F. 2.1 Simulation study with correlated regressors Only for L4, there is a decrease in power, for Jeﬀreys’ prior from 0.89 to 0.66 and for the robust g-prior from 0.9 to 0.66. On the other hand, the number of false positives increases for Jeﬀreys’ prior from 37 to 78 and for the robust g-prior from 28 to 73. It has to be expected that the performance of logic regression becomes a little worse under correlation but GMJMCMC is still behaving very well for the scenario with a general correlation structure. Jeﬀreys’ prior and robust g-prior perform almost equally well with only a very slight advantage of the latter. Rejoinder General QTL Jef. R. g Jef. R. g L1 = X37 1.00 1.00 0.83 0.85 L2 = X2 ∧X9 0.98 0.99 0.82 0.81 L3 = X7 ∧X12 ∧X20 0.96 0.99 0.92 0.92 L4 = X4 ∧X10 ∧X17 ∧X30 0.66 0.66 0.20 0.24 Overall Power 0.90 0.91 0.69 0.71 FP 0.78 0.73 2.02 2.01 FDR 0.13 0.13 0.39 0.38 WL 9 6 108 98 316 Table 1: Results for the additional simulation scenarios with correlated binary covariates. Power for individual trees, overall power, the expected number of false positives (FP) and FDR are compared for GMJMCMC using either Jeﬀreys’ prior or the robust g- prior under the general correlation structure and the correlation structure from QTL mapping with back-cross design. The results in Table 1 for our second correlation structure from QTL mapping are based on the strict deﬁnition that only discoveries of trees from the data-generating model itself are counted as true positives. While there is some loss of power, the results for the ﬁrst three logic expressions are still quite satisfactory. Only for L4, the estimated power becomes unacceptably low. At the same time, the number of false positives, as well as the number of wrongly detected leaves, increases substantially. For QTL mapping, the correlation between neighboring markers often is so strong, that it becomes extremely diﬃcult to distinguish between them. For that reason, in simulation studies for QTL- mapping, one often takes the approach that the detection of a marker strongly correlated with a QTL is still counted as a true positive. If we take such an approach and consider markers within a range of 15 cM as correct representatives of a leaf from the data generating model then we get slightly better results. 2.1 Simulation study with correlated regressors In particular, the number of wrong leaves goes down from 108 to 50 for Jeﬀreys’ prior and from 98 to 58 for the robust g- prior. Extending the window for deﬁning true positives would further reduce the number of wrongly detected leaves. 3 Prior related aspects Clarte and Robert (2020) criticize several aspects of our choice of priors. We fully agree that the use of improper priors is debatable and should be done with great care. We had stated this explicitly already in the original article. From a theoretical point of view, our preference would be mixtures of g-priors. As a representative, the robust g-prior is implemented within our package. However, given the strong popularity of the BIC criterion, we wanted to study the performance of this choice as well. Our description of BIC as an approximation of the marginal likelihood under Jeﬀreys’ prior could indeed have included a discussion of its weak points. As Clarte and Robert (2020) remark, the good performance of the BIC choice is most likely connected with applying the Laplace 317 A. Hubin, G. Storvik, and F. Frommlet approximation of the marginal likelihood. However, in the case of the Gaussian linear model, the approximation is exact. approximation of the marginal likelihood. However, in the case of the Gaussian linear model, the approximation is exact. The main empirical point, though, is that in all our examples from the original manuscript, the BIC measure as an approximation for the marginal density performed better than the analytical expression under the robust g-prior, both in terms of eval- uation metrics and speed. Under the correlated designs provided in this rejoinder, the robust g-prior slightly outperforms Jeﬀreys’ prior in terms of evaluation metrics but the BIC choice still performs rather well. Moreover, the running time of GMJMCMC under Jeﬀreys’ prior (having all of the tuning parameters of the algorithm ﬁxed) is still signiﬁcantly shorter. Note that the main contributions of our approach are: a) introducing novel model priors and b) a new search algorithm, whilst for the choice of the parameter priors and the calculation of the marginal densities we are using already established procedures. For example, our approach is fully compatible with integrated nested Laplace approx- imations (INLA) (Rue et al., 2009) and all of the parameter priors available there can be used. More generally, the R-package we have developed allows the users to easily specify their own method of calculating the marginal likelihood (whatever they prefer and/or what is available for their speciﬁc model: analytical integration, Monte Carlo based approximation, or other approximations) for their own choice of parameter priors. This ﬂexibility allows extending the method easily to broader classes of logic regression models. 3 Prior related aspects In Section 5, for instance, we describe an extension to latent Gaussian models with both logic and non-logic covariates, where alternative types of parameter priors are possible and the marginal likelihood is computed via integrated nested Laplace approximations (INLA) (Rue et al., 2009). Also note, that in both Bayarri et al. (2012) and Li and Clyde (2018), priors on models are indirectly obtained through priors on the regression parameters. In our approach, we include speciﬁc priors on model complexity as well. This is done via equations (3) and (4) in the main paper. The theoretical properties of combining model and parameter priors deﬁnitely require further distinguished research, which, we feel, lies slightly outside the scope of this rejoinder. Rejoinder whether the covariates have any association with the response variable at all and one only then applies logic regression. This is, of course, a viable approach which can be easily adopted. In practice, this could save resources by avoiding to run computationally costly inference on BLR. Whilst Bogdan et al. (2020) are rather reserved with respect to the proposed algo- rithm, we believe that most of their concerns are actually based on misunderstandings of the algorithm and we are glad to have the opportunity to clarify some of these points. The question of correlated regressors has been addressed in Section 2.1 of this rejoinder, where we have seen that GMJMCMC works reasonably well even when regressors are heavily dependent. Furthermore, Bogdan et al. (2020) were wondering about a lack of treatment for tautologies. This can be easily addressed because, in fact, our implemen- tation of GMJMCMC is taking care of Boolean equivalence already when generating new trees. In particular, as we discuss in Section 2.3 of the paper, “for all three op- erators it holds that if the newly generated tree is already present in St then it is not considered for St+1 but rather a new replacement tree is proposed instead.” What we do in practice is to check whether newly generated trees have correlation ±1 with any tree within St, which for suﬃciently large sample size will correspond to logic equivalence. Consequently, tautologies within a GMJMCMC chain are simply not allowed. Bogdan et al. (2020) also wonder why the sum in the denominator of (0.2) contains only Mfin = 10000 models based on d trees from the ﬁnal stage of the algorithm. This is indeed one of the implemented options in our package (though Mfin does not have to be 10000). The reason for this choice is to avoid having either too large and/or too densely ﬁlled hash tables (as a data structure), both of which become quite slow to handle. Whilst this introduces some undesired limitations, it remains an important pragmatic decision to make. The number of logic trees grows exponentially with the number of leaves involved and the number of models grows exponentially with the number of logic trees. Hence, even for the small examples with p = 50, the size of a hash table including all visited models and their statistics can become prohibitively large to be used in practice. 4 Algorithmic aspects Given that one of the main contributions of this manuscript was the development of the GMJMCMC algorithm, it is no surprise that many comments of the discussants were concerned with the algorithm. We will start with replying to some questions which are simple to answer, then give a more detailed recap of the MJMCMC algorithm (Hubin and Storvik, 2018) and ﬁnally discuss some questions on the parameter settings of GMJMCMC. Ruczinski et al. (2020) wondered whether covariates which are not logic can be easily combined with Boolean combinations in the model. The answer is yes. We will discuss this extension in Section 5.2 and provide an example in the tutorial in Section 6.3. Ruczinski et al. (2020) also suggested a simple two-stage approach where one ﬁrst checks 318 Rejoinder (2020) promote using synchronization between the cores via priority queues. Whilst we ﬁnd the idea interesting, we are a little sceptical whether it would actually work. When compared to embarrassing parallelization, syn- chronization between the processes in practice often slows down the inference instead of speeding it up (Chai and Bose, 1993; Kukanov, 2008). There, of course, a lot depends on the back-end used for implementation. We currently do not have the capacity to try this approach ourselves, but we would like, by all means, to encourage Bogdan et al. (2020) or other future researchers of BLR to test this idea. We would be very happy if using synchronization via priority queues could lead to an objectively better and faster inference algorithm for BLR than GMJMCMC. Rejoinder That would be even more acute for larger p’s. As an alternative, one could use the best NH models from all T generations, where NH is ﬁnite and of reasonable size. But in this case, when the hash table is ﬁlled, the worst models must be deleted to allow new ones to be included. In practice, this strategy would become extremely slow. One has to read from, write to and delete from the almost full hash table, which will be also very large. One would either have to create some novel hashing/dehashing functions which make this approach eﬃcient or devise an alternative data structure which is especially designed for the problem at hand. Given the complexity of enumerating logic expressions due to logic equivalence and due to the super-exponential growth of the number of models with respect to the number of leaves involved, we would expect this to be a ground breaking task in the ﬁeld of algorithms and data structures. Bogdan et al. (2020) raised the question of why we need MJMCMC for the ﬁnal population of GMJMCMC when for d = 15 full enumeration is feasible. The simple answer is that for many applications one needs much larger d to obtain reliable results, see for example the remark after Theorem 1 of Hubin et al. (2020a) and also Figure 1, panel 3, from the sensitivity analysis of Hubin et al. (2020a). For larger d, a full enu- meration will no longer be possible, whilst we would like to oﬀer a generally functioning 319 A. Hubin, G. Storvik, and F. Frommlet algorithm. Bogdan et al. (2020) additionally say: “Also, a huge random reduction of the ﬁnal model space leads to substantially diﬀerent results for diﬀerent parallel runs of the algorithm. Therefore the authors aggregate results from diﬀerent runs using a weighting scheme speciﬁed in equation (15) of their paper. In our opinion it seems more reasonable to estimate the posterior probabilities of diﬀerent models simply by including all models visited in diﬀerent runs in denominator of (0.2).” We agree that in principle this is a reasonable approach, which we, in fact, suggested in Section 2.3 of our paper. There, however, we also discussed the drawback that this approach is computationally more costly because one has to transfer a large amount of information from diﬀerent models between the cores. Finally, Bogdan et al. Rejoinder marginal likelihoods p(Y \|M) are available for a given M, and then use MJMCMC to explore p(M\|Y ). By Bayes formula p(M\|Y ) = p(Y \|M)p(M) M ′∈Ω p(Y \|M ′)p(M ′). (2) (2) In order to calculate p(M\|Y ) we have to iterate through the whole model space Ω, which becomes computationally infeasible for large p. The ordinary Monte Carlo estimate is based on a number of MJMCMC samples M (i), i = 1, . . . , W: In order to calculate p(M\|Y ) we have to iterate through the whole model space Ω, which becomes computationally infeasible for large p. The ordinary Monte Carlo estimate is based on a number of MJMCMC samples M (i), i = 1, . . . , W: p(M\|Y ) = 1 W W i=1 I(M (i) = M) d −−−−→ W →∞p(M\|Y ), (3) (3) where I(·) is the indicator function. An alternative named the renormalized model (RM) estimate by Clyde et al. (2011), is where I(·) is the indicator function. An alternative named the renormalized model (RM) estimate by Clyde et al. (2011), is p(M\|Y ) = p(Y \|M)p(M) M ′∈V p(Y \|M ′)p(M ′)I(M ∈V), (4) (4) where now V is the set of all models visited at least once during the MJMCMC run. Assuming the Markov chain eventually will visit all possible models, also p(M\|Y ) will converge to p(M\|Y ). Note that this estimate also can utilize all models that are visited, not only those that have been accepted. This answers the comment of Bogdan et al. (2020), who presumed that we include only models accepted by MJMCMC into V. Although both (4) and (3) are asymptotically consistent, (4) will often be the preferable estimator since the convergence of the MCMC based approximation (3) is typically much slower, see Clyde et al. (2011). We now describe the MJMCMC algorithm in more detail. We aim at approximating p(M\|Y ) by means of searching for some subspace V of Ω which makes the approxima- tion (4) as precise as possible. Models with high values of p(Y \|M) are important to be included. This means that modes and near modal values of marginal likelihoods are particularly important for the construction of V ⊂Ω and missing them can dramatically inﬂuence our estimates. Note that these considerations are equally important for the standard MCMC estimate (3). 4.1 Mode jumping Markov Chain Monte Carlo Both Bogdan et al. (2020) and Clarte and Robert (2020) seem to be slightly confused with respect to the MJMCMC algorithm (Hubin and Storvik, 2018), which we did not describe in detail in Hubin et al. (2020a). We thus brieﬂy discuss the main ideas of MJMCMC to clarify certain misunderstandings. In Hubin and Storvik (2018), a proper MCMC algorithm for the search through a ﬁxed limited model space was proposed. The algorithm deals with the multimodality in the space of models through mode jumping proposals. The mode jumping MCMC (MJMCMC) algorithm relies upon the idea of making smart moves between local ex- trema with a reasonable frequency. Local MCMC is performed in the absolute majority of steps. For the rest, a large move in the model space (which is likely to hit a model with very low posterior probability) is made, followed by local optimization. The goal of the latter step is to reach a local optimum in a diﬀerent part of the model space. Then the proposal is randomized around this optimum and the transition to the proposed model is either accepted or rejected according to a Metropolis-Hastings acceptance probability. The convergence properties of the suggested Markov chain is proven through a reﬁne- ment of the results of Tjelmeland and Hegstad (2001). Its limiting distribution is shown to correspond to the marginal model posterior probabilities. Further extensions of the algorithm allowing for parallel computing and using mixtures of proposals were also suggested. MJMCMC is described in more detail in Algorithm 1 below, where we consider M = (γ1, . . . γp) to be associated with models in the given discrete model space Ω (here γj ∈{0, 1} indicates whether covariate xj is included in the model). We assume that 320 Algorithm 1 Mode jumping MCMC. 1: Generate a large jump M ∗ 0 according to a proposal distribution ql(M ∗ 0 \|M). 0 2: Perform a local optimization, deﬁned through M ∗ k ∼qo(M ∗ k\|M ∗ 0 ) 0 2: Perform a local optimization, deﬁned through M ∗ k ∼qo(M ∗ k\|M ∗ 0 ). p , g k q ( k\| 0 ) 3: Perform a small randomization to generate the proposal M ∗∼qr(M p , g k q ( k\| 0 ) 3: Perform a small randomization to generate the proposal M ∗∼qr(M ∗\|M ∗ k). 3: Perform a small randomization to generate the proposal M ∗∼qr(M ∗\|M ∗ k). 4: Generate backwards auxiliary variables M0 ∼ql(M0\|M ∗), Mk ∼qo(Mk\|M0). 5: Put 3: Perform a small randomization to generate the proposal M qr(M \|Mk). 4: Generate backwards auxiliary variables M0 ∼ql(M0\|M ∗), Mk ∼qo(Mk\|M0). 5: Put 4: Generate backwards auxiliary variables M0 ∼ql(M0\|M ∗), Mk ∼qo(Mk\|M0). 5: Put M ′ = M ∗ with probability rmh(M, M ∗; Mk, M ∗ k); M otherwise, where r∗ mh(M, M ∗; Mk, M ∗ k) = min 1, p(M ∗\|y)qr(M\|Mk) p(M\|y)qr(M ∗\|M ∗ k) . (5) (5) Algorithm 1 describes in detail the mode jumping step within the MJMCMC algo- rithm. In the ﬁrst step, a large change in the model space is made through the proposal distribution ql. This will typically lead to a model with little support in the data, so in step 2 a local optimization is performed in order to obtain a better model. Due to the need for a proper Metropolis-Hastings probability derived through a backwards move (step 4), a randomization, through qr, of the local optima is needed for the reverse move back to the original model to be possible. Step 5 speciﬁes the acceptance probability which is shown in Hubin and Storvik (2018) to satisfy the detailed balance equation with respect to p(M\|Y ). Algorithm 1 Mode jumping MCMC. Hopefully, this detailed discussion of MJMCMC fully resolves the confusion of Clarte and Robert (2020), who, in their discussion, presume the following: “While we have not read the refered article on MJMCMC in detail, a ﬁrst comment is that the name itself is somewhat unsuitable, as indeed the algorithm does not sample from a distribution but only explores its surface.” We would like to emphasize that the MJMCMC is not incorporating any of the ideas of Tabu search algorithms (Glover et al., 1995), which are not allowing to return to the previously visited models. This should also clarify another misleading presumption by Clarte and Robert (2020): “In a self-avoiding mode, keeping track of all the previous states visited by the chain ensures that those states will never be visited again. As we are in a discrete setting, this implies that once a mode has been visited the algorithm is constrained to eventually visit another mode, even if the potential between the modes is almost zero.” Rejoinder The main diﬀerence is that when using (3) the number of times a speciﬁc model is visited is important, for (4) it is enough that a model is visited at least once. In this context, the denominator of (4) becomes an extremely relevant measure for the quality of the search. It should be as large as possible in order to capture the probability mass from all the local optima of the posterior distribution, whilst at the same time the size of V should be low in order to save computational time. 321 A. Hubin, G. Storvik, and F. Frommlet 4.2 Parameter settings The choice of the tuning parameters for the algorithm is deﬁnitely an important problem as indicated by Ruczinski et al. (2020) and Clarte and Robert (2020). Whilst there is not (and cannot be) any uniformly best choice of tuning parameters of GMJMCMC, we will try to brieﬂy indicate some strategies allowing to manually choose reasonable values of the most important tuning parameters of the algorithm. Regarding the choice of the population size d and the maximal number of variables in a model kmax, we give some guidance in Remark 1 after Theorem 1 in Hubin et al. (2020a): “When d1 > 0 (which is the Ninit covariates with largest marginal inclusion probability in S1), some restrictions on the possible search spaces are introduced. However, when d −d1 ≥kmax, any model of maximum size kmax will eventually be visited. If d −d1 < kmax, then every model of size up to d −d1 plus some of the larger models will eventually be visited, although the model space will get some additional constraints. In practice, it is more important that d−d1 ≥k∗, where k∗is the size of the true model. Unfortunately, neither k∗nor d1 are known in advance, and one has to make reasonable choices of kmax and d depending on the problem one analyses.” Also, note that we provide some sensitivity analysis of d in Section 3.1 of the main article. Regarding the maximal depth of logic expressions Cmax, one should use some prior knowledge on the complexity of logic expressions. It also depends upon the individual hypotheses the researcher has. At the same time, using unreasonably large Cmax is prohibitive computationally and also unrealistic in terms of power to detect too complex trees. When combining two Boolean expressions, ﬁrst a decision is made whether it will be combined through an and or an or operation (with Pand specifying the probability for and) and thereafter a decision is made whether the logic not is applied to it (with probability Pnot). In our experience, the actual values of these tuning parameters will not inﬂuence the result very much with respect to ﬁnding the right expressions within the equivalence classes. However, simpler expressions (within the equivalence classes) are usually obtained when choosing somewhat larger Pand and somewhat smaller Pnot. We recommend the choice Pand = 0.9 and Pnot = 0.1. Convergence of GMJMCMC The MJMCMC algorithm, in the setting of BLR, only gives convergence within each of the restricted search spaces (populations) that it considers. We apply the MJMCMC as an inner iteration within the GMJMCMC algorithm where the space of models is dynamically modiﬁed. Given that the movement within and between the search spaces is irreducible with respect to the whole model space, which is shown in Theorem 1 of Hubin et al. (2020a), the GMJMCMC provides the estimates equivalent to (4). They also converge towards the right model probabilities. This fully resolves another concern from Rejoinder 322 Clarte and Robert (2020) who stated that the renormalized estimator of the marginal posterior model probabilities “does not provide the theoretical security of (asymptotic) unbiasedness that is attained with MCMC method.” Clarte and Robert (2020) who stated that the renormalized estimator of the marginal posterior model probabilities “does not provide the theoretical security of (asymptotic) unbiasedness that is attained with MCMC method.” 4.2 Parameter settings The tuning parameter ρmin is used to determine which variables should be removed from the current population with probability one minus the current approximation of the marginal inclusion probability of these variables. ρmin should be chosen in such a way that it is on the one hand possible to get rid of unimportant trees, while at the same time avoiding the deletion of potentially important trees. Concerning the question of Ruczinski et al. (2020) on the choice of Mfin and Tmax and the resulting chain length, we provided some guidance in Theorem A.1 in Section A.2 of Hubin et al. (2020b). There, we proved convergence guarantees also for ﬁxed Tmax and Mfin when increasing the number of parallel chains of GMJMCMC. Thus, apparently, there exists a natural A. Hubin, G. Storvik, and F. Frommlet 323 A. Hubin, G. Storvik, and F. Frommlet trade oﬀ: the more chains one can aﬀord running in parallel the fewer resources could be used within each chain and vice versa – the less parallel chains one runs – the larger Tmax and Mfin are required. The choice of the tuning parameters for the examples from Hubin et al. (2020a) are provided in Section A.1 of Hubin et al. (2020b). These values might be considered for problems of similar dimensionality, eﬀect sizes and correlations between covariates. At the same time, we cannot provide a strict stopping criterion for GMJMCMC or a general rule for the choice of its parameters. Experimental tuning for diﬀerent applications might be beneﬁcial. If one has enough computational resources, grid search or an adaptation of Bayesian optimization for the tuning parameters of GMJMCMC (Snoek et al., 2012) can be considered. Alternatively, one might consider some kind of adaptive learning of the algorithm’s tuning parameters similarly to Hubin (2019). More details on these possibilities are beyond the scope of this rejoinder. 5 Various extensions of BLR and GMJMCMC In this section, we brieﬂy present extensions of the logic regression model. Some of these extensions are further discussed in the tutorial of Section 6 of the rejoinder. A more detailed description of the proposed extensions, including theoretical support and real applications, are material for a future publication. 5.1 Predictions with BLR As mentioned in the discussion section of Hubin et al. (2020a), our method is directly applicable to prediction as well. In particular, the standard Bayesian model averaging can be easily applied. Thus, one can approximate the posterior probability of some parameter/variable Δ via model averaging by ˆp(Δ \| Y ) = M∈V p(Δ \| M, Y )ˆp(M \| Y ), (6) (6) where Δ might be, for example, the predictor of unobserved data based on a speciﬁc set of covariates. Given estimates of model posterior probabilities, other prediction pro- cedures such as the median probability model (Barbieri and Berger, 2004) or the most probable model can be also easily adopted, yielding: where Δ might be, for example, the predictor of unobserved data based on a speciﬁc set of covariates. Given estimates of model posterior probabilities, other prediction pro- cedures such as the median probability model (Barbieri and Berger, 2004) or the most probable model can be also easily adopted, yielding: ˆp(Δ \| Y ) = p(Δ \| M ∗, Y ), (7) (7) where M ∗is the selected median probability or the most probable a posteriori model. 5.2 BLR with non-binary covariates Responding to a question from Ruczinski et al. (2020), we can allow non-binary ﬁxed eﬀects to be included in the model. For this extension, we simply replace equation (2) Rejoinder 324 324 Rejoinder in Hubin et al. (2020a) with: in Hubin et al. (2020a) with: in Hubin et al. (2020a) with: h (μ (X)) = α + q j=1 γjβjLj + q+q′ j=q+1 γjβjzj−q, (8) (8) where zl, l ∈{1, . . . , q′} are non-binary covariates which are not allowed to form logic expressions. In this formulation of the Bayesian logic regression, the model includes q+q′ possible components. The priors on the additional components γj, j ∈{q +1, . . . , q +q′} are of form (4) from Hubin et al. (2020a) with c(zj−q) = 1, j ∈{q + 1, . . . , q + q′}. This results in the following joint model prior: where zl, l ∈{1, . . . , q′} are non-binary covariates which are not allowed to form logic expressions. In this formulation of the Bayesian logic regression, the model includes q+q′ possible components. The priors on the additional components γj, j ∈{q +1, . . . , q +q′} are of form (4) from Hubin et al. (2020a) with c(zj−q) = 1, j ∈{q + 1, . . . , q + q′}. This results in the following joint model prior: p(M) ∝a q+q′ j′=q+1 γj′ q j=1 aγjc(Lj). (9) (9) In terms of model inference, the GMJMCMC is adopted, where modiﬁcations, mutations and reductions are only allowed for the Boolean terms. In terms of model inference, the GMJMCMC is adopted, where modiﬁcations, mutations and reductions are only allowed for the Boolean terms. 5.3 BLR with non-binary covariates and latent Gaussian variables We also mentioned in Hubin et al. (2020a) that it is straight-forward to extend our approach for generalized linear mixed models. Here, we will formally describe this ex- tension by including both ﬁxed eﬀects for non-binary covariates and latent Gaussian variables, which can be used to model correlation structures between observations (in space and time) and over-dispersion. For this extension, we further update equation (8), h (μ (X)) = β0 + q j=1 γjβjLj + q+q′ j=q+1 γjβjzj−q + r k=1 δik, (10) (10) (10) where zl, l ∈{1, . . . , q′} are non-binary covariates which are not allowed to form logic expressions and δk = (δ1k, . . . , δnk) ∼Nn (0, Σk) are latent Gaussian variables. The latent Gaussian variables with covariance matrices Σk allow to model diﬀerent correla- tion structures between individual observations (e.g. auto-regressive models or various other spatio-temporal models). The matrices typically depend only on a few parameters, so that in practice one has Σk = Σk(ψk). Whilst the model priors (9) are still valid, parameter priors here need to be adjusted as β\|γ ∼Npγ(0, Ipγe−ψβγ ), (11) ψk ∼πk(ψk). (12) (11) (12) (11) (11) (12) (12) Here, all kind of hyper-parameters of priors compatible with INLA (Rue et al., 2009) can in principle be chosen. This allows to eﬃciently compute the marginal likelihoods of individual models using the INLA approach (Rue et al., 2009; Hubin and Storvik, 2016). 325 A. Hubin, G. Storvik, and F. Frommlet 6 A tutorial on GMJMCMC for BLR Finally, we provide a brief tutorial on how to apply our approach in practice. Our code should be run under Linux. One would need to incorporate some sort of extra hacks (see https://bit.ly/37tf3cm) to be able to run the code under Windows (due to the limitations of the standard parallel::mclapply R function which is applied within the library). 6.1 Installing the packages We start by preparing the R environment for running our approach to BLR. The R- script below will install all packages that are needed to run the code. Depending on which R packages you have already installed, running this script might take a while. Then we install the EMJMCMC package from GitHub. One might want to restart R before proceeding to have a clean environment. After having the package installed we can load EMJMCMC. One might want to restart R before proceeding to have a clean environment. After having the package installed we can load EMJMCMC. Additionally, we will need the following three packages for the tutorial, which you might have to install from CRAN. 326 Rejoinder We obtain the following results using the robust g prior: 6.2 Running BLR with weakly correlated covariates We ﬁrst generate some binary data with the general correlation structure from the ﬁrst scenario of the simulation study above. The following code generates the heat-map of Figure 1 which illustrates the non-trivial correlations of the simulated binary variables. Next, we simulate the responses according to Scenario 4 from Hubin et al. (2020a), but with correlated binary covariates. Before performing logic regression with GMJMCMC one might like to have a look at the documentation of the R function LogicRegr: Before performing logic regression with GMJMCMC one might like to have a look at the documentation of the R function LogicRegr: 327 A. Hubin, G. Storvik, and F. Frommlet The following code runs inference on BLR with 32 parallel threads of GMJMCMC, where we are ﬁrst using the robust g-prior and then Jeﬀreys’ prior. Depending on the cluster each of these might run for some time from several minutes to more than half an hour. If you are running the code on a home PC or a laptop, please reduce ncores parameter to something reasonable for your machine (e.g. set ncores = 3). We obtain the following results using the robust g prior: We obtain the following results using the robust g-prior: We obtain the following results using the robust g-prior: and rather similar results with the Jeﬀreys’ prior: 328 6.3 Additional non-binary ﬁxed eﬀects and predictions Further, many of the input options used are explained in the help pages of pinferunemjmcmc. If one is not interested in predictions, runemjmcmc.params$save.beta = F, predict = F and test.data = NULL should be set (this will decrease inference time for the same training data sample and other tuning parameters ﬁxed). The function pinferunemjmcmc has more capabilities than performing logic regres- sion. First, one might want to see its arguments: The following call of pinferunemjmcmc performs logic regression using 30 cores. Note that the non-binary covariate is not a part of the formula passed to the function, but is rather speciﬁed through runemjmcmc.params$latnames = “I(age)”. Also, one might expect this to run slightly longer than previous examples, particularly because keeping track of the β coeﬃcients for prediction takes some additional time. Further, many of the input options used are explained in the help pages of pinferunemjmcmc. If one is not interested in predictions, runemjmcmc.params$save.beta = F, predict = F and test.data = NULL should be set (this will decrease inference time for the same training data sample and other tuning parameters ﬁxed). The following call of pinferunemjmcmc performs logic regression using 30 cores. Note that the non-binary covariate is not a part of the formula passed to the function, but is rather speciﬁed through runemjmcmc.params$latnames = “I(age)”. Also, one might expect this to run slightly longer than previous examples, particularly because keeping track of the β coeﬃcients for prediction takes some additional time. Further, many of the input options used are explained in the help pages of pinferunemjmcmc. If one is not interested in predictions, runemjmcmc.params$save.beta = F, predict = F and test.data = NULL should be set (this will decrease inference time for the same training data sample and other tuning parameters ﬁxed). The following call of pinferunemjmcmc performs logic regression using 30 cores. Note that the non-binary covariate is not a part of the formula passed to the function, but is rather speciﬁed through runemjmcmc.params$latnames = “I(age)”. Also, one might expect this to run slightly longer than previous examples, particularly because keeping track of the β coeﬃcients for prediction takes some additional time. Further, many of the input options used are explained in the help pages of pinferunemjmcmc. 6.3 Additional non-binary ﬁxed eﬀects and predictions Ruczinski et al. (2020) asked whether it would be possible to include covariates in the model which are not a part of the logic expressions. Furthermore, Schwender and Ickstadt (2020) are interested in whether the model can be easily used for predictions. These options are currently not implemented in the LogicRegr function, which we would like to keep as simple as possible. At the same time, these tasks can be easily performed by a general call of the EMJMCMC::pinferunemjmcmc function which is available in our package. This routine is however much more advanced and requires, at this time, expert knowledge to be used. First, we will generate an additional Poisson distributed covariate age which is then used as an additional additive eﬀect in the data generating logic regression model. For the sake of brevity we perform the analysis here only with Jeﬀreys’ prior. We will not only perform model inference but also show how to make predictions with the EMJMCMC package. To this end, we will randomly divide the data into a training set (900 observations) and a testing set (100 observations). The function pinferunemjmcmc has more capabilities than performing logic regres- sion. First, one might want to see its arguments: The following call of pinferunemjmcmc performs logic regression using 30 cores. Note that the non-binary covariate is not a part of the formula passed to the function, but is rather speciﬁed through runemjmcmc.params$latnames = “I(age)”. Also, one might expect this to run slightly longer than previous examples, particularly because keeping track of the β coeﬃcients for prediction takes some additional time. Further, many of the input options used are explained in the help pages of pinferunemjmcmc. If one is not interested in predictions, runemjmcmc.params$save.beta = F, predict = F and test.data = NULL should be set (this will decrease inference time for the same training data sample and other tuning parameters ﬁxed). The function pinferunemjmcmc has more capabilities than performing logic regres- sion. First, one might want to see its arguments: The following call of pinferunemjmcmc performs logic regression using 30 cores. Note that the non-binary covariate is not a part of the formula passed to the function, but is rather speciﬁed through runemjmcmc.params$latnames = “I(age)”. Also, one might expect this to run slightly longer than previous examples, particularly because keeping track of the β coeﬃcients for prediction takes some additional time. 6.3 Additional non-binary ﬁxed eﬀects and predictions If one is not interested in predictions, runemjmcmc.params$save.beta = F, predict = F and test.data = NULL should be set (this will decrease inference time for the same training data sample and other tuning parameters ﬁxed). 329 A. Hubin, G. Storvik, and F. Frommlet A. Hubin, G. Storvik, and F. Frommlet Below, a list of the logic expressions and non-logic covariates that were found to be of importance is listed. There, we clearly see that all features from the data-generative model are detected without any false positive discoveries. two criteria, RMSE = n Y ∗ i are responses in the test he size of the test data set BLR in this example with a 330 Rejoinder We want to compare the performance of BLR in this example with a simple standard ap- proach, namely ridge regression (Zou and Hastie, 2005), combined with model selection according to AIC. In the script below, we run ridge regression and perform prediction on the test data set. We ﬁnally compute the evaluation metrics for prediction based on the expectations of the data-generative (true) model for the test data: We clearly see that for this speciﬁc example logic regression signiﬁcantly outperforms the ridge regression baseline with respect to both RMSE and MAE. This is not surprising given that the data generative process has multiple non-linear eﬀects. Moreover, the predictions obtained by the BLR model are extremely close to the predictions from the means of the data generative model. 331 A. Hubin, G. Storvik, and F. Frommlet 8 Conclusions We would like to thank once again all of the discussants for their valuable and insightful feedback. We are happy to have provoked so many questions, comments and remarks. We hope that we managed to shed light on the majority of them in this rejoinder. Moreover, we provided some useful extension of Bayesian logic regression method here. The discussions also motivate multiple directions for further research, which are outside the scope of this rejoinder. However, we hope this research will be in future performed in close collaboration with the discussants. 7 Comparison with other approaches Several other approaches were mentioned by the discussants. Ruczinski et al. (2020) mentioned that simulated annealing for logic regression could be equipped with a pe- nalized likelihood criterion following from the priors used in our setting. Schwender and Ickstadt (2020) pointed out certain similarities of GMJMCMC with Genetic Program- ming for Association Studies as well as logic Feature Selection. Bogdan et al. (2020) mentioned the recently developed non-reversible MCMC algorithms as well as parallel tempering MCMC algorithms. It would be most interesting to compare all these diﬀer- ent algorithms with GMJMCMC but we believe this would need substantial additional eﬀort and goes far beyond the scope of this rejoinder. We leave these possibilities open as topics for further research. References Barbieri, M. M. and Berger, J. O. (2004). “Optimal predictive model selection.” The Annals of Statistics, 32(3): 870–897. MR2065192. doi: https://doi.org/10.1214/ 009053604000000238. 323 Bayarri, M. J., Berger, J. O., Forte, A., Garc´ıa-Donato, G., et al. (2012). “Criteria for Bayesian model choice with application to variable selection.” The Annals of Statis- tics, 40(3): 1550–1577. MR3015035. doi: https://doi.org/10.1214/12-AOS1013. 317 Bogdan, M., Miasojedow, B., and Wallin, J. (2020). “Discussion of “A Novel Algorithmic Approach to Bayesian Logic Regression” by A. Hubin, G. Storvik and F. Frommlet.” Bayesian Analysis. 313, 318, 319, 320, 331 Broman, K. W., Wu, H., Sen, ´S., and Churchill, G. A. (2003). “R/qtl: QTL mapping in experimental crosses.” Bioinformatics, 19(7): 889–890. 314 Chai, J. S. and Bose, A. (1993). “Bottlenecks in parallel algorithms for power system stability analysis.” IEEE Transactions on Power Systems, 8(1): 9–15. 319 Chen, Z. (2016). Statistical methods for QTL mapping. Chapman and Hall/CRC. MR3241239. 314 332 Rejoinder Rejoinder Clarte, G. and Robert, C. (2020). “A discussion on “A Novel Algorithmic Approach to Bayesian Logic Regression”.” Bayesian Analysis. 316, 319, 321, 322 Clyde, M. A., Ghosh, J., and Littman, M. L. (2011). “Bayesian adaptive sampling for variable selection and model averaging.” Journal of Computational and Graphical Statistics, 20(1): 80–101. MR2816539. doi: https://doi.org/10.1198/jcgs.2010. 09049. 320 Fritsch, A. (2006). “A Full Bayesian Version of Logic regression for SNP Data.” Ph.D. thesis, Diploma Thesis. 313 Glover, F., Kelly, J. P., and Laguna, M. (1995). “Genetic algorithms and tabu search: hybrids for optimization.” Computers & Operations Research, 22(1): 111–134. 321 Hubin, A. (2019). “An adaptive simulated annealing EM algorithm for inference on non- homogeneous hidden Markov models.” In Proceedings of the International Conference on Artiﬁcial Intelligence, Information Processing and Cloud Computing, 1–9. 323 Hubin, A. and Storvik, G. (2016). “Estimating the marginal likelihood with Integrated nested Laplace approximation (INLA).” arXiv:1611.01450v1. 324 Hubin, A. and Storvik, G. (2018). “Mode jumping MCMC for Bayesian variable selection in GLMM.” Computational Statistics & Data Analysis, 127: 281–297. MR3820324. doi: https://doi.org/10.1016/j.csda.2018.05.020. 312, 317, 319, 321 Hubin, A., Storvik, G., and Frommlet, F. (2020a). “A novel algorithmic approach to Bayesian Logic Regression.” Bayesian Analysis. 312, 313, 314, 318, 319, 321, 322, 323, 324, 326 Hubin, A., Storvik, G., and Frommlet, F. (2020b). “Supplementary material for “A novel algorithmic approach to Bayesian Logic Regression”.” URL https:// projecteuclid.org/download/suppdf 1/euclid.ba/1545296448. 322, 323 Hubin, A., Storvik, G., Grini, P., and Butenko, M. (2020c). “A Bayesian binomial regression model with latent Gaussian processes for modelling DNA methylation.” Austrian Journal of Statistics, 49–50. 313 Joe, H. (2006). “Generating random correlation matrices based on partial correlations.” Journal of Multivariate Analysis, 97(10): 2177–2189. MR2301633. doi: https://doi. org/10.1016/j.jmva.2005.05.010. 314 Kukanov, A. (2008). “Why a simple test can get parallel slowdown.” URL https:// software.intel.com/en-us/blogs/2008/03/04/why-a-simple-test-can-get- parallel-slowdown. 319 Leisch, F., Weingessel, A., and Hornik, K. (1998). “On the generation of correlated artiﬁcial binary data.” 314 Li, Y. and Clyde, M. A. (2018). “Mixtures of g-priors in generalized linear models.” Journal of the American Statistical Association, 113(524): 1828–1845. MR3902249. doi: https://doi.org/10.1080/01621459.2018.1469992. 317 A. Hubin, G. Storvik, and F. Frommlet A. Hubin, G. Storvik, and F. Frommlet 333 Ruczinski, I., Kooperberg, C., and LeBlanc, M. (2020). “Invited comment on Article by Hubin, Storvik and Frommlet.” Bayesian Analysis. 313, 317, 322, 323, 328, 331 Ruczinski, I., Kooperberg, C., and LeBlanc, M. (2020). “Invited comment on Article by Hubin, Storvik and Frommlet.” Bayesian Analysis. Rejoinder 313, 317, 322, 323, 328, 331 Rue, H., Martino, S., and Chopin, N. (2009). “Approximate Bayesian inference for latent Gaussian models by using integrated nested Laplace approximations.” Journal of the Royal Statistical Society, 71(2): 319–392. MR2649602. doi: https://doi.org/ 10.1111/j.1467-9868.2008.00700.x. 317, 324 Schwender, H. and Ickstadt, K. (2020). “Discussion of a novel algorithmic approach to Bayesian logic regression.” Bayesian Analysis. 315, 328, 331 Snoek, J., Larochelle, H., and Adams, R. P. (2012). “Practical Bayesian optimization of machine learning algorithms.” In Advances in Neural Information Processing Systems, 2951–2959. 323 Tjelmeland, H. and Hegstad, B. K. (2001). “Mode jumping proposals in MCMC.” Scan- dinavian Journal of Statistics, 28(1): 205–223. MR1844357. doi: https://doi.org/ 10.1111/1467-9469.00232. 319 Zou, H. and Hastie, T. (2005). “Regularization and variable selection via the elastic net.” Journal of the Royal Statistical Society: Series B (Statistical Methodology), 67(2): 301–320. MR2137327. doi: https://doi.org/10.1111/j.1467-9868.2005.00503. x. 330
https://openalex.org/W2587260557	https://research-information.bris.ac.uk/ws/files/102517782/art_3A10.1186_2Fs12874_017_0306_8.pdf	English	null	Utilising identifier error variation in linkage of large administrative data sources	BMC Medical research methodology	2,017	cc-by	7,183	Goldstein, H., Hagger-Johnson, G., Gilbert, R., & Harron, K. (2017). Utilising identifier error variation in linkage of large administrative data sources. BMC Medical Research Methodology, 17, Article 23. https://doi.org/10.1186/s12874-017-0306-8 Publisher's PDF, also known as Version of record License (if available): CC BY Link to published version (if available): 10.1186/s12874-017-0306-8 Link to publication record on the Bristol Research Portal PDF-document This is the final published version of the article (version of record). It first appeared online via BioMed Central at http://bmcmedresmethodol.biomedcentral.com/articles/10.1186/s12874-017-0306-8. Please refer to any applicable terms of use of the publisher. Goldstein, H., Hagger-Johnson, G., Gilbert, R., & Harron, K. (2017). Utilising identifier error variation in linkage of large administrative data sources. BMC Medical Research Methodology, 17, Article 23. https://doi.org/10.1186/s12874-017-0306-8 Goldstein, H., Hagger-Johnson, G., Gilbert, R., & Harron, K. (2017). Utilising identifier error variation in linkage of large administrative data sources. BMC Medical Research Methodology, 17, Article 23. https://doi.org/10.1186/s12874-017-0306-8 Abstract Background: Linkage of administrative data sources often relies on probabilistic methods using a set of common identifiers (e.g. sex, date of birth, postcode). Variation in data quality on an individual or organisational level (e.g. by hospital) can result in clustering of identifier errors, violating the assumption of independence between identifiers required for traditional probabilistic match weight estimation. This potentially introduces selection bias to the resulting linked dataset. We aimed to measure variation in identifier error rates in a large English administrative data source (Hospital Episode Statistics; HES) and to incorporate this information into match weight calculation. Methods: We used 30,000 randomly selected HES hospital admissions records of patients aged 0–1, 5–6 and 18–19 years, for 2011/2012, linked via NHS number with data from the Personal Demographic Service (PDS; our gold- standard). We calculated identifier error rates for sex, date of birth and postcode and used multi-level logistic regression to investigate associations with individual-level attributes (age, ethnicity, and gender) and organisational variation. We then derived: i) weights incorporating dependence between identifiers; ii) attribute-specific weights (varying by age, ethnicity and gender); and iii) organisation-specific weights (by hospital). Results were compared with traditional match weights using a simulation study. Results: Identifier errors (where values disagreed in linked HES-PDS records) or missing values were found in 0.11% of records for sex and date of birth and in 53% of records for postcode. Identifier error rates differed significantly by age, ethnicity and sex (p < 0.0005). Errors were less frequent in males, in 5–6 year olds and 18–19 year olds compared with infants, and were lowest for the Asian ethic group. A simulation study demonstrated that substantial bias was introduced into estimated readmission rates in the presence of identifier errors. Attribute- and organisational-specific weights reduced this bias compared with weights estimated using traditional probabilistic matching algorithms. Conclusions: We provide empirical evidence on variation in rates of identifier error in a widely-used administrative data source and propose a new method for deriving match weights that incorporates additional data attributes. Our results demonstrate that incorporating information on variation by individual-level characteristics can help to reduce bias due to linkage error. Key words: Data linkage, Record linkage, Administrative data, Linkage error, Linkage evaluation, Hospital admission challenges for identifying the same individual in different sources [1]. © The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. University of Bristol – Bristol Research Portal General rights This document is made available in accordance with publisher policies. Please cite only the published version using the reference above. Full terms of use are available: http://www.bristol.ac.uk/red/research-policy/pure/user-guides/brp-terms/ Harron et al. BMC Medical Research Methodology (2017) 17:23 DOI 10.1186/s12874-017-0306-8 Open Access * Correspondence: Katie.harron@lshtm.ac.uk 1London School of Hygiene and Tropical Medicine, 15-17 Tavistock Place, London WC1 H 9SH, UK Full list of author information is available at the end of the article Abstract Typographical errors, missing values and identifiers that change over time can prevent records from matching and lead to linkage error (false-matches and missed-matches) [2, 3]. Even low error rates can lead to biased results, particularly when records from particular types of individuals or organisations are less likely to link successfully than others. Such ‘differential’ linkage can lead to a form of bias in analysis, for Utilising identifier error variation in linkage of large administrative data sources Katie Harron1* , Gareth Hagger-Johnson2, Ruth Gilbert3 and Harvey Goldstein4 Katie Harron1* , Gareth Hagger-Johnson2, Ruth Gilbert3 and Harvey Goldstein4 Methods Data Hospital Episode Statistics (HES) is an administrative data source containing information on all admissions to NHS hospitals in England. Linkage of HES is coordi- nated through NHS Digital (previously known as the Health and Social Care Information Centre) [17]. The HES extract used for this study had previously been linked with a reference (gold-standard) dataset of re- cords extracted from the Personal Demographic Service (PDS), which is also coordinated by NHS Digital (http:// systems.digital.nhs.uk/demographics/pds). PDS contains the latest demographic details corresponding to a given NHS number. PDS also contains historical information such as previous addresses and is used for the NHS number tracing service (known as the Demographics Batch Service) and to provide identifiers for the NHS Patient Spine. Linkage with PDS reference data allowed us to quantify identifier errors. In this study, we define identifier error as discrepancies between PDS and HES, e.g. where identifiers had been recorded incorrectly, had legitimately changed over time (e.g. postcode) or were missing in HES. g There are several problems associated with the calcu- lation of probabilistic match weights using the trad- itional approach. Firstly, match weights are calculated assuming that identifier errors occur randomly within a dataset, and that the probability of an identifier error is unrelated to any other characteristic (age, ethnicity etc.) [11]. However, this assumption is often invalid: data quality is often associated with individual-level charac- teristics and can also vary on an organisational level [15]. These associations are typically ignored, unless these characteristics are incorporated into a blocking scheme with match weights being produced separately for each block. Secondly, match weights are typically cal- culated by summing the logarithms of m- and u- probability ratios across identifiers. This requires the as- sumption that identifier errors are independent (i.e. agreement on year of birth is independent of agreement on forename) - an assumption that often fails and can lead to misclassification of record pairs [16]. For the purposes of this study, we defined our true (reference) match status by agreement or disagreement of NHS number between HES and PDS. We used a ran- dom sample of 10,000 record pairs from HES inpatient data linked with PDS, for the financial year 1st April 2011 to 31st March 2012, for each of three cohorts de- fined by date of birth: i) infants aged <1 year; ii) children aged 5–6 years; and iii) young adults aged 18–19 years. Background Linkage of administrative data is an important tool for service evaluation and research, as individual-level infor- mation can be combined in a relatively cost-effective and timely manner compared with conventional data collection models. Most administrative data sources were not developed with linkage in mind, posing unique Harron et al. BMC Medical Research Methodology (2017) 17:23 Page 2 of 9 Page 2 of 9 example when specific groups of records are misclassi- fied or excluded from the linked dataset [4–8]. likely to be representative of other administrative sources, i.e. those where identifiers are input using a range of IT sys- tems, and so information on identifier error rates will be relevant to linkage of other large administrative data sources. Secondly, we develop methods to estimate match weights without relying on the independence assumption, and incorporating individual or organisational-level attri- butes, and evaluate these weights as alternatives to trad- itional probabilistic match weights. For linkage of data sources that do not contain a reli- able unique identifier, probabilistic methods are com- monly used [9, 10]. Probabilistic linkage makes use of variables such as sex, date of birth and postcode to cre- ate a match weight for classifying records as matches or non-matches. Match weights are traditionally based on the Fellegi-Sunter approach using conditional probabil- ities derived from estimated rates of errors in identifiers: the probability that identifiers agree given records be- long to the same subject (m-probability), and the prob- ability that identifiers agree given records belong to different subjects (u-probability) [11, 12]. Conditional probabilities can be derived from ‘training’ data, i.e. a sub-sample of data where the true match status for each record pair is known (supervised matching) [13]. If no training data are available, probabilities are typically esti- mated using statistical methods, such as the expectation-maximisation (EM) algorithm [14]. Identifier error rates u−probability ¼ uφ ¼ P agreeonsex and date of birth; disagreement on postcode U j Þ ¼ We estimated identifier error rates for sex, date of birth and postcode, based on the number of times these iden- tifiers disagreed in matched HES-PDS records. We mod- elled the risk of identifier error using logistic regression with a set of attribute predictors recorded in HES (ethni- city, age and sex). We used a multi-level model with hospital as a random effect to explore organisational- level variation. Dependence between pairwise identifiers was also tested using multi-level logistic regression models using Stata [18]. P φ ¼ 110 f g U j ð Þ Match weights were then derived as: Match weights were then derived as: W ¼ log2 mφ uφ W ¼ log2 mφ uφ 3. Attribute-specific and organisational-specific match weights We derived attribute-specific match weights using the procedures described above, but now for each combination of characteristics as recorded in PDS (age cohort, sex, ethnic group, N combinations = 36). This process is distinct from blocking, in that agreement on any of these attributes is not required for linkage (and attribute-specific weights can be cal- culated for variables not used within the linkage, e.g. ethnic group). Organisational-specific match weights were derived by calculating m- and u-probabilities separately for each hospital (N hospitals= 388). Attribute-specific and organisational-specific match weights were calculated in the traditional manner (i.e. assuming independence between identifiers), as it was not possible to stratify each agreement pattern by age, sex, ethnicity due to low numbers. Methods Data For each age cohort, the set of matches was created by identifying the PDS record associated with the NHS number on each HES record (n = 10,000 matches). The set of non-matches was created by identifying all PDS records with different NHS numbers to each HES record. This resulted in (10,000 × 10,000)-10,000 = 99,000,000 non-matches for each age cohort. However, the majority of these non-matches did not agree on any identifier, or only agreed on sex, and so were excluded from consideration. This resulted in around 30,000 non- matches for each age cohort. One approach to overcome these problems is to estimate match weights jointly over a set of identifiers (the agree- ment pattern), thus overcoming the need for independence between identifiers. It is also possible to calculate match weights allowing dependence on individual- and/or organisational-level covariates. Although characteristics of the identifying variables, such as the frequency of common or rare surnames, are often incorporated into match weight calculation, this has not been the case for individual charac- teristics that are not used for matching (e.g. ethnicity) or at an organisational level (e.g. by hospital). The present study aims first to provide empirical evidence on the associations between identifier error rates and individual characteristics in a national administrative data source (Hospital Episode Statistics; HES). Quality of identifier recording in HES is The data used for this study comprised patterns of agreement/disagreement between date of birth, sex and Page 3 of 9 Harron et al. BMC Medical Research Methodology (2017) 17:23 postcode in HES-PDS linked pairs, but contained no ac- tual identifiers. Agreement patterns were aggregated by age cohort, sex and ethnic group. m−Probability ¼ mφ ¼ P agree on sex and date of birth; disagreement on postcode M j Þ ¼ m−Probability ¼ mφ ¼ P agree on sex and date of birth; disagreement on postcode M j Þ ¼ P φ ¼ 110 f g M j ð Þ 2. Match weights incorporating dependence between identifiers Each HES-PDS record pair was associated with an agreement pattern φ representing agreement or dis- agreement on the joint set of three identifiers {sex, date of birth, postcode}. For binary agreement (agree = 1; disagree = 0), there are 23 = 8 possible agreement patterns for sex, date of birth and post- code: {1,1,1}, {1,1,0} … and {0,0,0} etc. Conditional probabilities were derived jointly over all identifiers for each observed agreement pattern, e.g. for agree- ment on sex, date of birth and disagreement on postcode, represented as {110}: Aim We performed a simulation study to determine the ef- fect of the identifier-independence assumption and the value of incorporating attribute information into match weight calculation. Our scenario was linkage of hospital admissions records containing sex, date of birth, post- code, and NHS number. The aim was to estimate re- admission rates by linking multiple hospital records for the same individual over time. Where there was a match between hospital records, this indicated that an individ- ual had been admitted multiple times within the study period. Individuals with only a single hospital record and no matches were admitted only once during the study year. W ¼ X k log2 mk uk ¼ log2 msex usex þ log2 mdob udob þ log2 mpostcode upostcode 2. Match weights incorporating dependence between identifiers Probabilistic match weights 1. Traditional probabilistic match weights (assuming independence between identifiers) We derived conditional probabilities for sex, date of birth and postcode based on the observed error rates for each identifier. Probabilities were derived from the number of times an identifier agreed or disagreed in pairs of matched HES-PDS records, e.g. for sex: 1. Traditional probabilistic match weights (assuming independence between identifiers) We derived conditional probabilities for sex, date of birth and postcode based on the observed error rates for each identifier. Probabilities were derived from the number of times an identifier agreed or disagreed in pairs of matched HES-PDS records, e.g. for sex: m−probability ¼ msex ¼ P agree on sexjM ð Þ u−probability ¼ usex ¼ P agree on sexjU ð Þ m−probability ¼ msex ¼ P agree on sexjM ð Þ u−probability ¼ usex ¼ P agree on sexjU ð Þ where M represents a match and U represents a non-match. Missing values were treated as di Simulation study Aim Match weights were then derived by summing the log-ratio of m- and u-probabilities over all k identi- fiers, i.e. Identifier error rates Identifier errors (including missing values) were found in 0.11% of records for sex and date of birth, and in 53% of records for postcode. In these data, there was no evi- dence of dependence between postcode and date of birth or sex (p = 0.266 and 0.187 respectively from the multi- level logistic regression model). Although the error rate for date of birth was low, errors in this variable were more likely to occur in records where there was also an error in sex (p = 0.021). Since by design, the original HES-PDS extract only in- cluded records that agreed on NHS number, we intro- duced NHS number identifier error rates representative of those observed in the literature [20, 21]. We used sev- eral scenarios to determine the effect of different NHS number error rates on results: The probability of identifier error (disagreement of identifier values between HES and PDS) differed signifi- cantly according to age (p < 0.0001), ethnicity (p = 0.0005) and sex (p < 0.0001) (Fig. 1). Identifier errors oc- curred less frequently in records from females compared with males (odds ratio 0.84; 95% CI 0.81-0.86); and were lowest for Asian ethnicity (odds ratio 0.89; 95% CI 0.84- 0.94 compared with White ethnicity). Across all identi- fiers, errors occurred less frequently in 5–6 year olds and 18–19 year olds compared with infants (odds ratios 0.39; 95% CI 0.37-0.40 and 0.37; 95% CI 0.36-0.39 re- spectively). However, patterns differed according to the identifier: sex was more likely to be correct in 18–19 year olds than infants, but the pattern was reversed for date of birth (Fig. 1). 1. NHS number was randomly missing or incorrect in 30% of records 2. NHS number was randomly missing or incorrect in 0.5% of records. 3. NHS number was missing or incorrect in 30% of records overall, but was twice as likely to contain errors if there were errors in any of the other identifiers (sex, date of birth or postcode). 4. NHS number was missing or incorrect in 30% of records overall, but errors were distributed with the same pattern as errors in ethnicity (as observed in the HES-PDS extract). Multi-level logistic regression showed there was sub- stantial variation on an organisational level although no particular hospital provider had a significantly higher error rate than the overall mean (Fig. 2). Identifier error rates For each simulation, records were rank ordered by match weight, and a cut-off threshold for classifying records as matches was chosen by determining the maximum weight or probability that would not ex- ceed a false-match rate of 1% (or 99% specificity). It was possible to fix this threshold since the true match status was known in the simulated data, although this would not be possible in real data. Data generating mechanism For each simulation, we created our ‘matches’ by ran- domly sampling agreement patterns (with replacement) from matched pairs in the HES-PDS extract, retaining distributions of age, sex and ethnicity from the original data. We created our ‘non-matches’ by sampling Page 4 of 9 Harron et al. BMC Medical Research Methodology (2017) 17:23 Harron et al. BMC Medical Research Methodology (2017) 17:23 Page 4 of 9 Comparisons Results from three approaches were averaged over 500 simulated datasets and compared with those from traditional match weights: i) match weights incorpor- ating dependence between identifiers (based on agree- ment patterns), ii) attribute-specific match weights (based on 36 different combinations of characteristics) and iii) organisational-specific match weights (based on 388 hospitals). We compared sensitivity (i.e. the proportion of true matches that were identified) be- tween methods and compared estimated readmission rates from each method with the ‘true’ readmission rate within 12 months (8.8%) in the simulated data. We assessed the performance of each method by measuring bias, i.e. the percentage difference between estimated and true readmission rates. Probabilistic match weights Absolute values differed for traditional match weights, match weights incorporating dependence between iden- tifiers, and attribute-specific match weights (Table 1). The ordering of weights (and therefore of record pairs) was the same using both traditional weights and weights incorporating dependence. However, for attribute- specific weights, ordering differed according to individ- ual characteristics. For example, for 0–1 year olds, agree- ment on date of birth and sex only had a higher weight than agreement on sex and postcode only, but these weights were reversed for the older age groups. Variation in attribute-specific match weights reflected underlying identifier error rates. For example, the match weight for agreement on date of birth and sex but disagreement on postcode was 9.2 for infants, but 7.9 for 5–6 year olds, reflecting the fact that postcode was more likely to be missing in infant records. Results agreement patterns from non-matches in the HES-PDS extract. Sampling of matches and non-matches was stratified by age, sex and ethnicity, in order to reflect dif- ferences in readmission rates observed in the litera- ture.[19] This approach avoided any distributional assumptions about identifier error rates for date of birth, sex or postcode, and also preserved associations between identifiers and individual characteristics. Simulation study Sensitivity of linkage varied from 79% using traditional match weights and match weights incorporating depend- ence, to 97% using attribute-specific match rates. With an Harron et al. BMC Medical Research Methodology (2017) 17:23 Page 5 of 9 Fig. 1 Percentage of HES-PDS linked records with disagreeing or missing identifiers according to age, ethnicity and sex. The larger identifier error rates in postcode reflect that postcode was missing for 83% of records for infants aged 0–1 years Fig. 1 Percentage of HES-PDS linked records with disagreeing or missing identifiers according to age, ethnicity and sex. The larger identifier error rates in postcode reflect that postcode was missing for 83% of records for infants aged 0–1 years Fig. 1 Percentage of HES-PDS linked records with disagreeing or missing identifiers according to age, ethnicity and sex. rates in postcode reflect that postcode was missing for 83% of records for infants aged 0–1 years error rate of 30% in NHS number, all methods underesti- mated the ‘true’ overall readmission rate of 8.8%, except for the organisation-specific match weights (Table 2). Traditional match weights and match weights incorporat- ing dependence provided similar results; organisational- and attribute-specific match weights performed best overall. traditional weights or weights incorporating dependence, 3% using attribute-specific weights, 0.1% using organisational-specific weights). Errors in date of birth were highest in records with missing ethnicity, lower in the White group compared to Mixed, Asian, or Black, and there were no errors in the ‘Other’ category (Fig. 1). When this distribution was ap- plied to NHS number errors, bias varied accordingly: lit- tle bias was introduced to the estimated readmission rate for the ‘Other’ group, but estimates for the Missing group were substantially biased (Figs. 3–4). Attribute- or organisational-specific weights performed well at hand- ling these dependencies, with an overall bias of 1% and 0.2% respectively, Bias in estimated readmission rates was highest when NHS number errors were more likely to occur in records with at least one other identifier error (21% bias using Fig. 2 Variation in identifier error rates by hospital provider (n = 167). Each dot represents one hospital (hospitals with <500 matches were excluded). Inner lines = 95% control limits; outer lines = 99.8% control limits Discussion Record pairs with no agreement on any identifiers, or where only sex agreed (agreement patterns {000} and {010}), were assumed to be non-matches and excluded underestimated, due to low sensitivity when fixing the false-match rate at 1%. In practice, false-match rates are often lower than 1%, corresponding with a lower sensi- tivity (where false-matches are avoided, to the expense of missed-matches) and a greater risk of under- estimated readmission rates [15, 27]. Bias was greatest for Mixed, Asian, or Black ethnic groups, meaning that relative comparisons by ethnicity would be biased using match weights derived by the traditional method. How- ever, we show that attribute- or organisational-specific match weights, incorporating information on variation in identifier errors, can substantially reduce bias associ- ated with linkage error. Additional methods for handling linkage error, such as incorporating match weights into analysis in a multiple imputation framework, could be used to reduce bias further [23, 28]. Incorporating information on individual or organisa- tional characteristics, or dependence between identi- fiers, into match weight estimation is a relatively simple process, given a large enough sample from which to estimate the relevant parameters. In prac- tice, detailed information on identifier error rates is not always available and parameters are often derived from a sample of data. Where a large enough sample on which to base estimates of error rates is not avail- able, it would be possible to incorporate characteris- tics into latent class models such as the Expectation- Maximisation (EM) algorithm, which can be used to estimate conditional probabilities for the traditional Fellegi-Sunter approach [14, 29]. The value of incorp- orating information on record attributes is likely to be most evident in linkage of large-scale administra- tive datasets, particularly where records are grouped, for example by organisation or region. However, our study used a relatively simple design of linkage within one longitudinal dataset, and further evaluation is re- quired to understand performance and practicalities of the method in large, complex linkages involving multiple files. aMatch probability = N matches/Total record pairs individual characteristics or organisational variation were effective at reducing bias associated with errors in linkage, particularly when errors are distributed non-randomly. There is limited evidence on how the failure of the as- sumption of independence affects linkage quality over and above the calculation of match weights. Tromp et al. Discussion Our study provides empirical evidence on variation in identifier error rates by individual characteristics in a widely-used and extensive administrative data source. This information will be valuable for other researchers assessing the feasibility of linkage with administrative data sources, particularly where no training data are available, as the identifier error rates observed in HES will provide an appropriate starting point for estimat- ing m- and u-probabilities in other similar datasets. We provide methods for incorporating dependence between identifiers, and variation in identifier errors by individual and organisational-level characteristics, into match weight calculation. Our simulation study demonstrated that match weights incorporating Fig. 2 Variation in identifier error rates by hospital provider (n = 167). Each dot represents one hospital (hospitals with <500 matches were excluded). Inner lines = 95% control limits; outer lines = 99.8% control limits Harron et al. BMC Medical Research Methodology (2017) 17:23 Page 6 of 9 Page 6 of 9 Harron et al. BMC Medical Research Methodology (2017) 17:23 Table 1 Traditional match weights, match weights incorporating dependence between identifiers, and attribute- specific match weights according to agreement pattern {date of birth, sex, postcode}. Record pairs with no agreement on any identifiers, or where only sex agreed (agreement patterns {000} and {010}), were assumed to be non-matches and excluded Agreement pattern {date of birth, sex, postcode} 001 100 011 101 110 111 N Matches 1 21 18 12 15,924 14,009 N Non- matches 259 414,307 248 4 415,888 10 Match probabilitya 0.0039 0.0001 0.0726 0.7500 0.0369 0.9993 Traditional match weight 5.3 −1.0 9.6 14.9 8.6 23.7 Match weight assuming dependence −0.5 −1.2 9.3 17.8 8.6 27.6 Attribute-specific match weight: Sex Female −1.7 −1.7 8.7 17.3 8.7 27.7 Male 0.4 −0.9 9.7 18.1 8.5 27.5 Age 0–1 years −0.5 0.1 8.6 18.5 9.2 27.6 5–6 years −0.2 −2.0 9.6 18.2 7.9 28.0 18–19 years −1.5 −2.6 9.5 16.2 8.3 27.2 Ethnicity Missing 2.7 0.3 10.8 19.5 8.5 27.6 White −1.3 −1.6 8.8 17.4 8.6 27.6 Mixed 1.8 −0.4 10.9 19.4 8.8 28.6 Asian −0.4 −2.3 10.4 16.9 8.6 27.8 Black 1.6 0.9 9.6 19.1 8.9 27.1 Other 1.4 −0.2 10.5 19.0 8.8 28.1 Organisational- specific match weight (mean) 5.7 1.3 12.3 20.7 8.1 25.4 aMatch probability = N matches/Total record pairs Table 1 Traditional match weights, match weights incorporating dependence between identifiers, and attribute- specific match weights according to agreement pattern {date of birth, sex, postcode}. Discussion In addition, we based our extract on date of birth, and so excluded all records where date of birth was missing. We also used a one year study period, and therefore would not have captured changes in postcode over time. Inspection of PDS reveals that 55% of chil- dren have at least two postcodes in their first year of life and 69% have at least two postcodes by age 5/6 (19% have four or more different postcodes by this age). In our simulation study, we fixed our threshold at a false- match rate of 1%. In practice, choice of appropriate thresholds can be difficult, and is typically chosen based on a sample of manually-reviewed records, or using syn- thetic data [35]. However, our study was limited by the assumption that agreement on NHS number between HES and PDS indi- cated that records belonged to the same individual. In reality, NHS number is not always a reliable identifier for linkage [20]. If well-completed NHS number is indi- cative of good data quality more generally, we may have underestimated identifier error rates through our study design. In addition, we based our extract on date of birth, and so excluded all records where date of birth was missing. We also used a one year study period, and therefore would not have captured changes in postcode over time. Inspection of PDS reveals that 55% of chil- dren have at least two postcodes in their first year of life and 69% have at least two postcodes by age 5/6 (19% have four or more different postcodes by this age). In our simulation study, we fixed our threshold at a false- match rate of 1%. In practice, choice of appropriate thresholds can be difficult, and is typically chosen based on a sample of manually-reviewed records, or using syn- thetic data [35]. be ignored [32, 33]. In our study, ordering of record pairs based on weights incorporating dependence be- tween identifiers was the same as with traditional match weights, mainly due to a lack of strong dependence be- tween errors in sex, postcode and date of birth observed in HES. However, incorporating dependence into match weights may become more important in data where obvious dependencies do exist, although handling de- pendence between a large number of identifiers may be- come impractical. Discussion (2008) found that dependence between highly corre- lated identifiers (such as expected birth and actual date of birth) had a negative impact on match weights and that this resulted in an incorrect ranking of record pairs ordered by match weight [16]. Similarly, Herzog et al. (2010) found that match weights assigned to non- dependent identifiers were too low in the presence of dependent identifiers [30]. Methods for accounting for dependence between identifiers have also been shown to improve the quality of linkage [31]. Others believe that the impact of dependence between identifiers is small, and that the failure of the independence assumption can Results from our simulation study support a large body of literature showing that substantial bias can be introduced into results of analyses based on data con- taining linkage errors [22–25]. This is particularly im- portant when error is non-random, i.e. dependent on individual-level characteristics, and when there are a large number of missed-matches (e.g. with linkage re- quiring exact matching of identifiers). Evidence from previous studies highlights that the most vulnerable groups are those most likely to be affected by linkage error [2, 26]. In our study, readmission rates estimated using linkage with traditional match weights were Harron et al. BMC Medical Research Methodology (2017) 17:23 Page 7 of 9 Table 2 Simulation study results: estimated readmission rates. The ‘true’ readmission rate was 8.8% NHS number error distribution in simulated data 30%, random 0.5%, random 30%, associated with other identifier errors 30%, associated with ethnicity Traditional match weight % readmitted 7.4 7.4 6.9 7.4 Standard error 0.002 0.002 0.002 0.002 % bias −15.9 −15.9 −21.3 −15.7 Match weight incorporating dependence % readmitted 7.4 7.4 6.9 7.4 Standard error 0.002 0.002 0.002 0.002 % bias −16.0 −16.0 −21.4 −15.8 Attribute-specific match weight % readmitted 8.7 8.7 8.7 8.8 Standard error 0.002 0.002 0.002 0.002 % bias −0.6 −0.6 −0.9 −0.2 Organisation-specific match weight % readmitted 8.8 8.8 8.8 8.8 Standard error 0.002 0.002 0.002 0.002 % bias 0.2 0.2 −0.1 0.2 However, our study was limited by the assumption that agreement on NHS number between HES and PDS indi- cated that records belonged to the same individual. In reality, NHS number is not always a reliable identifier for linkage [20]. If well-completed NHS number is indi- cative of good data quality more generally, we may have underestimated identifier error rates through our study design. Competing interests The authors have no competing interests to declare. The authors have no competing interests to declare. Abbreviations HES: Hospital Episode Statistics; NHS: National Health Service; PDS: Personal Demographic Service Funding Katie Harron is funded by the Wellcome Trust (grant number 103975/Z/14/ Z). This work was also supported by the Administrative Data Research Centre for England funded by the ESRC (grant number ES/L007517/1) and the initial work for this study was undertaken with funding from the ESRC NCRM (grant number ES/F035098/1). This work was discussed at the Isaac Newton Institute for Mathematical Sciences, Cambridge, supported by EPSRC grant no EP/K032208/1. Discussion A major strength of this study was the use of a large, generalizable administrative data source that is fre- quently linked with other datasets and used for commis- sioning and monitoring of the NHS in England and for research. Our study demonstrates the usefulness of the PDS as a reference dataset. There is a lack of published information available on PDS but it holds potential for developing a better understanding of the mechanisms underlying identifier errors, for improving data linkage methods, and for validating identifiers in HES [34]. Fig. 3 Simulation study results: estimated readmission rates by ethnicity, according to NHS number error rate distribution Harron et al. BMC Medical Research Methodology (2017) 17:23 Page 8 of 9 Fig. 4 Simulation study results: absolute bias in estimated readmission rates for NHS number error associated with ethnicity. Results for traditional match weights fall behind those for weights incorporating dependence between identifiers ig. 4 Simulation study results: absolute bias in estimated readmission rates for NHS number error associated with ethnicity. match weights fall behind those for weights incorporating dependence between identifiers dy results: absolute bias in estimated readmission rates for NHS number error associated with ethnicity. Results for traditiona hind those for weights incorporating dependence between identifiers Conclusions for enabling this work by allowing GHJ to access data within NHS Digital. The authors would also like to thank the Isaac Newton Institute for Mathematical Sciences, Cambridge, for support and hospitality during the Data Linkage and Anonymisation programme, where initial findings were presented and discussed. Incorporating information on individual characteristics or organisational variation into match weight calculation can reduce bias associated with errors in linkage, particularly when errors are distributed non-randomly. Continued im- provement of linkage methods will allow more efficient exploitation of administrative data sources, reduce bias as- sociated with linkage of imperfect identifiers and improve the reliability and transparency of analysis based on linked data. This will improve the ability of those working in gov- ernment and health policy, who frequently use research data generated from administrative data sources to inform health policy, to make informed decisions on patient care and health systems. Evaluation of services for specific age or ethnic groups can be important for policy, but as our study shows, results for specific groups can be biased if as- sociated linkage error is not addressed. Careful consider- ation should be given to the trade-off between bespoke linkage strategies for each study (that prioritise the quality of linkage) versus routine linkage systems that maximise efficiency and security. In order for data users to under- stand the limitations of linked data sources, it is vital that information on linkage quality and error rates are made available on release of linked data. Data providers need to improve transparency about data processing before during and after linkage. Availability of data and materials Access to Hospital Episode Statistics and the Personal Demographic Service requires approval from NHS Digital; these data cannot be made publicly available. Simulated aggregate datasets are available from the corresponding author on request. Ethics approval Hospital Episode Statistics were made available by NHS Digital (at the time of the study, named the Health and Social Care Information Centre). As the analysis was a service evaluation to improve the quality of service provided by NHS Digital, which did not directly involve participants in research, the study was exempt from UK ethics approval. GHJ conducted initial analyses internally at NHS Digital; aggregate results tables were then shared with co- authors (with no small cell sizes). The study design and results were shared with NHS Digital staff during meetings between January and May 2016. Author contributions KH and HG conceived the study; GHJ performed initial analyses of data, KH wrote the first draft of the manuscript, analysed aggregate data, and performed the simulation study. RG and HG critically reviewed the manuscript. All authors read and approved the final manuscript. HES: Hospital Episode Statistics; NHS: National Health Service; PDS: Personal Demographic Service References 1. Benchimol EI, et al. The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) Statement. PLoS Med. 2015; 12(10):e1001885. 30. Herzog TH, Scheuren F, Winkler WE. Record linkage. WIREs Comput Stat. 2010;2(5):535–43. 31. Daggy J, et al. A practical approach for incorporating dependence among fields in probabilistic record linkage. BMC Med Res Methodol. 2013;13(1):97. 2. Bohensky M, et al. Data linkage: A powerful research tool with potential problems. BMC Health Serv Res. 2010;10(1):346–52. 2. Bohensky M, et al. Data linkage: A powerful research tool with potential problems. BMC Health Serv Res. 2010;10(1):346–52. 32. Herzog T, Scheuren F, Winkler W. Data quality and record linka techniques. New York: Springer Verlag; 2007. 3. Harron K, et al. Opening the black box of record linkage. J Epidemiol Community Health. 2012;66(12):1198. 3. Harron K, et al. Opening the black box of record linkage. J Epidemiol Community Health. 2012;66(12):1198. 33. Winkler WE. The state of record linkage and current research problems. 1999. 4. Neter J, Maynes E, Ramanathan R. The effect of mismatching on the measurement of response error. J Am Stat Assoc. 1965;60(312):1005–27. 4. Neter J, Maynes E, Ramanathan R. The effect of mismatching on the measurement of response error. J Am Stat Assoc. 1965;60(312):1005–27. 34. NHS England. Understanding the impact of data quality on data 35. Winglee M, Valliant R, Scheuren F. A case study in record linkage. Surv Methodol. 2005;31(1):3–11. 5. Lariscy JT. Differential Record Linkage by Hispanic Ethnicity and Age in Linked Mortality Studies: Implications for the Epidemiologic Paradox. J Aging Health. 2011;23(8):1263–84. 6. Jasilionis D, et al. Ethnic mortality differentials in Lithuania: contradictory evidence from census-linked and unlinked mortality estimates. J Epidemiol Community Health. 2011;66(6):e7. 7. Schmidlin K, et al. Impact of unlinked deaths and coding changes on mortality trends in the Swiss National Cohort. BMC Med Inform Decis Mak. 2013;13(1):1–11. 8. Ford JB, Roberts CL, Taylor LK. Characteristics of unmatched maternal and baby records in linked birth records and hospital discharge data. Paediatr Perinat Epidemiol. 2006;20(4):329–37. 8. Ford JB, Roberts CL, Taylor LK. Characteristics of unmatched maternal and baby records in linked birth records and hospital discharge data. Paediatr Perinat Epidemiol. 2006;20(4):329–37. 9. Clark D. Practical introduction to record linkage for injury research. Inj Prev. 2004;10(3):186–91. 10. Sayers, A. et al. Probabilistic record linkage. Int J Epidemiol. 2015;45(3):954– 964. 11. Fellegi IP, Sunter AB. A theory for record linkage. J Am Stat Assoc. 1969; 64(328):1183–210. 12. Acknowledgements h h ld l k The authors would like to thank Wataru Suzuki and Trevor Anders from NHS Digital for their help with this collaboration. We are grateful to NHS Digital Page 9 of 9 Page 9 of 9 Page 9 of 9 Harron et al. BMC Medical Research Methodology (2017) 17:23 Harron et al. BMC Medical Research Methodology (2017) 17:23 Harron et al. BMC Medical Research Methodology (2017) 17:23 Received: 26 September 2016 Accepted: 2 February 2017 Received: 26 September 2016 Accepted: 2 February 2017 Received: 26 September 2016 Accepted: 2 February 2017 28. Goldstein H, Harron K, Wade A. The analysis of record-linked data using multiple imputation with data value priors. Stat Med. 2012;31(28):3481–93. 29. Winkler WE. Chapter 2: Probabilistic linkage. In: Harron K, Dibben C, Goldstein H, editors. Methodological Developments in Data Linkage. London: Wiley; 2015. References Jaro M. Probabilistic linkage of large public health data files. Stat Med. 1995; 14(5–7):491–8. 13. Christen P. A two-step classification approach to unsupervised record linkage. In Proceedings of the sixth Australasian conference on Data mining and analytics-Volume 70. Darlinghurst: Australian Computer Society, Inc.; 2007. 14. Yancey W. Improving EM algorithm estimates for record linkage parameters. in Joint Statistical Meetings - Section on Survey Research Methods. 2004. 15. Hagger-Johnson G, et al. Identifying false matches in anonymised hospital administrative data without patient identifiers. Health Serv Res. 2014;50(4):1162–78. 15. Hagger-Johnson G, et al. Identifying false matches in anonymised hospital administrative data without patient identifiers. Health Serv Res. 2014;50(4):1162–78. 16. Tromp M, et al. Ignoring dependency between linking variables and its impact on the outcome of probabilistic record linkage studies. J Am Med Inform Assoc. 2008;15(5):654–60. 16. Tromp M, et al. Ignoring dependency between linking variables and its impact on the outcome of probabilistic record linkage studies. J Am Med Inform Assoc. 2008;15(5):654–60. 17. Health and Social Care Information Centre. Methodology for creation of the HES Patient ID (HESID). 2014. Author details 1 24. Brenner H, Schmidtmann I. Effects of record linkage errors on disease registration studies. Method Inf Med. 1998;37(1):69–74. 1London School of Hygiene and Tropical Medicine, 15-17 Tavistock Place, London WC1 H 9SH, UK. 2Administrative Data Research Centre for England, UCL, 222 Euston Road, London NW1 2DA, UK. 3Administrative Data Research Centre for England and UCL Great Ormond Street Institute of Child Health, 30 Guilford Street, London WC1 N 1EH, UK. 4University of Bristol, 25. Krewski D, et al. The effect of record linkage errors on risk estimates in cohort mortality studies. Surv Methodol. 2005;31(1):13–21. cohort mortality studies. Surv Methodol. 2005;31(1):13–21. 26. Bohensky M. Chapter 4: Bias in data linkage studies. In: Harron K, Dibben C, Goldstein H, editors. Methodological Developments in Data Linkage. London: Wiley; 2015. Administrative Data Research Centre for England and UCL Great Ormond Street Institute of Child Health, 30 Guilford Street, London WC1 N 1EH, UK. 27. Hagger-Johnson G, et al. Data linkage errors in hospital administrative data when applying a pseudonymisation algorithm to paediatric intensive care records. BMJ Open. 2015;5(8):e008118. Submit your next manuscript to BioMed Central and we will help you at every step: Submit your next manuscript to BioMed Central and we will help you at every step: 18. Stata. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP; 2015. • We accept pre-submission inquiries • Our selector tool helps you to ﬁnd the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit p y y p 19. Wijlaars L.P.M.M. et al. Who comes back with what: a retrospective database study on reasons for emergency readmission to hospital in children and young people in England. Arch Dis Child, 2016. in press. 20. Aldridge RW, et al. Accuracy of Probabilistic Linkage Using the Enhanced Matching System for Public Health and Epidemiological Studies. PLoS One. 2015;10(8):e0136179. 21. Hippisley-Cox J. Validity and completeness of the NHS Number in primary and secondary care: electronic data in England 1991–2013. In Project Report. University of Nottingham; 2013. 22. Moore CL, et al. A new method for assessing how sensitivity and specificity of linkage studies affects estimation. PLoS One. 2014;9(7):e103690. 22. Moore CL, et al. A new method for assessing how sensitivity and specificity of linkage studies affects estimation. PLoS One. 2014;9(7):e103690. 23. Harron K, et al. Evaluating bias due to data linkage error in electronic healthcare records. BMC Med Res Methodol. 2014;14(1):36. 23. Harron K, et al. Evaluating bias due to data linkage error in electronic healthcare records. BMC Med Res Methodol. 2014;14(1):36.
W3024239115.txt	https://www.researchsquare.com/article/rs-17638/latest.pdf	en		NON-STEROIDAL ANTI-INFLAMMATORY DRUG-INDUCED ENTEROPATHY AS A MAJOR RISK FACTOR FOR SMALL BOWEL BLEEDING	Endoscopy	2,020	cc-by	4,936	Non-steroidal anti-inflammatory drug-induced enteropathy as a major risk factor for small bowel bleeding: A retrospective study Doo-Ho Lim Ulsan University Hospital Kyoungwon Jung Kosin University Gospel Hospital Seung Bum Lee (  sblee@uuh.ulsan.kr ) Ulsan University Hospital https://orcid.org/0000-0002-5880-5659 In Kyu Park Ulsan University Hospital Hee Jeong Cha Ulsan University Hospital Jae Ho Park Ulsan University Hospital Byung Gyu Kim Ulsan University Hospital Seok Won Jung Ulsan University Hospital In Du Jeong Ulsan University Hospital Jae Hyun Kim Kosin University Gospel Hospital Sung Eun Kim Kosin University Gospel Hospital Won Moon Kosin University Gospel Hospital Moo In Park Kosin University Gospel Hospital Seun Ja Park Kosin University Gospel Hospital Research article Page 1/17 Keywords: Capsule endoscopy, Nonsteroidal anti-inflammatory drugs, Obscure gastrointestinal bleeding, Iron deficiency anemia, Angioectasia Posted Date: May 26th, 2020 DOI: https://doi.org/10.21203/rs.3.rs-17638/v2 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Version of Record: A version of this preprint was published at BMC Gastroenterology on June 8th, 2020. See the published version at https://doi.org/10.1186/s12876-020-01329-5. Page 2/17 Abstract Background: Small bowel (SB) bleeding accounts for 5% of all gastrointestinal (GI) bleeding cases and 80% of obscure GI bleeding cases. Although angioectasia is the common etiology of SB bleeding, nonsteroidal anti-inflammatory drug (NSAID)-induced SB lesions are also reported as a major cause in studies from Eastern countries. Herein, we assessed the frequency of occurrence of NSAID-induced SB lesions in Korean patients with obscure GI bleeding. Methods: We retrospectively analyzed medical records of all consecutive patients aged ≥18 years who underwent capsule endoscopy from March 2018 to February 2019 at Ulsan University Hospital and Kosin University Gospel Hospital. Results: Of the 83 subjects (all Korean; mean age ± standard deviation: 59 ± 18 years; age range: 18–84 years; men: n=52; women: n=31), 55 (66.2%) had stool with clear blood and 28 (33.8%) had normal stool with iron deficiency anemia. The detection rate of SB bleeding and lesions using capsule endoscopy was 72.3% (60 of 83 patients). A significantly higher frequency (40 of 51) of ulcerative/erosive lesions than other causes was observed in patients with inactive bleeding but visible SB lesions. As a result, NSAIDinduced enteropathy accounted for 30.1% of 83 patients with obscure GI bleeding (25 of the all 60 SB bleeding cases). Conclusions: Contrary to what is reported for patients in Western countries, this study in Korean patients showed an improved diagnostic yield of capsule endoscopy for obscure GI bleeding and that NSAIDinduced enteropathy was the most common etiology of SB bleeding. Aggressive small intestine examination is required for patients with unexplained GI bleeding. Background Obscure gastrointestinal (GI) bleeding is defined as persistent or recurrent bleeding of unknown origin despite repetitive diagnostic testing, including upper GI endoscopy and colonoscopy.1 Small bowel (SB) bleeding accounts for 80% of obscure GI bleeding cases and 5% of all GI bleeding cases. Unexplained iron deficiency anemia (IDA) manifests in 30–40% of obscure GI bleeding cases.2 Capsule endoscopy is the first-line examination in cases of obscure GI bleeding and the third diagnostic test after negative upper and lower endoscopy results in cases of ongoing overt bleeding.3-6 Whereas angioectasia is the most common etiology of SB bleeding in Western countries,4, 7, 8 nonsteroidal anti-inflammatory drug (NSAID)-induced SB lesions are the major cause reported in Japanese studies.9, 10 NSAIDs are widely prescribed in most clinical conditions; however, it is well known that these drugs can cause GI complications.11 Although various studies have reported an association between NSAID intake and GI adverse effects,12, 13 the role of NSAIDs in SB bleeding remains to be elucidated. Therefore, we aimed to assess the frequency of NSAID-induced SB lesions in Korean patients with obscure GI bleeding, who underwent capsule endoscopy. Page 3/17 Methods Study population We retrospectively evaluated the medical records of all consecutive patients aged ≥18 years with obscure GI bleeding who underwent capsule endoscopy from March 2018 to February 2019 at Ulsan University Hospital and Kosin University Gospel Hospital. The indication for the test included bloody stool in patients with an unidentified source of bleeding and lesions during diagnostic evaluations, including upper GI endoscopy, colonoscopy, and abdominal computed tomography (CT) in the last 3 months.5 We also included patients who had no visible bloody stool, showed persistent or repeated exacerbation of IDA despite iron supplementation for >6 months, and had no hemorrhagic lesions identified during the above screening methods in the last 3 months.14 If CT-scan had been performed in the last 1 year, in some patients, the test was not performed again when anemia was confirmed. Rebleeding was defined as recurrent anemia (≥2 g/dL decrease in the hemoglobin level), overt melena/ hematochezia, or occult GI bleeding during the follow-up period.15 This study was approved by the Institutional Review Boards (IRB of the Ulsan University Hospital (IRB No. 2019-10-014) and the Kosin University Gospel Hospital (IRB No. 2019-10-001)). The requirement for informed consent from patients was waived because patient records and information were de-identified prior to analysis. Capsule endoscopy Capsule endoscopy was performed using Pillcam® (SB3, Given Imaging Ltd., Yoqneam, Israel) and MiroCam® (MC1200, IntroMedic Ltd., Seoul, Korea) devices. Given that both pieces of equipment have 12 hours of battery life, all endoscopies were performed for 12 hours. All patients fasted for 12 hours and received 40 mg of oral simethicone before the procedure to prevent air bubbles from forming.16 Moreover, we used 2 L polyethylene glycol solution at least 2–16 hours before examination to improve the quality of the SB image.17, 18 Patients with bloody stool were examined in the hospital in all cases; based on the time at which the blood stool was noted, the capsule endoscope was able to start testing within a median of 51 hours (inter-quartile range, 24 to 96 hours). Those who were referred due to anemia had an outpatient examination if there was no bloody stool. To prevent capsule retention, we checked the presence of small bowel stenosis on the CT scan that was performed at least 1 year before the capsule endoscopy. Five gastroenterologists with extensive experience in GI endoscopy reviewed and extensively discussed all capsule video images to reach a diagnosis. An erosion was defined as a roundish area of mucosal disruption smaller than a diameter equivalent to 1 circular fold of the mucosa (also called the valves of Kerckring, valvulae conniventes, or plicae circulares).19 An ulcerative lesion was defined as a mucosal penetration with a diameter larger than 1 circular fold of the mucosa.15 The etiology of ulcerated lesions was determined based on clinical information and endoscopic findings. We examined NSAIDinduced enteropathy based on the following criteria: history of NSAID use within the previous 1 month; endoscopic findings, including ulcers, erosions, scar changes, or luminal stenosis; improvement of the clinical course and/or endoscopic findings after cessation of NSAID use; and exclusion of other Page 4/17 etiologies, including infection, inflammatory bowel disease, or malignancy.12, 20, 21 There is no criterion regarding the duration of NSAID use to diagnose NSAID-induced enteropathy. However, because SB injuries were observed in 68% of healthy volunteers taking NSAIDs for only 1-2 weeks,20, 22 we chose to include patients who had history of NSAID use within the previous 1 month, as was done in a previous study.21 Statistical analysis Continuous variables were compared using Student's t-test and categorical variables were analyzed with a chi-square or Fisher’s exact test. A two-tailed p-value <0.05 was considered statistically significant. All statistical analyses were performed using the SPSS statistical package for Windows, Version 24.0 (SPSS Inc., Chicago, IL, USA). Results Sample analysis We assessed 83 subjects (52 men and 31 women) who underwent capsule endoscopy during the study period (see Table 1 for patient characteristics). The subjects were Korean and aged 18–84 years (mean ± standard deviation: 59 ± 18 years). Of these, 55 (66.2%) patients had clear bloody stool and 28 (33.8%) had normal stool; all had IDA (Figure 1). The mean small bowel transit time was 6.39 hours (range, 2.4211.38 hours). Capsule retention did not occur in this study. Most patients (92.4%) showed scores of 3 or 2 on the Boston bowel preparation scale.23 A total of 37 (44.6%) patients had a history of taking NSAIDs/low-dose aspirin (Table 2). The brand name and dosage of NSAIDs were not clearly identified in the medical records. Musculoskeletal disorders were the most common indication for taking NSAIDs, while coronary artery disease and other cardiovascular diseases were the main indications for taking aspirin. The duration of NSAID use varied from 2 weeks to 10 years prior to the capsule endoscopy. An association with SB bleeding was considered only when NSAIDs/low-dose aspirin was taken within the previous 1 month. Of the 37 patients, 33 had SB lesions; of these, 25 patients were diagnosed with NSAID-induced enteropathy without any other cause of SB bleeding. The main treatments for NSAIDinduced enteropathy were discontinuation of NSAIDs or the replacement of NSAIDs with drugs with a low risk of bleeding. Among 37 patients with a history of low-dose aspirin or NSAID medication, only 3 who were treated with selective cyclooxygenase (COX)-2 inhibitors had normal stool. Most patients with active bleeding during endoscopy had angioectasia (8 of 9 patients). A fecal occult blood test was performed in patients with normal stool (16 of 28) and the results were negative in 12 of these patients, among which 7 had no SB lesions. A significantly higher frequency (40 of 51) of ulcerative/erosive lesions was observed in patients with inactive bleeding but visible SB lesions than other causes; among these patients, 25 had a history of low-dose aspirin or NSAID medication. As a result, NSAID-induced enteropathy accounted for 30.1% of 83 patients with obscure GI bleeding (25 of all 60 SB bleeding cases). Four cases of previously undiagnosed Crohn's disease and 1 case of tuberculosis were identified. Page 5/17 Other uncommon etiologies associated with unknown GI bleeding included radiation ileitis (n=3) and SB polyps (n=3). Evaluation of rebleeding and treatment Prior to capsule endoscopy, 58 (69.9%) patients had anemia with hemoglobin <10 g/dL. SB bleeding was predominant in >50% (17 of 28) of patients with normal stool. All patients with SB bleeding due to angioectasia showed bloody stool, while ulcerative lesions were prevalent in patients with normal stool. Rebleeding occurred in 7 patients during the 12-month follow-up period (Table 3). Two patients with Crohn's disease who presented with persistent IDA required specific treatments: One patient was given thiopurine, but the anemia continued and so thiopurine was replaced with methotrexate; no recurrence of anemia has been reported since then. In the other case, administration of immunosuppressive drugs, including thiopurine, failed continuously, mainly due to leukocytopenia associated with persistent IDA; this patient underwent resection of a short segment of the small bowel in which ulcers were observed. After the surgery, anemia has not recurred so far. One patient who was considered to have Crohn’s disease initially, presented with multiple lymphadenopathy during the follow-up period and was finally diagnosed with tuberculosis. Among 2 patients with radiation ileitis, 1 underwent surgical treatment due to recurrent life-threatening bloody stool, whereas in the other patient with persistent IDA, medical treatment was continued with prolonged oral iron replacement. A patient with a rebleeding episode from angioectasia received surgery. Lastly, one patient who received NSAIDs to treat fibromyalgia for >2 years presented with an ulcerative lesion, had unresolved anemia, and showed recurrent abdominal pain, which lasted for >6 months, despite adequate iron administration and stopping NSAIDs. Capsule endoscopy showed multiple small bowel ulcers and an additional abdominal CT at the recurrence of bleeding revealed multiple strictures of the SB, which was not detected previous CT-scan performed 8 months ago; eventually segmental resection of the SB was performed, through which 5 sites of stenosis in the distal ileum were identified (Figure 2). No recurrence of anemia or abdominal pain was present 6 months postoperatively. Discussion Capsule endoscopy for unexplained GI bleeding has shown diagnostic yields of 57–62%,24 with the most common diagnosis being angioectasia (50%), followed by ulcers (26.8%) and tumors (8.8%), as suggested in a recent systematic literature review.14 The present study revealed a detection rate of SB bleeding and lesions using capsule endoscopy of 72.3% (60 of 83 patients), of which NSAID-induced enteropathy accounted for 25 of all 60 SB bleeding cases. The diagnostic yield of small bowel lesions in our study was higher than in a previous nationwide Korean study of capsule endoscopy,25 but the prevalence of ulcerative/erosive lesions was comparable (40/60 vs 106/157). Similar to our results, a Korean study using balloon-assisted enteroscopy found that the most common type of SB lesions associated with obscure GI bleeding were mucosal injury (56%), followed by vascular lesions (18.7%).26 Because of limited availability of data regarding the use of NSAIDs/low-dose aspirin, we could not compare the prevalence of NSAID-induced enteropathy with the results from these studies. However, Page 6/17 considering the meaningful association between NSAID use and SB injury presented in previous studies,22, 27 NSAID-induced enteropathy could be the major cause of obscure GI bleeding in Eastern countries. Contrary to previous reports from Western countries, we observed a higher occurrence rate of ulcerative/erosive lesions than angioectasia in patients with obscure GI bleeding, supporting the implication of low-dose aspirin or NSAID medications in the disease etiology. A recent systematic review and meta-analysis suggested that the optical timing of capsule endoscopy would be within 2 days, to improve the diagnostic yield.28 For patients with clear bloody stool in our study, the capsule endoscope was able to start testing within a median of 51 hours (inter-quartile range, 24 to 96 hours). In the case of persistent melena or hematochezia enough to show symptoms, the patients usually visited the emergency room and in these patients, the capsule endoscopy could be performed within at least 48 hours, but if patients showed intermittent blood stool or if symptoms related to bleeding or anemia were not clear, capsule endoscopy would not be conducted within 2 days. In real practice, further improvement is needed in this respect. NSAIDs are frequently used anti-inflammatory analgesic agents that represent 7.7% of worldwide prescriptions, of which 90% are prescribed to elderly (>65 years) patients.29 The mechanism of NSAIDinduced enteropathy is supposed to be mediated through COX inhibition.30 Administration of low-dose aspirin (an irreversible nonselective COX inhibitor) is also associated with SB mucosal injuries; large erosions or ulcers were reported in 60% of healthy volunteers who took 100 mg of low-dose enteric-coated aspirin.31, 32 In the present study, a history of low-dose aspirin or NSAID use was common in patients with obscure GI bleeding (44.6%), showing a higher frequency of SB ulcerative lesions than of other sources of lesions (angioectasia, n=6; upper or lower GI bleeding, n=6). Prostaglandins (PG) play an important role in regulating GI blood flow and mucus production; therefore, NSAID-induced suppression of PG production has been implicated in SB damage.33, 34 Previously, COX-1 inhibition was regarded to be dominantly related to GI mucosal injuries. However, in a recent animal model study, damage to the SB developed only when both COX-1 and COX-2 were inhibited.35 This result indicates that COX-2-derived PGs also play an important role in the maintenance of tissue integrity and repair of mucosal injury. However, clinical research has shown conflicting results. Several studies have shown an improved GI safety profile with selective COX-2 inhibitors compared to nonselective NSAIDs,36, 37 while others studies indicated no significant differences in SB injuries between these NSAIDs.38, 39 In the current study, among 37 patients with a history of low-dose aspirin or NSAID medications, the 3 treated with selective COX-2 inhibitors had normal stool, suggesting favorable GI outcomes with selective COX-2 inhibitor therapy. Considering that selective COX-2 inhibitors are not completely safe for the SB, further long-term studies with a larger sample size are warranted to establish the safety profile of the drug in the SB. Furthermore, the impact of capsule endoscopy on clinical outcomes remains controversial despite reports of SB mucosal damage in 70% of patients taking NSAIDs,22, 40 because it remains unclear whether SB mucosal injuries contribute to significant bleeding.41 Although patients with NSAID-induced SB injury Page 7/17 show low frequency of severe bleeding in the SB,42 rebleeding rates of 21–35% have been reported in patients with SB ulcerations during a mean follow-up period of 17.1–29.7 months.15, 43 These reports suggest a clinical implication of SB ulcers, which cannot be ignored. The most effective method of preventing NSAID-induced enteropathy is discontinuation of NSAIDs if possible.12 Previously, there was no strategy to prevent NSAID-induced enteropathy.13, 44 However, a recent study reported the effectiveness of misoprostol in the treatment of SB ulcer bleeding associated with aspirin.45 On the contrary, lesions that induce stenosis, which may not be treated with medication alone, require endoscopic or surgical interventions.46 In the present study, a patient suffering from fibromyalgia developed SB stricture after NSAID medications for >2 years and eventually underwent surgical resection. This study has several limitations. First, it was a retrospective analysis with a small sample size, with insufficient power to detect a significant effect. Second, because balloon-assisted enteroscopy was not routinely performed, pathological findings could not confirm SB ulcers. Third, the short follow-up period prevented the adequate assessment of risk factors for rebleeding. Finally, the fecal occult blood test could not be performed in 12 of 28 patients who presented with normal stool, thereby limiting the interpretation of the results. Conclusions This study showed an improved diagnostic yield of capsule endoscopy for obscure GI bleeding and reaffirmed that NSAID-induced enteropathy is the most common etiology of SB bleeding in Korean patients. Therefore, aggressive clinical management, including SB capsule endoscopy, should be considered for patients with unexplained GI bleeding or drug-refractory iron deficiency anemia, particularly during aspirin or NSAID medications. Declarations Acknowledgement Not Applicable. Author contributions Study coordination and drafting of the manuscript: DHL and KJ. Study coordination and critical review of the manuscript: DHL and LSB. Review of capsule endoscopy images: KJ, LSB, JHP, JHK, and SEK. Data supply and approval of the final version of this manuscript: IKP, HJC, BGK, IDJ, and MW. Data analysis and approval of the final version of this manuscript: SWJ, MIP, and SJP. All authors read and approved the final manuscript. Funding Page 8/17 No funding was obtained for the study. Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on request. Ethics approval and consent to participate This study was approved by the Institutional Review Board (IRB, the local ethical committee) of the Ulsan University Hospital (IRB No. 2019-10-014) and the Kosin University Gospel Hospital (IRB No. 2019-10001). Consent for publication Not applicable. The requirement for informed consent from patients was waived by the IRB because patient records and information were de-identified. Competing interests The authors declare no competing interests. Abbreviations GI, gastrointestinal; SB, small bowel; NSAIDs, nonsteroidal anti-inflammatory drugs; COX, cyclooxygenase; PG, prostaglandin; CT, computed tomography References 1. Leighton JA, Goldstein J, Hirota W, et al. Obscure gastrointestinal bleeding. Gastrointest Endosc 2003;58:650-5. 2. Singh A, Baptista V, Stoicov C, et al. Evaluation of small bowel bleeding. Curr Opin Gastroenterol 2013;29:119-24. 3. Gerson LB, Fidler JL, Cave DR, et al. ACG Clinical Guideline: Diagnosis and Management of Small Bowel Bleeding. Am J Gastroenterol 2015;110:1265-87; quiz 1288. 4. Pennazio M, Santucci R, Rondonotti E, et al. Outcome of patients with obscure gastrointestinal bleeding after capsule endoscopy: report of 100 consecutive cases. Gastroenterology 2004;126:64353. 5. Raju GS, Gerson L, Das A, et al. American Gastroenterological Association (AGA) Institute technical review on obscure gastrointestinal bleeding. Gastroenterology 2007;133:1697-717. 6. Iddan G, Meron G, Glukhovsky A, et al. Wireless capsule endoscopy. Nature 2000;405:417. Page 9/17 7. Saurin JC, Delvaux M, Gaudin JL, et al. Diagnostic value of endoscopic capsule in patients with obscure digestive bleeding: blinded comparison with video push-enteroscopy. Endoscopy 2003;35:576-84. 8. Ell C, Remke S, May A, et al. The first prospective controlled trial comparing wireless capsule endoscopy with push enteroscopy in chronic gastrointestinal bleeding. Endoscopy 2002;34:685-9. 9. Yamamoto H, Kita H, Sunada K, et al. Clinical outcomes of double-balloon endoscopy for the diagnosis and treatment of small-intestinal diseases. Clin Gastroenterol Hepatol 2004;2:1010-6. 10. Niikura R, Yamada A, Nagata N, et al. New predictive model of rebleeding during follow-up of patents with obscure gastrointestinal bleeding: A multicenter cohort study. J Gastroenterol Hepatol 2016;31:752-60. 11. Sostres C, Gargallo CJ, Lanas A. Nonsteroidal anti-inflammatory drugs and upper and lower gastrointestinal mucosal damage. Arthritis Res Ther 2013;15 Suppl 3:S3. 12. Shin SJ, Noh CK, Lim SG, et al. Non-steroidal anti-inflammatory drug-induced enteropathy. Intest Res 2017;15:446-455. 13. Lim YJ, Yang CH. Non-steroidal anti-inflammatory drug-induced enteropathy. Clin Endosc 2012;45:138-44. 14. Liao Z, Gao R, Xu C, et al. Indications and detection, completion, and retention rates of small-bowel capsule endoscopy: a systematic review. Gastrointest Endosc 2010;71:280-6. 15. Aoki T, Yamada A, Hirata Y, et al. Etiology and long-term rebleeding of endoscopic ulcerative lesions in the small bowel in patients with obscure gastrointestinal bleeding: A multicenter cohort study. J Gastroenterol Hepatol 2018;33:1327-1334. 16. Albert J, Gobel CM, Lesske J, et al. Simethicone for small bowel preparation for capsule endoscopy: a systematic, single-blinded, controlled study. Gastrointest Endosc 2004;59:487-91. 17. Yang L, Wang X, Gan T, et al. Polyethylene Glycol for Small Bowel Capsule Endoscopy. Gastroenterol Res Pract 2017;2017:7468728. 18. Dai N, Gubler C, Hengstler P, et al. Improved capsule endoscopy after bowel preparation. Gastrointest Endosc 2005;61:28-31. 19. Graham DY, Opekun AR, Willingham FF, et al. Visible small-intestinal mucosal injury in chronic NSAID users. Clin Gastroenterol Hepatol 2005;3:55-9. 20. Hayashi Y, Yamamoto H, Kita H, et al. Non-steroidal anti-inflammatory drug-induced small bowel injuries identified by double-balloon endoscopy. World J Gastroenterol 2005;11:4861-4. 21. Matsumoto T, Kudo T, Esaki M, et al. Prevalence of non-steroidal anti-inflammatory drug-induced enteropathy determined by double-balloon endoscopy: a Japanese multicenter study. Scand J Gastroenterol 2008;43:490-6. 22. Maiden L, Thjodleifsson B, Theodors A, et al. A quantitative analysis of NSAID-induced small bowel pathology by capsule enteroscopy. Gastroenterology 2005;128:1172-8. Page 10/17 23. Lai EJ, Calderwood AH, Doros G, et al. The Boston bowel preparation scale: a valid and reliable instrument for colonoscopy-oriented research. Gastrointest Endosc 2009;69:620-5. 24. Xin L, Liao Z, Jiang YP, et al. Indications, detectability, positive findings, total enteroscopy, and complications of diagnostic double-balloon endoscopy: a systematic review of data over the first decade of use. Gastrointest Endosc 2011;74:563-70. 25. Min YW, Kim JS, Jeon SW, et al. Long-term outcome of capsule endoscopy in obscure gastrointestinal bleeding: a nationwide analysis. Endoscopy 2014;46:59-65. 26. Jeon SR, Kim JO, Kim HG, et al. Changes over time in indications, diagnostic yield, and clinical effects of double-balloon enteroscopy. Clin Gastroenterol Hepatol 2012;10:1152-6. 27. Maiden L. Capsule endoscopic diagnosis of nonsteroidal antiinflammatory drug-induced enteropathy. J Gastroenterol 2009;44 Suppl 19:64-71. 28. Uchida G, Nakamura M, Yamamura T, et al. Systematic review and meta-analysis of the diagnostic and therapeutic yield of small bowel endoscopy in patients with overt small bowel bleeding. Dig Endosc 2020. 29. Tai FWD, McAlindon ME. NSAIDs and the small bowel. Curr Opin Gastroenterol 2018;34:175-182. 30. Davies NM, Saleh JY, Skjodt NM. Detection and prevention of NSAID-induced enteropathy. J Pharm Pharm Sci 2000;3:137-55. 31. Iwamoto J, Mizokami Y, Saito Y, et al. Small-bowel mucosal injuries in low-dose aspirin users with obscure gastrointestinal bleeding. World J Gastroenterol 2014;20:13133-8. 32. Shiotani A, Haruma K, Nishi R, et al. Randomized, double-blind, pilot study of geranylgeranylacetone versus placebo in patients taking low-dose enteric-coated aspirin. Low-dose aspirin-induced small bowel damage. Scand J Gastroenterol 2010;45:292-8. 33. Robert A, Asano T. Resistance of germfree rats to indomethacin-induced intestinal lesions. Prostaglandins 1977;14:333-41. 34. Fang WF, Broughton A, Jacobson ED. Indomethacin-induced intestinal inflammation. Am J Dig Dis 1977;22:749-60. 35. Sigthorsson G, Simpson RJ, Walley M, et al. COX-1 and 2, intestinal integrity, and pathogenesis of nonsteroidal anti-inflammatory drug enteropathy in mice. Gastroenterology 2002;122:1913-23. 36. Mizukami K, Murakami K, Yamauchi M, et al. Evaluation of selective cyclooxygenase-2 inhibitorinduced small bowel injury: randomized cross-over study compared with loxoprofen in healthy subjects. Dig Endosc 2013;25:288-94. 37. Inoue T, Iijima H, Arimitsu J, et al. Amelioration of small bowel injury by switching from nonselective nonsteroidal anti-inflammatory drugs to celecoxib in rheumatoid arthritis patients: a pilot study. Digestion 2014;89:124-32. 38. Maiden L, Thjodleifsson B, Seigal A, et al. Long-term effects of nonsteroidal anti-inflammatory drugs and cyclooxygenase-2 selective agents on the small bowel: a cross-sectional capsule enteroscopy study. Clin Gastroenterol Hepatol 2007;5:1040-5. Page 11/17 39. Maehata Y, Esaki M, Morishita T, et al. Small bowel injury induced by selective cyclooxygenase-2 inhibitors: a prospective, double-blind, randomized clinical trial comparing celecoxib and meloxicam. J Gastroenterol 2012;47:387-93. 40. Maiden L, Elliott T, McLaughlin SD, et al. A blinded pilot comparison of capsule endoscopy and small bowel histology in unresponsive celiac disease. Dig Dis Sci 2009;54:1280-3. 41. Niikura R, Yamada A, Maki K, et al. Associations between drugs and small-bowel mucosal bleeding: Multicenter capsule-endoscopy study. Dig Endosc 2018;30:79-89. 42. Chan FK, Lanas A, Scheiman J, et al. Celecoxib versus omeprazole and diclofenac in patients with osteoarthritis and rheumatoid arthritis (CONDOR): a randomised trial. Lancet 2010;376:173-9. 43. Shinozaki S, Yamamoto H, Yano T, et al. Long-term outcome of patients with obscure gastrointestinal bleeding investigated by double-balloon endoscopy. Clin Gastroenterol Hepatol 2010;8:151-8. 44. Park SC, Chun HJ, Kang CD, et al. Prevention and management of non-steroidal anti-inflammatory drugs-induced small intestinal injury. World J Gastroenterol 2011;17:4647-53. 45. Kyaw MH, Otani K, Ching JYL, et al. Misoprostol Heals Small Bowel Ulcers in Aspirin Users With Small Bowel Bleeding. Gastroenterology 2018;155:1090-1097 e1. 46. Slesser AA, Wharton R, Smith GV, et al. Systematic review of small bowel diaphragm disease requiring surgery. Colorectal Dis 2012;14:804-13. Tables Table 1. Clinical characteristics of patients Page 12/17 Characteristic Overall Bloody stool Normal stool (n=83) (n=55) (n=28) 59 ± 18 61 ± 19 56 ± 15 0.185 Sex, male 52 (62.7) 37 (67.3) 15 (53.6) 0.222 Aspirin/NSAID 37 (44.6) 25 (45.5) 12 (42.9) 0.822 Hemoglobin (g/dL) 8.7 ± 3.0 8.7 ± 2.8 8.9 ± 3.5 0.785 Albumin (g/dL) 3.8 ± 0.6 3.7 ± 0.7 3.9 ± 0.6 0.191 Small bowel bleeding 60 (72.3) 43 (78.2) 17 (60.7) 0.093 Angioectasia 14 (23.3) 14 (32.6) 0 (0.0) NSAID-induced enteropathy 25 (41.7) 15 (34.9) 10 (58.8) Other causes† 21 (35.0) 14 (32.6) 7 (41.2) Jejunum 16 (26.7) 14 (36.2) 2 (11.8) Ileum 38 (63.3) 26 (60.5) 12 (70.6) 6 (10.0) 3 (7.0) 3 (17.6) Age, yr. p-value Specific lesion* 0.024 Site of bleeding* Indeterminate 0.169 The denominator is 60 patients with confirmed small bowel lesion. † Other causes of small bowel bleeding, including Crohn’s disease, intestinal tuberculosis, radiation ileitis, small bowel polyps, and small bowel lesions with unknown causes. Categorical and continuous variables are presented as number (%) and mean ± SD, respectively. SD, standard deviation; NSAID, nonsteroidal anti-inflammatory drug; Hb, hemoglobin. Table 2. Details of NSAIDs/low-dose aspirin Page 13/17 N=37 (100%) (NSAID=17 / Low-dose aspirin=21) Indication Coronary artery disease 11 (29.7) Other cardiovascular disease 9 (24.3) Musculoskeletal disease 15 (40.5) Unknown 3 (8.1) Duration <1 month 2 (5.4) 1-6 months 11 (29.7) >6 months 22 (59.5) Unknown 2 (5.4) Concurrent medication Clopidogrel 6 (16.2) Warfarin/NOAC 7 (18.9) Other anti-platelet agent 3 (8.1) None 21 (56.8) Site of bleeding Small bowel 33 (89.2) Stomach 2 (5.4) Undetermined 2 (5.4) Etiology of small bowel bleeding† NSAID-induced enteropathy 25 (75.8) Angioectasia 7 (21.2) Others 1 (3.0) Initial management of NSAID-induced enteropathy‡ Discontinue NSAID/low-dose aspirin 20 (80.0) Continue NSAID/low-dose aspirin 2 (8.0) + Mucosal protective agent§ Unknown 3 (12.0) * One patient had a history of both NSAID and low dose aspirin. Page 14/17 † The denominator is 33 patients with confirmed small bowel bleeding on capsule endoscopy. ‡ The denominator is 25 patients with small bowel bleeding associated with NSAID/low-dose aspirin. § Mucosal protective agent indicates rebamipide, sucralfate, or sodium alginate. NSAID, nonsteroidal anti-inflammatory drug; NOAC, non-vitamin K antagonist oral anticoagulant. Table 3. Rebleeding cases No. Indication Cause of for capsule Final diagnosis Timing of Specific Treatment small bowel rebleeding therapy outcome bleeding after Thiopurine Persistent endoscopy 1 Bloody diagnosis Ulcer stool 2 Crohn’s 23 days disease Bloody Diffuse scar Radiation stool change ileitis* IDA‡ 87 days Surgery No further bleeding 3 Bloody Ulcer stool Crohn’s 117 days disease 5- Persistent aminosalicylic IDA‡ acid 4 5 6 7 Bloody Ulcer with Radiation stool stricture ileitis* IDA Ulcer Intestinal Tbc† IDA Bloody 49 days Ulcer with NSAID-induced stricture enteropathy Angioectasia Angioectasia 146 days 15 days Iron Follow-up replacement loss Anti-Tbc Recovered medication IDA Surgery Recovered IDA 330 days stool Surgery No further bleeding Page 15/17 * Indication for radiation therapy was uterine cervical cancer. † The patient who was considered as having Crohn’s disease initially, presented with multiple lymphadenopathy and was finally diagnosed with tuberculosis. ‡ Patient No.1 was given methotrexate instead of thiopurine and no recurrence of anemia has been reported since then; Patient No.3 underwent resection of a short segment of the small bowel and anemia has not recurred so far. IDA, iron deficiency anemia; Tbc, tuberculosis; NSAID, nonsteroidal anti-inflammatory drug. Figures Figure 1 Representative capsule endoscopy images of small bowel lesions. A. Angioectasia with active bleeding. B. Angioectasia with no bleeding. C. Ulcer suggestive of Crohn’s disease. D. Ulcer suggestive of NSAIDinduced enteropathy. E. Erosion suggestive of NSAID-induced enteropathy. F. Radiation ileitis. G. SB polyp (GIST). IDA, iron deficiency anemia; OB, occult blood; SB, small bowel; UGI, upper gastrointestinal; LGI, Page 16/17 lower gastrointestinal; NSAIDs, nonsteroidal anti-inflammatory drugs; RTx, radiotherapy; GIST, gastrointestinal stromal tumor. Figure 2 Surgical intervention in a patient with small-bowel stricture. A. Abdominal computed tomography showing a stricture in the distal ileum. B. Gross findings after bowel resection with multiple stricture sites. C. Surgical resection exposing the inside of the small bowel. D. Capsule endoscopy image showing a semi-circular ulcer with active hemorrhage. E. Histologic findings show ulceration including diffuse loss of villi, mucosal and submucosal neutrophilic exudates, and transmural inflammation (Hematoxylin and Eosin x40). Page 17/17
https://openalex.org/W4295679100	https://www.nature.com/articles/s41467-022-32797-w.pdf	English	null	The clinical drug candidate anle138b binds in a cavity of lipidic α-synuclein fibrils	Nature communications	2,022	cc-by	12,197	The clinical drug candidate anle138b binds in a cavity of lipidic α-synuclein ﬁbrils Leif Antonschmidt 1,10, Dirk Matthes2,10, Rıza Dervişoğlu 1,9,10, Benedikt Frieg3, Christian Dienemann 4, Andrei Leonov1,5, Evgeny Nimerovsky1, Vrinda Sant1, Sergey Ryazanov1,5, Armin Giese5,6, Gunnar F. Schröder 3,7, Stefan Becker 1, Bert L. de Groot 2 , Christian Griesinger 1,8 & Loren B. Andreas 1 Leif Antonschmidt 1,10, Dirk Matthes2,10, Rıza Dervişoğlu 1,9,10, Benedikt Frieg3, Christian Dienemann 4, Andrei Leonov1,5, Evgeny Nimerovsky1, Vrinda Sant1, Sergey Ryazanov1,5, Armin Giese5,6, Gunnar F. Schröder 3,7, Stefan Becker 1, Bert L. de Groot 2 , Christian Griesinger 1,8 & Loren B. Andreas 1 Aggregation of amyloidogenic proteins is a characteristic of multiple neuro- degenerative diseases. Atomic resolution of small molecule binding to such pathological protein aggregates is of interest for the development of ther- apeutics and diagnostics. Here we investigate the interaction between α- synuclein ﬁbrils and anle138b, a clinical drug candidate for disease modifying therapy in neurodegeneration and a promising scaffold for positron emission tomography tracer design. We used nuclear magnetic resonance spectroscopy and the cryogenic electron microscopy structure of α-synuclein ﬁbrils grown in the presence of lipids to locate anle138b within a cavity formed between two β-strands. We explored and quantiﬁed multiple binding modes of the com- pound in detail using molecular dynamics simulations. Our results reveal stable polar interactions between anle138b and backbone moieties inside the tubular cavity of the ﬁbrils. Such cavities are common in other ﬁbril structures as well. The presence of α-synuclein ﬁbrils in inclusion bodies is a hallmark of neurodegenerative diseases such as Parkinson’s Disease1, Multiple System Atrophy2,3 and Dementia with Lewy bodies4,5. α-Synuclein aggregation is thought to be involved in cellular pathology and spreading between cells6. To date, no disease-modifying treatment is available for any of those diseases7. Hence, atomic resolution struc- tures of small molecule binding to such pathological protein aggre- gates are of interest for the development of therapeutics and diagnostics8,9. To address this issue, we studied the binding of the small molecule anle138b to α-synuclein ﬁbrils. diseases8,10–14. It was tested successfully in healthy volunteers15 and is currently undergoing phase Ib testing in Parkinson’s disease (NCT04685265). On a molecular level, anle138b was shown to interfere with α-synuclein oligomerization in vitro, inhibiting pore formation in membranes as well as cytochrome c leakage from mitochondria and depopulating α-synuclein aggregates in vivo8,14. Due to its highly lipophilic nature, it readily crosses the blood-brain-barrier11. Article https://doi.org/10.1038/s41467-022-32797-w nature communications 1NMR-based Structural Biology, Max Planck Institute for Multidisciplinary Sciences, Göttingen, Germany. 2Department of Theoretical and Computational Biophysics, Max Planck Institute for Multidisciplinary Sciences, Göttingen, Germany. 3Institute of Biological Information Processing (IBI-7: Structural Bio- chemistry), Forschungszentrum Jülich, Jülich, Germany. 4Department of Molecular Biology, Max Planck Institute for Multidisciplinary Sciences, Göttingen, Germany. 5MODAG GmbH, Mikroforum Ring 3, 55234 Wendelsheim, Germany. 6Center for Neuropathology and Prion Research, Ludwig- Maximilians-University Munich, Munich, Germany. 7Physics Department, Heinrich Heine University Düsseldorf, Düsseldorf, Germany. 8Cluster of Excellence “Multiscale Bioimaging: From Molecular Machines to Networks of Excitable Cells” (MBExC), University of Göttingen, Göttingen, Germany. 9Present address: Max Planck Institute for Chemical Energy Conversion, Mülheim an der Ruhr, Germany. 10These authors contributed equally: Leif Antonschmidt, Dirk Matthes, Rıza Dervişoğlu. e-mail: bgroot@gwdg.de; cigr@mpinat.mpg.de; land@mpinat.mpg.de Bound anle138b changes the local structure g To probe potential changes in the α-synuclein protoﬁlament structure induced by anle138b, we recorded room temperature MAS NMR spectra. Speciﬁcally, we acquired 2D 13C,13C correlation spectra with dipolar assisted rotational recoupling (DARR) to obtain correlations between carbon atoms, and 3D (H)CANH spectra to obtain F94 Y39 L38 G86 I88 G68 a h 15N 13C O O NH N Br 4 5 1 2 3 Agitation 96 h anle138b 90° 70 60 50 40 30 20 10 180 170 13C Chemical Shift (ppm) 15N Chemical Shift (ppm) 200 170 (U)13C αS 200 170 13C,15N Gly 200 170 13C,15N Ile 200 170 13C,15N Leu & 13C,15N Tyr 200 170 13C Phe b c d e f i 0 0.4 0.6 0.8 1.0 0.2 tmix=20.41 ms S0 ΔS × 10 70 60 50 40 30 20 10 13C Chemical Shift (ppm) NH O N H O N H O N H O HO N H O 50% 15N g 13C,15N Gly 13C,15N Ile 13C,15N Leu 13C Phe 13C,15N Tyr αSyn-monomer αSyn-fibril SUVs I112 (C ) & Lipid-CH2 I88 (C ) Fig. 1 \| Identiﬁcation of anle138b binding inside α-synuclein ﬁbrils by DNP- enhanced MAS NMR. a Isotope labeling strategy used for the preparation of ﬁbril samples for studies by DNP-enhanced MAS NMR. 13C-labeling of amino acids indi- cated by colored circles: unifom- (gray), Gly- (magenta), Ile- (blue), Leu- (yellow), Tyr- (red), and Phe-labeling (green). 15N-labeling on anle138b indicated by orange circles. b–f Slices of 2D NHHC spectra of α-synuclein ﬁbrils prepared from protein with b uniform 13C-labeling and c–f amino-acid-speciﬁc isotope labeling on c Gly, d Ile, e Leu and Tyr, and f Phe in the presence of 1,2-15N-anle138b. The spectral region shown contains only cross-peaks between pyrazole NH of anle138b (195 ppm) and protein carbon atoms. Proton–proton mixing was 200 μs. g Chemical structures of amino acids used for speciﬁc isotope labeling with 13C- labeled nuclei indicated by colored circles (same as in a) and 15N-labeling indicated by orange circles. h Negative stain EM micrograph of ﬁbrils grown in the presence of lipids (scale bar 100 nm) alongside cryo-EM structure of α-synuclein protoﬁlament L220 with color coding of speciﬁc isotope-labeled residues. i Frequency selective REDOR spectrum (carrier frequency at 26 ppm, Supplementary Table 2) of Ile-labeled α-synuclein ﬁbrils in the presence of 1,2-15N-anle138b. The difference spectrum (ΔS) is scaled up 10-fold with respect to the reference spectrum (S0) for clarity. Bound anle138b changes the local structure Fibrils were prepared at a lipid to protein to anle138b molar ratio of 10:1:0.5. All spectra were recorded at 600 MHz at 100 K (see Supplementary Table 1) in the presence of TEMTriPol-1. 70 60 50 40 30 20 10 0 13C Chemical Shift (ppm) 15N Chemical Shift (ppm) 200 170 (U)13C αS 200 170 13C,15N Gly 200 170 13C,15N Ile 200 170 13C,15N Leu & 13C,15N Tyr 200 170 13C Phe b c d e f i 0 0.4 0.6 0.8 1.0 0.2 tmix=20.41 ms S0 ΔS × 10 70 60 50 40 30 20 10 13C Chemical Shift (ppm) NH O N H O N H O N H O HO N H O 50% 15N g 13C,15N Gly 13C,15N Ile 13C,15N Leu 13C Phe 13C,15N Tyr I112 (C ) & Lipid-CH2 I88 (C ) acids used for speciﬁc isotope labeling with 13C- labeled nuclei indicated by colored circles (same as in a) and 15N-labeling indicated by orange circles. h Negative stain EM micrograph of ﬁbrils grown in the presence of lipids (scale bar 100 nm) alongside cryo-EM structure of α-synuclein protoﬁlament L220 with color coding of speciﬁc isotope-labeled residues. i Frequency selective REDOR spectrum (carrier frequency at 26 ppm, Supplementary Table 2) of Ile-labeled α-synuclein ﬁbrils in the presence of 1,2-15N-anle138b. The difference spectrum (ΔS) is scaled up 10-fold with respect to the reference spectrum (S0) for clarity. Fibrils were prepared at a lipid to protein to anle138b molar ratio of 10:1:0.5. All spectra were recorded at 600 MHz at 100 K (see Supplementary Table 1) in the presence of TEMTriPol-1. a 15N 13C O O NH N Br 4 5 1 2 3 Agitation 96 h anle138b 70 60 50 40 30 20 10 180 170 13C Chemical Shift (ppm) 15N Chemical Shift (ppm) 200 170 (U)13C αS 200 170 13C,15N Gly 200 170 13C,15N Ile 200 170 13C,15N Leu & 13C,15N Tyr 200 170 13C Phe b c d e f HO 15N g αSyn-monomer αSyn-fibril SUVs a 15N Chemical Shift (ppm) F94 Y39 L38 G86 I88 G68 F94 Y39 L38 G86 I88 G68 h 90° L h 90° h i Fig. 1 \| Identiﬁcation of anle138b binding inside α-synuclein ﬁbrils by DNP- acids used for speciﬁc isotope labeling with 13C- labeled nuclei indicated by colored circles (same as in a) and 15N-labeling indicated by orange circles. Article with selective 13C-labeling on these residues and used 2D 13C,13C RFDR for the assignment of resonances (Supplementary Fig. 2a, c–f). 2D NHHC spectra of these ﬁbrils in the presence of 15N-labeled anle138b (Fig. 1a, b–g, Supplementary Fig. 2b, and Supplementary Table 1) establish Gly as responsible for the peak at 44.8 ppm and Ile as another part of the binding site (28 ppm peak). Interaction with other lipophilic residues, Leu, Tyr, and Phe, or the amine of lysine (Fig. 1e, f and Supplementary Fig. 1b), could not be detected. The interaction of the compound with Cγ of an Ile (Fig. 1d) establishes the binding site at I88, since the only other Ile at position 112 can be excluded based on its distinct chemical shift and broad line shape (Supplementary Fig. 2d). The sidechain of I88 is facing towards the protoﬁlament interior and is part of a glycine-rich sequence that encloses a tubular hydrophobic cavity along α-synuclein protoﬁlament L220 (Fig. 1h, PDB ID 8A4L). Rotational echo double resonance (REDOR) experi- ments conﬁrm close proximity of anle138b to I88 (Fig. 1i). The observed interactions at I88 and Gly, as well as spectroscopic fea- tures of G68 and G86 (vide infra), therefore indicate binding of anle138b inside this cavity. brain lipid metabolism, and neuronal survival because of its afﬁnity to lipid membranes.17 In Parkinson’s disease-related Lewy bodies, α- synuclein ﬁbrils are found to colocalize with lipids, indicating the importance of lipids in the α-synuclein aggregation cascade18. We recently studied the aggregation of α-synuclein grown in the presence of negatively charged liposomes consisting of 1-palmitoyl-2-oleoyl-sn- glycero-3-phosphate (POPA) and 1-palmitoyl-2-oleoyl-sn-glycero-3- phosphocholine (POPC) and observed a segmental folding process19. Concordantly, we resolved the resulting protoﬁlament fold L2 of lipid- bound α-synuclein ﬁbrils using cryogenic electron microscopy (cryo-EM)20. The clinical drug candidate anle138b binds in a cavity of lipidic α-synuclein ﬁbrils Its tight binding (Kd = 190 ± 120 nM) to α-synuclein ﬁbrils16 also makes it an attractive starting point for the development of diagnostic positron emission tomography (PET) tracer molecules9. g p y α-Synuclein is a 140 amino-acid intrinsically disordered protein, which has been suggested to participate in synaptic vesicle exocytosis, y Anle138b [3-(1,3-benzodioxol-5-yl)-5-(3-bromophenyl)-1H-pyr- azole] shows efﬁcacy in animal models for several aggregation related Nature Communications\| (2022) 13:5385 Nature Communications\| (2022) 13:5385 1 Article https://doi.org/10.1038/s41467-022-32797-w Anle138b binds to ﬁbril interior We located the binding site by isotope labeling anle138b with 15N, and directly detecting its interaction with ﬁbrils grown from 13C-labeled full-length α-synuclein (Fig. 1a). Such interactions were recorded in two-dimensional (2D) NHHC nuclear magnetic resonance (NMR) spectra with signal enhancement via dynamic nuclear polarization (DNP) and magic-angle spinning (MAS) at 100 K21. During ﬁbrilliza- tion, we used anle138b doped liposomes consisting of POPA and POPC (1:1). Proton–proton mixing in these experiments results in cross-peaks between nearby nuclei up to a distance of about 6 Å22 (Fig. 1b and Supplementary Fig. 1). The most prominent cross-peak at 44.8 ppm is consistent with the pyrazole NH of anle138b binding near Cβ of Phe, Leu, Tyr, or Ile or Cα of Gly. We prepared α-synuclein ﬁbrils Bound anle138b changes the local structure Bound anle138b changes the local structure g Chemical structures of amino Nature Communications\| (2022) 13:5385 2 Article https://doi.org/10.1038/s41467-022-32797-w 38 60 55 20 30 35 15 50 20 55 25 60 70 75 65 55 13C Chemical Shift (ppm) 13C Chemical Shift (ppm) S87 S42 T72 T75 T64 T59 T44 T92 V82 I88 A85 A69 34 b Ibound/Ifree Residue Number CSP d e 13C Chemical Shift (ppm) V82 A85 I88 a 15N Chemical Shift (ppm) 52 48 130 125 68 64 125 120 115 48 44 115 110 V74 T54 T81 A76 A53 A78 A91 A56 A90 A69 H50 Q79 Q62 K60 G51 G41 G67 G73 G93 G47 G86 G68 T81 0 0.5 0 1 40 50 60 70 80 90 S87 G68 A69 V70 G86 G84 K80 c Fig. 2 \| Chemical shift perturbations comparing α-synuclein ﬁbrils grown in the presence and absence of anle138b. a Excerpts of carbon-nitrogen projections of 3D (H)CANH spectra of lipidic α-synuclein ﬁbrils. Signal attenuation (Ibound/Ifree) and chemical shift perturbation (CSP) are observed comparing spectra of ﬁbrils with anle138b (green, 800 MHz with 55 kHz MAS) and without anle138b (gray, 950 MHz with 100 kHz MAS). b Relative peak intensities from 3D (H)CANH spectra and c, average CSP from DARR and (H)CANH spectra induced by anle138b inside α- synuclein ﬁbrils plotted per residue. Black dotted line represents a 2-residue moving average; highlighted areas indicate the cavity in protoﬁlament L2. d Excerpts of 2D 13C,13C correlation spectra (20 ms DARR mixing) of α-synuclein ﬁbrils with (red) and without (gray) anle138b acquired at 850 MHz with 17 kHz MAS. e Residues enclosing the cavity in α-synuclein protoﬁlament L2 color-coded by nuclei experiencing signiﬁcant CSP (blue), signal attenuation (green), or both (copper). Source data are provided as a Source Data ﬁle. Bound anle138b changes the local structure The Cγ of I88 showed close and persistent contacts to the anle138b pose with outward-facing pyrazole nitrogen atoms, while the Cα of either G67, G68, G84 or G86 did so for both anle138b poses in the simulations (Fig. 3b, c). This is most consistent with NHHC and REDOR spectra for outward-facing anle138b (Fig. 1c, d, i). Since short distances dominate REDOR data, the presence of the inward-facing pose, with larger separation of anle138b nitrogen atoms and I88 Cγ is compatible as well. We thus surmise that the outward-facing pose is present in the ﬁbrils, as well as an undetermined amount of inward-facing anle138b. From the MD simulations, we observed a stepwise displacement of anle138b within the internal cavity and along the stacked β-strand layers of the protoﬁlament (Fig. 3d and Supplementary Fig. 6d) on the tens to hundreds of ns timescale (Fig. 3e, h, i, l and Supplementary Movie S1), consistent with the disappearance of some NMR signals at ambient temperatures and the use of cryogenic DNP conditions to localize the compound. Relative free-energy binding proﬁles (Fig. 3f, g, j, k and Supplementary Fig. 7c) show multiple local energy minima as a func- tion of anle138b insertion depth within the protoﬁlament core for all tested binding poses. However, for inward-facing anle138b binding poses signiﬁcantly higher energy barriers are found (~20 kJ/mol, Fig. 3f). In concordance with simulations on model systems23, anle138b showed characteristic polar interactions with exposed backbone epi- topes (Fig. 4a). Polar ladders of backbone atoms lining the cavity (Fig. 4a, c and Supplementary Figs. S6e, S8) interact with anle138b either through hydrogen bonds via the pyrazole ring (outward-facing) or halogen bonds involving the bromine atom (inward-facing) (see “Methods”/Supplementary Information). The common pattern of correlations between protein backbone Cα, NH and HN. The 2D DARR spectra with long mixing (Supplementary Fig. 3) show similar protein long-range contacts in the presence and absence of anle138b, con- ﬁrming an overall conserved α-synuclein protoﬁlament fold. The pre- sence of anle138b results in chemical shift perturbation (CSP) for 80KTV82, A85, and I88 (Fig. 2c and Supplementary Fig. 4a–c) as well as attenuation for residues 67GGAV70 and 82VEGAGSI88 as seen from 3D (H)CANH (Fig. 2a, b and Supplementary Fig. 4d) and 2D DARR spectra (Fig. 2d). This observation indicates increased dynamics at the affected residues compared to the absence of anle138b. Peak-broadening of Cα of G68 as well as G86 (Fig. Bound anle138b changes the local structure h Negative stain EM micrograph of ﬁbrils grown in the presence of lipids (scale bar 100 nm) alongside cryo-EM structure of α-synuclein protoﬁlament L220 with color coding of speciﬁc isotope-labeled residues. i Frequency selective REDOR spectrum (carrier frequency at 26 ppm, Supplementary Table 2) of Ile-labeled α-synuclein ﬁbrils in the presence of 1,2-15N-anle138b. The difference spectrum (ΔS) is scaled up 10-fold with respect to the reference spectrum (S0) for clarity. Fibrils were prepared at a lipid to protein to anle138b molar ratio of 10:1:0.5. All spectra were recorded at 600 MHz at 100 K (see Supplementary Table 1) in the presence of TEMTriPol-1. acids used for speciﬁc isotope labeling with 13C- labeled nuclei indicated by colored circles (same as in a) and 15N-labeling indicated by orange circles. h Negative stain EM micrograph of ﬁbrils grown in the presence of lipids (scale bar 100 nm) alongside cryo-EM structure of α-synuclein protoﬁlament L220 with color coding of speciﬁc isotope-labeled residues. i Frequency selective REDOR spectrum (carrier frequency at 26 ppm, Supplementary Table 2) of Ile-labeled α-synuclein ﬁbrils in the presence of 1,2-15N-anle138b. The difference spectrum (ΔS) is scaled up 10-fold with respect to the reference spectrum (S0) for clarity. Fibrils were prepared at a lipid to protein to anle138b molar ratio of 10:1:0.5. All spectra were recorded at 600 MHz at 100 K (see Supplementary Table 1) in the presence of TEMTriPol-1. Fig. 1 \| Identiﬁcation of anle138b binding inside α synuclein ﬁbrils by DNP enhanced MAS NMR. a Isotope labeling strategy used for the preparation of ﬁbril samples for studies by DNP-enhanced MAS NMR. 13C-labeling of amino acids indi- cated by colored circles: unifom- (gray), Gly- (magenta), Ile- (blue), Leu- (yellow), Tyr- (red), and Phe-labeling (green). 15N-labeling on anle138b indicated by orange circles. b–f Slices of 2D NHHC spectra of α-synuclein ﬁbrils prepared from protein with b uniform 13C-labeling and c–f amino-acid-speciﬁc isotope labeling on c Gly, d Ile, e Leu and Tyr, and f Phe in the presence of 1,2-15N-anle138b. The spectral region shown contains only cross-peaks between pyrazole NH of anle138b (195 ppm) and protein carbon atoms. Proton–proton mixing was 200 μs. Bound anle138b changes the local structure 13C Chemical Shift (ppm) 52 48 130 125 48 44 115 110 A76 A53 A78 A91 A56 A90 A69 H50 Q79 Q62 K60 G51 G41 G67 G73 G93 G47 G86 G68 b Ibound/Ifree Residue Number CSP 0 0.5 0 1 40 50 60 70 80 90 c a 15N Chemical Shift (ppm) 68 64 125 120 115 V74 T54 T81 c 60 70 75 65 55 13C Chemical Shift (ppm) S87 S42 T72 T75 T64 T59 T44 T92 d 20 30 35 15 I88 e 30 35 20 25 hift (ppm) A85 A69 ) V82 A85 I88 T81 S87 G68 A69 V70 G86 G84 K80 20 30 35 15 20 25 hift (ppm) I88 A85 A69 3 0 20 5 25 Shift (ppm) A85 A69 38 60 55 5 5 13C Chemica V82 34 e 20 13C Chemical Shift (ppm) Fig. 2 \| Chemical shift perturbations comparing α-synuclein ﬁbrils grown in the presence and absence of anle138b. a Excerpts of carbon-nitrogen projections of 3D (H)CANH spectra of lipidic α-synuclein ﬁbrils. Signal attenuation (Ibound/Ifree) and chemical shift perturbation (CSP) are observed comparing spectra of ﬁbrils with anle138b (green, 800 MHz with 55 kHz MAS) and without anle138b (gray, 950 MHz with 100 kHz MAS). b Relative peak intensities from 3D (H)CANH spectra and c, average CSP from DARR and (H)CANH spectra induced by anle138b inside α- synuclein ﬁbrils plotted per residue. Black dotted line represents a 2-residue moving average; highlighted areas indicate the cavity in protoﬁlament L2. d Excerpts of 2D 13C,13C correlation spectra (20 ms DARR mixing) of α-synuclein ﬁbrils with (red) and without (gray) anle138b acquired at 850 MHz with 17 kHz MAS e Residues enclosing the cavity in α-synuclein protoﬁlament L2 color-coded by nuclei experiencing signiﬁcant CSP (blue), signal attenuation (green), or both (copper). Source data are provided as a Source Data ﬁle. independent of the initial binding pose and tautomeric form of anle138b (Supplementary Fig. 6a). Intentionally placing the compound at the edge of the protoﬁlament leads to spontaneous insertion into the internal binding site, demonstrating favorable binding (Supple- mentary Fig. 7a, b). The steric hindrance imposed by nonpolar side- chain groups (A69, T81, V82, and I88) surrounding the cavity did not allow for rotations of the molecule (Supplementary Fig. 6b). Bound anle138b changes the local structure 2a) and particularly Cγ of I88 (Fig. 2d) cor- roborate the contacts observed under DNP conditions. As these resi- dues belong to the two stretches of the α-synuclein sequence which establish the aforementioned cavity in protoﬁlament L2 (Fig. 2e), we conclude that anle138b is embedded within α-synuclein ﬁbrils, while causing only local structural changes. Nature Communications\| (2022) 13:5385 Polar contacts govern internal binding The rather narrow cavity of the α-synuclein protoﬁlament L2 permits four principal internal binding poses for the planar anle138b molecule. Inside the protoﬁlament cavity, the pyrazole nitrogen atoms of anle138b can face either “inward” or “outward” and the bromophenyl ring pointing “up” or “down” with respect to the direction of the long axis of the protoﬁlament (Fig. 3a). Here, the protoﬁlament axis was deﬁned by the direction of the NH vector of every 2nd amino acid in the folded, N-terminal part of the protoﬁlament, starting with the residue G36. All-atom molecular dynamics (MD) simulations (see “Methods” and Supplementary Table 3) were employed to probe if these anle138b-binding poses are indeed consistent with the observations made by NMR. To do so, both tautomers of the anle138b molecule were considered and simulated in the dominant conformer as deter- mined by NMR, respectively (Supplementary Fig. 5). During 40 inde- pendent, 1-µs long MD simulations, no unbinding events from the cavity of the α-synuclein protoﬁlament L2 structure were observed Nature Communications\| (2022) 13:5385 3 https://doi.org/10.1038/s41467-022-32797-w Article a g e 0 0.2 0.4 0.6 0.8 1.0 time [µs] initial anle138b locations initial anle138b locations i j b Gly Cα 3 4 5 6 7 I88 G67 G68 G84 d layer 2 layer 4 layer 3 layer 5 c 90° protofilament axis dmin (Nanle138b) [Å] fraction 0 0.05 0.1 3 4 5 6 7 dmin (Nanle138b) [Å] fraction 0 0.05 0.1 0 5 10 15 20 25 insertion depth x [Å] f 0 10 20 G(x) [kJ/mol] dwell time [ns] 0 5 10 15 20 25 0 0.2 0.4 0.6 0.8 1.0 time [µs] 0 5 10 15 20 25 insertion depth x [Å] 0 10 20 G(x) [kJ/mol] 0 5 10 15 20 25 500 1000 0 0.2 0.4 0.6 0.8 0 norm. count dwell time [ns] 500 1000 0 0.2 0.4 0.6 0.8 0 norm. count G86 I88 G67 G68 G84 G86 Ile Cγ1 Gly Cα Ile Cγ1 67GGAV70 80KTVEGAGSI88 h t(max) t(max) t0 t0 t0 t0 k l Fig. 3 \| Anle138b binds dynamically inside the α-synuclein protoﬁlament. a Starting structures for MD simulations for the four principal poses of anle138b: Pyrazole nitrogen atoms of anle138b (indicated by blue arrows) are oriented either to residues 67GGAV70 (inward-facing, yellow) or to residues 80KTVEGAGSI88 (out- ward-facing, green) inside the cavity (cyan) of α-synuclein protoﬁlament L2. Polar contacts govern internal binding Further studies are needed to determine whether the internal binding mode of anle138b in lipidic ﬁbrils, is also a prominent feature of anle138b interactions with early aggregates23. itrogen atoms of anle138b and protein A69, or G86 and S87 (Fig. 4b and Supple- th 2D NHHC spectra (Fig. 1c and Supple- d inside the protoﬁlament, anle138b mole- uctural ﬂuctuations of residues around the ons. All β-strand layers in proximity to the n facing outward, (Fig. 4d) exhibit a higher s 67GGA69, and 80KTVEGAGSI88 consistent SP observed by NMR for these perturbed ffect is also seen for layers not in contact e). e138b molecules starting outside the α- vity demonstrate the afﬁnity of anle138b for d pronounced colocalization of the com- g sites on the protoﬁlament exterior (Sup- 0). In contrast to the stable internal binding, ansient, explaining the absence of exterior ig. 1). served in the α-synuclein protoﬁlament L2 s such as PDB IDs 6SSX/6RT024, 7NCI, and 7NCH25, (Supplementary Fig. 11) as well as many other amyloid ﬁbrils26,27 are a unique feature of the repetitive ﬁbril architecture. This level of commonality across different ﬁbrils could have implications for positron emission tomography radiotracer development. So far, tracer targets are usually identiﬁed in exterior binding sites similar to those reported for ﬁbrils of HET-s28, amyloid-β (Aβ)29, and tau30,31. While theoretical studies have predicted the existence of binding sites in the core of ﬁbrillar aggregates32, experimental data did not allow for a high-resolution structure33. End-to-end ﬁlling of tubular cavities in ﬁbrils by hydrophobic molecules has been proposed before34. We ﬁnd selective binding inside and translational motion of anle138b along the α-synuclein ﬁbril cavity, consistent with this model. Our results emphasize that expanding radiotracer target identiﬁcation to ﬁbril cavities as well as structure-based reﬁnement of cavity binding mole- cules, such as anle138b, could prove to be beneﬁcial. Fibrils ampliﬁed from Parkinson’s Disease patient's brain exhibit a similar fold to the lipidic α-synuclein L2 ﬁbrils and support the relevance of the proto- ﬁlaments in the present study35,36. Further studies are needed to determine whether the internal binding mode of anle138b in lipidic ﬁbrils, is also a prominent feature of anle138b interactions with early aggregates23. We presented a high-resolution structural model for anle138b binding in a lipidic α-synuclein ﬁbril. We explored how the linear Moreover, when bound inside the protoﬁlament, anle138b mole- cules markedly alter the structural ﬂuctuations of residues around the cavity during MD simulations. Polar contacts govern internal binding All β-strand layers in proximity to the compound, especially when facing outward, (Fig. 4d) exhibit a higher averaged RMSF for residues 67GGA69, and 80KTVEGAGSI88 consistent with the attenuation and CSP observed by NMR for these perturbed regions (Fig. 2). A similar effect is also seen for layers not in contact with the compound (Fig. 4e). Simulations with anle138b molecules starting outside the α- synuclein protoﬁlament cavity demonstrate the afﬁnity of anle138b for the lipid environment and pronounced colocalization of the com- pound around lipid binding sites on the protoﬁlament exterior (Sup- plementary Figs. S9 and S10). In contrast to the stable internal binding, such binding events are transient, explaining the absence of exterior contacts in NMR spectra (Fig. 1). Polar contacts govern internal binding 11) as well as many other amyloid ﬁbrils26,27 are a unique feature of the repetitive ﬁbril architecture. This level of commonality across different ﬁbrils could have implications for positron emission tomography radiotracer development. So far, tracer targets are usually identiﬁed in exterior binding sites similar to those reported for ﬁbrils of HET-s28, amyloid-β (Aβ)29, and tau30,31. While theoretical studies have predicted the existence of binding sites in the core of ﬁbrillar aggregates32, experimental data did not allow for a high-resolution structure33. End-to-end ﬁlling of tubular cavities in ﬁbrils by hydrophobic molecules has been proposed before34. We ﬁnd selective binding inside and translational motion of anle138b along the α-synuclein ﬁbril cavity, consistent with this model. Our results emphasize that expanding radiotracer target identiﬁcation to ﬁbril cavities as well as structure-based reﬁnement of cavity binding mole- cules, such as anle138b, could prove to be beneﬁcial. Fibrils ampliﬁed from Parkinson’s Disease patient's brain exhibit a similar fold to the lipidic α-synuclein L2 ﬁbrils and support the relevance of the proto- ﬁlaments in the present study35,36. Further studies are needed to determine whether the internal binding mode of anle138b in lipidic ﬁbrils, is also a prominent feature of anle138b interactions with early aggregates23. 7NCH25, (Supplementary Fig. 11) as well as many other amyloid ﬁbrils26,27 are a unique feature of the repetitive ﬁbril architecture. This level of commonality across different ﬁbrils could have implications for positron emission tomography radiotracer development. So far, tracer targets are usually identiﬁed in exterior binding sites similar to those reported for ﬁbrils of HET-s28, amyloid-β (Aβ)29, and tau30,31. While theoretical studies have predicted the existence of binding sites in the core of ﬁbrillar aggregates32, experimental data did not allow for a high-resolution structure33. End-to-end ﬁlling of tubular cavities in ﬁbrils by hydrophobic molecules has been proposed before34. We ﬁnd selective binding inside and translational motion of anle138b along the α-synuclein ﬁbril cavity, consistent with this model. Our results emphasize that expanding radiotracer target identiﬁcation to ﬁbril cavities as well as structure-based reﬁnement of cavity binding mole- cules, such as anle138b, could prove to be beneﬁcial. Fibrils ampliﬁed from Parkinson’s Disease patient's brain exhibit a similar fold to the lipidic α-synuclein L2 ﬁbrils and support the relevance of the proto- ﬁlaments in the present study35,36. Polar contacts govern internal binding count t(max) k j k i g e 0 0.2 0.4 0.6 0.8 1.0 time [µs] initial anle138b locations 0 5 10 15 20 25 insertion depth x [Å] f 0 10 20 G(x) [kJ/mol] 0 5 10 15 20 25 h g 0 f 0 10 20 G(x) [kJ/mol] 0 5 10 15 20 25 h e 0 0.2 0.4 0.6 0.8 1.0 time [µs] initial anle138b locations 0 5 10 15 20 25 insertion depth x [Å] f g e h Fig. 3 \| Anle138b binds dynamically inside the α-synuclein protoﬁlament. the simulation. d Snapshots and e, i trajectories (representative trajectories g \| y y y p a Starting structures for MD simulations for the four principal poses of anle138b: Pyrazole nitrogen atoms of anle138b (indicated by blue arrows) are oriented either to residues 67GGAV70 (inward-facing, yellow) or to residues 80KTVEGAGSI88 (out- ward-facing, green) inside the cavity (cyan) of α-synuclein protoﬁlament L2. The bromophenyl moiety can point up or down the protoﬁlament (small panels). A color code denotes the individual simulation sets for each binding pose in all following panels (inward-facing—yellow, outward-facing—green). b, c Orientation (left) and distributions of minimal distances (right) between nitrogen atoms of anle138b and Cγ1 of I88 (light-blue), and Cα of G67, G68, G84, G86 (magenta) as described in a. Dashed lines indicate the initial positions of anle138b at the start of as bold lines; others as thin lines for clarity) illustrate discrete translational motion as bold lines; others as thin lines for clarity) illustrate discrete translational motion along the protoﬁlament axis and alignment of anle138b with individual β-strand layers. Relative free-energy binding proﬁles G(x) as a function of anle138b insertion depth x for combined simulation sets of f inward-facing and j outward-facing pose. Data are presented as mean values ± SEM (depicted by shading). g, k Isosurfaces of corresponding anle138b occupancy. h, l Cumulative histogram of dwell times cal- culated from anle138b displacements inside the tubular cavity for h inward-facing (n = 24) and l outward-facing (n = 218) pose. Data are presented as mean values ± SEM (indicated by error bars). Water and lipid molecules omitted for clarity in all renderings. close contacts between nitrogen atoms of anle138b and protein backbone atoms of G68, A69, or G86 and S87 (Fig. 4b and Supple- mentary Fig. 6e) agrees with 2D NHHC spectra (Fig. 1c and Supple- mentary Fig. 1b). 7NCH25, (Supplementary Fig. Polar contacts govern internal binding The bromophenyl moiety can point up or down the protoﬁlament (small panels). A color code denotes the individual simulation sets for each binding pose in all following panels (inward-facing—yellow, outward-facing—green). b, c Orientation (left) and distributions of minimal distances (right) between nitrogen atoms of anle138b and Cγ1 of I88 (light-blue), and Cα of G67, G68, G84, G86 (magenta) as described in a. Dashed lines indicate the initial positions of anle138b at the start of the simulation. d Snapshots and e, i trajectories (representative trajectories as bold lines; others as thin lines for clarity) illustrate discrete translational motion along the protoﬁlament axis and alignment of anle138b with individual β-strand layers. Relative free-energy binding proﬁles G(x) as a function of anle138b insertion depth x for combined simulation sets of f inward-facing and j outward-facing pose. Data are presented as mean values ± SEM (depicted by shading). g, k Isosurfaces of corresponding anle138b occupancy. h, l Cumulative histogram of dwell times cal- culated from anle138b displacements inside the tubular cavity for h inward-facing (n = 24) and l outward-facing (n = 218) pose. Data are presented as mean values ± SEM (indicated by error bars). Water and lipid molecules omitted for clarity in all renderings. a 67GGAV70 a protofilament axis b Gly Cα 3 4 5 6 7 I88 G67 G68 G84 c dmin (Nanle138b) [Å] fraction 0 0.05 0.1 G86 Ile Cγ1 t0 t0 c I88 G67 G68 G84 G86 d 3 4 5 6 7 dmin (Nanle138b) [Å] fraction 0 0.05 0.1 Gly Cα Ile Cγ1 t0 t0 b c d dwell time [ns] 500 1000 0 0.2 0.4 0.6 0.8 0 norm. count t(max initial anle138b locations i j 0 0.2 0.4 0.6 0.8 1.0 time [µs] 0 5 10 15 20 25 insertion depth x [Å] 0 10 20 G(x) [kJ/mol] 0 5 10 15 20 25 x) k l dwell time [ns] 500 1000 0 0.2 0.4 0.6 0.8 0 norm. Discussion We presented a high-resolution structural model for anle138b binding in a lipidic α-synuclein ﬁbril. We explored how the linear Long tubular cavities, as observed in the α-synuclein protoﬁlament L2 and other α-synuclein folds such as PDB IDs 6SSX/6RT024, 7NCI, and Nature Communications\| (2022) 13:5385 4 https://doi.org/10.1038/s41467-022-32797-w Article Fig. 4 \| Anle138b forms polar interactions with the protein backbone that alter local structural ﬂuctuations in the protoﬁlament. a Polar contacts between anle138b atoms (indicated by arrows) and protein backbone atoms (blue circles— amide nitrogen; red circles—carbonyl oxygen) observed in MD simulations for inward- (yellow) and outward-facing (green) internal binding poses. b Contact probabilities for polar interactions with backbone atoms of individual residues for polar interactions deﬁned in a. Scale bars (right) indicate contact probabilities. c Tubular cavity viewed up close down the long axis of the α-synuclein Fig. 4 \| Anle138b forms polar interactions with the protein backbone that alter local structural ﬂuctuations in the protoﬁlament. a Polar contacts between anle138b atoms (indicated by arrows) and protein backbone atoms (blue circles— amide nitrogen; red circles—carbonyl oxygen) observed in MD simulations for inward- (yellow) and outward-facing (green) internal binding poses. b Contact probabilities for polar interactions with backbone atoms of individual residues for polar interactions deﬁned in a. Scale bars (right) indicate contact probabilities. c Tubular cavity viewed up close down the long axis of the α-synuclein protoﬁlament (shown without anle138b for clarity). Prominent contact sites in the backbone scaffold are represented by circles (colors as in a). d, e Average root mean squared ﬂuctuations (RMSF) for d the four consecutive β-strands in contact with anle138b and e β-strands without close polar interactions. Average RMSF without anle138b present shown for comparison (black lines). β-strands of the ﬁbril ends were not considered. Data are presented as mean values ± SEM (depicted by shading). Source data are provided as a Source Data ﬁle. protoﬁlament (shown without anle138b for clarity). Prominent contact sites in the backbone scaffold are represented by circles (colors as in a). d, e Average root mean squared ﬂuctuations (RMSF) for d the four consecutive β-strands in contact with anle138b and e β-strands without close polar interactions. Average RMSF without anle138b present shown for comparison (black lines). β-strands of the ﬁbril ends were not considered. Data are presented as mean values ± SEM (depicted by shading). Source data are provided as a Source Data ﬁle. Fig. Preparation of isotope-labeled anle138b arrangement of polar groups inside the continuous tubular cavity formed along the cross-β spine of the protoﬁlament facilitates small molecule binding. It will be of signiﬁcant interest whether the parti- cular features of glycine enriched tight and hydrophobic cavities as reported here, can also be exploited for selective compound binding to ﬁbrillar aggregates of other amyloidogenic proteins such as Aβ, tau, and prion protein which also bear these cavities. The general information for the synthesis of compounds was descri- bed previously8. Preparation of [1,2-15N2]-3-(1,3-benzodioxol-5-yl)-5-(3-bromo- phenyl)-1H-pyrazole Discussion 4 \| Anle138b forms polar interactions with the protein backbone that alter local structural ﬂuctuations in the protoﬁlament. a Polar contacts between anle138b atoms (indicated by arrows) and protein backbone atoms (blue circles— amide nitrogen; red circles—carbonyl oxygen) observed in MD simulations for inward- (yellow) and outward-facing (green) internal binding poses. b Contact probabilities for polar interactions with backbone atoms of individual residues for polar interactions deﬁned in a. Scale bars (right) indicate contact probabilities. c Tubular cavity viewed up close down the long axis of the α-synuclein protoﬁlament (shown without anle138b for clarity). Prominent contact sites in the backbone scaffold are represented by circles (colors as in a). d, e Average root mean squared ﬂuctuations (RMSF) for d the four consecutive β-strands in contact with anle138b and e β-strands without close polar interactions. Average RMSF without anle138b present shown for comparison (black lines). β-strands of the ﬁbril ends were not considered. Data are presented as mean values ± SEM (depicted by shading). Source data are provided as a Source Data ﬁle. Preparation of α-synuclein ﬁbrils Samples of α-synuclein ﬁbrils were prepared as previously reported19. In brief, vesicles were prepared by mixing 1-palmitoyl-2-oleoyl-sn-gly- cero-3-phosphocholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero-3- phosphate (POPA), sodium salt and anle138b each dissolved in chloroform respectively and evaporating the solvent under a N2- stream and lyophilized overnight. SUVs were obtained by repeated sonication of a solution of 1.5 mM POPC, 1.5 mM POPA, and 150 μM anle138b. Vesicles were incubated with 70 μM α-synuclein in buffer (50 mM HEPES, 100 mM NaCl, pH 7.4) at a lipid to protein ratio of 10:1 and a compound to protein ratio of 1:2 and subjected to repeated cycles of 30-s sonication (20 kHz) at 37 °C followed by an incubation period of 30 min. After 96 h, ﬁbrils were harvested by centrifuging at 152,460 × g (TLA-100.3 rotor in an Optima™MAX-TL) for 1 h at 4 °C. The supernatant was removed, and the resulting pellet was washed with fresh buffer and centrifuged again 212,940 × g for 10 min at 18 °C. Excess moisture was removed, and the pellet was immediately packed into rotors for ssNMR. Preparation of [3,4-13C2]-3-(1,3-benzodioxol-5-yl)-5-(3-bromo- phenyl)-1H-pyrazole Preparation of [3,4-13C2]-3-(1,3-benzodioxol-5-yl)-5-(3-bromo- phenyl)-1H-pyrazole [1,2-13C2]-1-(1,3-benzodioxol-5-yl)ethanone. Starting material [1,2- 13C2]-1-(1,3-benzodioxol-5-yl)ethanone was synthesized by Friedel- Crafts reaction of 1,3-benzodioxole and [1,2-13C2]-acetic acid (99% 13C) according to procedure described in39. In an atmosphere of nitrogen, triﬂic anhydride (5.0 g, 17.7 mmol) was added to [1,2-13C2]-acetic acid (99% 13C) (1 g, 16.1 mmol) with cooling (10 °C bath temperature). After stirring at room temperature for 5 min, 1,3-benzodioxole (1.97 g, 16.1 mmol) was added, and the stirring at room temperature was continued for 30 min. The mixture was quenched with saturated NaHSO4 solution (30 mL) and extracted with EtOAc (30 mL). The separated organic layer was washed with water, brine, dried over Na2SO4, and concentrated under reduced pressure. The residue was puriﬁed by ﬂash chromatography on silica gel (n-hexane/EtOAc = 3/1) to afford the title compound (1.63 g, 9.8 mmol, 61%) as a white solid. 1H NMR (400 MHz, CDCl3) d = 7.55 (ddd, J = 8.1, 4.1 1.7 Hz, 1H), 7.44 (dd, J = 3.6, 1.7 Hz, 1H), 6.85 (d, J = 8.1 Hz, 1H), 6.04 (s, 2H), 2.54 (dd, J = 127.5, 5.9 Hz, 3H). ssNMR spectroscopy For ssNMR studies, ﬁbrils were grown from (U)-13C15N-α-synuclein in the presence of vesicles of POPA and POPC (1:1) containing 3,4- 13C-anle138b at a L/P ratio of 5:1. For room temperature experiments, both samples with and without the compound were prepared at the same time and a single sample each was used for all 3D (H)CANH and 2D 13C13C-DARR experiments. Fibrils were packed into ssNMR rotors by cutting off the bottom of the tube and centrifuging the pellet directly into the rotor of choice through a custom-made ﬁlling device made from a truncated pipette tip. Last, the sample was centrifuged into the rotor in an ultracentrifuge packing device for 30 min at 98,381×g in a SW 32 Ti rotor in an Optima™L-80 XP Ultracentrifuge (both Beckman Coulter)40. 3D (H)CANH experiments41 on ﬁbrils in the presence of anle138b were recorded on an 800 MHz Bruker Avance III HD spectrometer at a magnetic ﬁeld of 18.8 T equipped with a 1.3-mm magic-angle spinning (MAS) HCN probe and MAS at 55 kHz. The temperature of the cooling gas was set to 250 K, resulting in an estimated sample tempera- ture of 20 °C. [3,4-13C2]-3-(1,3-benzodioxol-5-yl)-5-(3-bromophenyl)-1H-pyr- azole. The compound was synthesized according to the procedure described in8. To a solution of [1,2-13C2]-1-(1,3-benzodioxol-5-yl)etha- none (830 mg, 5.0 mmol) and methyl 3-bromobenzoate (1.35 g, 6.3 mmol) in DMSO (12 mL) and THF (3 mL) was added sodium hydride (60% in oil, 6.3 mmol, 252 mg). The reaction mixture was stirred at 20 °C for 24 h, then poured into 60 mL of ice water containing 85% H3PO4 (0.6 mL) and stirred for 1 h. The resulting precipitate was ﬁl- tered off, washed with water (3.10 mL), MeOH (5 mL), hexane (10 mL), and air-dried to give the crude intermediate [1,2-13C2]-3-(1,3-benzo- dioxol-5-yl)-5-(3-bromophenyl)propane-1,3-dione (1.57 g) as a yellow solid which was used for the next step without puriﬁcation. To a sus- pension of this in EtOH (20 mL) hydrazine hydrate (500 mg, 10 mmol) was added. The reaction mixture was stirred at 78 °C for 15 h, cooled down, and concentrated in vacuum. The crude product was crystal- lized from n-BuOH (10 mL) and vacuum dried to give 1.12 g of the title compound as a white solid (3.2 mmol, 65% over two steps). Preparation of isotope-labeled protein α-Synuclein was recombinantly produced in E. coli strain BL21(DE3) as previously described37. Uniformly 15N- and 13C,15N-labeled samples were expressed in a minimal medium supplemented with 15NH4Cl and 13C6-D- glucose (Cambridge Isotope Laboratories and Sigma Aldrich). For the production of amino-acid-speciﬁc forward-labeled protein (13C,15N- Leu/Tyr, 13C-Phe, 13C,15N-Ile), the labeled amino acids were added to the minimal medium 1 h before induction of protein expression. For for- ward labeling with 13C- and 13C,15N-Gly, protein expression was per- formed with the glycine auxotrophic E. coli strain DL39 (DE3) GlyA following a published protocol38. The protein was ﬁnally dialyzed against buffer (50 mM HEPES, 100 mM NaCl, pH 7.4) to obtain a 0.3 mM solution, and the resulting solution was stored at −80 °C until use. A mixture of [15N2] hydrazine sulfate (2.57 g, 19.5 mmol), EtOH (20 mL), water (1 mL), and 40% KOH aqueous solution (5.6 g, 40 mmol) was stirred in a sealed pressure ﬂask at 80 °C for 1 h. The mixture was cooled, 3-(1,3-benzodioxol-5-yl)-5-(3-bromophenyl)propane-1,3-dione8 (2.43 g, 7 mmol) was added, the reaction vessel was sealed again, and the heterogeneous mixture was stirred at 80 °C for 15 h. The cooled mixture was diluted with water (100 mL), the resulting precipitate was ﬁltered off, washed with water, and air-dried. The crude product was crystallized from n-BuOH (10 mL) and vacuum dried to give the title compound as a white solid (1.95 g, 5.65 mmol, 80%). 1H NMR (400 MHz, DMSO-d6 + 0.5% TFA) d = 8.05 (s, 1H), 7.84 (d, J = 7.8 Hz, 1H), Nature Communications\| (2022) 13:5385 5 Article https://doi.org/10.1038/s41467-022-32797-w CH3CN in 30 min), RT 14.2 min and mass 345.1 (100%), 347.1 (98%), [M + H]+. TLC (SiO2, n-hexane/EtOAc = 2/1) Rf 0.39, m.p. 195–197 °C. CH3CN in 30 min), RT 14.2 min and mass 345.1 (100%), 347.1 (98%), [M + H]+. TLC (SiO2, n-hexane/EtOAc = 2/1) Rf 0.39, m.p. 195–197 °C. CH3CN in 30 min), RT 14.2 min and mass 345.1 (100%), 347.1 (98%), [M + H]+. TLC (SiO2, n-hexane/EtOAc = 2/1) Rf 0.39, m.p. 195–197 °C. 7.52 (d, J = 8.1 Hz, 1H), 7.45-7.40 (m, 2H), 7.40-7.34 (m, 1H), 7.25 (d, J = 3.1 Hz, 1H), 7.01 (d, J = 8.1 Hz, 1H), 6.06 (s, 2H). 13C NMR (100.6 MHz, DMSO-d6 + 0.5% TFA) d = 147.9, 147.4, 146.5, 146.3, 133.9, 131.1, 130.6, 127.7, 124.4, 124.2, 122.3, 119.3, 108.8, 105.8, 101.3, 100.1. Preparation of isotope-labeled protein LC MS (RP18- 100Å, gradient 50% CH3CN/50% H2O to 100% CH3CN in 30 min), RT 14.2 min and mass 345.1 (100%), 347.2 (98%), [M + H]+. TLC (SiO2, n- hexane/EtOAc = 2/1) Rf 0.39, m.p. 195–197 °C. Preparation of α-synuclein ﬁbrils 13C NMR (100.6 MHz, CDCl3) d = 196.2 (d, J = 43.1 Hz, 13C- enriched), 151.7, 148.2, 132.1 (dd, J = 53.9, 14.1 Hz, 13C-enriched), 124.7, 108.0, 107.8, 101.8, 26.4 (d, J = 43.1 Hz, 13C-enriched). LC MS (RP18- 100Å, gradient 50% CH3CN/50% H2O to 100% CH3CN in 30 min), RT 6.2 min and mass 166.9 [M + H]+. TLC (SiO2, n-hexane/EtOAc = 3/1) Rf 0.5, m.p. 87–89 °C. DNP-enhanced ssNMR and bonds between all atoms in protein, lipid, and ligand molecules were constrained using the P-LINCS61 algorithm. Water molecules were constrained using SETTLE62. The integration timestep was set to 4 fs. Neighbor lists were updated with the Verlet list scheme50,63. The long- range electrostatic interactions were treated using the Particle Mesh Ewald (PME) method64,65, with a 1.0 nm real-space cutoff. The van der Waals interactions were cut off at 1.0 nm. A dispersion correction for energy and pressure was applied. For production runs, the tem- perature was kept at 300 K by applying the velocity-rescaling66 algo- rithm. The pressure was held constant at 1 bar by using the Parrinello–Rahman barostat67. All simulations were carried out using periodic boundary conditions for the simulation box. For DNP-enhanced ssNMR, ﬁbrils were grown as for ssNMR measure- ments, but with isotopic labeling (see the section “Preparation of isotope-labeled protein”) as indicated in the text. TEMTriPol-144 was synthesized based on published protocols45. A solution of 13C-depleted d8-Glycerol (70%-vol in water), containing 4–5 mM TEMTriPol was added to the resulting aggregate-lipid mixture and mixed thoroughly. Samples were packed into 3.2 mm zirconia ssNMR rotors via a custom-made ﬁlling device made from a truncated pipette tip. Finally, the sample was centrifuged into the rotor in an ultracentrifuge packing device for 30 min at 98,381 × g in a SW 32 Ti rotor in an Optima™L-80 XP Ultracentrifuge. All DNP spectra were recorded on a 600 MHz Bruker Avance III HD spectrometer at a magnetic ﬁeld of 14.1 T equipped with a 3.2 mm low temperature (LT) HCN MAS probe. MAS frequencies are indicated in Supplementary Table 1. For DNP radical-proton transfers, 395 GHz microwave irradiation from a gyrotron oscillator was delivered to the sample through a corrugated waveguide. Samples were cooled with a second-generation BRUKER liquid nitrogen cold cabinet, operating at approx. 100 K. Setup of α-synuclein protoﬁlament structure models. Simulations of α-synuclein β-sheet constructs were carried out based on struc- ture coordinates from the α-synuclein ﬁbril polymorph L2A (PDB ID 8A4L; EMDB ID EMD-15148)20 constituted by the residues 33-96, and the resolved parts of the N-terminus (residues 14-24). For simplicity, we refer to the protoﬁlament structure of polymorph L2A as L2, since it is structurally identical to the co-occurring polymorph L2B. The simulation model (see Supplementary Table 2) consisted of one protoﬁlament structure with ten β-strand layers stacked on top of each other, respectively. DNP-enhanced ssNMR A rhombic dodecahedron box was used for all simulations with the distance between the solute and the box edge of at least 2.5 nm. The resulting box size dimensions con- verged to ~12.8 × 12.8 × 12.8 nm3 after initial equilibration prior to starting the production runs. For each simulation system, the titratable groups of the protein structure were protonated accord- ing to their standard protonation states at pH 7, and the N- and C-termini of the peptides were capped with acetyl and N-methyl groups, respectively. Counterions (Na+, Cl−) were added to yield an ionic strength of 150 mM and to neutralize the net system charge. In all simulations, 100 POPC molecules were present and initially placed randomly in the solvent. The DNP ssNMR experiments were performed with a recycle delay of 2–4 s (1.2 × T1) for each sample and radio-frequency power levels of 100, 72, and 42 kHz for 1H, 13C, and 15N, respectively. During acquisition, during REDOR, as well as during radiofrequency driven recoupling (RFDR), SPINAL6446 decoupling was used with 100 kHz power on 1H. 2D 13C,13C spectrum, 1024 increments were acquired using 1.3 ms RFDR mixing using an xy-8 phase cycling and 83 kHz pi-pulses. 1H,13C cross-polarization (CP) was applied for 1.5 ms with a 90–100% ramp on the proton channel. For NHHC spectra, 32 increments were acquired with 200 μs of proton–proton mixing (Supplementary Table 1). The 1H,13C CP was set to 300 μs to limit transfers to the direct proton environment, and 1H,15N CP was applied for 800 μs with an 80–100% ramp on the proton channel. Frequency selective 13C,15N REDOR experiments47 were performed with 180° equivalent Q3 shape pulses on both 13C and 15N concurrently (Supplementary Table 2). The total time of the experiment was 4.67 h with a recycle delay of 2 s (4096 Scans). In total, the simulation systems contained ~155,000 atoms, including ~41,000 water molecules and 316 ions. The MD production runs. All production runs were preceded by a multi-step energy minimization and equilibration of the simulation systems. Initial velocities were taken according to the Maxwell- Boltzmann distribution at 300 K. Simulations were carried out in the presence and absence of anle138b. Two principal conditions were chosen for simulations with α-synuclein protoﬁlament models and anle138b: Solution NMR spectroscopy For determination of the conformation of anle138b we prepared a 20 mM solution of vesicles consisting of POPC and POPA (1:1) con- taining 1 mM anle138b in buffer (50 mM HEPES, 100 mM NaCl, pH 7.4) with 10% D2O and 100 μM 2,2-dimethyl-2-silapentane-5-sulfo- nate sodium salt (IUPAC: 3-(trimethylsilyl)propane-1-sulfonate, sodium salt). NOESY spectra were acquired with 72 scans using 256 increments in the indirect dimension and a relaxation delay of 2.3 s. The mixing time was set to 3 ms. Proton 90° ﬂip pulses were 14.4 μs. Water suppression was achieved by presaturation at a power level of 100 Hz. Experiments were recorded on a Bruker 900-MHz spec- trometer. The temperature during measurements was kept at 298 K. 2D datasets were processed in NMRPipe48 and analyzed in NMRFAM-Sparky43. Spectral traces were ﬁtted to a mixed Gaussian and Lorentzian function using the line shapes tool in Bruker Topspin 4.0.8. A. One anle138b molecule was modeled inside the internal cavity such that its long axis aligned with the protoﬁlament axis, and the solvent-facing atoms of the compound aligned with the surface of the protoﬁlament tip. The compound was inserted without interatomic overlaps to the side-chain atoms of the cavity enclosing residues, in a total of four different poses: with the pyrazole nitrogens oriented toward residues 67GGAV70 (inward- facing) or residues 81TVEGAGSI88 (outward-facing), and with an up or down positioning of the bromophenyl moiety of anle138b with respect to direction of the closest edge of the protoﬁlament. Each initial pose was energy minimized and subsequently equilibrated for 1 ns with position restraints acting on the protein-heavy atoms. B. Ten anle138b molecules were placed randomly in the solvent and at least 1 nm away from any protein-heavy atom. ssNMR spectroscopy 1H NMR (400 MHz, DMSO-d6 + 0.5% TFA) d = 8.05 (s, 1H), 7.85 (d, J = 7.7 Hz, 1H), 7.52 (d, J = 8.1 Hz, 1H), 7.35–7.41 (m, 3H), 7.26 (dd, J = 176.3, 6.2 Hz, 1H, 13C-enriched), 7.01 (d, J = 8.1 Hz, 1H), 6.07 (s, 2H). 13C NMR (100.6 MHz, DMSO-d6 + 0.5% TFA) d = 147.9 (d, J = 5.8 Hz, 13C-enriched), 147.4, 146.6 (d, J = 61.0 Hz, 13C-enriched), 146.3 (d, J = 62.7 Hz, 13C-enriched), 133.8, 131.1, 130.7, 127.7, 124.4 (d, J = 62.5 Hz, 13C-enriched), 124.2, 122.3, 119.3, 108.9 (d, J = 3.9 Hz, 13C-enriched), 105.8, 101.3, 100.1 (d, J = 62.7 Hz, 13C- enriched). LC MS (RP18-100Å, gradient 50% CH3CN /50% H2O to 100% 3D (H)CANH spectra on ﬁbrils in the absence of anle138b were recorded on a 950 MHz Bruker Avance III HD spectrometer at a mag- netic ﬁeld of 22.3 T equipped with a 0.7 mm HCDN probe and MAS at 100 kHz. The temperature of the cooling gas was set to 250 K, resulting in an estimated sample temperature of 20 °C. p p 2D 13C,13C-DARR spectra were acquired on an 850 MHz Avance III spectrometer with a 3.2-mm MAS HCN probe with mixing times of 200 ms at a magnetic ﬁeld of 20.0 T and MAS at 17 kHz. The tem- perature of the cooling gas was set to 265 K, resulting in an estimated sample temperature of 20 °C. 3D (H)CANH spectra were acquired in three blocks of 32 h in the presence and four blocks of 19 h in the absence of anle138b. All 2D 13C,13C DARR spectra were acquired in 34 blocks of 7.2 h in the presence and 25 equivalent blocks in the presence of anle138b. All related blocks were corrected for linear drift of the static magnetic ﬁeld using an in- house program executed from the command line in Bruker Topspin 4.0.842. The drift-corrected blocks were then averaged and processed as one spectrum. Spectra were analyzed using CcpNmr Analysis and NMRFAM-Sparky43. Peak intensities were determined by integrating cross-peaks using the Lorentzian ﬁt algorithm in Sparky, not allowing peak center motion. Assignments on α-synuclein ﬁbrils in the absence of anle138b had previously been reported19. Chemical shift perturba- tions were calculated as the average of HN, NH, Cα, and Cβ chemical shifts. Nature Communications\| (2022) 13:5385 6 Article https://doi.org/10.1038/s41467-022-32797-w Nature Communications\| (2022) 13:5385 References G. & Goedert, M. Synucleinopathies: past, present and future. Neuropathol. Appl Neurobiol. 42, 3–5 (2016). 6. 7. Cole, T. A. et al. α-Synuclein antisense oligonucleotides as a disease-modifying therapy for Parkinson’s disease. JCI Insight 6, e135633 (2021). 8. Wagner, J. et al. Anle138b: a novel oligomer modulator for disease- modifying therapy of neurodegenerative diseases such asprion and Parkinson’s disease. Acta Neuropathol. 125, 795–813 (2013). 9. Kuebler, L. et al. [11C]MODAG-001—towards a PET tracer targeting α-synuclein aggregates. Eur. J. Nucl. Med Mol. Imag. 48, 1759–1772 (2021). Anle138b-binding mode classiﬁcation. Based on the previously described heuristic classiﬁcation of anle138b interaction patterns to oligomeric forms of peptide aggregates in terms of atomic contact numbers and the presence of hydrogen bonds23, the following binding modes to ﬁbrillar aggregates were deﬁned: (I) “unbound”, no atomic contacts; (II) “partially bound”, fewer than half of the anle138b atoms in contact with peptide heavy atoms, mostly describing transient binding events; (III) “no HB or XB”, half or more of the anle138b atoms in contact with peptide heavy atoms, no HB or XB interactions to the backbone or side-chain atoms; (IV) “polar interactions”, at least one HB or XB between anle138b and the protoﬁlament structure; (IVa) “pyr- azole HB”, at least one HB between anle138b (via the pyrazole ring) and the protoﬁlament structure; (IVb) “benzodioxole HB”, at least one HB between anle138b (via the benzodioxole ring) and the protoﬁlament structure; (IVc) “bromophenyl XB”, at least one XB between anle138b (via the bromophenyl ring) and the protoﬁlament structure. 10. Levin, J. et al. The oligomer modulator anle138b inhibits disease progression in a Parkinson mouse model even with treatment started after disease onset. Acta Neuropathol. 127, 779–780 (2014). 11. Wagner, J. et al. Reducing tau aggregates with anle138b delays disease progression in a mouse model of tauopathies. Acta Neu- ropathol. 130, 619–631 (2015). 12. Martinez Hernandez, A. et al. The diphenylpyrazole compound anle138b blocks Aβ channels and rescues disease phenotypes in a mouse model for amyloid pathology. EMBO Mol. Med. 10, 32–47 (2018). 13. Heras-Garvin, A. et al. Anle138b modulates α-synuclein oligomer- ization and prevents motor decline and neurodegeneration in a mouse model of multiple system atrophy. Mov. Disord. 34, 255–263 (2019). 14. Wegrzynowicz, M. et al. Depopulation of dense α-synuclein aggregates is associated with rescue of dopamine neuron dys- function and death in a new Parkinson’s disease model. Acta Neu- ropathol. 138, 575–595 (2019). Article Article Analysis and visualization. The MD simulation trajectories were pre- processed and analyzed by tools provided by the GROMACS 2018.3 simulation software package49,50, as well as custom-made bash, awk, and python scripts. Renderings of atomic coordinates were car- ried out with the PyMOL molecular visualization system68. Graphs and ﬁgures were created using gnuplot 5.2 and inkscape 1.1. [https://doi.org/10.17617/3.R4FJQG]. Assigned chemical shift data (HN, Cα, Cβ, and C′) for α-synuclein ﬁbrils were deposited in the BMRB as updates under the accession number 50585. Atomic coordinates for L2A-ﬁbrils of αSyn used for MD simulations and visualization in this manuscript are available in the Protein Data Bank under the accession code 8A4L. Previously published structures of αSyn ﬁbrils used for a comparison of tubular ﬁbril cavities are available in the Protein Data Bank under the accession codes 6SSX, 7NCI, and 7NCH. All other data generated or analyzed during this study are included in this published article (and its supplementary information ﬁles). Source data are provided with this paper. [https://doi.org/10.17617/3.R4FJQG]. Assigned chemical shift data (HN, Cα, Cβ, and C′) for α-synuclein ﬁbrils were deposited in the BMRB as updates under the accession number 50585. Atomic coordinates for L2A-ﬁbrils of αSyn used for MD simulations and visualization in this manuscript are available in the Protein Data Bank under the accession code 8A4L. Previously published structures of αSyn ﬁbrils used for a comparison of tubular ﬁbril cavities are available in the Protein Data Bank under the accession codes 6SSX, 7NCI, and 7NCH. All other data generated or analyzed during this study are included in this published article (and its supplementary information ﬁles). Source data are provided with this paper. Relative free-energy binding proﬁle. For each simulation frame, the center of mass distance between the pyrazole ring of the anle138b molecule and the protoﬁlament tip (edge β-strand layer) was deter- mined (see e.g., Fig. 3e). The distances from each simulation trajectory set were combined and sorted into 1-d bins. The logarithm of the probability p = xi,j/xtotal was calculated for each bin i, j, where xi,j is the number of frames in the bin and xtotal is the total number of frames. Multiplying a constant factor (-RT, with T = 300 K) to p(x) results in G(x) in units of kJ/mol. Reporting summary Further information on research design is available in the Nature Research Reporting Summary linked to this article. 15. Levin, J. et al. Safety, tolerability and pharmacokinetics of the oli- gomer modulator anle138b with exposure levels sufﬁcient for therapeutic efﬁcacy in a murine Parkinson model: a randomised, double-blind, placebo-controlled phase 1a trial. eBioMedicine 80, 104021 (2022). Nature Communications\| (2022) 13:5385 Molecular dynamics (MD) simulations The GROMACS 2018.3 simulation software package49,50 was used to set up and carry out the MD simulations. For all simulations, the AMBER99SB*-ILDN51–53 force ﬁeld and the ion parameters in ref. 54 for use with the TIP3P water model55 were employed. A modiﬁed version56 of the all-atom Slipids force ﬁeld57–59 was used for the description of POPC lipids. The parameterization scheme for the small molecule compound anle138b is detailed elsewhere23. In order to speed up the simulations, aliphatic hydrogen atoms were converted to virtual sites60 The total simulation time for this study adds up to 100 μs, resulting in 50 μs sampling of the internally bound state of anle138b. From the individual simulation trajectories, snapshots were collected every 100 ps. Only the last 75% of the trajectory data were used for subsequent analyses as a measure to discard the initial equilibration phase and structural relaxation of the systems. Nature Communications\| (2022) 13:5385 7 https://doi.org/10.1038/s41467-022-32797-w References 1. Spillantini, M. G. et al. α-Synuclein in Lewy bodies. Nature 388, 839–840 (1997). 2. Graham, J. G. & Oppenheimer, D. R. Orthostatic hypotension and nicotine sensitivity in a case of multiple system atrophy. J. Neurol. Neurosurg. Psychiatry 32, 28 (1969). Contact analysis. The frequency of short-range interatomic interac- tions between the anle138b small molecules, POPC lipids, and the peptide aggregates were quantiﬁed with the g_contacts program69. An anle138b or POPC contact was considered as formed, if any heavy atom of the respective molecules was within a cutoff of 0.4 nm from any heavy atom of a protein residue. In addition, and according to the isotope labeling scheme in the NMR experiments, if the 1,2-N atoms of the anle138b molecule were found within a cutoff of 0.5 nm from any backbone carbon or nitrogen atom of a protein residue, a contact was considered formed. Contacts were averaged over time and over the individual trajectories of each simulation set. If not explicitly stated, only contacts to the core β-strand layers (neglecting the top β-strands on both protoﬁlament edges) were considered for analysis. Hydrogen bond (HB) energies were calculated according to the empirical formula provided in ref. 70. For the deﬁnition of halogen bonds (XB) the fol- lowing criteria were adopted from ref. 71: (a) distance between halogen atom and carbonyl oxygens ≤0.36 nm; (b) angle between XB and halogen–carbon bond ≥150°. 3. Fanciulli, A. & Wenning, G. K. Multiple-system atrophy. N. Engl. J. Med 372, 249–263 (2015). 4. Gibb, W. R. G., Esiri, M. M. & Lees, A. J. Clinical and pathological features of diffuse cortical Lewy body disease (Lewy body dementia). Brain 110, 1131–1153 (1987). 5. Gómez-Tortosa, E., Newell, K., Irizarry, M. C., Sanders, J. L. & Hyman, B. T. alpha-Synuclein immunoreactivity in dementia with Lewy bodies: morphological staging and comparison with ubiquitin immunostaining. Acta Neuropathol. 99, 352–357 (2000). 6. Spillantini, M. G. & Goedert, M. Synucleinopathies: past, present and future. Neuropathol. Appl Neurobiol. 42, 3–5 (2016). 7 Cole T A et al α-Synuclein antisense oligonucleotides as a 5. Gómez-Tortosa, E., Newell, K., Irizarry, M. C., Sanders, J. L. & Hyman, B. T. alpha-Synuclein immunoreactivity in dementia with Lewy bodies: morphological staging and comparison with ubiquitin immunostaining. Acta Neuropathol. 99, 352–357 (2000). 6 S ill ti i M G & G d t M S l i thi t t immunostaining. Acta Neuropathol. 99, 352 357 (2000). 6. Spillantini, M. Data availability NMR spectra raw data generated in this study, Negative stain EM micrographs for αSyn ﬁbrils in the presence of anle138b, and mass spectrometry data for the synthesis of selectively labeled α-synuclein and 13C- and 15N-labeled anle138b, respectively, have been deposited in the open research data repository Edmond at [https://doi.org/10. 17617/3.9C6TEW]. The molecular dynamics data generated in this study have been deposited in the Edmond data repository at 16. Deeg, A. A. et al. Anle138b and related compounds are aggregation speciﬁc ﬂuorescence markers and reveal high afﬁnity binding to α- synuclein aggregates. Biochim Biophys. Acta 1850, 1884–1890 (2015). Nature Communications\| (2022) 13:5385 8 Article https://doi.org/10.1038/s41467-022-32797-w 17. Sulzer, D. & Edwards, R. H. The physiological role of alpha-synuclein and its relationship to Parkinson’s disease. J. Neurochem 150, 475–486 (2019). 39. Khodaei, M. M., Alizadeh, A. & Nazari, E. Tf2O as a rapid and efﬁcient promoter for the dehydrative Friedel–Crafts acylation of aromatic compounds with carboxylic acids. Tetrahedron Lett. 48, 4199–4202 (2007). 18. Shahmoradian, S. H. et al. Lewy pathology in Parkinson’s disease consists of crowded organelles and lipid membranes. Nat. Neu- rosci. 22, 1099–1109 (2019). 40. Böckmann, A. et al. Characterization of different water pools in solid-state NMR protein samples. J. Biomol. NMR 45, 319 (2009). 19. Antonschmidt, L. et al. Insights into the molecular mechanism of amyloid ﬁlament formation: Segmental folding of α-synuclein on lipid membranes. Sci. Adv. 7, eabg2174 (2021). 41. Barbet-Massin, E. et al. Rapid proton-detected NMR assignment for proteins with fast magic angle spinning. J. Am. Chem. Soc. 136, 12489–12497 (2014). 42. Najbauer, E. E. & Andreas, L. B. Correcting for magnetic ﬁeld drift in magic-angle spinning NMR datasets. J. Magn. Reson 305, 1–4 (2019). 20. Frieg, B. et al. The 3D structure of lipidic ﬁbrils of α-synuclein. Preprint at https://www.biorxiv.org/content/10.1101/2022.03.02. 481946v1 (2022). 43. Lee, W., Tonelli, M. & Markley, J. L. NMRFAM-SPARKY: enhanced software for biomolecular NMR spectroscopy. Bioinformatics 31, 1325–1327 (2014). 21. Lange, A., Luca, S. & Baldus, M. Structural constraints from proton- mediated rare-spin correlation spectroscopy in rotating solids. J. Am. Chem. Soc. 124, 9704–9705 (2002). 22. Loquet, A., Gardiennet, C. & Böckmann, A. Protein 3D structure determination by high-resolution solid-state NMR. C. R. Chim. 13, 423–430 (2010). 44. Mathies, G. et al. Efﬁcient dynamic nuclear polarization at 800 MHz/ 527 GHz with trityl-nitroxide biradicals. Angew. Chem. 54, 11770–11774 (2015). 23. Matthes, D., Gapsys, V., Griesinger, C. & de Groot, B. L. Data availability Resolving the atomistic modes of Anle138b inhibitory action on peptide oli- gomer formation. ACS ChemNeurosci 8, 2791–2808 (2017). 45. Liu, Y., Villamena, F. A., Rockenbauer, A., Song, Y. & Zweier, J. L. Structural factors controlling the spin–spin exchange coupling: EPR spectroscopic studies of highly asymmetric trityl–nitroxide bir- adicals. J. Am. Chem. Soc. 135, 2350–2356 (2013). 24. Guerrero-Ferreira, R. et al. Two new polymorphic structures of human full-length alpha-synuclein ﬁbrils solved by cryo-electron microscopy. eLife 8, e48907 (2019). 46. Fung, B. M., Khitrin, A. K. & Ermolaev, K. An improved broadband decoupling sequence for liquid crystals and solids. J. Magn. Reson 142, 97–101 (2000). 25. Lövestam, S. et al. Seeded assembly in vitro does not replicate the structures of α-synuclein ﬁlaments from multiple system atrophy. FEBS Open Bio. 11, 999–1013 (2021). 47. Jaroniec, C. P., Tounge, B. A., Herzfeld, J. & Grifﬁn, R. G. Frequency selective heteronuclear dipolar recoupling in rotating solids: accurate 13C−15N distance measurements in uniformly 13C,15N-labeled peptides. J. Am. Chem. Soc. 123, 3507–3519 (2001). 26. Wang, T., Jo, H., DeGrado, W. F. & Hong, M. Water distribution, dynamics, and interactions with Alzheimer’s β-amyloid ﬁbrils investigated by solid-state NMR. J. Am. Chem. Soc. 139, 6242–6252 (2017). 48. Delaglio, F. et al. NMRPipe: a multidimensional spectral proces- sing system based on UNIX pipes. J. Biomol. NMR 6, 277–293 (1995). 27. Boyer, D. R. et al. Structures of ﬁbrils formed by α-synuclein her- editary disease mutant H50Q reveal new polymorphs. Nat. Struct. Mol. Biol. 26, 1044–1052 (2019). 49. Pronk, S. et al. GROMACS 4.5: a high-throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 29, 845–854 (2013). 28. Schütz, A. K. et al. The amyloid–Congo Red interface at atomic resolution. Angew. Chem. Int Ed. 50, 5956–5960 (2011). 50. Abraham, M. J. et al. GROMACS: high performance molecular simulations through multi-level parallelism from laptops to super- computers. SoftwareX 1-2, 19–25 (2015). 29. Niu, Z. et al. Mapping the binding interface of PET tracer molecules and Alzheimer disease Aβ ﬁbrils by using MAS solid-state NMR spectroscopy. ChemBioChem 21, 2495–2502 (2020). 51. Hornak, V. et al. Comparison of multiple Amber force ﬁelds and development of improved protein backbone parameters. Proteins: Struct., Funct., Bioinforma. 65, 712–725 (2006). 30. Shi, Y. et al. Cryo-EM structures of tau ﬁlaments from Alzheimer’s disease with PET ligand APN−1607. Acta Neuropathol. 141, 697–708 (2021). 52. Best, R. B. & Hummer, G. Data availability Optimized molecular dynamics force ﬁelds applied to the helix−coil transition of polypeptides. J. Phys. Chem. B 113, 9004–9015 (2009). 31. Duan, P. et al. Binding sites of a positron emission tomography imaging agent in Alzheimer’s β-amyloid ﬁbrils studied using 19F solid-state NMR. J. Am. Chem. Soc. 144, 1416–1430 (2022). 53. Lindorff-Larsen, K. et al. Improved side-chain torsion potentials for the Amber ff99SB protein force ﬁeld. Proteins: Struct., Funct., Bioinforma. 78, 1950–1958 (2010). 32. Kuang, G., Murugan, N. A. & Ågren, H. Mechanistic insight into the binding proﬁle of DCVJand α-synuclein ﬁbril revealed by multiscale simulations. ACS ChemNeurosci 10, 610–617 (2019). 54. Smith, D. E. & Dang, L. X. Computer simulations of NaCl association in polarizable water. J. Chem. Phys. 100, 3757–3766 (1994). 33. Pujols, J. et al. Small molecule inhibits α-synuclein aggregation, disrupts amyloid ﬁbrils, and prevents degeneration of dopaminer- gic neurons. Proc. Natl Acad. Sci. USA 115, 10481–10486 (2018). 55. Jorgensen, W. L., Chandrasekhar, J., Madura, J. D., Impey, R. W. & Klein, M. L. Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 79, 926–935 (1983). 34. Prade, E. et al. Structural mechanism of the interaction of Alzheimer disease Aβ ﬁbrils with the non-steroidal anti-inﬂammatory drug (NSAID) sulindac sulﬁde. J. Biol. Chem. 290, 28737–28745 (2015). 56. Melcr, J., Bonhenry, D., Timr, Š. & Jungwirth, P. Transmembrane potential modeling: comparison between methods of constant electric ﬁeld and ion imbalance. J. Chem. Theory Comput. 12, 2418–2425 (2016). 35. Frieg, B. et al. Quaternary structure of patient-homogenate ampli- ﬁed α-synuclein ﬁbrils modulates seeding of endogenous α- synuclein. Commun. Biol. (2022). In press. 57. Jämbeck, J. P. M. & Lyubartsev, A. P. An extension and further validation of an all-atomistic force ﬁeld for biological membranes. J. Chem. Theory Comput. 8, 2938–2948 (2012). 36. Fan, Y. et al. Different structures and pathologies of α-synuclein ﬁbrils derived from preclinical and postmortem patients of Parkin- son’s disease. Preprint at https://www.biorxiv.org/content/10.1101/ 2021.11.02.467019v1 (2021). 58. Jämbeck, J. P. M. & Lyubartsev, A. P. Derivation and systematic validation of a reﬁned all-atom force ﬁeld for phosphatidylcholine lipids. J. Phys. Chem. B 116, 3164–3179 (2012). 37. Hoyer, W. et al. Dependence of α-synuclein aggregate morphology on solution conditions. J. Mol. Biol. 322, 383–393 (2002). 59. Jämbeck, J. P. M. & Lyubartsev, A. P. Another piece of the mem- brane puzzle: extending slipids further. J. Chem. Theory Comput. 9, 774–784 (2013). 38. Lin, B. et al. Reprints and permission information is available at http://www.nature.com/reprints 71. Kolář, M. H. & Hobza, P. Computer modeling of halogen bonds and other σ-hole interactions. Chem. Rev. 116, 5155–5187 (2016). Publisher’s note Springer Nature remains neutral with regard to jur- isdictional claims in published maps and institutional afﬁliations. Funding g Open Access funding enabled and organized by Projekt DEAL. 63. Verlet, L. Computer “Experiments” on classical ﬂuids. I. Thermo- dynamical properties of Lennard-Jones molecules. Phys. Rev. 159, 98–103 (1967). Acknowledgements We thank M. Wegstroth and K. Giller for excellent technical help in protein expression and puriﬁcation. We also thank G. Heim for the acquisition of negative stain EM images. We thank J. Schimpfhauser, J. Bienert, and V.N. Belov from the facility for Synthetic Chemistry at the Max Planck Institute for Multidisciplinary Sciences, Göttingen for syn- thesizing the TEMTriPol-1 radical. This work was supported by the Max Planck Society and the Deutsche Forschungsgemeinschaft (DFG, Ger- man Research Foundation) under Germany’s Excellence Strategy-EXC 2067/1-390729940 as well as the Emmy Noether program (grant AN1316/1-1 to L.B.A.). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. https://doi.org/10.1038/s41467-022-32797-w G.F.S. provided αSyn ﬁbril atomic coordinates and discussion thereof. D.M. carried out MD simulations. L.A. and D.M. prepared ﬁgures and wrote the initial draft. S.R. and A.G. contributed to the interpretation of data in the medical context. B.L.dG., C.G., and L.B.A. supervised the project. All authors contributed to the writing of the manuscript. 60. Feenstra, K. A., Hess, B. & Berendsen, H. J. C. Improving efﬁciency of large time-scale molecular dynamics simulations of hydrogen-rich systems. J. Comput Chem. 20, 786–798 (1999). 61. Hess, B. P-LINCS: a parallel linear constraint solver for molecular simulation. J. Chem. Theory Comput. 4, 116–122 (2008). 62. Miyamoto, S. & Kollman, P. A. Settle: an analytical version of the SHAKE and RATTLE algorithm for rigid water models. J. Comput Chem. 13, 952–962 (1992). Competing interests 64. Darden, T., York, D. & Pedersen, L. Particle mesh Ewald: an N⋅log(N) method for Ewald sums in large systems. J. Chem. Phys. 98, 10089–10092 (1993). A.G., A.L., and S.R. are employed by MODAG GmbH, which licensed WO/2010/000372 in which anle138b is described. A.G. and C.G. are shareholders of MODAG GmbH. The remaining authors declare no competing interests. A.G., A.L., and S.R. are employed by MODAG GmbH, which licensed WO/2010/000372 in which anle138b is described. A.G. and C.G. are shareholders of MODAG GmbH. The remaining authors declare no competing interests. 65. Essmann, U. et al. A smooth particle mesh Ewald method. J. Chem. Phys. 103, 8577–8593 (1995). Additional information 66. Bussi, G., Donadio, D. & Parrinello, M. Canonical sampling through velocity rescaling. J. Chem. Phys. 126, 014101 (2007). Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-022-32797-w. Supplementary information The online version contains supplementary material available at 67. Parrinello, M. & Rahman, A. Polymorphic transitions in single crys- tals: a new molecular dynamics method. J. Appl. Phys. 52, 7182–7190 (1981). https://doi.org/10.1038/s41467-022-32797-w. Correspondence and requests for materials should be addressed to Bert L. de Groot, Christian Griesinger or Loren B. Andreas. Correspondence and requests for materials should be addressed to Bert L. de Groot, Christian Griesinger or Loren B. Andreas. 68. Schrodinger, LLC. The PyMOL Molecular Graphics System, Version 2.0. (2015). 69. Blau, C. & Grubmuller, H. g_contacts: fast contact search in bio- molecular ensemble data. Comput. Phys. Commun. 184, 2856–2859 (2013). Peer review information Nature Communications thanks Ashok Das- mahapatra, Alexey Potapov and the other anonymous reviewer(s) for their contribution to the peer review of this work. Peer review information Nature Communications thanks Ashok Das- mahapatra, Alexey Potapov and the other anonymous reviewer(s) for their contribution to the peer review of this work. 70. Espinosa, E., Molins, E. & Lecomte, C. Hydrogen bond strengths revealed by topological analyses of experimentally observed electron densities. Chem. Phys. Lett. 285, 170–173 (1998). Data availability 17O NMR studies of yeast ubiquitin in aqueous solution and in the solid state. ChemBioChem 22, 826–829 (2021). Nature Communications\| (2022) 13:5385 9 9 Article https://doi.org/10.1038/s41467-022-32797-w Author contributions L.A., D.M., B.L.dG, C.G., and L.B.A. conceptualized the project. S.B. oversaw protein expression and puriﬁcation. A.L. synthesized isotope-labeled anle138b. L.A. and V.S. prepared ﬁbril samples. L.A., R.D., E.N., and L.B.A. performed NMR experiments. B.F., C.D., and © The Author(s) 2022 © The Author(s) 2022 10 Nature Communications\| (2022) 13:5385 Nature Communications\| (2022) 13:5385
https://openalex.org/W2480008629	https://europepmc.org/articles/pmc4965028?pdf=render	English	null	The Hog Cycle of Law Professors: An Econometric Time Series Analysis of the Entry-Level Job Market in Legal Academia	PloS one	2,016	cc-by	12,156	RESEARCH ARTICLE OPEN ACCESS Citation: Engel C, Hamann H (2016) The Hog Cycle of Law Professors: An Econometric Time Series Analysis of the Entry-Level Job Market in Legal Academia. PLoS ONE 11(7): e0159815. doi:10.1371/ journal.pone.0159815 The Hog Cycle of Law Professors: An Econometric Time Series Analysis of the Entry-Level Job Market in Legal Academia Christoph Engel, Hanjo Hamann Max Planck Institute for Research on Collective Goods, Bonn, Germany Christoph Engel, Hanjo Hamann Max Planck Institute for Research on Collective Goods, Bonn, Germany * engel@coll.mpg.de * engel@coll.mpg.de Editor: Julia Lane, New York University, UNITED STATES Editor: Julia Lane, New York University, UNITED STATES Copyright: © 2016 Engel, Hamann. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Abstract The (German) market for law professors fulfils the conditions for a hog cycle: In the short run, supply cannot be extended or limited; future law professors must be hired soon after they first present themselves, or leave the market; demand is inelastic. Using a comprehen- sive German dataset, we show that the number of market entries today is negatively corre- lated with the number of market entries eight years ago. This suggests short-sighted behavior of young scholars at the time when they decide to prepare for the market. Using our statistical model, we make out-of-sample predictions for the German academic market in law until 2020. a1111 Introduction What do pigs and law professors have in common? They come in cycles! German academics who observe historical fluctuations on the market for law professors are occasionally overheard mumbling, tongue in cheek: It’s a hog cycle. In this paper we read this sigh as a testable proposi- tion. We show that the German market for law professors fulfils the theoretical conditions for the emergence of a hog cycle, and test empirically whether academics allow themselves to be governed by the brute logic of supply and demand, or whether their job market behavior is mit- igated by additional forces that are not part of the neoclassical model. This endeavor contrib- utes to a growing literature which addresses the fact that even today we still “know surprisingly little about the market for law professors” [1]:4. The Hog Cycle of Law Professors Fig 1. Hog Cycle That Started the Literature ([2]: 10). English translation of German original: Hog Prices in Berlin July 1896 to June 1914 (Pigs with Live Weight 80–100 kg) in Mark [German currency of the time] per 100 kg live weight. doi:10.1371/journal.pone.0159815.g001 Fig 1. Hog Cycle That Started the Literature ([2]: 10). English translation of German original: Hog Prices in Berlin July 1896 to June 1914 (Pigs with Live Weight 80–100 kg) in Mark [German currency of the time] per 100 kg live weight. Fig 1. Hog Cycle That Started the Literature ([2]: 10). English translation of German original: Hog Prices in Berlin July 1896 to June 1914 (Pigs with Live Weight 80–100 kg) in Mark [German currency of the time] per 100 kg live weight. doi:10.1371/journal.pone.0159815.g001 Fig 1. Hog Cycle That Started the Literature ([2]: 10). English translation of German original: Hog Prices in Berlin July 1896 to June 1914 (Pigs with Live Weight 80–100 kg) in Mark [German currency of the time] per 100 kg live weight. doi:10.1371/journal.pone.0159815.g001 doi:10.1371/journal.pone.0159815.g001 early 20th century Berlin (Germany) which ignited academic research on this issue. As one sees, apart from a positive time trend (in the long run, Germans spent more money on buying pork), pork prices were fluctuating cyclically, resulting from cycles of over- and undersupply. p ) p p g y y g y pp y Theoreticians call the underlying logic a cobweb [3, 4]. Three conditions must be fulfilled for a cycle to emerge: (1) in the short run production cannot react to changes in demand; (2) the product in question is perishable; (3) demand in the short run does not increase if there is more supply [3]. Pigs meet all three conditions: (1) Newborn piglets need time to mature; (2) there is the optimal moment for slaughter; (3) if pork is cheap, people hardly eat more. Assume the demand for pork drops unexpectedly, say since consumers dread swine flu. Farmers have to lower prices if they want to sell any pork. Since this meant that profit was low, this year they breed fewer pigs. If these pigs are ready for slaughter, supply is small and prices are high. Farm- ers react by breeding a larger next generation. When that reaches the market, supply is in excess of demand. What originally was an exogenous shock has triggered a cycle. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 On pigs. . . As its name suggests, the concept of “hog cycles” hails from agricultural economics. It describes the idea that once farmers observe a high pork price, they tend to start breeding pigs all at the same time, which means that one year later (when pigs are ready for slaughter) the market will be swamped with pork. Since pork now exceeds demand, its price drops, thus incentivizing farmers to reduce (or cease) their breeding activity, which in turn produces a swing in the opposite direction: A year later, too few pigs are available for slaughter, so demand exceeds sup- ply, prices explode and the cycle restarts. Fig 1 illustrates this principle with the example from Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: The authors have no support or funding to report. Competing Interests: The authors have declared that no competing interests exist. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 1 / 22 . . . and professors Given how “little is empirically known about what explains success in the market for law pro- fessors” [1]:1, professors-to-be have to take decisions in a complex market based on very few observables. The most salient such observable is the number of positions available when the candidate considers embarking on her academic career. If there are presently more open posi- tions than professors on the market, post-docs may be tempted to start a career in academia. Years later, however, they may find that many others had the same intuition, thereby creating an oversupply which in turn discourages junior researchers and will soon lead to law schools having problems to fill their vacancies. If this were true, we would observe a hog cycle. Of course, future law professors are not hogs, tenure is not slaughter, and hiring a new assistant professor is not breeding. But in theory, the market for law professors fulfils the three condi- tions, outlined above, which the emergence of a cycle requires. Note a difference between the proverbial market for hogs, and the market for law professors, though. If there is excess supply, the price of meat may be reduced. If there is excess demand, the price of meat may increase. This channel for short-term adjustment is shut off in the Ger- man market for law professors. Prices, i.e. salaries, are defined by statute, and do not react to market fluctuations. In the meat market, the inefficiency results from a mismatch between investment (in rearing piglets) and prices. In the market for professors, the inefficiency results from a mismatch between investment (in habilitation) and the probability of being hired as a professor. Yet as we explain in the following paragraphs, the conditions for a cobweb are no less fulfilled. (1) Becoming a law professor takes time. In the German system, scholars are eligible only if they have earned their doctor juris (Dr. iur., roughly J.S.D. equivalent, not just a J.D.) and have subsequently passed “habilitation” (from Latin habilitare, to qualify, [25]). Habilitation is a for- mal procedure by which the entire professorate of the law school testifies that a candidate is competent to teach law. It requires writing a second book (traditionally called Habilitationss- chrift, habilitation thesis), having a decent list of publications, and convincing the entire law school during a talk and discussion. Legally, candidates may pass habilitation without having held a post-doc position. This holds whenever supply and demand react symmetrically, i.e., the elasticity of demand is similar to the elasticity of supply. If supply is more elastic than demand, the cycle turns into a self-rein- forcing spiral. If demand is more elastic than supply, the cycle eventually vanishes [3]. Of course, there is no cycle to start with if a sufficiently large fraction of farmers anticipate this and adapt their breeding policy [5, 6]. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 2 / 22 The Hog Cycle of Law Professors These mechanisms are now well understood, thanks to a wealth of academic research. The first papers date back almost a century [2, 7] and were immediately translated into practical advice for farmers, urging them to adopt an anti-cyclical investment policy [8]. Later, Nobel winners Ronald Coase [9, 10] and Paul Samuelson [11] have contributed to this literature, as well as a long list of agricultural economists [12–19]. Similar phenomena have been docu- mented on the market for potatoes [3, 20], for real estate [21, 22], for oil [23], and for nurses [24]. So how about bright young academics? PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 . . . and professors The candidate could just submit her thesis, plus her list of publica- tions, to her law school of choice and ask to be habilitated. Yet this legal option is almost never used. One reason is financial: Outside university, there are practically no jobs or stipends for future law professors. The other, more important reason concerns habilitation itself: Getting the whole law school to approve the candidate is a political exercise. The candidate needs strong backing by her supervisor. Effectively, law schools see this as an occasion to express approval or disapproval with the supervisor’s own scholarly work. The candidate is regarded as the supervisor’s pupil, and support for the supervisor’s academic program is at stake. If law schools do not receive this signal of credibility, odds for success are very low. This implies that, when she accepts a candidate as a post-doc, the supervisor engages her own academic reputa- tion. Consequently, preparing for the academic market is practically a joint decision by the supervisor and the candidate. 3 / 22 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 The Hog Cycle of Law Professors This explains why the standard career path follows the apprentice system. Some esteemed law professor hires the future colleague as a post-doc (Habilitand). Once the post-doc has been hired, it is rare that she leaves academia before finishing habilitation. (There are no statistics on the drop out rate, but our own experience and hearsay suggest that it is negligible). Having spent additional time at the university would not give her an advantage on the market for non- academic lawyers. On the contrary, she would be perceived as less inclined to do practical work, if not poorly equipped for the challenges of legal practice. Holding a post-doc position is thus perceived as a negative signal by the market for non-academic lawyers. Most post-docs would also regard it as personal failure if they did not pass habilitation; they stick to their habil- itation plans not least for reasons of self-esteem. The typical post-doc position is for the dura- tion of six years, but finishing the thesis often takes a bit longer (and post-docs and their supervisors creatively need to find her a living for that extra period). . . . and professors Once a Privatdozent has taken up a position in legal practice, returning to academia is close to impossible. This makes future law professors a perishable commodity. Note that future law professors also have very little room for changing the moment when they enter the market, in anticipation of changing market conditions. Speeding up the writing process is a frequent plan, but it hardly ever suc- ceeds. Postponing habilitation is usually not possible either, since positions expire. (3) To date there are only two private law schools in Germany (BLS in Hamburg, EBS in Wiesbaden). The remaining 42 law schools [26] and other departments that occasionally hire lawyers (e.g., business or technology) are funded by the state. This helped law schools during the financial crisis, as their budgets were basically unaffected. Yet it also means that law schools cannot react to an oversupply of excellent candidates by creating new positions. Moreover, pro- fessors are public servants, with salaries fixed by statute. Therefore law schools cannot react to excess supply by lowering prices either. Sadly, if there are more good candidates than the mar- ket can take, the market does not clear. Those who have not been hired must change their pro- fession. Most of them join law firms. The third condition for a cycle is thus also fulfilled: Demand is not elastic enough to parry fluctuations of supply. (3) To date there are only two private law schools in Germany (BLS in Hamburg, EBS in Wiesbaden). The remaining 42 law schools [26] and other departments that occasionally hire lawyers (e.g., business or technology) are funded by the state. This helped law schools during the financial crisis, as their budgets were basically unaffected. Yet it also means that law schools cannot react to an oversupply of excellent candidates by creating new positions. Moreover, pro- fessors are public servants, with salaries fixed by statute. Therefore law schools cannot react to excess supply by lowering prices either. Sadly, if there are more good candidates than the mar- ket can take, the market does not clear. Those who have not been hired must change their pro- fession. Most of them join law firms. The third condition for a cycle is thus also fulfilled: Demand is not elastic enough to parry fluctuations of supply. Given that these theoretical conditions are fulfilled, one spark might have been enough to ignite the cycle. . . . and professors According to statistical data reported by Germany’s Federal Bureau of Statistics, the time passed between earning the doctoral degree and passing habilitation was 7.76 years on average over all years for which data is available (1960 to 1989). There is thus a time lag of about eight years between the decision to prepare a new scholar for the market and the moment when she can apply for a position. The short run supply of law professors is inelastic. (2) By passing habilitation, a young scholar becomes a Privatdozent, i.e., an academic enti- tled to teach, but not usually on a salaried position. Since the law school that granted habilita- tion is legally prohibited from hiring this candidate itself, the candidate must go on the market and hope to be hired by a different law school. Most Privatdozenten are on the market for a year or two. If they have not found a position within three or four years, odds are low that they ever will. The law schools take the fact that a Privatdozent has not been hired by other law schools as a signal that something must be wrong with this candidate. Once a Privatdozent has taken up a position in legal practice, returning to academia is close to impossible. This makes future law professors a perishable commodity. Note that future law professors also have very little room for changing the moment when they enter the market, in anticipation of changing market conditions. Speeding up the writing process is a frequent plan, but it hardly ever suc- ceeds. Postponing habilitation is usually not possible either, since positions expire. (2) By passing habilitation, a young scholar becomes a Privatdozent, i.e., an academic enti- tled to teach, but not usually on a salaried position. Since the law school that granted habilita- tion is legally prohibited from hiring this candidate itself, the candidate must go on the market and hope to be hired by a different law school. Most Privatdozenten are on the market for a year or two. If they have not found a position within three or four years, odds are low that they ever will. The law schools take the fact that a Privatdozent has not been hired by other law schools as a signal that something must be wrong with this candidate. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 Hypothesis Given the theoretical considerations outlined above, we have reason to expect the occurrence of a cycle in the German job market for legal academics, with a duration between six and nine years (eight years being the typical time needed until habilitation). This leads to our hypothesis: The supply of German law professors exhibits negative autocorrelation with a lag between d Given the theoretical considerations outlined above, we have reason to expect the occurrence of a cycle in the German job market for legal academics, with a duration between six and nine years (eight years being the typical time needed until habilitation). This leads to our hypothesis: years (eight years being the typical time needed until habilitation). This leads to our hypothesis: The supply of German law professors exhibits negative autocorrelation with a lag between six and nine years. y g y g yp yp The supply of German law professors exhibits negative autocorrelation with a lag between six and nine years. To the best of our knowledge, we are the first to put this hypothesis forward, let alone test it. There is one predecessor paper on the German market for law professors, covering the period from 1949 to 1969 [28], but it only reports aggregate descriptive statistics; a more recent empir- ical study of the US job market for law professors [1] did not gather time series data. In the remainder of the paper, we describe our data (2.) and explore our hog cycle hypothe- sis by testing for negative autocorrelation and adding control variables (3.). Finally, we discuss the robustness of our model using additional specifications (4.) and conclude by using our empirical model to predict the supply of law professors for another ten years (5.). . . . and professors In the past, the German market for law professors was hit by exogenous shocks on multiple occasions. The two most prominent shocks happened to be positive ones. In the 1960s and early 70s, the German government decided to invest heavily in university education. No less than 15 new law schools were founded in Augsburg, Bayreuth, Bielefeld, Bochum, Bre- men, Düsseldorf, Gießen, Hagen, Hamburg II, Hannover, Konstanz, Mannheim, Osnabrück, Passau and Regensburg. Many additional law professors were needed to fill these new posi- tions. The next shock came with reunification. The former German Democratic Republic had five law schools (Berlin, Halle, Jena, Leipzig and Potsdam), and four more were established shortly after reunification (Dresden, Frankfurt/Oder, Greifswald and Rostock). Since most East PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 4 / 22 The Hog Cycle of Law Professors German law professors were believed to be too close to the Communist regime, most of them soon lost their positions. They were replaced with candidates from the West. In both periods, demand heavily exceeded supply. In theory, either shock would have been sufficient to start a hog cycle. German law professors were believed to be too close to the Communist regime, most of them soon lost their positions. They were replaced with candidates from the West. In both periods, demand heavily exceeded supply. In theory, either shock would have been sufficient to start a hog cycle. In 2002, the position of a “junior professor” was introduced to the German academic system to supplant traditional “habilitation” procedures. Law schools, however, were reluctant to offer such positions: Their number increased from 1 when first introduced to merely 20 at the end of our sample (German Federal Bureau of Statistics schedule H201.250), meaning that no more than 25 lawyers had ever held a junior professor position within our sampling window. Even five years later, in 2014, there were less than 40 positions in total, and virtually none was reported to come with a tenure track option [27]. Passing habilitation remains the only de facto possibility to enter the market for professors. Once a scholar has passed habilitation, she is no longer eligible for junior professorship, so these positions cannot absorb excess supply either. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 Materials and Methods We want to explain the number of habilitations in a given year with the number of habilitations six to nine years prior. We thus need a time series, which we generate from data about individ- ual habilitations. Ideally we would like to know, for each candidate who has passed habilitation, her name (to double-check with alternative sources), the sub-discipline (private, public, or criminal law), and, most importantly, the year when the candidate passed habilitation. We can get these infor- mation from two partly overlapping sources: A statute from 1969 (see Federal Law Gazette BGBl. vol. I, pp. 265) created the German National Library and obliges all publishing houses to submit at least one copy of each newly published book to that library. Based on these submissions, the “German National Bibliogra- phy” has subsequently been compiled and published. If the book in question is a habilitation thesis, this is noted in the bibliography, together with the year of the habilitation procedure. To the extent possible, the statute on the German National Library was applied retroactively. Data for habilitations seem reasonably reliable from 1960 onwards. Unfortunately, sampling revealed that this data is partly incomplete. To complete the data, we exploit the fact that a Who’s Who of German academics, the Kürschners Gelehrtenkalender, was recently made available electronically (www.degruyter.com/view/db/kdgo). In this digest, professors of all disciplines at universities of German language self-report biographical 5 / 22 The Hog Cycle of Law Professors sketches. Since its web interface allowed merely the exporting of names, not the year of habili- tation, the sub-discipline and the university, we had to complete these data points by hand. Hand coding was also necessary to identify persons who had been invited to contribute to the digest without having passed habilitation. This mainly concerns honorary professors, i.e., prac- titioners who receive the honorary title of a professor in exchange for teaching students free of charge. We have removed such entries from our dataset. We have matched these two datasets by name and removed all duplicates. The resulting gross dataset contains 2071 data points. We thus have complete data of 2071 habilitations for the time between 1876 and 2009. (Data from 2010 onwards is still too incomplete to include.) However, we truncate our dataset in 1960, before which year we cannot rely on the records to be complete. Materials and Methods Since sub-disciplines do not make a pronounced difference, in the remainder of the paper we work with the complete dataset. Fig 3 demonstrates that the development in the sub-disciplines was very similar. The only remarkable difference concerns the early spike (which was more pronounced in private law) and the recent spike (which was more pronounced in public law). Furthermore, habilitation data in criminal law is noisier, which may be attributed to the comparatively small number of law professors organized in this association. The overall shape of the curve is nonetheless simi- lar. Since sub-disciplines do not make a pronounced difference, in the remainder of the paper we work with the complete dataset. Materials and Methods This leaves us with 1993 data points, which are available (with control variables included) as S1 File. There is a predecessor publication reporting descriptive statistics of habili- tations in law from World War II until 1969 [28]. It relies on a survey, but unfortunately reports only aggregates. We therefore cannot use this publication to extend our time series fur- ther into the past. In our regressions, we work with the total number of habilitations per year. Our original dataset comprises 50 observations. Depending on the length of the lag, it is reduced to the number of years for which we can observe the lag. Given the inevitably small number of obser- vations, it is all the more remarkable that we find very robust results. Finally, to identify sub-disciplines, we have double-checked our data with lists compiled by the associations of private-law professors (website of Zivilrechtslehrervereinigung, www.zlv- info.de, accessed on 1 Oct 2010), of criminal-law professors (courtesy of Prof. Bernd Schüne- mann in private correspondence on 20 Jan 2011), and of public-law professors (website of Ver- einigung der Deutschen Staatsrechtslehrer, http://vdstrl.zar-muenster.de, accessed on 1 Oct 2010). Fig 2 shows that the production of law professors has not been smooth. There was a first spike in the early 70s and a second shortly after the year 2000. In the intermediate period, and possibly also in the current years, there is a much smaller supply of law professors. Note that visual inspection of the raw data already suggests a similarity with the hog cycle graph in Fig 1. Fig 2 shows that the production of law professors has not been smooth. There was a first spike in the early 70s and a second shortly after the year 2000. In the intermediate period, and possibly also in the current years, there is a much smaller supply of law professors. Note that visual inspection of the raw data already suggests a similarity with the hog cycle graph in Fig 1. Fig 3 demonstrates that the development in the sub-disciplines was very similar. The only remarkable difference concerns the early spike (which was more pronounced in private law) and the recent spike (which was more pronounced in public law). Furthermore, habilitation data in criminal law is noisier, which may be attributed to the comparatively small number of law professors organized in this association. The overall shape of the curve is nonetheless simi- lar. Autocorrelation If there is a hog cycle of law professors, the supply of professors today must be negatively corre- lated with the supply of professors x years ago. As a first test for our hypothesis, we estimate time series regressions with an autocorrelation term. Specifically we estimate ð1Þ yt ¼ a þ rytx þ εt ð1Þ where yt is the number of habilitations in year t, α is a constant term, yt-x is the number of habilitations x years ago, and εt is residual error. The coefﬁcient of interest is ρ, which we expect to be negative and signiﬁcant. The peaks in Fig 2 suggest that the data might be hetero- skedastic, which is why we estimate Huber-White robust standard errors. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 6 / 22 The Hog Cycle of Law Professors Autocorrelation requires the coefficient for the number of habilitations x years ago to be sig- nificantly different from zero. Fig 4 compresses the findings from 17 time series regressions with different lags. It illustrates two findings: First, all coefficients for lags between length 8 and 21 are negative, i.e., below the zero line. We thus do find a hog cycle. If there have been many tutionalised Sub-disciplines. Total number of habilitations per year. Fig 2. Production of Law Professors. Total number of habilitations per year. doi:10.1371/journal.pone.0159815.g002 Fig 2. Production of Law Professors. Total number of habilitations per year. doi:10.1371/journal.pone.0159815.g002 Fig 2. Production of Law Professors. Total number of habilitations per year. doi:10.1371/journal.pone.0159815.g002 doi:10.1371/journal.pone.0159815.g002 Autocorrelation requires the coefficient for the number of habilitations x years ago to be sig- nificantly different from zero. Fig 4 compresses the findings from 17 time series regressions with different lags. It illustrates two findings: First, all coefficients for lags between length 8 and 21 are negative, i.e., below the zero line. We thus do find a hog cycle. If there have been many Fig 3. Development in the Three Institutionalised Sub-disciplines. Total number of habilitations per year. doi:10.1371/journal.pone.0159815.g003 Fig 3. Development in the Three Institutionalised Sub-disciplines. Total number of habilitations per year. doi:10 1371/journal pone 0159815 g003 doi:10.1371/journal.pone.0159815.g003 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 7 / 22 The Hog Cycle of Law Professors Fig 4. Autocorrelation (absent controls). Dots are coefficients of the lag of the number of habilitations x years ago from regression (1), bars represent the standard error of this regressor. doi:10.1371/journal.pone.0159815.g004 Fig 4. doi:10.1371/journal.pone.0159815.t001 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 Autocorrelation Autocorrelation (absent controls). Dots are coefficients of the lag of the number of habilitations x years ago from regression (1), bars represent the standard error of this regressor. doi:10.1371/journal.pone.0159815.g004 doi:10.1371/journal.pone.0159815.g004 habilitations in the past, the statistical model predicts few habilitations today, and vice versa. This is what we expected. However, the error bars indicate that only lags from nine to 16 years are significantly different from zero. For the lag at the upper end of our expectation, i.e., a lag of nine years, the autocorrelation is weakly significant (p = .051), see Table 1. This model predicts that there are 52.82 new law pro- fessors on the market every year, minus .258 times the number of habilitations nine years ago. Since this number has never been zero, the predicted number of habilitations is always smaller than 52.82 –but the more so, the larger the supply in the past. The negative correlation is pro- nounced. Four habilitations in the past seem to deter one habilitation now. Table 1. Pure Autocorrelation. lag9 habil -.258+ cons 52.816* N 41 p model .051 R2 .0585 Table 1. Pure Autocorrelation. lag9 habil -.258+ cons 52.816* N 41 p model .051 R2 .0585 depvar: # habilitations in year x, robust standard errors * p < .001, p < .01, * p < .05, + p < .1 doi:10.1371/journal.pone.0159815.t001 8 / 22 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 The Hog Cycle of Law Professors Our data thus support the theoretical prediction of negative autocorrelation, for a lag at the upper end of our expectation. Yet deeper lags are also significant, and there are also statistical reasons to revisit the estimation. Even the best-performing model with one lag explains only 13% of the variance. Obviously a lot of the fluctuation in the supply of law professors is still unexplained. More importantly even, as long as we miss those additional factors that determine the development of habilitations over time, the coefficients of that time lag(s) in Table 1 risk being inconsistent due to omitted variable bias. In the next section, we introduce control vari- ables to mitigate this bias. Control Variables Fig 5 shows that, despite the fluctuations, the number of habilitations grows over time. (The red line graphs the predicted number of habilitations if we regress the output of law faculties only on the respective year.) Over the 50 years of observation, legal academia has been growing. The German apprenticeship model of academic careers takes considerable time. Data by the German Federal Bureau of Statistics (schedule 4.15.2a) show that between 1980 and 2008, the mean age at habilitation was 39.76 (standard deviation .92), so future law professors are around the age of 40 upon entering the market. One should expect that there are more candidates the larger the respective birth cohort is. The number of those who bring the right talents should be more or less proportional to the number of those born forty years before. Consider Fig 6, which depicts data on birth cohorts from the German Federal Bureau of Statistics. Fig 6 illustrates why birth cohorts are relevant for understanding the production of law pro- fessors. The data contain several structural breaks: one drop in birth cohorts, presumably caused by the economic crisis in the early 30s (playing itself out in the early 70s), and an even more pro- nounced dip, which resulted from the end of World War II (and played itself out in the mid- 1980s). Conversely, steep increases resulted from the baby booms in the early Nazi years (becom- ing effective in the late 70s) and during the recovery after World War II (peaking in the early 2000s). Overall, there is the clear negative trend characteristic of modern affluent societies. Fig 5. Time Trend. doi:10.1371/journal.pone.0159815.g005 PLOS ONE \| DOI:10 1371/journal pone 0159815 July 28 2016 9 / 22 Fig 5. Time Trend. doi:10.1371/journal.pone.0159815.g005 doi:10.1371/journal.pone.0159815.g005 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 9 / 22 The Hog Cycle of Law Professors Fig 6. Development of Birth Cohorts over Time. doi:10.1371/journal.pone.0159815.g006 Fig 6. Development of Birth Cohorts over Time. doi:10.1371/journal.pone.0159815.g006 The right panel of Fig 6 is even more important. It shows that birth numbers also exhibit autocorrelation. The autocorrelation is negative in the short run, and most pronounced around eleven years, and positive in the long run, with a peak at 27 years. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 Control Variables Now the original time series regressions from Fig 4 cannot discriminate between the two channels of negative autocorrela- tion: the one resulting from parents’ decisions on family planning, and the other resulting from the future candidate’s decisions on her own career. Controlling for the size of birth cohorts will isolate the latter effect. The supply of professors need not be proportional to the size of the birth cohort. It would only be proportional if the composition of professions in the population were stable over time. Now, in the period of observation, Germany has experienced what is often referred to as the “education revolution”. While at the beginning of our time series only little more than 5% of a cohort passed high-school exams (Abitur) and were therefore eligible for university education, this number grew steadily and has reached almost 25% today [29]. If the supply of professors is proportional to those members of a birth cohort who go to university, rather than to the total size of the cohort, we still have an omitted variable. We therefore use the size of the student cohort of which the future law professor is part as an additional control variable. Specifically, we use the number of students who were enrolled in German universities 20 years ago. We take this data from the German Federal Bureau of Statis- tics. The lag of 20 years is motivated by the fact that a typical law professor is around 40 when she is on the market, and that Germans typically enter university around age 20. If we add the time trend, the birth cohort 40 years ago, and the student cohort 20 years ago to the specification of (1), only shorter lags of the number of habilitations turn out significant. The best performing model now has a highly plausible lag of eight years (Fig 7). The regressions in Table 2 show that, once we control for the size of the student cohort, the effect of the birth cohort becomes immaterial. In model 2, the regressor of the birth cohort is insignificant. Adding this control variable does not explain any additional variance. Two more findings are remarkable. With these regressions, we explain a huge proportion of the variance, namely more than 75%. Moreover, once we control for the size of the student cohort, the over- all time trend becomes negative. Student cohorts grow faster than calendar time. depvar models 1 and 2: # habilitations in year x, robust standard errors; depvar model 3: student cohort, 20 years ago; time trend is 0 in 1960, and grows by 1 every year; student cohort is # of newly inscribed students in all German universities, 20 years ago; birth cohort is # of births (in 1,000), 40 years ago * p < .001, doi:10.1371/journal.pone.0159815.t002 Control Variables PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 10 / 22 The Hog Cycle of Law Professors Fig 7. Adding Time Trend, Birth Cohort and Student Cohort. Dots are the coefficients of the lag of the number of habilitations x years ago, bars represent the standard error of this regressor. doi:10.1371/journal.pone.0159815.g007 Fig 7. Adding Time Trend, Birth Cohort and Student Cohort. Dots are the coefficients of the lag of the number of habilitations x years ago, bars represent the standard error of this regressor. doi:10.1371/journal.pone.0159815.g007 Actually we can even be more sophisticated. In model 3 of Table 2, we regress student cohorts 20 years ago on birth cohorts 40 years ago and the time trend. Both coefficients are sig- nificant, and have the expected signs: student cohorts grow over time, and they grow faster than birth cohorts (which is why, even when controlling for birth cohorts, the time trend remains significant and negative). In the last step, using the procedure introduced by [30], we Table 2. Adding Time Trend and Student Cohort. model 1 model 2 model 3 depvar # habil # habil student cohort time trend -2.772* -2.366* 2.210* student cohort lag20 1.659* 1.522* birth cohort lag40 .011 .021* lag8 habil -.335* -.405 cons 68.087* 53.138 -44.277* N 42 42 42 p model < .001 < .001 < .001 R2 .7738 .7808 .9353 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 11 / 22 The Hog Cycle of Law Professors can test whether birth cohorts have a significant indirect effect on the number of habilitations, mediated by student cohorts (for background see [31]). This turns out to be the case (z of indi- rect effect 3.867, p = .00011). Thus birth cohorts matter chiefly because they affect how many students go to university, which in turn affects how many graduates can become professors. Alternative Career Options Arguably, when they consider preparing themselves for a career in academia, promising young lawyers consider attractive alternative career options. Recall that, once they have become a post-doc, it is very unusual for lawyers to leave university and go to practice (unless forced to do so since, post-habilitation, they cannot find a job at a university). Therefore the relevant moment for exploring alternative career paths is after having earned their Dr. juris, and before taking up a post-doc position. In the German system, future law professors typically consider one of two alternative career paths. Those who are interested in earning money would join one of the big law firms. Those more interested in advancing justice, serving their country, or maybe having a couple of chil- dren soon (note that the system of maternity leave is most generous for public servants), would rather join the judiciary. In Germany, lawyers may become judges right after having finished their legal education. After the age of 32, the judiciary normally no longer hires, meaning that lawyers stand little chance of becoming judges should they pass habilitation, but are not hired by a law school. All over the country, there are more than 20,000 judges. Careers in the judi- ciary tend to be slow, but can ultimately lead to the Supreme Court or to one of the prestigious appellate courts. Salaries are rather modest and fixed by statute. Given that we have already seen that a lag of eight years is most relevant, we also check for career opportunities eight years before a generation of candidates is on the market. Fig 8. Number of Judges in Germany (in 1,000). doi:10.1371/journal.pone.0159815.g008 J l / Fig 8. Number of Judges in Germany (in 1,000). doi:10.1371/journal.pone.0159815.g008 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 12 / 22 The Hog Cycle of Law Professors Fig 9. GNP Development as a Proxy for earning Prospects in Law Firms. doi:10.1371/journal.pone.0159815.g009 Fig 9. GNP Development as a Proxy for earning Prospects in Law Firms. doi:10.1371/journal.pone.0159815.g009 Entry into the judiciary critically depends on hiring. As Fig 8 shows, until the mid-90s, the number of judges grew slowly, with the only peak resulting from reunification. We take this data from the biennial employment reports of the German Federal Ministry of Justice. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 The Hog Cycle of Law Professors Table 3. Adding Measures for Alternative Career Options. model 1 model 2 model 3 model 4 time trend -1.820* -2.469* -6.361* -5.029* # judges 9.026 -1.901 GNP 447.808*** 162.010* student cohort lag20 1.767* 1.520* lag8 habil -.258+ -.358* -.557 -.404* cons -41.831+ 86.836* -31.221* 40.544 N 42 42 42 42 p model < .001 < .001 < .001 < .001 R2 .3645 .7782 .3997 .7944 depvar: # habilitations in year x, robust standard errors; time trend is 0 in 1960, and grows by 1 every year; student cohort is # of newly inscribed students in all German universities, 20 years ago; # judges eight years ago; GNP eight years ago, in % of GNP 2009 * p < .001, ** p < .01, * p < .05, Table 3. Adding Measures for Alternative Career Options. depvar: # habilitations in year x, robust standard errors; time trend is 0 in 1960, and grows by 1 every year; student cohort is # of newly inscribed students in all German universities, 20 years ago; # judges eight years ago; GNP eight years ago, in % of GNP 2009 * p < 001 doi:10.1371/journal.pone.0159815.t003 effect of GNP in model 4 could reflect the expectation of future law professors that a more prosperous country will continue the expansion of the university system. Alternative Career Options As a proxy for earning prospects in the big law firms, we take the development of gross national product; unfortunately, no major law firm has been both willing and able to give us time series data on salaries for associates, or on the number of associates hired per year. As Fig 9 shows, GNP has been growing almost linearly year by year. The turnover of law firms should be positively correlated with the ability of potential clients to pay higher fees, which in turn should be correlated with GNP. The regressions in Table 3 make it clear that alternative career options are not an important determinant in the decision to prepare for an academic career in law. If we do not control for the size of the respective student cohort, in models 1 and 3, we do find a significant effect of these regressors, but it has an unanticipated sign, suggesting that alternative career options make rather than break the case for academia: If the judiciary is large at the moment when a young lawyer decides about her career, she is more likely to become a professor. Likewise, the larger the gross national product is at this point in time, the more she is likely to forego the earning opportunities in the law firms, and to start an academic career instead. The effect of the total number of judges disappears once we control for the size of the student cohort (model 2). The effect of GNP goes down to about a third of its previous size, but remains significant (models 4). The effect of the number of judges in model 1 partly picks up the effect of student cohorts, suggesting a correlation about the causes of which we could only speculate. The higher the GNP is, the more habilitations occur eight years later, as shown in model 3. The remaining 13 / 22 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 Available Positions in Academia When professors offer a promising student a post-doc position, and when this student decides to accept the position and to prepare for an academic career, it would be rational for both to consider the number of positions that will be available when the candidate is on the market, i.e., some eight years from then. If this number is constant, this is of course not a concern. But the two shocks that drive the peaks in Fig 2 result from the rapid expansion of public universi- ties. In the 1970s, policy makers wanted to educate a much larger fraction of the population. In the 1990s, reunification led to a high demand for new professors. In response to both shocks, individuals from some birth cohorts were much more likely to become professors than aca- demics who were a few years older or younger. In German public universities, the mandatory retirement age is, once again, 68 (with a sustained interlude where retirement age has been low- ered to 65). Due to this, the demand for law professors predictably peaks when one of the cohorts hired during one of the two shocks is up for retirement. Data about open positions is only available from 1992 on. Since we need an eight-year lag, this would reduce our time series to only nine years. We therefore work with the share of pro- fessors aged 60 and above. This data has been reported by the German Federal Bureau of Statis- tics since 1966 (series 11 schedule S.1 prior to 1990 and schedule 4.4 subsequently), although initially only for selected years. Descriptively it seems as if this information might have had some influence after the first peak. By contrast, the second peak seems to have been unaffected (Fig 10). Yet model 2 of Table 4 suggests that the number of habilitations is not driven by the antici- pated development of open positions eight years ahead. The regressor is insignificant. Adding this control variable explains less than an additional percent of the variance, in a model that already explains 88% (model 1). The coefficients of the remaining control variables change very little. Most importantly, the eight-year lag of the dependent variable becomes even more pronounced. We can therefore exclude that the hog cycle is spurious. The impression of a cycle PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 14 / 22 The Hog Cycle of Law Professors Fig 10. Available Positions in Academia Expected Number of Available Positions. doi:10.1371/journal.pone.0159815.g010 Fig 10. Expected Number of Available Positions. doi:10.1371/journal.pone.0159815.g010 Fig 10. Expected Number of Available Positions. doi:10.1371/journal.pone.0159815.g010 Table 4. Adding a Proxy for Available Positions in Academia. model 1 model 2 model 3 time trend -4.369* -4.615* -4.651* student cohort lag 20 2.190* 2.195* 2.076* fraction professors above 60 lag 8 82.481 absolute number professor above 60 lag 8 .016 lag8 habil -.370 -.462** -.655* cons 98.164* 101.176* 117.940*** N 28 28 28 p model < .001 < .001 < .001 R2 .8810 .8881 .8913 depvar: # habilitations in year x, robust standard errors; time trend is 0 in 1960, and grows by 1 every year; student cohort is # of newly inscribed students in all German universities, 20 years ago; model 1 repeats model 1 of Table 2, but cuts the time series to the 28 data points for which we have the additional control variable that we use in model 2; fraction/number of professors 8 years ago: missing data replaced by linear imputation, using last reported value + ((next reported value–last reported value)/(number of missing values))(position in this series of missing values) p < .001, p < .01, * p < 05 Table 4. Adding a Proxy for Available Positions in Academia. model 1 model 2 model 3 time trend -4.369* -4.615* -4.651* student cohort lag 20 2.190* 2.195* 2.076* fraction professors above 60 lag 8 82.481 absolute number professor above 60 lag 8 .016 lag8 habil -.370 -.462 -.655* cons 98.164* 101.176* 117.940*** N 28 28 28 p model < .001 < .001 < .001 R2 .8810 .8881 .8913 Table 4. Adding a Proxy for Available Positions in Academia. Table 4. Adding a Proxy for Available Positions in Academia. depvar: # habilitations in year x, robust standard errors; time trend is 0 in 1960, and grows by 1 every year; student cohort is # of newly inscribed students in all German universities, 20 years ago; model 1 repeats model 1 of Table 2, but cuts the time series to the 28 data points for which we have the additional control variable that we use in model 2; fraction/number of professors 8 years ago: missing data replaced by linear imputation, using last reported value + ((next reported value–last reported value)/(number of missing values))(position in this series of missing values) ** p < 001 doi:10.1371/journal.pone.0159815.t004 Model Selection The previous analysis indicates that a model controlling for the time trend and the student cohort twenty years ago, plus an autocorrelation term with a lag of eight years, best explains our data. This fully supports our hypothesis. Since the lag is highly significant and substantially negative in this model, this suggests that there is indeed a hog cycle, triggered by perceived job opportunities at the moment when the future professor and her supervisor decide to have her prepare for entering the market. We now undertake a number of robustness checks to corrobo- rate our result. As we have discussed near the end of Section 2 and illustrated in Fig 2, our dependent vari- able reacted positively to two strong exogenous shocks in the early 1970s and 2000s. If our hypothesis holds true, promising young lawyers should not only be influenced in their decision to prepare for an academic career if they observe such salient shocks. As a first robustness check, we thus control for the effect these two shocks have had per se (apart from their ability to trigger a cycle). Model 3 below enriches our model with dummy variables that neutralise the two peaks exhibited by our data in Fig 2. (Note that the second peak stretches over more than one year. Our dummy flags the first year in which the number of habilitations was close to maximum, but results do not change for any other year during that peak.) These additional control variables are indeed significant, but they do not affect our variable of interest, the eight- year lag. As another check of robustness, we examine a partial autocorrelation plot. Since we already know that there is a time trend and that we must control for the size of student cohorts, we do not base this plot on the raw data, but on the residuals from a regression that controls for these two variables (but of course not for the eight-year lag). Partial correlations measure the correla- tion of the number of habilitations in any given year with the number of habilitations n years ago, after the effect of all shorter lags has been partialled out. As one sees, the negative correla- tion with the number of habilitations eight years ago indeed stands out (see Fig 11). This demonstrates that we are justified in using the eight-year lag, rather than lags of alternative duration. Available Positions in Academia depvar: # habilitations in year x, robust standard errors; time trend is 0 in 1960, and grows by 1 every year; student cohort is # of newly inscribed students in all German universities, 20 years ago; model 1 repeats model 1 of Table 2, but cuts the time series to the 28 data points for which we have the additional control variable that we use in model 2; fraction/number of professors 8 years ago: missing data replaced by linear imputation, using last reported value + ((next reported value–last reported value)/(number of missing values))(position in this series of missing values) p < 001 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 15 / 22 The Hog Cycle of Law Professors does not result from the fact that those considering preparing for the academic market do something very sensible, namely react to predicted volatility of demand. The decision to embark on an academic career might also be influenced by the absolute number of professors close to retirement, rather than their share. This would imply that the future professors are less sensitive to the overall turnaround, and would rather care about how many opportunities they expect to exist for applying. Model 3 of Table 4 demonstrates that this information does not have a significant effect either. In this specification, the eight-year lag of the dependent variable becomes even more pronounced. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 Model Selection Fig 11 suggests that, additionally, there is positive autocorrelation with the year before. This is plausible. Habilitation procedure is complicated. Candidates starting in the same year may not all be able to finish their habilitation thesis at the same moment, and some law schools are better than others at organizing the formal procedure. As a first robustness check, we therefore estimate a model with an additional autocorrelation term for the one-year lag (Table 5, model 2). The additional lag is indeed significant, but adding this additional control variable does not affect the significance or the size of the negative eight-year lag. In our preferred model, we use the time trend for detrending the data, since this gives coeffi- cients a straightforward interpretation. In the time series literature, detrending is typically done through first differencing. If we use this procedure, we again support the significant nega- tive eight-year lag, model 4. 16 / 22 The Hog Cycle of Law Professors Fig 11. Partial Autocorrelation of Residuals. From OLS of # of habilitations in year t, controlling for # of new students in t-20 and the time trend. doi:10 1371/journal pone 0159815 g011 Fig 11. Partial Autocorrelation of Residuals. From OLS of # of habilitations in year t, controlling for # of new students in t-20 and the time trend. doi:10.1371/journal.pone.0159815.g011 le 5. Alternative Specifications of Preferred Model. model 1 model 2 model 3 model 4 model 5 model 6 var # habil # habil # habil # habil # habil # habil trend -2.772* -1.740* -1.252 -2.434* -3.044* ent cohort 1.659* 1.066 .829 .194 1.468* 1.744*** habil .352* .421 -.346* habil -.335* -.303 -.244 -.238 -.594* 1970 27.089* 32.770*** 21.249* 17.420* 1998 7.783 7.715 1.391 7.082 1) .380 habil, instrumented by lag10habil -.293 s 68.087* 46.003 35.648* 14.993 53.790* 71.142* 42 42 42 42 42 40 odel < .001 < .001 < .001 < .001 < .001 < .001 .7738 .8095 .8636 .5459 .8479 var: # habilitations in year x, robust standard errors; time trend is 0 in 1960, and grows by 1 every year; student cohort is # of newly inscribed students in erman universities, 20 years ago; year1970 is an indicator variable that is 1 for habilitations in 1970; year1998 is an indicator variable that is 1 for litations in 1998; ma(1) is a moving average term, of length 1 year p < .001, < 01 Table 5. depvar: # habilitations in year x, robust standard errors; time trend is 0 in 1960, and grows by 1 every year; student cohort is # of newly inscribed students in all German universities, 20 years ago; year1970 is an indicator variable that is 1 for habilitations in 1970; year1998 is an indicator variable that is 1 for habilitations in 1998; ma(1) is a moving average term, of length 1 year * < 001 doi:10.1371/journal.pone.0159815.t005 Model Selection Alternative Specifications of Preferred Model. model 1 model 2 model 3 model 4 model 5 model 6 depvar # habil # habil # habil # habil # habil # habil time trend -2.772*** -1.740* -1.252 -2.434* -3.044* student cohort 1.659* 1.066 .829 .194 1.468* 1.744*** lag1habil .352* .421 -.346* lag8habil -.335* -.303 -.244 -.238 -.594* year1970 27.089* 32.770*** 21.249* 17.420* year1998 7.783 7.715 1.391 7.082 ma(1) .380 lag8habil, instrumented by lag10habil -.293 cons 68.087* 46.003 35.648* 14.993 53.790* 71.142* N 42 42 42 42 42 40 p model < .001 < .001 < .001 < .001 < .001 < .001 R2 .7738 .8095 .8636 .5459 .8479 depvar: # habilitations in year x, robust standard errors; time trend is 0 in 1960, and grows by 1 every year; student cohort is # of newly inscribed students in all German universities, 20 years ago; year1970 is an indicator variable that is 1 for habilitations in 1970; year1998 is an indicator variable that is 1 for habilitations in 1998; ma(1) is a moving average term, of length 1 year * p < .001, ** p < .01, * p < .05 Table 5. Alternative Specifications of Preferred Model. Table 5. Alternative Specifications of Preferred Model. depvar: # habilitations in year x, robust standard errors; time trend is 0 in 1960, and grows by 1 every year; student cohort is # of newly inscribed students in all German universities, 20 years ago; year1970 is an indicator variable that is 1 for habilitations in 1970; year1998 is an indicator variable that is 1 for habilitations in 1998; ma(1) is a moving average term, of length 1 year *** p < 001 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 17 / 22 The Hog Cycle of Law Professors A model with an autocorrelation term assumes that the effect of a past shock only gradually fades away. Our explanation for the significance of the one-year lag would also fit a model with a moving average term, which is assumed to become immaterial after the period affected is over. Model 5 shows that the moving average term is indeed positive and significant but that, in this model, the p-value for the eight-year lag is even smaller. Model Selection Since we want to jointly esti- mate a moving average term and an autocorrelation term, in this model we represent the auto- correlation by the lag in residuals, not in the dependent variable (i.e., we use Stata command arima), and find that deeper moving average terms are insignificant. Finally, in model 6 we instrument the eight-year lag by the ten-year lag that our earlier results show to be clearly uncorrelated with the dependent variable, i.e., with the number of habilitations today (after controlling for the time trend and the size of the relevant student cohort). This way we guard against the possibility that omitted variables make our estimation of the eight-year lag inconsistent. Although in this model we only have the correlation between our instrument and the eight-year lag for explanation, the effect of the eight-year lag still is highly significant–and remains significant if we add the 11-year lag as an additional instru- ment. Note that this procedure is in the spirit of the Arellano Bond model. The Arellano Bond model is concerned with possible inconsistency resulting from patterned heterogeneity between individuals (cross sections) in a panel. Heterogeneity can be interpreted as an omitted variable. Instrumenting with lags of the dependent variable one has reason to believe to be uncorrelated with contemporaneous choices removes the resulting risk of inconsistency. We do not have a panel, so that inconsistency cannot result from unobserved heterogeneity. But past lags as internal instruments are also valid in a mere time series model. Yet, since we need not remove individual specific effects, we have no reason to first first-difference the data. Summing these additional tests up, we find that our main result, the eight-year lag, is very robust to changes in the specification of the statistical model. Note that all robustness checks leave the significant eight-year lag of the independent variable unaffected, even if we perform them on model 2 of Table 4. To avoid working with this considerably shorter time series (and partly imputed data for the percentage of professors above 60) we have presented robustness checks only on the model without this additional control variable, but the alternative robust- ness checks are available from the authors upon request. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 Conclusions (A) Left panel is based on the best-performing model of Fig 4, i.e., on mere autocorrelation with a 15-year lag. (B) Right panel is based on final model, i.e., on model 1 of Table 5. doi:10.1371/journal.pone.0159815.g012 doi:10.1371/journal.pone.0159815.g012 habilitation is taken by the law school of origin, which is prevented from hiring its own candi- dates. We are thus in a position to quantify precisely the supply of candidates for all the years since 1960. We find significant autocorrelation, but time lags alone explain little of the variance in our data. This changes if we control for the overall time trend and for the size of student cohorts. Fig 12 summarizes our results. A mere time series model (left panel) has a relatively poor fit, despite the fact that the lag of 15 years is highly significant. By contrast, the model con- trolling for the overall time trend and the size of student cohorts, with autocorrelation of eight years (right panel), has an excellent visual fit. In our final model (see Table 5, model 1), we find significant negative correlation between the number of habilitations eight years ago and today. The correlation coefficient is -.335, i.e., three more habilitations eight years ago approximately deter one habilitation today, and three fewer habilitations eight years ago encourage one more habilitation today. This suggests that, when a supervisor approaches a promising candidate, and when this candidate accepts the offered post-doc position, both are overly sensitive to current experiences on the academic market. If law schools currently have a hard time filling open positions, and if those who have just passed habilitation quickly find a good job, supervisors and candidates infer that there is demand for more law professors. By contrast, if one open position currently draws 50 or 60 applications, out of which many are serious candidates, supervisors are hesitant to invite prom- ising young lawyers to embark on such a hazardous career, or candidates turn down offers for that reason. Fig 12 also serves another purpose. We can use either model to make out-of-sample predic- tions for 10 years after the end of our time-series. Given our findings, one should not trust pre- dictions from the mere time series, but should also take student cohorts into account, and should control for the time trend. Conclusions Currently the entry-level job market in legal academia “remains largely a black box” for which “we do not know the factors that influence how law schools and candidates make decisions” [1]:5. To contemplate some such factors and to lift the lid on the black box of these job market decisions, we have shown that the market for law professors fulfils the conditions for a cobweb, namely inelasticity of supply and demand as well as perishability: Preparation for an academic career in law takes many years. Therefore supply cannot swiftly adapt to changes in demand. The (German) law faculties are hesitant to hire candidates who have been on the market for more than a small number of years. The faculties tend to read this as a signal that something must be wrong with this candidate. Therefore future law professors are a perishable commod- ity. Finally, the (German) law faculties live off public funds and must pay salaries that are fixed by statute. The faculties are therefore unable to respond to excess supply by hiring more profes- sors. Theory thus predicts a hog cycle: When there have been too few law professors in the past, there is too much supply today. When there have been too many law professors in the past, there is too little supply today. In this paper, we exploit the fact that, in the German university system, future law professors must formally qualify for the academic market by passing habilitation. The decision to grant 18 / 22 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 The Hog Cycle of Law Professors Fig 12. Out-of-sample Predictions. (A) Left panel is based on the best-performing model of Fig 4, i.e., on mere autocorrelation with a 15-year lag. (B) Right panel is based on final model, i.e., on model 1 of Table 5. doi:10 1371/journal pone 0159815 g012 Fig 12. Out-of-sample Predictions. (A) Left panel is based on the best-performing model of Fig 4, i.e., on mere autocorrelation with a 15-year lag. (B) Right panel is based on final model, i.e., on model 1 of Table 5. Fig 12. Out-of-sample Predictions. (A) Left panel is based on the best-performing model of Fig 4, i.e., on mere au panel is based on final model, i.e., on model 1 of Table 5. Fig 12. Out-of-sample Predictions. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 Conclusions Student cohorts of those researchers applying for open posi- tions at the law faculties between 2010 and 2020 are already known. Using this information, our model predicts another steep increase until 2018, with some 85 candidates. Shall we use this information to repeat the advice Fritz Baade and Hermann Abeking gave hog farmers in the 30s of the previous century [8]? They urged them to adopt an anti-cyclical 19 / 22 The Hog Cycle of Law Professors investment policy. In our case, this would mean that promising young lawyers should be hesi- tant to embark on an academic career (and supervisors should hesitate to encourage them) when present job opportunities are plentiful. If our regression has it right, when these lawyers are ready for the market of professors, good candidates will be in excess supply. There are two reasons why such conclusions ought to be treated with more caution, though. It is not clear how fast and how profoundly the demand side of this market is changing. Finance ministers may be forced, after the financial crisis, to cut back on all public expenses, university education included. Law faculties may be forced to react to severe cuts by structural reforms. Such reforms are likely to privilege cheaper new candidates over the promotion of more established professors. Moreover, an anti-cyclical reaction is only individually beneficial if most others behave cyclically. Theoretically, the resulting problem is well understood. It is known in game theory as the beauty contest, following John Maynard Keynes’ famous likening of the stock market “to those newspaper competitions in which the competitors have to pick out the six prettiest faces from a hundred photographs, the prize being awarded to the competitor whose choice most nearly corresponds to the average preferences of the competitors as a whole” [32]. Keynes anticipated what later got generalized as the theory of level-k-reasoning [33, 34]. It considers self-referential decision processes, where decision outcomes depend on the expected decisions of others in the same situation, and assumes that players determine best replies to each other’s behaviour by reasoning iteratively. For instance, let a simpleton with no regard for fellow play- ers’ strategies be defined as a level 0 player, then any player who anticipates and reacts to level- 0-reasoning would inhabit level 1. Even more sophisticated strategists would ascend to level 2 by anticipating level-1-reasoning, and so on. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 Acknowledgments We thank Jonathan Klick, Sven Fischer, Ioanna Grypari and Yoan Hermstrüwer for helpful comments on an earlier version of the paper. Author Contributions Analyzed the data: CE HH. Wrote the paper: CE HH. Conclusions In its general form, any level-k-strategy assumes that all other players use reasoning levels between 0 and (k-1). Empirical research suggests that most people only use a degenerate version of this iterated best reply mechanism, with level-1 and level-2 being most prominent (for a comprehensive survey, see [35]). While this literature suggests that the foresight of most people is rather limited, is it plausi- ble to assume that the large majority of future law professors, and of their supervisors, is not clear-sighted at all? Probably not, as the studies mentioned above even found a remarkable pro- portion of people adopting level-1-reasoning, depending on their respective training, time availability, effort, and judgment confidence. On the other hand, for future law professors the problem is exacerbated by the fact that they must predict a market some eight years ahead. Therefore the strategic uncertainty inherent in the beauty contest is compounded by the sto- chastic uncertainty inherent in all sorts of changes that may happen in the meantime. In principle, government could step in and limit the number of positions available for future candidates for habilitation. This number could correspond to the expected number of open positions eight years ahead. Yet for a number of reasons, this solution would be problematic. In many domains, government forecasts of future developments have widely gone wrong. If the number of positions available for habilitation is determined by law, there will be strong pres- sure for candidates indeed to be hired once they have passed habilitation. This would cut into the budget autonomy of the legislator. Most importantly, this solution would profoundly change universities. One would need a nationwide competition for the positions that are made available. Effectively, the decision who is going to be a professor would be taken after this indi- vidual has acquired her doctorate, not after habilitation. Essentially, Germany would have to shift to the tenure-track system. There is debate whether this system is preferable. But breaking the hog cycle of candidates is certainly not a sufficient reason for making this move. Conclusions We must therefore leave young lawyers considering an academic career with a problem of judgement: If they want to rely on our findings, they must also estimate how many of their potential competitors will try to second-guess the market at the moment when they expect to 20 / 22 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 The Hog Cycle of Law Professors enter it. Ultimately, the traditional piece of advice will remain best: Only prepare for an aca- demic career in law if your advisor truly believes in you, and if you prefer a life driven by aca- demic curiosity over the alternative options for a good lawyer. enter it. Ultimately, the traditional piece of advice will remain best: Only prepare for an aca- demic career in law if your advisor truly believes in you, and if you prefer a life driven by aca- demic curiosity over the alternative options for a good lawyer. S1 File. Comprehensive dataset in Stata.dta file format. (DTA) S1 File. Comprehensive dataset in Stata.dta file format. (DTA) PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 References 1. George TE, Yoon AH. The Labor Market for New Law Professors. Journal of Empirical Legal Studies. 2014; 11(1):1–38. 2. Hanau A. Die Prognose der Schweinepreise: Reimar Hobbing; 1928. 2. Hanau A. Die Prognose der Schweinepreise: Reimar Hobbing; 1928. 3. Ezekiel M. The Cobweb Theorem. The Quarterly Journal of Economics. 1938; 52(2):255–80. 4. Stein JL. Cobwebs, Rational Expectations and Futures Markets. The Review of Economics and Statis- tics. 1992; 74(1):127–34. 5. Schultz H. The Theory and Measurement of Demand: The University of Chicago Press; 1958. 6. Muth JF. Rational Expectations and the Theory of Price Movements. Econometrica: Journal of the Econometric Society. 1961; 29(3):315–35. 7. Haas GC, Ezekiel M. Factors Affecting the Price of Hogs: US Dept. of Agriculture; 1926. 8. Baade F, Abeking H. Schweinefibel: Oder, was jeder Bauer vor dem Decken seiner Sauen bedenken muss: Reichsforschungsstelle für landwirtschaftliches Marktwesen; 1930. 9. Coase R, Fowler RF. Bacon Production and the Pig-cycle in Great Britain. Economica. 1935; 2(6):142– 67. 10. Coase R, Fowler RF. The Pig-Cycle. A Rejoinder. Economica. 1935; 2(8):423–8. 11. Samuelson PA. An Economist's Non-linear Model of Self-generated Fertility Waves. Population Stud- ies. 1976; 30(2):243–7. doi: 10.1080/00324728.1976.10412732 PMID: 22077804 12. Harlow AA. The Hog Cycle and the Cobweb Theorem. Journal of Farm Economics. 1960:842–53. 13. Maki WR. Decomposition of the Beef and Pork Cycles. Journal of Farm Economics. 1962; 44(3):731– 43. 14. Larson AB. The Hog Cycle as Harmonic Motion. Journal of Farm Economics. 1964; 46(2):375–86. 15. Jelavich MS. Distributed Lag Estimation of Harmonic Motion in the Hog Market. American Journal of Agricultural Economics. 1973; 55(2):223–4. 16. Shonkwiler JS, Spreen TH. Statistical Significance and Stability of the Hog Cycle. Southern Journal of Agricultural Economics. 1986; 18(02). 17. Hayes DJ, Schmitz A. Hog Cycles and Countercyclical Production Response. American Journal of Agricultural Economics. 1987; 69(4):762–70. 18. Chavas J-P, Holt MT. On Nonlinear Dynamics. The Case of the Pork Cycle. American Journal of Agri- cultural Economics. 1991; 73(3):819–28. 19. Holt MT, Craig LA. Nonlinear Dynamics and Structural Change in the US Hog–Corn Cycle. A Time- varying STAR Approach. American Journal of Agricultural Economics. 2006; 88(1):215–33. 21 / 22 PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 The Hog Cycle of Law Professors 20. Simmons WM. An Economic Study of the US Potato Industry: US Dept. of Agriculture, Economic Research Service, Economic and Statistical Analysis Division; 1962. 21. Wheaton WC. Real Estate “Cycles”. Some Fundamentals. Real Estate Economics. 1999; 27(2):209– 30. 22. PLOS ONE \| DOI:10.1371/journal.pone.0159815 July 28, 2016 References Wernecke M, Rottke N, Holzmann C. Incorporating the Real Estate Cycle into Management Decisions. Evidence from Germany. Journal of Real Estate Portfolio Management. 2004; 10(3):171–86. 23. Krugman PR. The Oil-hog Cycle. The New York Times, November. 2001; 18:14. 24. AbuAlRub RF. Nursing Shortage in Jordan. What is the Solution? Journal of Professional Nursing. 2007; 23(2):117–20. PMID: 17383605 25. Köbler G. Lateinisches Abkunfts- und Wirkungswörterbuch für Altertum und Mittelalter. 2010. 26. Legal Tribune Online. Uni-Ranking Jura. Die besten Jura-Fakultäten Deutschlands. 2016. 27. Horstkotte H. Juniorprofessur an Juristischen Fakultäten. Lehrstuhl ohne Habilitation? Legal Tribune Online. 2015. 28. Rüthers B. Die rechtswissenschaftlichen Habilitationen in der Bundesrepublik zwischen 1945 und1969. Juristenzeitung. 1972; 27:185–91. 29. GESIS. 2007. 29. GESIS. 2007. 30. Sobel ME. Asymptotic Confidence Intervals for Indirect Effects in Structural Equation Models. Sociolog- ical Methodology. 1982; 13:290–312. 31. Wood RE, Goodman JS, Beckmann N, Cook A. Mediation Testing in Management Research: A Review and Proposals. Organizational Research Methods. 2008; 11:270–95. 32. Keynes JM. The General Theory of Employment, Interest and Money. New York: Harcourt Brace; 1936. xii, 403 p. 33. Nagel R. Unraveling in Guessing Games:.An Experimental Study. American Economic Review. 1995; 85(5):1313–26. 34. Stahl DO, Wilson P. On Players Models of Other Players. Theory and Experimental Evidence. Games and Economic Behavior. 1995; 10(1):218–54. 35. Bosch-Domenech A, Montalvo JG, Nagel R, Satorra A. One, Two, (Three), Infinity,. . .: Newspaper and Lab Beauty-Contest Experiments. American Economic Review. 2002; 92(5):1687–701. 22 / 22
https://openalex.org/W2945105163	https://www.intechopen.com/citation-pdf-url/67149	English	null	Applications of Tannin Resin Adhesives in the Wood Industry	IntechOpen eBooks	2,020	cc-by	7,935	Abstract Tannin is extracted from natural sustainable materials. It is widely used to prepare tannin resin adhesives owing to its naturally occurring phenolic structure. This chapter aims to introduce the resources and structures of tannin, existing reactions that are involved in the synthesis of tannin resin, and the applications of tannin resin adhesives in the wood industry. Additionally, the advancements in the research based on the use of tannin resins in manufacturing plywood, par- ticleboard, wood preservation, decoration paper impregnation, structural glulam, impregnated fibers, and other wooden products are reviewed. Herein, the main limitations encountered during the application of tannin resin adhesives and the future key research points are identified. Finally, the potential applications of tan- nin resin adhesives in the wood industry have been discussed. Keywords: tannins, resins, adhesives, wood industry, applications Applications of Tannin Resin Adhesives in the Wood Industry Xiaojian Zhou and Guanben Du 1. Introduction The use of adhesives dates back to approximately 3000 years ago. Several types of adhesives based on specific applications have been developed, particularly for the manufacturing of wood and paper products, among other products. Therefore, thousands of adhesive products have been developed. Factors that affect the selec- tion of the adhesives are cost, assembly process, bonding strength, and durability. The fabrication of wood-based panel products involves a “preparation and recombination of wood unit” process wherein wood adhesives play a crucial role. Adhesives play a vital role in wood processing because their quality has a direct impact on the performance of the final wood product. p p p Synthetic and natural resins are the most commonly used adhesives in the wood industry. Some examples of synthetic resins are urea-formaldehyde resin; phenolic resin; melamine formaldehyde resin; and copolycondensation resin, which include phenol-urea-formaldehyde resin (PUF) and melamine–urea-formaldehyde resin (MUF). Some examples of natural resins are soy protein adhesive, tannin resin, lignin adhesive, and starch adhesive. Although synthetic resin has high weathering resistance and mechanical strength, its raw materials are derived from nonrenewable petrochemical products that are volatile and expensive. Additionally, these products emit formaldehyde, which is toxic and carcinogenic. g The awareness of environmental protection and personal health has been emphasized in recent years. Therefore, natural resins with renewable resources as the main materials have attracted considerable amount of attention. Research 1 Tannins - Structural Properties, Biological Properties and Current Knowledge and application of the tannin resin have been highly successful in some countries because its phenolic structure enables its use as adhesives and as a partial or com- plete substitute for phenols in adhesives. This chapter provides a comprehensive discussion of the situation of the existing tannin resources, reaction mechanisms involved in the synthesis of tanning resins, and general application of tannin resins in the wood industry. This information could provide ideas for the scholars and broaden the application scope of tannin resins in the wood industry. pp p y The production of tannin for leather manufacturing peaked immediately after World War II and has progressively declined. Tannin adhesives were first successfully commercialized in South Africa in the early 1970s. Subsequently, mimosa tannin adhesives were used instead of synthetic phenolic adhesives to manufacture particle- board and plywood for external and marine applications. Tannin resin adhesives have been used in Australia, Zimbabwe, Chile, Argentina, Brazil, and New Zealand [1]. 2. Tannin resources Tannins are extracted from agroforestry biomaterials, such as wood, bark, leaves, and fruits, by the water extraction method. Tannins can be categorized as hydrolyzable tannin or condensed polyflavonoid tannin. The latter is one of the main objects of wood adhesive research and accounts for 90% of the global tannin output. The annual industrial output of tannin reaches up to 200,000 tons. The distribution of tannin resources in the world has regional characteristics. For example, black wattle tannin is mainly manufactured in Brazil, South Africa, India, and other countries. Quebracho tannin is mainly manufactured in Argentina. Chestnut tannin is mainly manufactured in Italy and Slovenia. Pine bark tannin is mainly manufactured in Chile and Turkey. Oak tannin is mainly manufactured in Poland. Tannin from grape residues, such as skins and seeds, is mainly manu- factured in France. In China, tannin is mainly synthesized from larch, poplar, and acacia bark. 3. Tannin structures Hydrolyzable tannin comprises different types of unit structures, including gallic, digallic, and ellagic acids (see Figure 1), as well as sugar esters, which usually exist in the form of glucose [2, 3]. Condensed tannin comprises monoflavonoids or flavonoid units that have under- gone various degrees of polymerization. These units are associated with their precur- sors, such as flavanes-3-ol and flavanes-3,4-diol, among other flavonoids [4, 5]. Each flavonoid contains two types of phenolic nuclei, which are A- and B-ring, as shown in Figure 2. The A-ring includes resorcinol and phloroglucinol, whereas the B-ring includes pyrogallol and catechol, among other rare phenols. The A-rings of different tannins possess different chemical structures. The A-rings of tannins extracted from mimosa/wattle, quebracho, Douglas fir, and spruce include resorcinol, whereas those of pine include phloroglucinol. p p g The main polyphenolic pattern is represented using flavonoid analogs that are based on the resorcinol A-ring and pyrogallol B-ring (I type in Figure 3). This unit structure accounts for 70% of tannin. Unit structure II constitutes 25% of tannin and comprises a resorcinol A-ring and catechol B-ring (II type in Figure 3). The remaining 5% is a mixture of phloroglucinol-pyrogallol (III type in Figure 3) and phloroglucinol-catechol (IV type in Figure 3) flavonoids. These four patterns consti- tute 65–80% of mimosa bark extract. The remaining components are non-tannins, 2 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Figure 1. Unit structures of hydrolyzable tannin. Figure 2. Main structure of condensed tannin. cations of Tannin Resin Adhesives in the Wood Industry http://dx.doi.org/10.5772/intechopen.86424 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Figure 1. g Unit structures of hydrolyzable tannin. g Unit structures of hydrolyzable tannin. Figure 2. Figure 2. g Main structure of condensed tannin. g Main structure of condensed tannin. Figure 3. Main units of condensed tannin. g 3 Main units of condensed tannin. g Main units of condensed tannin. which are simple carbohydrates, hydrocolloid gums, and nitrogen compounds, i.e., amino and imino acids. Gums and pectins are the most important components of tannins and have a significant effect on the viscosity of the extract despite their low concentration, i.e., 3–6%. These non-tannin substances can attenuate wood failure and can decrease the water resistance of glued products. 3. Tannin structures Pine tannin mainly presents two patterns: one is represented by phloroglucinol A-ring and catechol B-ring structures (V type in Figure 3) and the other is repre- sented by phloroglucinol A-ring and phenol B-ring structures (VI type in Figure 3). 3 Tannins - Structural Properties, Biological Properties and Current Knowledge Flavonoid units can be bound through their 4,6- and/or 4,8-linkages to form polyflavonoids. Wattle-extracted tannin comprises 4–5 flavonoid units joined together through 4,6-linkages. Each unit of wattle-extracted tannin has an average mass number of 1250. The average mass number of quebracho tannin and pine is 1784 and approximately 4300, respectively. Pine tannin is phloroglucinolic in nature and its flavonoid units are joined together through 4,8-interflavonoid link- ages. Linear polymeric tannins have only 4,6- (V) or 4,8-linkages (VI). However, 4,6- and 4,8-linkages may simultaneously exist in the presence of resorcinolic and phloroglucinolic A-rings. This phenomenon results in the synthesis of angular rather than linear polymeric tannins (VII). Matrix-assisted laser desorption/ ionization time-of-flight revealed that mimosa tannin is highly branched owing to the presence of high proportions of angular units in its structure. On the contrary, quebracho tannin is almost completely linear. These structural differences contrib- ute to the considerable differences in the viscosity of tannin water solutions [6]. 4. Synthesis of tannin resin adhesives The low reactivity of hydrolyzable tannins with formaldehyde limits their application in the wood industry, which can be attributed to their simple phenolic structures (Figure 1). Tannin extracts usually contain sugars and gums, which are not involved in the synthesis of resin adhesives. Commercially available tannin extracts from black wattle and hardwood typically contain 70–80% of natural phenolic polymers, whereas those obtained from pine contain only 50–60% of natural phenolic poly- mers. Sugar dilutes the actual solid content, thus affecting the final properties of resins. Gum considerably affects the strength of the resin and water resistance of the adhesive. Due to the presence of non-tannin components, unmodified tannin adhesive is unsuitable for the production of wood products with high requirements. Therefore, tannin adhesives must be modified. Normally, the viscosity of tannin resin adhesives is higher than that of synthetic resins at the same concentration due to (1) the presence of high molecular weight tannins in the extract and (2) the existence of hydrogen bonding and electrostatic interactions between tannin and tannin, tannin and gum, and gum and gum. Effective methods for decreasing the viscosity of tannin extracts in aqueous solu- tions include the following: (1) acid or alkaline hydrolysis of high molecular weight carbohydrates, e.g., with acetic anhydride, maleic acid anhydride, or NaOH [7, 8]; (2) addition of small amounts of hydrogen bond breakers (e.g., 3% urea based on the solid content of the extract); and (3) destruction of heterocyclic ether in tannin molecules through sulfite or bisulfite treatment. Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 unreactive and may only be activated via anion formation at a relatively high pH [9, 10]. Hence, the B-ring does not participate in polymerization except at a high pH (pH = 10). However, the reactivity between the A-ring and formaldehyde influences pot life because it is too fast to control. unreactive and may only be activated via anion formation at a relatively high pH [9, 10]. Hence, the B-ring does not participate in polymerization except at a high pH (pH = 10). However, the reactivity between the A-ring and formaldehyde influences pot life because it is too fast to control. In general, only the A-ring structure participates in crosslinking to build networks in tannin resin adhesives (Figure 4). However, owing to their size and shape, tannin molecules become immobile at low levels of condensation with formaldehyde. Thus, a large distance between the available reactive sites for further methylene bridge formation results in the incomplete polymerization of tannin resin adhesives. Incomplete polymerization, in turn, results in the forma- tion of weak and brittle adhesives. Bridging agents with long molecules, such as phenolic and amino-plastic resins [10, 11], have been used to overcome this limitation by bridging the distances that are too large for interflavonoid methylene to bridge. g Catechol and catecholic B-ring do not react with formaldehyde at a pH value less than 10. Adding zinc acetate to the reaction mixture induces the B-ring to react with formaldehyde at low pH values, the optimum pH being in the range of 4.5–5.5, as shown by the high amount of formaldehyde being consumed. This finding implies that the further crosslinking of the tannin-formaldehyde network could be achieved through the participation of the B-ring in the reaction in the presence of zinc acetate. Strength can be improved through the addition of zinc acetate at economi- cally acceptable levels (5–10% in resin solids). Nevertheless, improved strength is not comparable with the strength of fortified tannin resin. Crosslinking is sometimes performed through the addition of isocyanate. The highly reactive diphenylmethane diisocyanate (MDI) can be used to assist the participation of B-ring in the crosslinking reaction [12]. Additionally, the reaction between polymeric diphenylmethane diisocyanate (pMDI) and carbohydrates or hydrocolloid gums can help in increasing the bonding strength of wood products. 4.1 Reaction of tannin with aldehyde Tannin being phenolic in nature undergoes the same alkali- or acid-catalyzed reaction with formaldehyde experienced by phenols. Alkali-catalyzed reactions are predominantly used in industrial applications. Nucleophilic centers on the A-ring of any flavonoid unit tend to be more reactive than those on the B-ring. Thus, the reac- tion for inducing polymerization between formaldehyde and tannin mainly occurs on the A-ring through methylene bridge linkages. The A-ring of the condensed tan- nin molecules contains flavonoid units that possess one highly reactive nucleophilic center each. The reactivity of the resorcinol A-ring (e.g., wattle) toward formal- dehyde is comparable with that of resorcinol. On the contrary, the phloroglucinol A-ring (e.g., pine) behaves as phloroglucinol. Pyrogallol or the catechol B-ring are 4 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 The reaction rate of wattle and pine tannins with formaldehyde is slowest in the pH range of 4.0–4.5 and 3.3–3.9, respectively. Figure 4. Reaction mechanism of tannin with formaldehyde. Figure 4. Reaction mechanism of tannin with formaldehyde. Figure 4. Reaction mechanism of tannin with formaldehyde. 5 5 Tannins - Structural Properties, Biological Properties and Current Knowledge Formaldehyde is a major aldehyde used for the synthesis, setting, and curing of tannin resin adhesives. It is normally used as a liquid formalin solution or in the form of the polymer paraformaldehyde, which is capable of fairly rapid depoly- merization under alkaline conditions. The formaldehyde reaction with tannin can be controlled by the addition of alcohols to the system. Under these circumstances, some of the formaldehydes are stabilized by the formation of hemiacetals, such as the formation CH2[OH][OCH3], if methanol is used. When the adhesive is cured at an elevated temperature, the alcohol is driven off and formaldehyde is progressively released from the hemiacetal. These effects minimize formaldehyde volatilization when the reactants reach curing temperature and extend the pot life of the adhesive. Hexamethylenetetramine (hexamine) may also be added to tannin resins owing to its formaldehyde-releasing action under heat. Although hexamine is unstable in acidic environments, formaldehyde is liberated under alkaline conditions when heated. This effect indefinitely extends pot life at the room temperature. However, in most cases, hexamine does not decompose formaldehyde and ammonia in the presence of chemical species with highly reactive nucleophilic sites, such as melamine, resorcinol, and condensed flavonoid tannins. Instead, unstable interme- diate fragments can be reacted with highly reactive nucleophilic sites, such as tannin or melamine, among others, to form amino methylene bridges before yielding formaldehyde. Any species with a strong negative charge under alkaline conditions can react with the intermediate species formed by the decomposition of hexamine far more readily than formaldehyde. This characteristic accounts for the capability of wood adhesive formulations based on hexamine to render bonded panels with extremely low formaldehyde emission [13]. y y In the absence of highly reactive species with strong negative charges, hex- amine decomposition proceeds rapidly and results in formaldehyde formation. Formaldehyde emissions from wood particleboards bonded with pine and wattle tannin-based adhesives with paraformaldehyde, hexamine, and tris(hydroxyl) nitromethane hardeners have been measured using the perforator method. All particleboards manufactured using wattle tannin systems with three different hardeners satisfied grade E1 requirements. 4.2 Acidic and alkaline hydrolysis and autocondensation Tannin is subjected to two competing reactions when heated in the presence of strong mineral acids: (1) degradation leading to anthocyanidin and catechin formation and (2) condensation as a result of the hydrolysis of heterocyclic rings (p-hydroxybenzyl ether links). The created p-hydroxybenzyl carbonium ions condense randomly with nucleophilic centers on other tannin units to form phlo- baphenes. Other modes of condensation such as free radical coupling of B-ring catechol units cannot be excluded in the presence of atmospheric oxygen [17]. The interflavonoid bonds of condensed tannins with phloroglucinolic A-rings are susceptible to cleavage under even mild alkaline conditions. This characteristic could increase the reactivity with aldehydes. Increased reactivity and autoconden- sation can be introduced through heterocyclic ring opening. g y g p g A drastic increase in the reactivity can be attributed to the liberation of the phloroglucinol species of intermediate products. Model compounds have been used to demonstrate that alkaline-catalyzed rearrangements increase tannin reactivity. Nevertheless, some researches have considered model compounds to demonstrate that tannin structural rearrangements can increase or decrease reactivity toward aldehydes. y The autocondensation reactions that are characteristic of polyflavonoid tan- nins have recently been utilized in adhesive preparation processes, i.e., adhesive hardening in the absence of aldehyde. Autocondensation reactions are based on the opening of the O1–C2 bond of the flavonoid repeat unit and the subsequent condensation of the reactive center formed at C2 with free C6 or C8 sites of a flavonoid unit on another tannin chain under alkaline or acidic conditions (Figure 5). Although this reaction increases the viscosity considerably, gelling does not generally take place. Normally, gelling occurs (1) in the presence of a small amount of dissolved silica (silicic acid or silicates) catalyst or some other catalysts and (2) on a lignocellulosic surface. y g In the case of highly reactive pine tannin, cellulose catalysis is sufficient to induce hardening and to produce boards with strengths that satisfy the relevant standard requirements for interior-grade panels. The addition of dissolved silica or silicate catalyst to low-reactive tannins, such as mimosa and quebracho, is the best approach to achieve the required panel strength. The amount of silicic acid or silicates affects gelling. Gelling accelerates as silicate content increases and stabilizes after reaching a certain value. Although tannin resin adhesive that was manufactured through autocondensation increases the dry strength of panels, the strength of the resulting crosslinking is insufficient for exterior-graded panels [18]. Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 On the contrary, only particleboards made with pine tannin and hexamine hardener satisfied grade E1 requirements. This tendency was attributed to the curing mechanism of the hardener, the reactiv- ity of the tannin molecule toward formaldehyde, and rapid reactivity of pine tannin toward formaldehyde [13, 14]. y Formaldehyde is substituted with other aldehydes given that the methylene linkages may be too short to form cross-linkages. Pizzi and Scharfetter have shown that furfural-aldehyde is an efficient cross-linking agent and an excellent plasticizer for tannin resin adhesives [15, 16]. The complete replacement of formaldehyde with other aldehydes is unfeasible owing to their slow reactivity with tannins. For example, the water resistance of cured tannin-formaldehyde networks was improved by substituting 10–30% of formaldehyde with other aldehydes with satu- rated hydrocarbon chains but not by the cosmetic addition of water repellents such as waxes. Tannin adhesives prepared and/or set and/or cured with other adhesives only or with mixtures of formaldehyde and high proportions of other aldehydes yielded cured bonds weaker than those obtained with formaldehyde alone or its mixtures with furfural. The metal ion effect on phenol-formaldehyde reactions can be applied to con- densed tannins of the flavonoid type with some degree of success. The acceleration effect of the metal ions follows the order of PbII, ZnII, CdII, NiII > MnII, MgII, CuII, CoII > MnIII, FeIII ≫ BeII, AlIII > CrIII, CoIII. 6 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Tannins - Structural Properties, Biological Properties and Current Knowledge Tannins - Structural Properties, Biological Properties and Current Knowledge Tannins - Structural Properties, Biological Properties and Current Knowledge autocondensation and Lewis acid-induced polyflavonoid autocondensation have different mechanisms but involve similar subsequent reactions [23]. Figure 5. Autocondensation of tannin resin. Figure 5. g 5 Autocondensation of tannin resin. g 5 Autocondensation of tannin resin. g 5 Autocondensation of tannin resin. autocondensation and Lewis acid-induced polyflavonoid autocondensation have different mechanisms but involve similar subsequent reactions [23]. autocondensation and Lewis acid-induced polyflavonoid autocondensation have different mechanisms but involve similar subsequent reactions [23]. 4.2 Acidic and alkaline hydrolysis and autocondensation Aldehyde curing agents should be added for the preparation of exterior-graded panels. Nevertheless, hardening through tannin autocondensa- tion without any aldehyde addition is also possible. The mechanism of polyfla- vonoid autocondensation has been examined using carbon-13 nuclear magnetic resonance and electron-spin resonance spectroscopy, among others [19–21]. Zinc acetate also appears to induce a similar type of autocondensation reac- tion that is slower than that induced by an aldehyde. The reaction induced by zinc acetate mainly occurs at high curing temperatures. Consequently, the effect of zinc acetate is too weak to hinder interflavonoid bond cleavage and pyran ring opening in procyanidins. Therefore, in the presence of zinc acetate, the autocondensation of prodelphinidins to prodelphinidins and prodelphinidins to procyanidins will occur, whereas that of procyanidins to procyanidins will never or will rarely occur [22]. p y p y y The autocondensation of polyflavonoid tannin is facilitated by the reaction that occurs on cellulose and lignocellulosic substrates. Cellulose-induced polyflavonoid 7 7 4.3 Sulfite reaction Tannin sulfonation is one of the most useful reactions in flavonoid chemistry and can be particularly useful for the preparation of tannin resin adhesives. The drastic differences between the sulfite treatment products of resorcinol A-ring type tannins (e.g., black wattle tannins) and those of resorcinol A-ring type tannins (e.g., pine tannins) are mainly attributed to the different stabilities of the linkage bonds between their units relative to those of heterocyclic ether bonds. When sodium bisulfite is used to treat black wattle tannins, heterocyclic ether bonds first open because of the relative stability of the connecting bonds between units. Then, sulfonate is added to C-2. In this situation, tannin molecules are negligibly degraded. 8 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 The reaction of 5,7-dihydroxy proanthocyanidins with sulfite ions under normal pH conditions proceeds through the cleavage of the interflavonoid bond with the formation of flavan-4- or proanthocyanidin-4-sulfonates, as indicated by the scheme shown in Figure 6. Figure 6. Sulfonation of tannins. Figure 6. g Sulfonation of tannins. g Sulfonation of tannins. Sulfonated products can be obtained from phloroglucinolic tannins without the opening of the etherocyclic ring because interflavonoid bonds are easily cleaved. Flavan-2,4-disulfonates are also formed readily. The involvement of interflavonoid bond cleavage in the sulfonation of phloro- glucinolic condensed tannins affects the utilization of these tannins because their molecular weights can be tailored to suit their applications such as wood adhesives. Additionally, sulfonation affords tannins with reduced viscosity and increased solubility through the following mechanisms: 1. The elimination of the water-repellent etherocyclic ether group. 2. The introduction of the hydrophilic sulfonate group and another hydroxyl group. 3. The reduction in polymer rigidity, steric hindrance, and intermolecular hydro- gen bonding through the opening of the etherocyclic ring. 4. The hydrolysis of hydrocolloid gums and interflavonoid bonds under acidic conditions. 4. The hydrolysis of hydrocolloid gums and interflavonoid bonds under acidic conditions. However, sulfonation may be disadvantageous because sulfonate groups promote sensitivity to moisture and thus aggravate the deterioration of adhe- sive. This problem could be solved through desulfonation. The desulfonation of 2,4,6-trihydroxybenzyl sulfonic acid and sodium epicatechin-(4β)-sulfonate is a facile reaction under mild alkaline conditions (i.e., pH > 8.0 and ambi- ent temperature). Hydroxyl benzyl sulfonic acids with resorcinol or phenol functionalities resist desulfonation at a pH value of 12 and a temperature of 90°C. 4.3 Sulfite reaction Therefore, sulfonation not only reduces molecular weight while improving the viscosity and solubility of tannin resin adhesives but also prevents sulfonic acid functionalization and affords aldehyde condensation products that are insoluble in water [24]. 5.1.1 Plywood Tannin resin adhesives are used to prepare plywood (Figure 7). However, adhe- sives manufactured with conventional formulations and technology fails to meet the requirements set for the exterior plywood adhesive. Thus, tannin resin adhe- sives must be modified by mixing them with other synthetic resins or organic or inorganic modifiers and by optimizing resin synthesis parameters and hot-pressing conditions. Additives can effectively solve the problems of tannin resin hydropho- bicity and formaldehyde release and can improve the physical and mechanical prop- erties, especially weather resistance, of the final wood products. Plywood products assembled with modified tannin resin adhesives meet the demand of exterior-grade plywood, have better properties than plywood assembled using phenolic resin, and have a certain commercial potential [26–30]. 5. Applications Tannin resin adhesives can be cured under high heat (thermosetting) or at room temperature (coldsetting) [25]. Thermoset tannin resin adhesives are used in the preparation of plywood, particleboard, wood preservation resin, and impregnated 9 Tannins - Structural Properties, Biological Properties and Current Knowledge resin, among other wood composites. Coldset tannin resin adhesives are used to manufacture glulam, laminating veneer lumber, and finger joints. Figure 8. Figure 8. Tannin resin adhesives for manufacturing particleboard. adhesives. Sometimes phenolic resin is also mixed with tannin resin. The properties of particleboards produced with 60% tannin resin still meet the requirements of exterior-grade application [34]. Tannin resin adhesives for particleboard production have high requirements for curing agents. Different types of tannin resin adhesives require different curing agents. The performances of tannin, i.e., tannin structure and curing agent selection, and the properties of the particleboard will be affected by hot-pressing conditions, including pressing temperature, time, and pressure. Selecting an appropriate curing agent can accelerate tannin resin curing, reduce formaldehyde emission, and most importantly, can ensure that the performances of the particle- board meet exterior-grade application requirements [35]. Kim et al. found that the reactive speeds of tannin resin adhesives for black wattle tannin followed the order of paraformaldehyde > hexamine > trinitromethane, whereas those for pine tannin followed the order of hexamine > paraformaldehyde > trinitromethane [36]. p y Additionally, modified tannin resin can be used to prepare particleboards from different sources, such as wheat straw [37], rice husk [38], cashew nut shell [39], and chestnut shell [40]. The elastic moduli, internal bonding strength, and water- absorbing thickness swelling of the prepared particleboards meet the requirements of European standards. 5.1.2 Particleboard The use of tannin resin adhesives in particleboard production has been accepted in many countries and is used in the manufacturing of industrial particleboard in many countries except in Asian countries. For example, mimosa tannin resin adhesives are used in industrial particleboard manufacture in South Africa and South America (Figure 8). g Although tannin resin adhesives have been successfully used for the production of interior and exterior particleboards, the synthesis processes and formulas of tannin resin are drastically different [31]. Generally, tannin resins polymerized with formaldehyde have higher weathering resistance than those polymerized with other nontoxic and nonvolatile aldehydes in accordance with the reactivity of tannin. Particleboards manufactured using tannin resin adhesives with formaldehyde contents that have been partially or completely replaced with acetaldehyde have very low formaldehyde emission or even free from formaldehyde release and have mechanical properties that meet the requirements of interior-grade application [32]. At the same time, particleboards prepared with tannin resin synthesized via an aldehyde-free process can also meet the requirements of interior-grade applications [33]. An appropriate amount of paraformaldehyde or curing agent must be added to increase the weathering resistance of particleboards prepared with tannin resin Figure 7. Tannin resin adhesives for plywood manufacturing: (a) profile image; (b) wood failure. Figure 7. Figure 7. Tannin resin adhesives for plywood manufacturing: (a) profile image; (b) wood failure. 10 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 adhesives. Sometimes phenolic resin is also mixed with tannin resin. The properties of particleboards produced with 60% tannin resin still meet the requirements of exterior-grade application [34]. Tannin resin adhesives for particleboard production have high requirements Figure 8. Tannin resin adhesives for manufacturing particleboard. Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 5.1.4 Wood-impregnated paper Melamine formaldehyde resin has been widely used in the decorative impreg- nated paper industry. The addition of small amounts of urea can drastically cut costs without affecting performance. Melamine resin-impregnated paper is widely used for the production of laminated wooden floors and panel furniture overlays [47]. Phenolic resin-impregnated paper has limited applications in the production of floor and furniture panels owing to its black color, but it is widely accepted and popular in some particular applications, such as the impregnation of architectural template veneer paper, owing to its good adhesive property and high strength [48]. g g The flavonoid phenol structure and properties of tannin are similar to those of phenols. Therefore, there are no theoretical constraints for using tannin resin in paper impregnation. Abdullah et al. [49, 50] synthesized a low-viscosity tannin resin, which was used to impregnate paper with a glue amount of 172 g/m2 and hot-pressed on wood-based panel substrates. The final overlay paper exhibited a very smooth surface, high wear resistance, scratch resistance, and water steam resistance. The performances of the optimized overlay paper were even superior to those of overlay paper impregnated with MUF resin (Figure 10). Similar to phenolic resins, overlay paper impregnated with tannin resin has potential use in template production. 5.1.3 Wood preservation Wood preservation is vital for protecting wooden products. Traditional approaches for wood preservation include the treatment of wood with various chemical agents to prevent attack by different organic microorganisms and insects. Although traditional wood preservatives confer good effects and strong durability, they inevitably introduce various other problems, such as environmental pollution and carcinogenic effects. Tannin is a natural fungicide and good preservative that can be used to prevent fungal and bacterial damage [41]. Most plant pathogens secrete enzymes that degrade cellulose and lignin. Tannin can effectively inhibit the activity of these enzymes and prevent the proliferation of pathogens by complexing with protein [42]. Pizzi and Conradie confirmed that the antifungal activity of wood treated with flavonoid tannin resins is twice as intense as that of neat wood [43]. Additionally, veneer treated with tannin resin modified with a small amount of boric acid avoided the fungal attack and exhibited high durability, mechanical strength, and fire resistance because tannin and boric acid can simultaneously inhibit bacterial and fungal growth (Figure 9). Meanwhile, the fixation of boric acid in wood with tannin resin and hexamine prevented loss and exerted a good preservative effect that met the requirements of the European standard EN 113 [44–46]. 11 Tannins - Structural Properties, Biological Properties and Current Knowledge 5.1.4 Wood-impregnated paper Figure 9. Tannin resin adhesives for wood preservation. Figure 9. Tannin resin adhesives for wood preservation. Figure 9. Tannin resin adhesives for wood preservation. Figure 9. Tannin resin adhesives for wood preservation. g 9 Tannin resin adhesives for wood preservation. 5.1.5 Fiber: veneer composites Natural fiber composites have been developed to overcome the limitations asso- ciated with petrochemical resources. They have extensive prospective applications in the automobile and aerospace industries because their raw materials are derived from biomass and they possess unique characteristics. Fibers impregnated with Figure 10. Overlay paper impregnated with (a) tannin resin; (b) MUF resin. Figure 10. Figure 10. Overlay paper impregnated with (a) tannin resin; (b) MUF resin. 12 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Figure 11. (a) Tannin resin impregnated fiber; (b) laminated composite with tannin resin impregnated fiber; (c) wood composites with veneer and tannin resin impregnated fiber; (d) tannin resin impregnated fiber overlay on the particleboard. Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Figure 11. Figure 11. (a) Tannin resin impregnated fiber; (b) laminated composite with tannin resin impregnated fiber; (c) wood composites with veneer and tannin resin impregnated fiber; (d) tannin resin impregnated fiber overlay on the particleboard. tannin resin can be used to synthesize different types of fibers or wood-based fiber composites (Figure 11). These composites possess high elastic moduli and tensile strength and good water-absorbing expansibility [51–54]. 5.1.6 Other wooden panels In addition to wooden panels, medium density fiberboard [55], oriented strand boards [56], wafer boards [57], container boards [58], and other furniture panels [59] could be prepared with tannin resin adhesives. 6. Conclusions Tannin has significant application prospects as a promising natural phenolic polymer. However, this raw material continues to exhibit limitations, such as reactivity, high viscosity, short pot life, and poor weather resistance, among others. Future works must address these problems. Research on tannin resin adhesives should focus not only on wood panels but also on other advanced wooden compos- ites. In addition, the industrialization of tannin resin adhesives in the field of wood manufacturing field is necessary. 5.2 Application of coldset tannin resin adhesives in glued wood products Adhesives for finger joint lumber and glulam must meet high standard requirements because of the rigorous application environment of these materi- als (Figure 12). These adhesives must possess high mechanical strength and Figure 12. Coldset tannin resin adhesives for wooden product application: (a) glulam; (b) laminates. Figure 12. Coldset tannin resin adhesives for wooden product application: (a) glulam; (b) laminates. Figure 12. Coldset tannin resin adhesives for wooden product application: (a) glulam; (b) laminates. 13 Tannins - Structural Properties, Biological Properties and Current Knowledge weather resistance. Acacia tree tannin has been widely used to prepare low- temperature curing resin adhesives for finger joint and laminated veneer lumbers [60]. Acacia tannin resin adhesive has a low curing temperature and excellent bonding performance. Additionally, the cost of acacia tannin is lower than that of phenol resorcinol formaldehyde resin. The performance and cost of tannin- resorcinol-formaldehyde resin adhesives must be balanced. The tannin content of such adhesives could reach up to 95% after optimization. The polymerization of resorcinol units is replaced by that of a large number of flavonoid tannin natural phenol units. Resin adhesives with high tannin contents can be cured at room temperature, can exhibit good performances, and can be used to produce veneer and finger joint lumbers and glulam [61, 62]. Other nonvolatile or nontoxic aldehydes, such as glyoxal and glutaraldehyde, are also used to synthesize coldset tannin resin adhesives to effectively reduce the problem of formaldehyde release. Although only a small amount of paraformaldehyde is used to cure tannin resin adhesives, the prepared wooden products demonstrate good mechanical strength and water resistance [63]. Acknowledgements This work is supported by the following grants and programs: 1. National Natural Science Foundation of China (NSFC 31760187); 2. Yunnan Provincial Applied and Basic Research Grants (2017FB060); 3. “Ten-thousand Program”- youth talent support program; and 4. Yunnan Provincial Reserve Talents for Middle & Young Academic and Technical Leaders. Conflict of interest There is no conflict of interest in this field. 14 © 2019 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms f th C ti C Att ib ti Li (htt // ti /li / Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Author details Xiaojian Zhou* and Guanben Du Yunnan Provincial Key Laboratory of Wood Adhesives and Glued Products, Southwest Forestry University, Kunming, China *Address all correspondence to: xiaojianzhou@hotmail.com Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Author details 15 Tannins - Structural Properties, Biological Properties and Current Knowledge References [1] Dunky M, Pizzl A. Chapter 23–Wood adhesives. In: Adhesion Science & Engineering. 2002. pp. 1039-1103 [10] Pizzi A. Hot-setting tannin-urea- formaldehyde exterior wood adhesives. Adhesives Age. 1977;12:27-29 [2] Tondi G. Tannin-based copolymer resins: Synthesis and characterization by solid state 13C NMR and FT-IR spectroscopy. Polymers. 2017;9(6):223 [11] Pizzi A, Merlin M. A new class of tannin adhesives for exterior particleboard. International Journal of Adhesion and Adhesives. 1981;1(5):261-264 [3] Roffael E, Dix B, Okum J. Use of spruce tannin as a binder in particleboards and medium density fiberboards (MDF). European Journal of Wood and Wood Products. 2000;58(5):301-305 [3] Roffael E, Dix B, Okum J. Use of spruce tannin as a binder in particleboards and medium density fiberboards (MDF). European Journal of Wood and Wood Products. 2000;58(5):301-305 [3] Roffael E, Dix B, Okum J. Use of spruce tannin as a binder in particleboards and medium density fiberboards (MDF). European Journal of Wood and Wood Products. 2000;58(5):301-305 [12] Pizzi A. Exterior wood adhesives by MDI crosslinking of polyflavonoid tannin B rings. Journal of Applied Polymer Science. 1980;25:2123-2127 [13] Pizzi A, Tekely P. Mechanism of polyphenolic tannin resin hardening by hexamethylenetetramine: CP-MAS 13C-NMR. Journal of Applied Polymer Science. 1995;56:1645-1650 [4] Drewes E, Roux DG. Condensed tannins. XV. Interrelations of flavonoid components in wattle-bark extract. The Biochemical Journal. 1963;87(1):167-172 [14] Kim S, Kim HJ. Evaluation of formaldehyde emission of pine and wattle tannin-based adhesives by gas chromatography. Holz als Roh- und Werkstoff. 2004;62:101-106 [5] Roux DG, Paulus E. Condensed tannins. VIII. The isolation and distribution of interrelated heartwood components of Schinopsis species. The Biochemical Journal. 1961;78:785-789 [5] Roux DG, Paulus E. Condensed tannins. VIII. The isolation and distribution of interrelated heartwood components of Schinopsis species. The Biochemical Journal. 1961;78:785-789 [15] Pizzi A, Scharfetter HO. The chemistry and development of tannin- based adhesives for exterior plywood. Journal of Applied Polymer Science. 1978;22:1745-1761 [15] Pizzi A, Scharfetter HO. The chemistry and development of tannin- based adhesives for exterior plywood. Journal of Applied Polymer Science. 1978;22:1745-1761 [6] Pasch H, Pizzi A, Rode K. MALDI-TOF mass spectrometry of polyflavonoid tannins. Polymer. 2001;42:7531-7539 [7] Pizzi A. Mechanism of viscosity variations during treatment of wattle tannins with hot NaOH. International Journal of Adhesion and Adhesives. 1998;4(1):213-214 [7] Pizzi A. Mechanism of viscosity variations during treatment of wattle tannins with hot NaOH. International Journal of Adhesion and Adhesives. 1998;4(1):213-214 [16] Baaka N, Ammar M, Saad MK, et al. References Properties of tannin-glyoxal resins prepared from lyophilized and condensed tannin. Journal of Fashion Technology & Textile Engineering. 2017;3(4):00110 [16] Baaka N, Ammar M, Saad MK, et al. Properties of tannin-glyoxal resins prepared from lyophilized and condensed tannin. Journal of Fashion Technology & Textile Engineering. 2017;3(4):00110 [8] Sowunmi S, Ebewele RO, Conner AH, et al. Differential scanning calorimetry of hydrolysed mangrove tannin. Polymer International. 2000;49:574-578 [17] Sun WD. Chemistry of Tannin. China: China Forestry Publishing House; 1988. pp. 368-388 [18] Pizzi A, Meikleham N. Induced accelerated autocondensation of polyflavonoid tannins for phenolic polycondensates. II. CP-MAS 13C- NMR of different tannins and models. Journal of Applied Polymer Science. 1995;55:1265-1269 [9] Pizzi A, Horak RM, Roux DG. Condensates of phenol, resorcinol, phloroglucinol, and pyrogallol as model compounds of flavonoid A- and B-rings with formaldehyde. Journal of Applied Polymer Science. 1979;24:1571-1578 16 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 for producing exterior grade plywood in China. Holz als Roh- und Werkstoff. 1994;52(2):113-118 for producing exterior grade plywood in China. Holz als Roh- und Werkstoff. 1994;52(2):113-118 [19] Garcia R, Pizzi A. Polycondensation and autocondensation networks in polyflavonoid tannins. I. Final networks. Journal of Applied Polymer Science. 1998;70:1083-1091 [19] Garcia R, Pizzi A. Polycondensation and autocondensation networks in polyflavonoid tannins. I. Final networks. Journal of Applied Polymer Science. 1998;70:1083-1091 [28] Moubarik A, Pizzi A, Allal A, Charrier F, Charrier B. Cornstarch and tannin in phenol–formaldehyde resins for plywood production. Industrial Crops and Products. 2009;30(2):188-193 [20] Masson E, Pizzi A, Merlin M. Comparative kinetics of the induced radical autocondensation of polyflavonoid tannins. II. Flavonoid units effects. Journal of Applied Polymer Science. 1997;64:243-265 [29] Pizzi A, Knauff C, Sorfa P. Effect of acrylic emulsions on tannin-based adhesives for exterior plywood. Holz als Roh- und Werkstoff. 1981;39(6):223-226 [29] Pizzi A, Knauff C, Sorfa P. Effect of acrylic emulsions on tannin-based adhesives for exterior plywood. Holz als Roh- und Werkstoff. 1981;39(6):223-226 [21] Garcia R, Pizzi A, Merlin A. Ionic polycondensation effects on the radical autocondensation of polyflavonoid tannins: An ESR study. Journal of Applied Polymer Science. 1997;65:2623-2633 [30] Gao Z, Yuan J, Wang X. Phenolated larch-bark formaldehyde adhesive with multiple additions of sodium hydroxide. Pigment & Resin Technology. 2007;36(5):279-285 [22] Pizzi A, Tekely P. 13C-NMR of Zn2+ acetate-induced autocondensation of polyflavonoid tannins for phenolic polycondensates. Journal of Applied Polymer Science. 1995;56:633-636 [31] Pizzi A. [27] Zhao L, Cao B, Wang F, Yazaki Y. Chinese wattle tannin adhesives suitable [25] Brosse N, Pizzi A. Tannins for wood adhesives, foams and composites: Preparation, Characterization, and Testing. In book: Bio-based wood Adhesives. 2017. pp. 197-220 [26] Efhamisisi D, Thevenon MF, Hamzeh Y, et al. Induced tannin adhesive by boric acid addition and its effect on bonding quality and biological performance of poplar plywood. ACS Sustainable Chemistry & Engineering. 2016;4(5):2734-2740 [22] Pizzi A, Tekely P. 13C-NMR of Zn2+ acetate-induced autocondensation of polyflavonoid tannins for phenolic polycondensates. Journal of Applied Polymer Science. 1995;56:633-636 References Friendly wood preservative system based on polymerized tannin resin- boric acid for outdoor application. Maderas-Ciencia Y Tecnologia. 2010;12(3):253-257 [38] Tesha JV, Ndazi B, Bisanda E. Properties of rice husk particleboards bonded by a blend of tannin-cnsl resin. Tanzania Journal of Engineering and Technology. 2018;25(2):55-64 [39] Bisanda E, Ogola W, Tesha J. Characterisation of tannin resin blends for particle board applications. Cement and Concrete Composites. 2003;25(6):593-598 [39] Bisanda E, Ogola W, Tesha J. Characterisation of tannin resin blends for particle board applications. Cement and Concrete Composites. 2003;25(6):593-598 [47] Taylor AR, Wulfekotter D. Modified melamine resin for use in decorative laminates: U.S. Patent 4,713,299. December 15, 1987 [48] Chun-gui DU. A study on top grade bamboo-wood composite concrete- form. Wood Processing Machinery. 2004;15(1):9-11 [40] Santos J, Antorrena G, Freire MS, et al. Environmentally friendly wood adhesives based on chestnut (Castanea sativa) shell tannins. European Journal of Wood and Wood Products. 2017;75(1):89-100 [49] Abdullah U, Zhou X, Pizzi A, Merlin A. Note on surface quality of plywood overlaid with mimosa (Acacia mearnsii) tannin and melamine urea formaldehyde impregnated paper: Effects of moisture content of resin impregnated papers before pressing on physical properties of overlaid panels. International Wood Products Journal. 2013;4(4):253-256 [49] Abdullah U, Zhou X, Pizzi A, Merlin A. Note on surface quality of plywood overlaid with mimosa (Acacia mearnsii) tannin and melamine urea formaldehyde impregnated paper: Effects of moisture content of resin impregnated papers before pressing on physical properties of overlaid panels. International Wood Products Journal. 2013;4(4):253-256 [41] Zucker WV. Tannins: Does structure determine function? An ecological perspective. The American Naturalist. 1983;121(3):335-365 [42] Laks PE, McKaig PA, Hemingway RW. Flavonoid biocides: Wood preservatives based on condensed tannins. Holzforschung-International Journal of the Biology, Chemistry, Physics and Technology of Wood. 1988;42(5):299-306 [42] Laks PE, McKaig PA, Hemingway RW. Flavonoid biocides: Wood preservatives based on condensed tannins. Holzforschung-International Journal of the Biology, Chemistry, Physics and Technology of Wood. 1988;42(5):299-306 [50] Abdullah UH, Pizzi A, Rode K, Delmotte L, Zhou X, Mansouri H. Mimosa tannin resins for impregnated paper overlays. European Journal of Wood and Wood Products. 2013;71(2):153-162 [50] Abdullah UH, Pizzi A, Rode K, Delmotte L, Zhou X, Mansouri H. Mimosa tannin resins for impregnated paper overlays. European Journal of Wood and Wood Products. 2013;71(2):153-162 [51] Nicollin A, Li X, Girods P, Pizzi A, Rogaume Y. Fast pressing composite using tannin-furfuryl alcohol resin and vegetal fibers reinforcement. Journal of Renewable Materials. [43] Pizzi A, Conradie W. A chemical balance/microdistribution theory- new CCA formulations for soft-rot control? Material und Organismen. 1986;21(1):31-46 [42] Laks PE, McKaig PA, Hemingway RW. Flavonoid biocides: Wood preservatives based on condensed tannins. Holzforschung-International Journal of the Biology, Chemistry, Physics and Technology of Wood. 1988;42(5):299-306 References Tannin-based adhesives. Journal of Macromolecular Science— Reviews in Macromolecular Chemistry. 1980;18(2):247-315 [23] Pizzi A, Meikleham N, Stephanou A. Induced accelerated autocondensation of polyflavonoid tannins for phenolic polycondensates. II. Cellulose effect and application. Journal of Applied Polymer Science. 1995;55:929-933 [32] Ballerini A, Despres A, Pizzi A. Non-toxic, zero emission tannin-glyoxal adhesives for wood panels. Holz als Roh- und Werkstoff. 2005;63(6):477-478 [33] Moubarik A, Mansouri HR, Pizzi A, Charrier F, Allal A, Charrier B. Corn flour-mimosa tannin-based adhesives without formaldehyde for interior particleboard production. Wood Science and Technology. 2013;47(4):675-683 [33] Moubarik A, Mansouri HR, Pizzi A, Charrier F, Allal A, Charrier B. Corn flour-mimosa tannin-based adhesives without formaldehyde for interior particleboard production. Wood Science and Technology. 2013;47(4):675-683 [24] Lei H. Résines à base de matériaux naturels et synthétiques destinées aux adhésifs pour le bois. In: Nancy 1. 2009 [25] Brosse N, Pizzi A. Tannins for wood adhesives, foams and composites: Preparation, Characterization, and Testing. In book: Bio-based wood Adhesives. 2017. pp. 197-220 [34] Lu Y, Shi Q. Larch tannin adhesive for particleboard. Holz als Roh- und Werkstoff. 1995;53(1):17-19 [34] Lu Y, Shi Q. Larch tannin adhesive for particleboard. Holz als Roh- und Werkstoff. 1995;53(1):17-19 [35] Pizzi A, Valenezuela J, Westermeyer C. Low formaldehyde emission, fast pressing, pine and pecan tannin adhesives for exterior particleboard. Holz als Roh- und Werkstoff. 1994;52(5):311-315 [35] Pizzi A, Valenezuela J, Westermeyer C. Low formaldehyde emission, fast pressing, pine and pecan tannin adhesives for exterior particleboard. Holz als Roh- und Werkstoff. 1994;52(5):311-315 [26] Efhamisisi D, Thevenon MF, Hamzeh Y, et al. Induced tannin adhesive by boric acid addition and its effect on bonding quality and biological performance of poplar plywood. ACS Sustainable Chemistry & Engineering. 2016;4(5):2734-2740 [36] Kim S, Lee Y, Kim H, Lee HH. Physico-mechanical properties of particleboards bonded with pine and wattle tannin-based adhesives. Journal [27] Zhao L, Cao B, Wang F, Yazaki Y. Chinese wattle tannin adhesives suitable 17 Tannins - Structural Properties, Biological Properties and Current Knowledge of Adhesion Science and Technology. 2003;17(14):1863-1875 and tannin adhesive. Annals of Forest Science. 2009;66(1):1 [37] Tabarsa T, Jahanshahi S, Ashori A. Mechanical and physical properties of wheat straw boards bonded with a tannin modified phenol–formaldehyde adhesive. Composites Part B: Engineering. 2011;42(2):176-180 [45] Thevenon M, Tondi G, Pizzi A. High performance tannin resin-boron wood preservatives for outdoor end-uses. European Journal of Wood and Wood Products. 2009;67(1):89-93 [46] Thévenon M, Tondi G, Pizzi A. [43] Pizzi A, Conradie W. A chemical balance/microdistribution theory- new CCA formulations for soft-rot control? Material und Organismen. 1986;21(1):31-46 [63] Zhou X, Pizzi A. Tannin– resorcinol–aldehyde cold-set wood adhesives with only formaldehyde as hardener. European Journal of Wood and Wood Products. 2013;71(4):537-538 [55] Xiao-dan H. The development on the black wattle tannin adhesive for exterior medium density fiberboard. China National Forest Products Industry. 2003;30(5):31-33 References 2013;1(4):311-316 [44] Moubarik A, Charrier B, Charrier F, Pizzi A, Allal A. Evaluation of decay resistance of wood products made from borax-impregnated wood and bonded with a formaldehyde-free cornstarch [52] Sauget A, Nicollin A, Pizzi A. Fabrication and mechanical analysis of mimosa tannin and commercial 18 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 Applications of Tannin Resin Adhesives in the Wood Industry DOI: http://dx.doi.org/10.5772/intechopen.86424 flax fibers biocomposites. Journal of Adhesion Science and Technology. 2013;27(20):2204-2218 [61] Gornik D, Hemingway R, Tišler V. Tannin-based cold-setting adhesives for face lamination of wood. European Journal of Wood and Wood Products. 2000;58(1):23-30 [61] Gornik D, Hemingway R, Tišler V. Tannin-based cold-setting adhesives for face lamination of wood. European Journal of Wood and Wood Products. 2000;58(1):23-30 [53] Zhu J, Zhu H, Immonen K, Brighton J, Abhyankar H. Improving mechanical properties of novel flax/ tannin composites through different chemical treatments. Industrial Crops and Products. 2015;67:346-354 [62] Mansouri HR, Pizzi A, Fredon E. Honeymoon fast-set adhesives for glulam and fingerjoints of higher natural materials content. European Journal of Wood and Wood Products. 2009;67(2):207-210 [54] Zhou X, Segovia C, Abdullah UH, et al. A novel fiber–veneer-laminated composite based on tannin resin. The Journal of Adhesion. 2017;93(6):461-467 [54] Zhou X, Segovia C, Abdullah UH, et al. A novel fiber–veneer-laminated composite based on tannin resin. The Journal of Adhesion. 2017;93(6):461-467 [63] Zhou X, Pizzi A. Tannin– resorcinol–aldehyde cold-set wood adhesives with only formaldehyde as hardener. European Journal of Wood and Wood Products. 2013;71(4):537-538 [55] Xiao-dan H. The development on the black wattle tannin adhesive for exterior medium density fiberboard. China National Forest Products Industry. 2003;30(5):31-33 [56] Pichelin F, Pizzi A, Frühwald A. OSB adhesives rate of strength development on single strands couples. European Journal of Wood and Wood Products. 2000;58(3):182-183 [57] Li G, Lei H, Shuang-bao Z, De-rong Z, Qiu-hui Z. Study on tannin resin Pinus tabulaeformis waferboard. Wood Processing Machinery. 2002;13(5):20-23 [58] Saayman HM, Brown CH. Wattle- base tannin-starch adhesives for corrugated containers. Forest Products Journal. 1977;27(4):21-25 [59] Lee J, Jeon J, Kim S. Green adhesives using tannin and Cashew nut shell liquid for environment-friendly furniture materials. Journal of Korean Furniture Society. 2011;22(3):219-229 [60] Pizzi A, Rossouw DT, Knuffel W, Singmin M. ‘Honeymoon’ phenolic and tannin-based fast-setting adhesive systems for exterior grade finger joints. Holzforschung und Holzverwertung. 1980;32(6):140-150 19
https://openalex.org/W2519594823	https://europepmc.org/articles/pmc5026333?pdf=render	English	null	Appraisal of Laparoscopic Distal Pancreatectomy for Left-Sided Pancreatic Cancer: A Large Volume Cohort Study of 152 Consecutive Patients	PloS one	2,016	cc-by	6,892	RESEARCH ARTICLE Background The aim of this study was to appraise the value of laparoscopic distal pancreatectomy (LDP) for left-sided pancreatic cancer based on a large volume cohort study. OPEN ACCESS Citation: Shin SH, Kim SC, Song KB, Hwang DW, Lee JH, Park K-M, et al. (2016) Appraisal of Laparoscopic Distal Pancreatectomy for Left-Sided Pancreatic Cancer: A Large Volume Cohort Study of 152 Consecutive Patients. PLoS ONE 11(9): e0163266. doi:10.1371/journal.pone.0163266 Methods We reviewed data for all consecutive patients undergoing LDP for left-sided pancreatic can- cer at Asan Medical Center (Seoul, Korea) between December 2006 and December 2014. Received: May 26, 2016 Accepted: September 5, 2016 Published: September 16, 2016 Copyright: © 2016 Shin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Sang Hyun Shin, Song Cheol Kim, Ki Byung Song, Dae Wook Hwang, Jae Hoon Lee, Kwang-Min Park, Young-Joo Lee Division of Hepato-Biliary and Pancreatic Surgery, Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, Seoul, South Korea Division of Hepato-Biliary and Pancreatic Surgery, Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, Seoul, South Korea a1111 drksc@amc.seoul.kr * drksc@amc.seoul.kr Results A total of 91 male and 61 female patients, with a median age of 62.7 years were included in this study. The median operative duration was 234 minutes. Pathological reports revealed the following: a median tumor size of 3.0 cm (range, 0.4–10.0), T stages (T1 in 7.9%, T2 in 5.3%, T3 in 86.8%, and no T4), the tumor differentiation (well differentiated in 16.4%, mod- erately differentiated in 75.4%, and poorly differentiated in 8.2%), and R0 resection in 126 patients (82.9%). After pancreatectomy, 96 patients (63.2%) received adjuvant chemother- apy, and the median time to chemotherapy was 30 days. The median length of hospital stay was 8 days (range, 5–31), and the median time to diet resumption was 1 day. Grade B or C postoperative pancreatic fistula occurred in 14 patients (9.2%) and grade II or III complica- tions occurred in 27 (17.7%). The median overall survival was 43.0 months. A Cox propor- tional hazards model showed that tumor size, N1 stage, combined resection, and incompleteness of planned adjuvant chemotherapy affect patient survival. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Appraisal of Laparoscopic Distal Pancreatectomy for Left-Sided Pancreatic Cancer: A Large Volume Cohort Study of 152 Consecutive Patients Sang Hyun Shin, Song Cheol Kim*, Ki Byung Song, Dae Wook Hwang, Jae Hoon Lee, Kwang-Min Park, Young-Joo Lee Introduction In the current era of minimally invasive surgery, laparoscopic or robotic surgeries have become a new standard paradigm for the various procedures performed to treat benign tumors, malig- nant tumors, and even transplantation [1–5]. The laparoscopic procedure for managing pan- creatic lesions has been standardized worldwide and its use has markedly increased since Gagner and colleagues [6] reported the first laparoscopic distal pancreatectomy (LDP) in 1996. The laparoscopic approach improves the visualization of retroperitoneal organs, and the absence of complicated anastomosis in distal pancreatectomy has led to LDP becoming the most commonly performed minimally invasive procedure in pancreatic surgery. Various previ- ous studies have reported that laparoscopic pancreatectomy shows similar or better results than open pancreatectomy in terms of postoperative outcomes [7–11]. However, these proce- dures were mainly performed in benign or low-grade malignant diseases, and the consensus on laparoscopic pancreatectomy for malignant diseases remains to be established. Our institute, which is a leading tertiary care hospital in South Korea, has extensive experi- ences with laparoscopic pancreatic surgery [10,12–15]. Our indications for laparoscopic pan- createctomy have been expanded to include more complicated diseases such as pancreatic cancer. We have already reported a comparative propensity score-matched analysis of laparo- scopic versus open distal pancreatectomy (ODP) for left-sided ductal adenocarcinoma [14]. That study was one of the largest studies to use propensity score-matching to investigate the outcomes of laparoscopic surgery for left-sided pancreatic ductal adenocarcinoma. In that report, LDP showed comparable oncologic outcomes to ODP and was associated with a shorter length of hospital stay and an earlier return to diet than the matched ODP group. In the present study, we analyzed the clinicopathological characteristics and postoperative outcomes of the largest series of patients who underwent LDP for resectable left-sided pancre- atic cancer in a single center. The objectives were to introduce our experiences and to appraise the value of LDP for left-sided pancreatic cancer. Conclusions LDP for left-sided pancreatic cancer is reasonable within selected indications. An interna- tional consensus on laparoscopic surgery for pancreatic cancer would be desirable and timely. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: This study was supported by a grant of the Korean Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (grant number: HI14C2640). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 1 / 13 Laparoscopic Distal Pancreatectomy for Cancer Operative Procedure First of all, current indications for LDP in our institute include the followings: (1) no distant metastasis, (2) no invasion to major vascular structures, (3) no involvement of adjacent organs, (4) no intraabdominal adhesion, and (5) no comorbidities precluding laparoscopic surgery. However, when unexpected involvement to adjacent organ is found during operation, we attempt laparoscopic en-bloc resection. Nowadays, there have been an increasing number of patients with locally advanced disease who undergo neoadjuvant therapy followed by surgical resection. These patients are candidates mainly for open surgery due to possibility of vascular reconstruction. When the patient is considered for LDP, the patient is informed about advan- tages and disadvantages of both ODP and LDP, and LDP is decided preoperatively when the patient agrees to receive a laparoscopic procedure. To transect the pancreas safely, rotated endoscopic linear staplers of various sizes (staple height, 3.5 to 4.2 mm) were used, depending on the thickness or hardness of the pancreas. After transecting the pancreas in the neck, pancreatosplenectomy is performed in an antegrade manner. Based on the concept of radical antegrade modular pancreatosplenectomy (RAMPS) [18], we perform en-bloc resection of peripancreatic retroperitoneal tissues to ensure that there is no residual tumor. The dissection plane of the RAMPS procedure is aiming at exposing the superior mesenteric artery, left side of the aorta, renal vessels, renal parenchyma, and adrenal gland. However, it is sometimes time-consuming to secure all planes of the RAMPS, so we sometimes modify the concept of the RAMPS according to the location or depth of the tumor. We concentrate on securing the retroperitoneal radial margin of the tumor rather than uni- formly exposing whole planes of RAMPS, especially when the tumor is located in the proximal body or far tail of the pancreas. Patients and Methods Patient Database Between December 2006 and December 2014, 462 consecutive patients with left-sided pancre- atic cancer underwent surgical resection at Asan Medical Center (Seoul, South Korea). Of the 462 patients, 169 (36.6%) underwent LDP. Their clinical, pathological, and surgical data were retrospectively reviewed using the electronic medical records of our institute. This study was approved by Asan Medical Center Institutional Review Board. Our institutional review board waived the need for written informed consent from the participants. Postoperative pancreatic fistulas (POPFs) and overall complications were assessed and graded based on the criteria of the International Study Group of Pancreatic Fistula (ISGPF) [16] and Clavien-Dindo compli- cation classification [17], respectively. Resection margin status was categorized as R0 or R1 PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 2 / 13 Laparoscopic Distal Pancreatectomy for Cancer (pancreatic transection or retroperitoneal margins). If the closest safe resection margin was less than 1 mm, it was categorized as an R1 retroperitoneal margin. Follow-up data were also obtained from the electronic medical records, and the duration of survival after surgery was measured from the time of surgery until death or the last visit to the outpatient department. All patients were preoperatively assessed using computed tomography (CT) and magnetic resonance cholangiopancreatography (MRCP). As a diagnostic strategy of our institute, most patients with PDAC preoperatively underwent 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) for initial cancer staging to find hidden metastasis. Endoscopic ultrasonography (EUS) and/or EUS-guided fine-needle aspiration biopsy were sometimes performed for accurate diagnosis in selected patients. After surgical resection, all patients underwent CT to assess surgical complications, including POPF, on the 3rd or 4th postoperative day. As postoperative surveillance, CT was performed and CA 19–9 levels were examined every 3 months in the first 2 postoperative years and then every 6 months in the sub- sequent years in all patients. If necessary, FDG-PET, chest CT, and/or biopsy were also per- formed to confirm recurrence. Patient Characteristics Of a total of 169 patients who underwent LDP for left-sided pancreatic cancer during the study period, 7 had stage IV disease, 6 were lost to follow-up, and 4 underwent open conversion. Thus, the remaining 152 patients were included in this study. The reasons for open conversion were as follows: invasion of the common hepatic artery, invasion of the superior mesenteric vein, massive bleeding, and severe adhesions caused by previous colon operation. The demo- graphic characteristics of all patients who underwent LDP are presented in Table 1. There were 91 women and 61 men, with a median age at diagnosis of 62.7 years (range, 30–88). Most patients were included in ASA score grade I or II because patients with poor general condition (ASA score grade IV or V) were not candidates for laparoscopic surgery. Regarding comorbid- ity, 51 patients had preoperative diabetes mellitus, 71 had cardiovascular disease, and 6 had pulmonary disease. The median preoperative body mass index was 24.0 kg/m2 (range, 16.6– 32.2). Operative Features The operative features of all 152 patients are listed in Table 2. The median operative duration was 234 minutes (range, 121–475). During LDP, 13 patients (8.6%) underwent laparoscopically combined resection of directly invaded organs to secure margins: wedge resection of the stom- ach in 7 patients, left hemicolectomy in 3, sleeve resection of the duodenal 4th portion in 2, and resection and anastomosis of the proximal jejunum in 1. Of the entire cohort, 4 patients (2.6%) received intraoperative blood transfusion. Statistical Analysis Categorical variables are expressed as number and percentage, and continuous variables are reported as median and range. The entire study cohort and those who underwent LDP for duc- tal adenocarcinoma between December 2006 and December 2014 were included in the analysis of the cumulative survival rates of the patient and the disease-free survival calculated using the Kaplan-Meier method. To estimate the factors affecting patient survival following LDP, the Cox proportional hazards model was performed. All statistical analyses were conducted using IBM SPSS version 18.0 (IBM SPSS). 3 / 13 PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 Laparoscopic Distal Pancreatectomy for Cancer PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 Pathological Features Pathological features are described in Table 3. Histological diagnoses of the resected left-sided pancreatic cancers, with a median size of 3.0 cm (range, 0.4–10.0), were of 130 ductal adenocar- cinomas (85.5%), 18 invasive intraductal papillary mucinous neoplasms (11.8%), and 4 mucin- ous cystadenocarcinomas (2.6%). Most patients (86.8%) had T3 cancer. The median number of Table 1. Demographic characteristics of all 152 patients who underwent LDP. Characteristics N or median % or range Sex Female 91 59.9 Male 61 40.1 Age, year 62.7 30–88 ASA score Grade I 38 25.0 Grade II 105 69.1 Grade III 9 5.9 Grade IV-V 0 0.0 Comorbidity DM 51 Cardiovascular 71 Pulmonary 6 BMI, kg/m2 24.0 16.6–32.2 DM, diabetes mellitus; BMI, body mass index doi:10.1371/journal.pone.0163266.t001 Table 1. Demographic characteristics of all 152 patients who underwent LDP. PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 4 / 13 Laparoscopic Distal Pancreatectomy for Cancer Table 2. Operative features. Features N or median % or range Operative duration, min 234 121–475 Combined resection, laparoscopically No 139 91.4 Yes 13 8.6 Wedge resection of stomach 7 Left hemicolectomy 3 Sleeve resection of duodenal 4th portion 2 Resection and anastomosis of proximal jejunum 1 Intraoperative transfusion No 148 97.4 Yes 4 2.6 doi:10.1371/journal.pone.0163266.t002 Table 3. Pathological features (n = 152). Features N or median % or range Histologic diagnosis Ductal adenocarcinoma 130 85.5 Invasive intraductal papillary mucinous neoplasm 18 11.8 Mucinous cystadenocarcinoma 4 2.6 Tumor size, cm 3.0 0.4–10.0 T stage T1 12 7.9 T2 8 5.3 T3 132 86.8 T4 0 0.0 N stage N0 87 57.2 N1 65 42.8 Number of harvested LN 11 0–42 Differentiation Well 22 16.4 Moderate 101 75.4 Poor 11 8.2 NA 18 - Resection margin status Negative (R0) 126 82.9 Pancreatic transection margin (R1) 1 0.7 Retroperitoneal margin (R1) a 25 16.4 Lymphovascular invasion Absent 107 70.4 Present 45 29.6 Perineural invasion Absent 47 30.9 Present 105 69.1 LN, lymph node; NA, not available a In 11 patients, the cancer didn’t penetrate the tangential margin, but their safety margin was less than 1mm and they were categorized as R1. 5 / 13 PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 Laparoscopic Distal Pancreatectomy for Cancer harvested lymph nodes (LNs) was 11 (range, 0–42), and 65 patients (42.8%) had a positive LN metastasis (N1 stage). R0 resection was achieved in 126 patients (82.9%) and R1 resection in 26 (17.1%). Postoperative Outcomes Postoperative outcomes are indicated in Table 4. After pancreatectomy, 96 patients (63.2%) received adjuvant chemotherapy, and their median time to chemotherapy was 30 days (range, 19–70). Of these 96 patients, 72 (75%) completed all planned adjuvant chemotherapy cycles. The median length of hospital stay was 8 days (range, 5–31), and the median time to water intake and a liquid diet were 1 and 2 days, respectively. Overall, POPF occurred in 48 patients (31.6%), and clinically relevant POPF (ISGPF grade B or C) was observed in 14 patients (9.2%). Complications of Clavien-Dindo grade II or higher occurred in 27 patients (17.7%). There was no in-hospital or 30-day mortality. Pathological Features In addition, there was microscopic involvement of the retroperitoneal margin in 25 patients (16.4%) and of the pancreatic transection margin in 1 case (0.7%). Lymphovascular and perineural invasion were present in 45 (29.6%) and 105 (69.1%) patients, respectively. y g p p a Of 96 patients, 72 (75%) completed full cycles of adjuvant chemotherapy. ISGPF, international study group of pancreatic ﬁstula PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 doi:10.1371/journal.pone.0163266.t004 Survival Analysis (B) In the pancreatic ductal adenocarcinoma cases (n = 130), the median survival was 37.0 months, and the 1-, 3-, and 5-year survival rates were 92.2%, 51.3%, and 39.3%, respectively. The 1-, 3, and 5-year disease-free survival rates were 63.3%, 46.0%, and 43.6%, respectively. doi:10.1371/journal.pone.0163266.g001 g ap a e e cu es o pat e t su a a d d sease ee su a ( ) t e e t e co o t (n = 152), the median patient survival was 43.0 months, and the 1-, 3-, and 5-year survival rates were 92.0%, 55.3%, and 44.7%, respectively. The 1-, 3, and 5-year disease-free survival rates were 66.7%, 50.8%, and 45.0%, respectively. (B) In the pancreatic ductal adenocarcinoma cases (n = 130), the median survival was 37.0 months, and the 1-, 3-, and 5-year survival rates were 92.2%, 51.3%, and 39.3%, respectively. The 1-, 3, and 5-year disease-free survival rates were 63.3%, 46.0%, and 43.6%, respectively. doi:10.1371/journal.pone.0163266.g001 doi:10.1371/journal.pone.0163266.g001 months and the 5-year patient survival and disease-free survival rates were 39.3% and 43.6%, respectively. To estimate the factors affecting patient survival following LDP, we performed multivariable analysis using the Cox proportional hazards model. In this analysis, several independent sur- vival factors were identified (Table 5). In the entire cohort, younger patients had relatively poorer survival (hazard ratio [HR] = 0.93). Larger tumor size (HR = 1.31), presence of lymph node metastasis (HR = 3.11), and combined resection of invaded adjacent organs (HR = 4.50) were included as dismal predictors of postoperative survival. Furthermore, when the patients received adjuvant treatment, failure to complete all planned cycles reduced survival, as expected (HR = 4.45). Survival Analysis In the entire cohort (n = 152; Fig 1A), the median survival was 43.0 months and the 1-, 3-, and 5-year patient survival rates were 92.0%, 55.3%, and 44.7%, respectively. The 1-, 3-, and 5-year disease-free survival rates were 66.7%, 50.8%, and 45.0%, respectively. When patients who underwent LDP for ductal adenocarcinoma (n = 130) were included in the analysis of the patient overall and disease-free survival outcomes (Fig 1B), the median survival was 37.0 Table 4. Postoperative outcomes. Outcome variables N or median % or range Adjuvant chemotherapy No 56 36.8 Yes a 96 63.2 Time to chemotherapy, day 30 19–70 Length of hospital stay, day 8 5–31 Time to restarting diet (water), day 1 1–6 Time to restarting diet (liquid diet), day 2 2–8 Pancreatic ﬁstula (ISGPF grade) None 104 68.4 Grade A 34 22.4 Grade B 9 5.9 Grade C 5 3.3 Complication Classiﬁcation (Clavien-Dindo) None 91 59.9 Grade I 34 22.4 Grade II 18 11.8 Grade III 9 5.9 Grade IV 0 0.0 Grade V 0 0.0 ISGPF, international study group of pancreatic ﬁstula a Of 96 patients, 72 (75%) completed full cycles of adjuvant chemotherapy. Table 4. Postoperative outcomes. 6 / 13 PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 Laparoscopic Distal Pancreatectomy for Cancer Fig 1. Kaplan-Meier curves of patient survival and disease-free survival. (A) In the entire cohort (n = 152), the median patient survival was 43.0 months, and the 1-, 3-, and 5-year survival rates were 92.0%, 55.3%, and 44.7%, respectively. The 1-, 3, and 5-year disease-free survival rates were 66.7%, 50.8%, and 45.0%, respectively. (B) In the pancreatic ductal adenocarcinoma cases (n = 130), the median survival was 37.0 months, and the 1-, 3-, and 5-year survival rates were 92.2%, 51.3%, and 39.3%, respectively. The 1-, 3, and 5-year disease-free survival rates were 63.3%, 46.0%, and 43.6%, respectively. doi:10.1371/journal.pone.0163266.g001 Fig 1. Kaplan-Meier curves of patient survival and disease-free survival. (A) In the entire cohor Fig 1. Kaplan-Meier curves of patient survival and disease-free survival. (A) In the entire cohort (n = 152), the median patient survival was 43.0 months, and the 1-, 3-, and 5-year survival rates were 92.0%, 55.3%, and 44.7%, respectively. The 1-, 3, and 5-year disease-free survival rates were 66.7%, 50.8%, and 45.0%, respectively. Discussion Laparoscopic pancreatectomy has been accepted as a standard procedure for benign or low- grade malignant disease. Previous studies showed that laparoscopic pancreatectomy has similar surgical outcomes to open pancreatectomy [7–13,19–21]. Although the laparoscopic technique Table 5. Cox proportional hazard model to estimate the factors affecting patients’ survival following LDP. Hazard ratio 95% Conﬁdence interval p value Age, year 0.932 0.891 to 0.975 0.002 Size, cm 1.311 1.002 to 1.714 0.049 N stage N0 Reference N1 3.113 1.330 to 7.286 0.009 Combined resection No Reference Yes 4.503 1.215 to 16.697 0.024 Planned adjuvant treatment Complete Reference Incomplete 4.448 1.847 to 10.713 0.001 doi:10.1371/journal.pone.0163266.t005 Table 5. Cox proportional hazard model to estimate the factors affecting patients’ survival following LDP. PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 7 / 13 Laparoscopic Distal Pancreatectomy for Cancer is not yet a commonly accepted approach for pancreatic cancer, several studies have reported the advantages and comparable outcomes of the laparoscopic approach [22–26]. We also reported, based on the extensive experience of our institute, that LDP for left-sided ductal ade- nocarcinoma showed several advantages over ODP and comparable oncologic outcomes [14]. In our present study, we have reported our experiences and appraised the value of LDP in all patients who underwent LDP for left-sided pancreatic cancer in our institute. There are three clinical implications of our current study. First, the surgical outcomes of LDP were similar to those of ODP and LDP reported in previous articles in terms of oncologic outcomes and the quality of the procedure [22–29]. As summarized in Table 6, all studies show comparable surgical outcomes. In our present study, the clinically significant complication rate was 17.7% (Clavien-Dindo grade II-V) and the clinically relevant POPF rate (ISGPF grade B or C) was 9.2%. Previous studies [9,14,21,22,24,30] have reported similar results, and LDP did not compromise surgical outcomes. Table 6. Summary of previously reported articles associated with operative outcomes of ODP and/or LDP for left-sided pancreatic cancer. First author Year (study years) No. of cases OP duration (min) Tumor size (cm) No. doi:10.1371/journal.pone.0163266.t006 ODP, open distal pancreatectomy; LDP, laparoscopic distal pancreatectomy; LN, lymph node; LOS, length of stay; NS, not stated; 5YSR, 5 year survival rate; 3YSR, 3 year survival rate; (M), Median; (m), mean a, based on the Revised Accordion Classiﬁcation b, based on the Clavien-Dindo Complication Classiﬁcation (grade I-V) cation Classiﬁcation (grade I V) ho underwent LDP for left-sided ductal adenocarcinoma b, based on the Clavien-Dindo Complication Classiﬁcation (grade I-V) , p (g ) c, estimated survival of 130 patients who underwent LDP for left-sided ductal adenocarcinoma n Classiﬁcation mplication Classiﬁcation (grade I-V) a, based on the Revised Accordion Classiﬁcation vised Accordion Classiﬁcation avien Dindo Complication Classiﬁcation (grade I V) b, based on the Clavien-Dindo Complication Classiﬁcation (grade I-V) c estimated survival of 130 patients who underwent LDP for left sided ductal adenocarcinoma P, open distal pancreatectomy; LDP, laparoscopic distal pancreatectomy; LN, lymph node; LOS, length of stay; NS, e; 3YSR, 3 year survival rate; (M), Median; (m), mean p (g ) al of 130 patients who underwent LDP for left-sided ductal adenocarcinoma Discussion of harvested LN N1 (%) R0 (%) Overall morbidity (%) LOS (day) Rate of adjuvant chemotherapy (%) Median survival (mo) Patient survival ODP Yamamoto et al29 2010 (1994– 2007) 73 (M) 345 (M) 2.8 NS 50 75.3 NS NS 34.2 NS 5YSR 30.0% Kooby et al23 2010 (2000– 2008) 189 (m) 230.4 (m) 4.5 (m) 12.5 NS 73 NS (m) 10.7 70 16 NS Mitchem et al27 2012 (1999– 2008) 47 (m) 243 (m) 4.4 (m) 18 55 81 83a (m) 11.3 NS 26 5YSR 35.5% Magge et al24 2013 (2002– 2010) 34 (m) 294 (M) 4.5 (M) 12 38 88 50b (M) 8 85 19 NS Rehman et al25 2014 (2008– 2011) 8 (M) 376 (M) 3.2 (M) 14 64 86 22 b (M) 12 64 52 NS Paye et al28 2015 (2004– 2009) 278 NS NS (M) 17 58.1 74.8 34.5 b NS 71.3 35 5YSR 29.5% Sharpe et al26 2014 (2010– 2011) 625 NS (M) 3.6 (M) 12 49 78 NS (M) 7 NS NS NS LDP Kooby et al23 2010 (2000– 2008) 23 (m) 238.4 (m) 3.5 (m) 13.8 NS 74 NS (m) 7.4 57 16 NS Magge et al24 2013 (2002– 2010) 28 (m) 317 (M) 3.7 (M) 11 57 86 39 b (M) 6 89 19 NS Rehman et al25 2014 (2008– 2011) 14 (M) 274 (M) 2.2 (M) 16 50 88 37 b (M) 8 50 33 NS Kawaguchi et al22 2014 (2002– 2013) 23 (m) 203 (m) 3.2 (m) 20 61 100 47 b (m) 17 NS 28 5YSR 33.0% Sharpe et al26 2014 (2010– 2011) 144 NS (M) 3.5 (M) 13 47 87 NS (M) 5 NS NS NS Current series 2015 (2006– 2014) 152 (M) 234 (M) 3.0 (M) 11 42.8 82.9 40.1 b (M) 8 63.2 37 c 5YSR 39.3%c ODP, open distal pancreatectomy; LDP, laparoscopic distal pancreatectomy; LN, lymph node; LOS, length of stay; NS, not stated; 5YSR, 5 year survival rate; 3YSR 3 year survival rate; (M) Median; (m) mean reported articles associated with operative outcomes of ODP and/or LDP for left-sided pancreatic cancer. PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 8 / 13 Laparoscopic Distal Pancreatectomy for Cancer International Study Group on Pancreatic Surgery (ISGPS) reported a consensus statement for lymphadenectomy in surgery for PDAC [31]. In this report, they recommended proper extent and number of a standard lymphadenectomy in pancreatic surgery. Discussion For cancers of the body and tail of the pancreas, removal of stations 10 (LNs at the splenic hilum), 11 (LNs along the splenic artery) and 18 (LNs along the inferior border of the body and tail of the pancreas) was standard, but there was no guideline for proper number of LNs retrieved during distal pan- createctomy. Ashfaq et al [32] studied the number of LNs required for accurate staging after distal pancreatectomy for pancreatic adenocarcinoma, and reported that at least 11 LNs should be examined to avoid understaging. The median or mean number of harvested LNs of LDP indicated in Table 6 ranges from 11 to 20. Although the median or mean values are not an absolute standard, the quality of LDP can be considered oncologically feasible and comparable to that of ODP. Additionally, a negative surgical margin is one of the important prognostic fac- tors for assessing oncologic adequacy. Although the RAMPS procedure was initially devised to achieve negative surgical margins and complete node dissections [18], the true oncologic and survival benefits have not achieved consensus [33,34]. Nevertheless, we believe that the concept of securing the retroperitoneal radial margin should be accepted, so we have generally followed it, sometimes modifying the procedure to reduce the time required. In this study, we achieved 82.9% R0 resection. Among the 25 patients who were categorized as having a positive retroper- itoneal margin (R1), the cancer in 11 patients did not microscopically penetrate the retroperi- toneal radial margin, but their safety margin was less than 1 mm (Table 3). In addition, we laparoscopically performed combined resection of invaded organs in 13 patients (8.6%). Yama- moto et al [29] reported that they performed combined resection of invaded organs, including the portal vein, in 20.5% patients during ODP. Although vascular resection and anastomosis is not an easy procedure in laparoscopic surgery, invasion to any other left-sided organs around the pancreas (stomach, colon, small intestine, and kidney) can be treated with appropriate laparoscopically obtainable surgical margins. The second clinical implication of our current study is improved recovery, earlier return to ordinary life and subsequent possibility of improved survival of patients undergoing laparo- scopic surgery. In our previous matched study [14] and other comparative studies The second clinical implication of our current study is improved recovery, earlier return to ordinary life and subsequent possibility of improved survival of patients undergoing laparo- scopic surgery. PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 Discussion In our previous matched study [14] and other comparative studies [9,11,21,25,26], LDP vs ODP showed that LDP was associated with a shorter operative time, shorter length of hospital stay, earlier return to diet and earlier return to ordinary life. These associated characteristics of the laparoscopic procedure meant that improved recovery from surgery led to more patients receiving adjuvant treatment in a shorter period. Previous studies of colon cancer [35] and ovarian cancer [36] have suggested that delayed initiation of adjuvant chemotherapy compromised overall survival. In the recent analysis of the European Study Group for Pancreatic Cancer-3 (ESPAC-3) trial, Valle et al [37] showed that completion of all 6 cycles of planned adjuvant chemotherapy was more predictive of survival than early initiation if chemotherapy was initiated within 12 weeks. However, they did not evaluate the survival of patients whose chemotherapy initiation was delayed beyond 12 weeks. Croome et al [38] reported that an adjuvant chemotherapy delay beyond 90 days (12 weeks) was a strong predic- tor of a dismal prognosis and that a significantly smaller proportion of patients had a delay of greater than 90 days in the laparoscopic pancreaticoduodenectomy group. In our present study, 63.2% of patients received adjuvant treatment. The median time to adjuvant chemother- apy was 30 days after surgery (range, 19–70). Two patients had a delay of more than 8 weeks (61 and 70 days), but no patient had a delay longer than 12 weeks. Additionally, we expect that the improved recovery following the laparoscopic procedure could lead to more patients com- pleting the planned adjuvant treatment. Of the patients receiving adjuvant treatment, 75% completed the planned chemotherapy cycle. When compared with the previous prospective tri- als that showed a 68% completion rate [37,39], more patients who completed all planned cycles PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 9 / 13 Laparoscopic Distal Pancreatectomy for Cancer in this study showed a higher probability of survival enhancement in the laparoscopic proce- dure. Our current data actually showed 5-year survival rates of the entire cohort (n = 152) and of patients with ductal adenocarcinoma (n = 130) of 44.7% and 39.3%, respectively. Acknowl- edging that patients who underwent LDP had relatively limited stages of cancer, we expected that this higher survival was attributable to not only tumor characteristics, but also to earlier recovery following laparoscopic procedure. S1 File. Data sheet for analysis. (PDF) S1 File. Data sheet for analysis. (PDF) Discussion Therefore, further study with high quality will be required to identify the correlation between early recovery and long-term oncologic outcomes in terms of receiving and completing the planned adjuvant treatment earlier. As the last implication of this study, we suggest relative indications and contraindications of LDP for left-sided pancreatic cancer. All types and all stages of left-sided pancreatic cancer can be treated by the laparoscopic procedure if the cancer belongs to a resectable category, even if the tumor invades adjacent organs. The instances of conversion to open surgery in this study could provide useful information for establishing the contraindications to LDP. When left- sided pancreatic cancer invades major vessels, such as the portal vein, superior mesenteric vein, celiac axis, and superior mesenteric artery, these vessels should be preserved by reconstruction. Therefore, LDP is not suitable for patients with major vascular invasion. Moreover, when there are severe adhesions caused by previous abdominal surgery, the laparoscopic approach might not be a safe choice. Our study has several limitations. There were selection biases caused by limitations in lapa- roscopic procedure. Although we reported in a previous comparative study [14] that there were no specific differences in the selection criteria in terms of clinicopathological parameters except for tumor size and concurrent resection of invaded organs, LDP had limitations in deal- ing with large sized cancer invading adjacent organs or locally advanced cancer. Relatively short follow-up duration of recent patients was also associated with another bias in calculating accurate survival. Additionally, although the median number of retrieved LN was similar with other studies and suggested statement, the median value showed that many patients had fewer than the recommended number of LNs. Among the studied patients, some patients with 0 or 1 of retrieved LN were preoperatively diagnosed to have premalignant disease. In these patients with incidental PDAC, additional operation for lymphadenectomy was not performed when resection margin was clear. This might be associated with a bias on accurate staging. Conclusions Acknowledging that this was a retrospective study and that the follow-up duration was rela- tively short, we appraised that LDP for left-sided pancreatic cancer was feasible or beneficial in terms of oncologic aspects. We believe that this approach is thoroughly reasonable within the selected indications and that it is time to establish an international consensus on laparoscopic surgery for pancreatic cancer. References 1. Barkun JS, Barkun AN, Sampalis JS, Fried G, Taylor B, Wexler MJ, et al. Randomised controlled trial of laparoscopic versus mini cholecystectomy. Lancet. 1992; 340(8828): 1116–1119. PMID: 1359210 2. Gagner M, Pomp A, Heniford BT, Pharand D, Lacroix A Laparoscopic adrenalectomy: lessons learned from 100 consecutive procedures. Ann Surg. 1997; 226(3): 238–247. PMID: 9339930 3. Gigot JF, Glineur D, Azagra JS, Goergen M, Ceuterick M, Morino M, et al. Laparoscopic liver resection for malignant liver tumors: preliminary results of a multicenter European study. Ann Surg. 2002; 236(1): 90–97. PMID: 12131090 4. Veldkamp R, Gholghesaei M, Bonjer HJ, Meijer DW, Buunen M, Jeekel J, et al. Laparoscopic resection of colon cancer: consensus of the European Association of Endoscopic Surgery (EAES). Surg Endosc. 2004; 18(8): 1163–1185. PMID: 15457376 5. Giessing M, Reuter S, Deger S, Tüllmann M, Hirte I, Budde K, et al. Laparoscopic versus open donor nephrectomy in Germany: impact on donor health-related quality of life and willingness to donate. Transplant Proc. 2005; 37(5): 2011–2015. PMID: 15964326 6. Gagner M, Pomp A, Herrera MF Early experience with laparoscopic resections of islet cell tumors. Sur- gery. 1996; 120(6): 1051–1054. PMID: 8957494 7. Mabrut JY, Fernandez-Cruz L, Azagra JS, Bassi C, Delvaux G, Weerts J, et al. Laparoscopic pancre- atic resection: results of a multicenter European study of 127 patients. Surgery. 2005; 137(6): 597–605. PMID: 15962401 8. Melotti G, Butturini G, Piccoli M, Casetti L, Bassi C, Mullineris B, et al. Laparoscopic distal pancreatec- tomy: results on a consecutive series of 58 patients. Ann Surg. 2007; 246(1): 77–82. PMID: 17592294 9. Vijan SS, Ahmed KA, Harmsen WS, Que FG, Reid-Lombardo KM, Nagorney DM, et al. Laparoscopic vs open distal pancreatectomy: a single-institution comparative study. Arch Surg. 2010; 145(7): 616– 621. doi: 10.1001/archsurg.2010.120 PMID: 20644122 10. Song KB, Kim SC, Park JB, Kim YH, Jung YS, Kim MH, et al. Single-center experience of laparoscopic left pancreatic resection in 359 consecutive patients: changing the surgical paradigm of left pancreatic resection. Surg Endosc. 2011; 25(10): 3364–3372. doi: 10.1007/s00464-011-1727-9 PMID: 21556993 11. Limongelli P, Belli A, Russo G, Cioffi L, D’Agostino A, Fantini C, et al. Laparoscopic and open surgical treatment of left-sided pancreatic lesions: clinical outcomes and cost-effectiveness analysis. Surg Endosc. 2012; 26(7): 1830–1836. doi: 10.1007/s00464-011-2141-z PMID: 22258300 12. Kim SC, Park KT, Hwang JW, Shin HC, Lee SS, Seo DW, et al. Writing – review & editing: SHS SCK JHL KMP YJL. Writing – review & editing: SHS SCK JHL KMP YJL. Acknowledgments This study was supported by a grant of the Korean Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea. (grant number: HI14C2640) The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. 10 / 13 PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 Laparoscopic Distal Pancreatectomy for Cancer Author Contributions Conceptualization: SHS SCK KBS DWH. Data curation: SHS SCK. Formal analysis: SHS SCK KBS DWH JHL. Investigation: SHS SCK KBS DWH JHL KMP YJL. Methodology: SHS SCK JHL KMP YJL. Project administration: SHS SCK KBS DWH JHL KMP YJL. Resources: SHS SCK KBS DWH JHL KMP YJL. Supervision: SHS SCK KMP YJL. Validation: SHS SCK KBS DWH. Writing – original draft: SHS SCK KBS DWH. Writing – review & editing: SHS SCK JHL KMP YJL. Writing – original draft: SHS SCK KBS DWH. Writing – review & editing: SHS SCK JHL KMP YJL. Conceptualization: SHS SCK KBS DWH. Validation: SHS SCK KBS DWH. Writing – original draft: SHS SCK KBS DWH. PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 References Comparative analysis of clinical out- comes for laparoscopic distal pancreatic resection and open distal pancreatic resection at a single insti- tution. Surg Endosc. 2008; 22(10): 2261–2268. doi: 10.1007/s00464-008-9973-1 PMID: 18528619 13. Kim SC, Song KB, Jung YS, Kim YH, Park DH, Lee SS, et al. Short-term clinical outcomes for 100 con- secutive cases of laparoscopic pylorus-preserving pancreatoduodenectomy: improvement with surgi- cal experience. Surg Endosc. 2013; 27(1): 95–103. doi: 10.1007/s00464-012-2427-9 PMID: 22752284 PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 11 / 13 Laparoscopic Distal Pancreatectomy for Cancer 14. Shin SH, Kim SC, Song KB, Hwang DW, Lee JH, Lee D, et al. A comparative study of laparoscopic vs open distal pancreatectomy for left-sided ductal adenocarcinoma: a propensity score-matched analy- sis. J Am Coll Surg. 2015; 220(2): 177–185. doi: 10.1016/j.jamcollsurg.2014.10.014 PMID: 25529901 15. Song KB, Kim SC, Hwang DW, Lee JH, Lee DJ, Lee JW, et al. Matched Case-Control Analysis Com- paring Laparoscopic and Open Pylorus-Preserving Pancreaticoduodenectomy in Patients With Peri- ampullary Tumors. Ann Surg. 2015; 262(1): 146–155. doi: 10.1097/SLA.0000000000001079 PMID: 25563866 16. Bassi C, Dervenis C, Butturini G, Fingerhut A, Yeo C, Izbicki J, et al. Postoperative pancreatic fistula: an international study group (ISGPF) definition. Surgery. 2005; 138(1): 8–13. PMID: 16003309 17. Dindo D, Demartines N, Clavien PA Classification of surgical complications: a new proposal with evalu- ation in a cohort of 6336 patients and results of a survey. Ann Surg. 2004; 240(2): 205–213. PMID: 15273542 18. Strasberg SM, Drebin JA, Linehan D Radical antegrade modular pancreatosplenectomy. Surgery. 2003; 133(5): 521–527. PMID: 12773980 19. Asbun HJ, Stauffer JA Laparoscopic vs open pancreaticoduodenectomy: Overall outcomes and sever- ity of complications using the accordion severity grading system. J Am Coll Surg. 2012; 215(6): 810– 819. doi: 10.1016/j.jamcollsurg.2012.08.006 PMID: 22999327 20. Correa-Gallego C, Dinkelspiel HE, Sulimanoff I, Fisher S, Viñuela EF, Kingham TP, et al. Minimally- invasive vs open pancreaticoduodenectomy: systematic review and meta-analysis. J Am Coll Surg. 2014; 218(1): 129–139. doi: 10.1016/j.jamcollsurg.2013.09.005 PMID: 24275074 21. Eom BW, Jang JY, Lee SE, Han HS, Yoon YS, Kim SW Clinical outcomes compared between laparo- scopic and open distal pancreatectomy. Surg Endosc. 2008; 22(5): 1334–1338. PMID: 18027035 22. Kawaguchi Y, Fuks D, Nomi T, Levard H, Gayet B Laparoscopic distal pancreatectomy employing radi- cal en bloc procedure for adenocarcinoma: Technical details and outcomes. Surgery. 2015; 157(6): 1106–1112. doi: 10.1016/j.surg.2014.12.015 PMID: 25704430 23. Kooby DA, Hawkins WG, Schmidt CM, Weber SM, Bentrem DJ, Gillespie TW, et al. PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 References A multicenter anal- ysis of distal pancreatectomy for adenocarcinoma: is laparoscopic resection appropriate? J Am Coll Surg. 2010; 210(5): 779–785. doi: 10.1016/j.jamcollsurg.2009.12.033 PMID: 20421049 24. Magge D, Gooding W, Choudry H, Steve J, Steel J, Zureikat A, et al. Comparative effectiveness of mini- mally invasive and open distal pancreatectomy for ductal adenocarcinoma. JAMA Surg. 2013; 148(6): 525–531. doi: 10.1001/jamasurg.2013.1673 PMID: 23426503 25. Rehman S, John SK, Lochan R, Jaques BC, Manas DM, Charnley RM, et al. Oncological feasibility of laparoscopic distal pancreatectomy for adenocarcinoma: a single-institution comparative study. World J Surg. 2014; 38(2): 476–483. doi: 10.1007/s00268-013-2268-2 PMID: 24081543 26. Sharpe SM, Talamonti MS, Wang E, Bentrem DJ, Roggin KK, Prinz RA, et al. The laparoscopic approach to distal pancreatectomy for ductal adenocarcinoma results in shorter lengths of stay without compromising oncologic outcomes. Am J Surg. 2015; 209(3): 557–563. doi: 10.1016/j.amjsurg.2014. 11.001 PMID: 25596756 27. Mitchem JB, Hamilton N, Gao F, Hawkins WG, Linehan DC, Strasberg SM Long-term results of resec- tion of adenocarcinoma of the body and tail of the pancreas using radical antegrade modular pancrea- tosplenectomy procedure. J Am Coll Surg. 2012; 214(1): 46–52. doi: 10.1016/j.jamcollsurg.2011.10. 008 PMID: 22192922 28. Paye F, Micelli Lupinacci R, Bachellier P, Boher JM, Delpero JR Distal pancreatectomy for pancreatic carcinoma in the era of multimodal treatment. Br J Surg. 2015; 102(3): 229–236. doi: 10.1002/bjs.9708 PMID: 25533421 29. Yamamoto J, Saiura A, Koga R, Seki M, Katori M, Kato Y, et al. Improved survival of left-sided pancreas cancer after surgery. Jpn J Clin Oncol. 2010; 40(6): 530–536. doi: 10.1093/jjco/hyq015 PMID: 20363769 30. Kooby DA, Chu CK Laparoscopic management of pancreatic malignancies. Surg Clin North Am. 2010; 90(2): 427–446. doi: 10.1016/j.suc.2009.12.011 PMID: 20362796 31. Tol JA, Gouma DJ, Bassi C, Dervenis C, Montorsi M, Adham M, et al. Definition of a standard lympha- denectomy in surgery for pancreatic ductal adenocarcinoma: a consensus statement by the Interna- tional Study Group on Pancreatic Surgery (ISGPS). Surgery. 2014; 156(3): 591–600. doi: 10.1016/j. surg.2014.06.016 PMID: 25061003 32. Ashfaq A, Pockaj BA, Gray RJ, Halfdanarson TR, Wasif N Nodal counts and lymph node ratio impact survival after distal pancreatectomy for pancreatic adenocarcinoma. J Gastrointest Surg. 2014; 18(11): 1929–1935. doi: 10.1007/s11605-014-2566-5 PMID: 24916590 12 / 13 PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 Laparoscopic Distal Pancreatectomy for Cancer 33. PLOS ONE \| DOI:10.1371/journal.pone.0163266 September 16, 2016 39. Neoptolemos JP, Stocken DD, Bassi C, Ghaneh P, Cunningham D, Goldstein D, et al. Adjuvant chemo- therapy with fluorouracil plus folinic acid vs gemcitabine following pancreatic cancer resection: a ran- domized controlled trial. JAMA. 2010; 304(10): 1073–1081. doi: 10.1001/jama.2010.1275 PMID: 20823433 References Latorre M, Ziparo V, Nigri G, Balducci G, Cavallini M, Ramacciato G Standard retrograde pancreatos- plenectomy versus radical antegrade modular pancreatosplenectomy for body and tail pancreatic ade- nocarcinoma. Am Surg. 2013; 79(11): 1154–1158. PMID: 24165249 34. Park HJ, Do You D, Choi DW, Heo JS, Choi SH Role of radical antegrade modular pancreatosplenect- omy for adenocarcinoma of the body and tail of the pancreas. World J Surg. 2014; 38(1): 186–193. doi: 10.1007/s00268-013-2254-8 PMID: 24166024 35. Biagi JJ, Raphael MJ, Mackillop WJ, Kong W, King WD, Booth CM Association between time to initia- tion of adjuvant chemotherapy and survival in colorectal cancer: a systematic review and meta-analy- sis. JAMA. 2011; 305(22): 2335–2342. doi: 10.1001/jama.2011.749 PMID: 21642686 36. Hofstetter G, Concin N, Braicu I, Chekerov R, Sehouli J, Cadron I, et al. The time interval from surgery to start of chemotherapy significantly impacts prognosis in patients with advanced serous ovarian carci- noma—analysis of patient data in the prospective OVCAD study. Gynecol Oncol. 2013; 131(1): 15–20. doi: 10.1016/j.ygyno.2013.07.086 PMID: 23877013 37. Valle JW, Palmer D, Jackson R, Cox T, Neoptolemos JP, Ghaneh P, et al. Optimal duration and timing of adjuvant chemotherapy after definitive surgery for ductal adenocarcinoma of the pancreas: ongoing lessons from the ESPAC-3 study. J Clin Oncol. 2014; 32(6): 504–512. doi: 10.1200/JCO.2013.50.7657 PMID: 24419109 38. Croome KP, Farnell MB, Que FG, Reid-Lombardo K, Truty MJ, Nagorney DM, et al. Total laparoscopic pancreaticoduodenectomy for pancreatic ductal adenocarcinoma: oncologic advantages over open approaches? Ann Surg. 2014; 260(4): 633–640. doi: 10.1097/SLA.0000000000000937 PMID: 25203880 39. Neoptolemos JP, Stocken DD, Bassi C, Ghaneh P, Cunningham D, Goldstein D, et al. Adjuvant chemo- therapy with fluorouracil plus folinic acid vs gemcitabine following pancreatic cancer resection: a ran- domized controlled trial. JAMA. 2010; 304(10): 1073–1081. doi: 10.1001/jama.2010.1275 PMID: 20823433 13 / 13
https://openalex.org/W4244915340	https://acp.copernicus.org/articles/19/1537/2019/acp-19-1537-2019.pdf	English	null	Chemical Characterisation of Water-soluble Ions in Atmospheric Particulate Matter on the East Coast of Peninsular Malaysia	null	2,018	cc-by	15,471	Atmos. Chem. Phys., 19, 1537–1553, 2019 https://doi.org/10.5194/acp-19-1537-2019 © Author(s) 2019. This work is distributed under the Creative Commons Attribution 4.0 License. Chemical characterisation of water-soluble ions in atmospheric particulate matter on the east coast of Peninsular Malaysia An atmospheric observation tower has been constructed on this coastline at the Bachok Marine Research Station. Daily PM2.5 samples were collected from the top of the observa- tion tower over a 3-week period, and ion chromatography was used to make time-resolved measurements of major at- mospheric ions present in aerosol. SO2− 4 was found to be the most dominant ion present and on average made up 66 % of the total ion content. Predictions of aerosol pH were made using the ISORROPIA II thermodynamic model, and it was estimated that the aerosol was highly acidic, with pH values ranging from −0.97 to 1.12. A clear difference in aerosol composition was found between continental air masses orig- inating from industrialised regions of East Asia and marine air masses predominantly inﬂuenced by the South China Sea. For example, elevated SO2− 4 concentrations and increased Cl−depletion were observed when continental air masses that had passed over highly industrialised regions of East Asia arrived at the measurement site. Correlation analyses of the ionic species and assessment of ratios between different Chemical characterisation of water-soluble ions in atmospheric particulate matter on the east coast of Peninsular Malaysia Naomi J. Farren1, Rachel E. Dunmore1, Mohammed Iqbal Mead2, Mohd Shahrul Mohd Nadzir3,4, Azizan Abu Samah5, Siew-Moi Phang5, Brian J. Bandy6, William T. Sturges6, and Jacqueline F. Hamilton1 1Wolfson Atmospheric Chemistry Laboratories, Department of Chemistry, University of York, York, YO10 5DD, UK 2Centre for Atmospheric Informatics and Emissions Technology, School of Energy, Environment and Agrifood/Environmental Technology, Cranﬁeld University, Cranﬁeld, UK 3Centre for Tropical Climate Change System (IKLIM), Institute of Climate Change, Universiti Kebangsaan Malaysia, 43600 Bangi, Selangor, Malaysia p gy, gy, nvironment and Agrifood/Environmental Technology, Cranﬁeld University, Cranﬁeld, UK 3Centre for Tropical Climate Change System (IKLIM), Institute of Climate Change, Universiti Ke 43600 Bangi, Selangor, Malaysia 4School of Environmental Science and Natural Resources, Faculty of Science and Technology, Universiti Kebangsaan Malaysia, 43600 Bangi, Selangor Darul Ehsan, Malaysia y Universiti Kebangsaan Malaysia, 43600 Bangi, Selangor Darul Ehsan, Malaysia 5 5Institute of Ocean and Earth Sciences, University of Malaya, Kuala Lumpur, Malaysia 6Centre for Ocean and Atmospheric Sciences, School of Environmental Sciences, Universi Correspondence: Jacqueline F. Hamilton (jacqui.hamilton@york.ac.uk) Received: 6 February 2018 – Discussion started: 24 April 2018 Received: 6 February 2018 – Discussion started: 24 April 2018 Revised: 13 January 2019 – Accepted: 18 January 2019 – Published: 6 February 2019 Revised: 13 January 2019 – Accepted: 18 January 2019 – Published: 6 February 2019 Abstract. Air quality on the east coast of Peninsular Malaysia is inﬂuenced by local anthropogenic and biogenic emissions as well as marine air masses from the South China Sea and aged emissions transported from highly pol- luted East Asian regions during the winter monsoon season. ions provided an insight into common sources and formation pathways of key atmospheric ions, such as SO2− 4 , NH+ 4 and C2O2− 4 . To our knowledge, time-resolved measurements of water-soluble ions in PM2.5 are virtually non-existent in ru- ral locations on the east coast of Peninsular Malaysia. Overall this dataset contributes towards a better understanding of at- mospheric composition in the Maritime Continent, a region of the tropics that is vulnerable to the effects of poor air qual- ity, largely as a result of rapid industrialisation in East Asia. Malaysia is inﬂuenced by local anthropogenic and biogenic emissions as well as marine air masses from the South China Sea and aged emissions transported from highly pol- luted East Asian regions during the winter monsoon season. 1 Introduction The tropical Maritime Continent, a region in Southeast Asia between 10◦S–20◦N and 90–150◦E, is a complex distri- bution of islands and peninsulas and incorporates countries such as Malaysia, Indonesia, the Philippines and Papua New Guinea (Neale and Slingo, 2003). It lies within a tropical warm pool that extends eastwards from the Indian Ocean to the Western Paciﬁc and is home to some of the warmest ocean temperatures in the world. Tropical regions such as the Maritime Continent are of central importance for the chemistry–climate system (Carpenter et al., 2010). For ex- ample, high photochemical activity in these regions means N. J. Farren et al.: Chemical characterisation of water-soluble ions Furthermore, the wind cir- culation system in the Maritime Continent is inﬂuenced by seasonal Asian monsoons, which are controlled by the natu- ral oscillation of the intertropical convergence zone (ITCZ). During the Northern Hemisphere winter, a large anticyclone forms over Siberia each year, creating strong northeasterly monsoon winds in the South China Sea (northeast monsoon). These strong northeasterlies can transport air masses from rapidly developing East Asian countries (e.g. China, Japan, Vietnam, and North and South Korea) across the South China Sea to the Maritime Continent (Zhang et al., 1997; Garreaud, 2001; Oram et al., 2017). In addition, cold-surge events oc- cur regularly throughout the winter monsoon season and last several days. Cold surges occur as a result of a southeast- erly movement of the anticyclone and are characterised by cold air masses over Southern China and strengthening of the northeasterly monsoon winds in the South China Sea (Zhang et al., 1997). The transport of pollution from East Asia to the tropics during the monsoon season, particularly during cold surges, means that rural areas such as the east coast of Peninsular Malaysia are potentially at an elevated risk of the detrimental effects of poor air quality. Figure 1. Location of the Bachok research station (BRT) and the global and regional GAW sites in the Maritime Continent: Danum Valley in Malaysia (DMV), Bukit Kototabang in Indonesia (BKT), Manila in Philippines (MNI), Songkhla in Thailand (SKH), Tanah Rata in Malaysia (TAR), Petaling Jaya in Malaysia (PJM) and Sin- gapore (SIN) (GAW-SIS, 2017). tre for Atmospheric Science (NCAS), UM and the Malaysian Meteorological Department (MMD). As part of this study, Dunmore et al. (2016) used a specialised multidimensional gas chromatography technique to accurately measure atmo- spheric mixing ratios of C5–C13 volatile organic compounds (VOCs) with a wide range of functionalities. Furthermore, Dominick et al. (2015) characterised the particulate matter in Bachok by studying the inﬂuence of northeasterly winds on the patterns of particle mass and particle number concentra- tion size distributions. Both studies highlighted the fact that the site is inﬂuenced by a mixture of local anthropogenic and biogenic emissions, clean marine air masses, and aged emis- sions transported from East Asia. p q y Tropical regions are highly important for atmospheric re- search, and whilst long-term atmospheric observations ex- ist (Robinson et al., 2014; Pyle et al., 2011), there are fewer measurements than in the mid and high latitudes. Published by Copernicus Publications on behalf of the European Geosciences Union. Published by Copernicus Publications on behalf of the European Geosciences Union. 1538 N. J. Farren et al.: Chemical characterisation of water-soluble ions Figure 1. Location of the Bachok research station (BRT) and the global and regional GAW sites in the Maritime Continent: Danum Valley in Malaysia (DMV), Bukit Kototabang in Indonesia (BKT), Manila in Philippines (MNI), Songkhla in Thailand (SKH), Tanah Rata in Malaysia (TAR), Petaling Jaya in Malaysia (PJM) and Sin- gapore (SIN) (GAW-SIS, 2017). that global atmospheric lifetimes of key atmospheric species, such as methane and ozone, are controlled by destruction rates in the tropics (Lawrence et al., 2001; Bloss et al., 2005). In terms of ocean productivity, the observed decrease in pri- mary productivity in low-latitude oceans has been linked to a reduced availability of nutrients for phytoplankton growth, caused by changes in upper-ocean temperature and stratiﬁ- cation (Behrenfeld et al., 2006). Furthermore, the wind cir- culation system in the Maritime Continent is inﬂuenced by seasonal Asian monsoons, which are controlled by the natu- ral oscillation of the intertropical convergence zone (ITCZ). During the Northern Hemisphere winter, a large anticyclone forms over Siberia each year, creating strong northeasterly monsoon winds in the South China Sea (northeast monsoon). These strong northeasterlies can transport air masses from rapidly developing East Asian countries (e.g. China, Japan, Vietnam, and North and South Korea) across the South China Sea to the Maritime Continent (Zhang et al., 1997; Garreaud, 2001; Oram et al., 2017). In addition, cold-surge events oc- cur regularly throughout the winter monsoon season and last several days. Cold surges occur as a result of a southeast- erly movement of the anticyclone and are characterised by cold air masses over Southern China and strengthening of the northeasterly monsoon winds in the South China Sea (Zhang et al., 1997). The transport of pollution from East Asia to the tropics during the monsoon season, particularly during cold surges, means that rural areas such as the east coast of Peninsular Malaysia are potentially at an elevated risk of the detrimental effects of poor air quality. that global atmospheric lifetimes of key atmospheric species, such as methane and ozone, are controlled by destruction rates in the tropics (Lawrence et al., 2001; Bloss et al., 2005). In terms of ocean productivity, the observed decrease in pri- mary productivity in low-latitude oceans has been linked to a reduced availability of nutrients for phytoplankton growth, caused by changes in upper-ocean temperature and stratiﬁ- cation (Behrenfeld et al., 2006). 2.2 Eluents and analytical standards Ultrapure Milli-Q water (18 M cm−1) from an ELGA Lab- Water puriﬁcation system was used to prepare all the re- quired eluents and analytical standards. A 20 mM solution of methanesulfonic acid was used as the eluent for cation ex- change chromatography, and for anion exchange chromatog- raphy, a solution of 8 mM Na2CO3 / 1 mM NaHCO3 was pre- pared. Using a variety of salts and organic acids, individ- ual analytical standards containing 500 ppm of each target ion (Cl−, NO− 2 , NO− 3 , PO3− 4 , SO2− 4 , CH3SO− 3 , C2O2− 4 , Na+, NH+ 4 , K+, Mg2+ and Ca2+) were prepared in Milli-Q wa- ter. The salts and organic acids were purchased from either Sigma-Aldrich Ltd. (Dorset, UK) or Fisher Scientiﬁc Ltd. (Loughborough, UK). In this study, measurements of water-soluble ions in atmo- spheric aerosol at a rural coastal location on the east coast of Peninsular Malaysia are presented. Analysis of temporal variation of different ionic species has been carried out and backward air mass trajectories have been used to determine the inﬂuence of air mass origin on aerosol composition. Cor- relation analyses of the ionic species and assessment of ratios between different ions have provided an insight into common sources and formation pathways of key atmospheric ions. This study contributes towards a better understanding of at- mospheric composition in the rural Bachok region, which is vulnerable to the effects of poor air quality, largely as a result of rapid industrialisation in East Asia. 2.3 Chromatographic analysis Chromatographic analysis was carried out using a Thermo Scientiﬁc Dionex ICS-1100 ion chromatography system equipped with an AS-DV autosampler. The column con- ﬁguration used for anion exchange consisted of an Ion- Pac AG14A guard column (4 mm × 50 mm) and an Ion- Pac AS14A analytical column (4 mm × 250 mm). Cation exchange chromatography was performed using an IonPac CG12A guard column (4 mm × 50 mm) and an IonPac CS12A analytical column (4 mm × 250 mm). ASRS 300 and CSRS 300 self-regenerating suppressors (4 mm) were used for anion and cation exchange respectively. All columns and suppressors were supplied from Thermo Scientiﬁc Dionex. The run times for the anion and cation separations were 18 and 15 min respectively. The suppressor current was 45 mA for anion exchange mode and 59 mA for cation exchange mode. For all separations, the instrument was operated in isocratic mode at a ﬂow rate of 1 mL min−1 and a column oven temperature of 30 ◦C. The injection volume was 100 µL and the data collection rate was 5 Hz. The system relied on a DS6 heated conductivity cell for ion detection, and all data were analysed using Thermo Scientiﬁc Chromeleon 7.1 Chromatography Data System software. N. J. Farren et al.: Chemical characterisation of water-soluble ions 1539 with a pore size of 0.22 µm (Millipore UK Limited, Watford, UK) and made up to a ﬁnal volume of 2.5 mL. Peninsular Malaysia, but the samples were collected at an ur- ban coastal city, Kuala Terengganu. This study used principal component analysis to determine the main sources of both ﬁne and coarse particles, which were found to be soil dust, marine aerosol, vehicle exhaust, secondary aerosol, road dust and biomass burning. In addition, Ismail et al. (2016) stud- ied PM10 concentrations in three major cities (Kota Bharu, Kuala Terengganu and Kuantan) on the east coast of Penin- sular Malaysia between 2006 and 2012. The study showed that during the northeast monsoon, the air arriving at the sites had originated from China and the Philippines and travelled over the South China Sea. During the southwest monsoon, the air came from Indonesia via the Strait of Malacca. Over the 6-year period, it was found that the atmospheric PM10 mass was directly proportional to the rate of urbanisation in each of the three cities. 2.1 Sample collection and extraction Thirty PM2.5 samples were collected at the Bachok Ma- rine Research Station (lat 6.00892, long. 102.42504) between 18 January and 7 February 2014. The samples were collected at the top of an atmospheric observation tower (18 m height) using a high-volume air sampler (Ecotech HiVol 3000, Victo- ria, Australia) operating at 1.13 m3 min−1 over 24 h sampling intervals. The tower is located on the coastline of the South China Sea and is within 100 m of the shore. A 3-day intensive measurement period was in operation between midday (local time) on 30 January 2014 and midday on 2 February 2014, in which ﬁlters were collected every 4–8 h. The quartz ﬁbre ﬁl- ters (20.3 cm × 25.4 cm) supplied by Whatman (Maidstone, UK) were pre-baked at 550 ◦C for a minimum of 12 h prior to sample collection. After sample collection, the ﬁlters were wrapped in aluminium foil and stored at −18 ◦C until anal- ysis. To prepare the samples for analysis, 5.7 cm2 of each sample was dissolved in 2 mL Milli-Q water and sonicated for 30 min at room temperature. The extract was ﬁltered us- ing a Millex-GP 33 mm diameter hydrophilic syringe ﬁlter N. J. Farren et al.: Chemical characterisation of water-soluble ions The Bachok Marine Research Station (lat 6.00892, long. 102.42504) has been set up on the east coast of Peninsular Malaysia and is ideally located for studying the outﬂow of these highly in- dustrialised regions and for investigating the interaction with cleaner air in the Southern Hemisphere. The research sta- tion, located approximately 30 km away from Kota Bharu, forms part of the Institute of Ocean and Earth Sciences at the University of Malaya (UM). An atmospheric observa- tion tower facing the South China Sea has been constructed at the research station; this has been built for the speciﬁc purpose of monitoring long-range transported pollution, air– sea exchange and coastal meteorology. The research station is working towards designation as a regional Global Atmo- spheric Watch (GAW) centre, which will be a valuable addi- tion to the network of other global and regional GAW sites in the Maritime Continent, as shown in Fig. 1 (GAW-SIS, 2017). To extend upon these studies, it is important to investigate atmospheric aerosol composition in the Bachok region. A better understanding of aerosol chemical composition is es- sential as aerosols play an important role in atmospheric pro- cesses and climate change. For example, aerosols can modify the global radiation budget both directly, by scattering and absorbing solar radiation, and indirectly, by altering cloud properties and lifetime (Charlson et al., 1991). The strength of these direct and indirect effects depends partly on the particle concentration and size distribution, but also on the chemical composition. There are a limited number of studies that focus on par- ticulate matter composition on the east coast of Peninsu- lar Malaysia, and to our knowledge the composition of ionic species has not been determined at any rural locations along this coastline. For example, Tahir et al. (2013) studied the composition of major elements and water-soluble ionic species in PM2.5 and PM10 samples on the east coast of In January and February 2014, an instrument demonstra- tion campaign was carried out to assess the capabilities of the new research station. This was funded by the Natural Environment Research Council (NERC) and UM and in- volved several UK universities, as well as the National Cen- Atmos. Chem. Phys., 19, 1537–1553, 2019 www.atmos-chem-phys.net/19/1537/2019/ 2.6 ISORROPIA II model Predictions of aerosol pH were made using the ISORROPIA- II thermodynamic equilibrium model. Although these mod- els produce better results when gas phase measurements such as NH3 and HNO3 are available, it is possible to use the ion measurements obtained in this study to make a prediction of aerosol pH (Fountoukis and Nenes, 2007). Calculations were made in reverse mode, in which known quantities are temper- ature, relative humidity, and particle phase concentrations of NH+ 4 , SO2− 4 , Na+, Cl−, NO− 3 , Ca2+, K+ and Mg2+. ISOR- ROPIA II assumes that the particles are internally mixed; this is a reasonable assumption for this study, as relative humid- ity was high (average = 77 %), and the aerosol arriving at the measurement site is often aged. The aerosol was assumed to be thermodynamically stable; i.e. the aerosol can exist as both solid and liquid, and salts are able to precipitate if the aqueous phase becomes saturated with respect to them. The ambient temperature and relative humidity data were taken from the measurements made nearby at the Sultan Ismail Pe- tra Airport. It is likely that these measurements are represen- tative of the Bachok research station, as further investigation of data from two other meteorological stations showed that temperature and relative humidity remain consistent along the coastline. N. J. Farren et al.: Chemical characterisation of water-soluble ions A receptor height of 10 m was chosen to represent the measurements made on the sampling tower. The trajectories were computed using the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model (Stein et al., 2015; Draxler, 1999; Draxler and Hess, 1998, 1997), and the data were analysed using the openair package in RStudio (Carslaw and Ropkins, 2012; Carslaw, 2015). N. J. Farren et al.: Chemical characterisation of water-soluble ions drophilic syringe ﬁlter with a pore size of 0.22 µm (Millipore UK Limited, Watford, UK) and made up to a ﬁnal volume of 2.5 mL. The recovery of the target ions from the ﬁlter pa- pers ranged from 74.5 % to 98.2 % for the target anions, and 78.3 % to 87.3 % for the target cations, with the exception of Ca2+ for which a recovery level of 123.3 % was calculated. Recovery tests were carried out in triplicate, and %RSDrec remained below 8 % for all the ions. The recovery of Ca2+ should not have exceeded 100 %, and the result may be at- tributed to inconsistencies in the amount of Ca2+ present on the blank ﬁlter, or due to Ca2+ contamination during sam- ple collection or storage. A 100 % recovery was assumed for Ca2+ during the data analysis process. Further details of the individual recovery levels and associated errors can be found in the Supplement (Table S1). Procedural blanks were also carried out using quartz ﬁbre ﬁlters (5.7 cm2), and blank sub- tractions were applied to any target ions found in detectable amounts. The blank peak areas for each ion and average blank contribution to ﬁeld samples over the entire sampling period are provided in the Supplement, Table S1. The main instrumental parameters of the IC system were evaluated and are also detailed in the Supplement (Table S2). Instrumen- tal limits of detection (LODs) and limits of quantiﬁcation (LOQs) were calculated according to the EPA protocol 40 CFR 136; multiplying the standard deviation (N = 10, 5 ng for cations, 25 ng for anions) by the t value (N = 10, 95 % conﬁdence interval) gave the LOD, and multiplying the stan- dard deviation by 10 gave the LOQ (EPA, 2017; Ripp, 1996). For anion exchange chromatography, LODs and LOQs were in the range 5.5–21.0 ng and 25.3–144.2 ng respectively. For cation exchange chromatography, the LODs ranged from 0.5 to 2.1 ng and the LOQs ranged from 2.5 to 6.1 ng. On aver- age, the instrument precision (%RSDins, n = 10) was 4.2 % for the target cations and 12.8 % for the target anions. Total errors were estimated by combining errors with the instru- ment and the recovery process and remained below 15.4 % for all ions except NO− 3 (22.6 %). throughout the entire measurement period. N. J. Farren et al.: Chemical characterisation of water-soluble ions 1540 drophilic syringe ﬁlter with a pore size of 0.22 µm (Millipore UK Limited, Watford, UK) and made up to a ﬁnal volume of 2.5 mL. The recovery of the target ions from the ﬁlter pa- pers ranged from 74.5 % to 98.2 % for the target anions, and 78.3 % to 87.3 % for the target cations, with the exception of Ca2+ for which a recovery level of 123.3 % was calculated. Recovery tests were carried out in triplicate, and %RSDrec remained below 8 % for all the ions. The recovery of Ca2+ should not have exceeded 100 %, and the result may be at- tributed to inconsistencies in the amount of Ca2+ present on the blank ﬁlter, or due to Ca2+ contamination during sam- ple collection or storage. A 100 % recovery was assumed for Ca2+ during the data analysis process. Further details of the individual recovery levels and associated errors can be found in the Supplement (Table S1). Procedural blanks were also carried out using quartz ﬁbre ﬁlters (5.7 cm2), and blank sub- tractions were applied to any target ions found in detectable amounts. The blank peak areas for each ion and average blank contribution to ﬁeld samples over the entire sampling period are provided in the Supplement, Table S1. The main instrumental parameters of the IC system were evaluated and are also detailed in the Supplement (Table S2). Instrumen- tal limits of detection (LODs) and limits of quantiﬁcation (LOQs) were calculated according to the EPA protocol 40 CFR 136; multiplying the standard deviation (N = 10, 5 ng for cations, 25 ng for anions) by the t value (N = 10, 95 % conﬁdence interval) gave the LOD, and multiplying the stan- dard deviation by 10 gave the LOQ (EPA, 2017; Ripp, 1996). For anion exchange chromatography, LODs and LOQs were in the range 5.5–21.0 ng and 25.3–144.2 ng respectively. For cation exchange chromatography, the LODs ranged from 0.5 to 2.1 ng and the LOQs ranged from 2.5 to 6.1 ng. On aver- age, the instrument precision (%RSDins, n = 10) was 4.2 % for the target cations and 12.8 % for the target anions. Total errors were estimated by combining errors with the instru- ment and the recovery process and remained below 15.4 % for all ions except NO− 3 (22.6 %). 2.4 Method validation Using isocratic elution methods for both cation and anion exchange chromatography, the target ions were successfully separated. Recovery tests were performed by spiking 5.7 cm2 of quartz ﬁbre ﬁlters (Whatman, Maidstone, UK) with 1 µg of each target ion (20 µL of a 50 ppm mixed ion solution). Prior to spiking, the ﬁlters were pre-baked at 550 ◦C for 6 h and wrapped in aluminium foil and stored at −18 ◦C until required. The spiked ﬁlters were dissolved in 2 mL of Milli- Q water and sonicated for 30 min at room temperature. The extract was ﬁltered using a Millex-GP 33 mm diameter hy- www.atmos-chem-phys.net/19/1537/2019/ Atmos. Chem. Phys., 19, 1537–1553, 2019 N. J. Farren et al.: Chemical characterisation of water-soluble ions N. J. Farren et al.: Chemical characterisation of water-soluble ions Figure 2. The 7-day HYSPLIT backward air mass trajectories centred on the Bachok Marine Research Station between 18 January and 7 February 2014. Each trajectory is coloured by the mean altitude (m). Trajectories with a mean height of 80 m or below are shown in panel (a) and trajectories with a mean height greater than 80 m are shown in panel (b). Figure 2. The 7-day HYSPLIT backward air mass trajectories centred on the Bachok Marine Research Station between 18 January and Figure 2. The 7-day HYSPLIT backward air mass trajectories centred on the Bachok Marine Research Station between 18 January and 7 February 2014. Each trajectory is coloured by the mean altitude (m). Trajectories with a mean height of 80 m or below are shown in panel (a) and trajectories with a mean height greater than 80 m are shown in panel (b). Figure 3. Time series of water-soluble ions (µg m−3) measured during the Bachok demonstration campaign (18 January to 7 February 2014). Figure 3. Time series of water-soluble ions (µg m−3) measured during the Bachok demonstration campaign (18 January to 7 February 2014). Figure 3. Time series of water-soluble ions (µg m−3) measured during the Bachok demonstration campaign ( Asia, Ashfold et al. (2017) used meteorological data from three locations that they believed to lie in the path of cold surges during the northeast monsoon, offering the best pos- sibility of observing a cold-surge inﬂuence on air pollution. These sites were in Kota Bharu (lat 6.141, long. 102.247), Kuala Terengganu (lat 5.308, long. 103.118) and Kemaman (lat 4.271, long. 103.428); this provides further conﬁrmation that data from the Sultan Ismail Petra Airport site are suit- able for evaluating broad-scale transport at the Bachok mea- surement site. Figure S2 in the Supplement shows average hourly wind speed and wind direction conditions across the entire duration of the measurement campaign (18 January to 6 February 2014). On most days, gentle southwesterlies from the land (Peninsular Malaysia) were observed in the early hours of the morning to around 11:00 LT (Malaysia time). At this stage, a dramatic shift in wind direction occurred, signiﬁcant continental inﬂuence from highly industrialised countries such as China, Japan, North and South Korea, and the island of Taiwan, whilst other air masses had a stronger marine inﬂuence from both the East China Sea and the South China Sea. N. J. Farren et al.: Chemical characterisation of water-soluble ions Although there was no reliable meteorological data recorded at the measurement site during the demonstration campaign, data from a nearby meteorological station were available. The station is located approximately 23 km away at the Sultan Ismail Petra Airport in Kota Bharu, as shown in Fig. S1 (Supplement). Whilst the meteorological data from the Airport will not be exactly representative of the mea- surement site, the patterns in wind direction are consistent with observations made by the ﬁeld scientists during the campaign. In addition, during a study of the inﬂuence of northeast monsoon cold surges on air quality in Southeast 2.5 Additional measurements The ﬁlter samples were collected at the Bachok atmospheric observation tower. The Bachok district, located in the state of Kelantan, is a rural area and the primary economic activ- ity comes from tobacco and kenaf plantations. Other agrarian activities in the wider Kelantan region include the produc- tion of rice and rubber as well as additional economic ac- tivities such as livestock rearing and ﬁshing. Figure 2 shows the 7-day backward air mass trajectories arriving at the mea- surement site during the demonstration campaign, with each trajectory coloured by the mean altitude (m). For compari- son with pollutant trends later in this study, the trajectories have been separated into two plots according to their mean height. As expected during the winter months, the strong an- ticyclone system known as the Siberian High led to the ar- rival of northeasterly onshore winds along the east coast of Peninsular Malaysia. Some of the air masses experienced a Individual VOCs were measured using a combined heart- cut and comprehensive two-dimensional gas chromatogra- phy system (GC −GC × GC); a detailed description of the instrument design is provided in a separate study (Dun- more et al., 2016). Measurements of NO and NO2 were per- formed using a two channel 42i-TL commercial gas analyser (Thermo Scientiﬁc, MA, USA), and SO2 measurements were made using a Thermo Scientiﬁc 43i SO2 analyser. Meteo- rological data from the nearest meteorological station, the Sultan Ismail Petra Airport (lat 6.17208, long. 102.29288), were accessed from the Integrated Surface Database (NOAA, 2003). Hourly measurements of wind direction, wind speed, air temperature, dew point, atmospheric pressure and rela- tive humidity were obtained. The 10-day backward air mass trajectories arriving at the sampling site were run every 3 h Atmos. Chem. Phys., 19, 1537–1553, 2019 www.atmos-chem-phys.net/19/1537/2019/ www.atmos-chem-phys.net/19/1537/2019/ 1541 N. J. Farren et al.: Chemical characterisation of water-soluble ions www.atmos-chem-phys.net/19/1537/2019/ Atmos. Chem. Phys., 19, 1537–1553, 2019 1542 N. J. Farren et al.: Chemical characterisation of water-soluble ions Table 1. Mean and maximum ion concentrations measured throughout the measurement period. The average percent mass contribution of each ion to the total measured ions is included as well as the percent of samples in which each target ion is found (%Qt). Ion Mean (ion) Maximum (ion) Mean percent mass of total %Qta %RSDb total µg m−3 µg m−3 measured ion content SO2− 4 10.7 20.8 65.6 100 11.2 NH+ 4 1.69 4.73 10.4 100 6.38 Na+ 1.13 2.60 6.95 100 6.88 Cl− 0.67 2.38 4.14 100 8.49 NO− 3 0.61 1.52 3.76 100 22.6 C2O2− 4 0.42 0.65 2.57 97 13.9 PO3− 4 0.36 2.34 2.22 93 15.4 K+ 0.38 0.67 0.67 100 6.35 Ca2+ 0.10 0.35 0.64 100 9.26 Mg2+ 0.10 0.21 0.61 100 6.72 CH3SO− 3 0.08 0.22 0.47 67 10.6 NO− 2 0.05 0.16 0.33 23 14.3 Total 16.2 27.0 – – – a Percentage of samples in which the target ion was above the LOQ. b Total error associated with each ion. 1542 N. J. Farren et al.: Chemical characterisation of water-soluble ions J Table 1. Mean and maximum ion concentrations measured throughout the measurement period. The average percent mass contribution of each ion to the total measured ions is included as well as the percent of samples in which each target ion is found (%Qt). Table 1. Mean and maximum ion concentrations measured throughout the measurement period. The average percent mass contribution of each ion to the total measured ions is included as well as the percent of samples in which each target ion is found (%Qt). Table 1. Mean and maximum ion concentrations measured throughout the measurement period. The average percent mass contribution of each ion to the total measured ions is included as well as the percent of samples in which each target ion is found (%Qt). 3.2 Composition of water-soluble ions in atmospheric aerosol (2) www.atmos-chem-phys.net/19/1537/2019/ crust (2) nssCa2+ = Ca2+ −ssCa2+ (3) ssCa2+ = ssNa+ · Ca2+ Na+ ! sea water (4) nssSO2− 4 = SO2− 4 −0.253 · ssNa+ (5) nssK+ = K+ −0.037 · ssNa+ (6) (1) www.atmos-chem-phys.net/19/1537/2019/ Ion Mean (ion) Maximum (ion) Mean percent mass of total %Qta %RSDb total µg m−3 µg m−3 measured ion content SO2− 4 10.7 20.8 65.6 100 11.2 NH+ 4 1.69 4.73 10.4 100 6.38 Na+ 1.13 2.60 6.95 100 6.88 Cl− 0.67 2.38 4.14 100 8.49 NO− 3 0.61 1.52 3.76 100 22.6 C2O2− 4 0.42 0.65 2.57 97 13.9 PO3− 4 0.36 2.34 2.22 93 15.4 K+ 0.38 0.67 0.67 100 6.35 Ca2+ 0.10 0.35 0.64 100 9.26 Mg2+ 0.10 0.21 0.61 100 6.72 CH3SO− 3 0.08 0.22 0.47 67 10.6 NO− 2 0.05 0.16 0.33 23 14.3 Total 16.2 27.0 – – – a Percentage of samples in which the target ion was above the LOQ. b Total error associated with each ion. in the crust and mean Ca2+/Na+ ratio in seawater have been estimated as 1.78 w/w and 0.038 w/w respectively (Bowen, 1979). Therefore it is possible to solve Eqs. (1)–(4) simulta- neously for ssNa+, nssNa+, ssCa2+ and nssCa2+ (Boreddy and Kawamura, 2015). Furthermore, the resulting estimate of ssNa+, which can be used as a sea spray marker, can also be used to predict the contribution of nssSO2− 4 and nssK+ in the aerosol, as shown in Eqs. (5) and (6) respec- tively. Whilst there may be some uncertainty in the mean Ca2+/Na+ crustal ratio, due to challenges associated with predicting the composition of the crust (Bowen, 1979), this approach provides more accuracy than simply using total Na+ as a sea spray marker. and until around 19:00 LT a strong onshore breeze from the northeast (South China Sea) was observed. From 20:00 LT to the early hours of each morning, a calmer sea breeze pre- dominantly from the east was seen. Hourly VOC measure- ments were conducted at the measurement site, and the de- velopment of a sea breeze at approximately 11:00 LT dramat- ically inﬂuenced the diurnal proﬁles of the measured species (Fig. S3, Supplement). In the morning, when the air being sampled was coming over the land, high levels of NO, NO2 and anthropogenic VOCs such as toluene and C10 aliphat- ics were observed; the main source of these species the was local burning of waste (Dunmore et al., 2016). When the sea breeze developed, the concentration of these species dropped signiﬁcantly. ssNa+ = Na+ −nssNa+ (1) nssNa+ = nssCa2+ · Ca2+ Na+ ! N. J. Farren et al.: Chemical characterisation of water-soluble ions Figure 4. Stacked bar chart to show the average mass composition of water-soluble ions in aerosol collected at the Bachok research station (a) and stacked bar charts to show the percentage of non-sea-salt and sea salt fractions of Ca2+, K+, Na+ and SO2− 4 (b, left to right). Figure 4. Stacked bar chart to show the average mass composition of water-soluble ions in aerosol collected at the Bachok research station (a) and stacked bar charts to show the percentage of non-sea-salt and sea salt fractions of Ca2+, K+, Na+ and SO2− 4 (b, left to right). Na+ and Cl−made up 11.1 % of the total ion content, and 72 % of the measured Na+ was attributed to ssNa+. The average concentrations of nssK+ and nssCa2+ were 0.30 and 0.09 µg m−3 respectively; these ions can be used as tracers for biomass burning (nssK+) and atmospheric dust (nssCa2+). NO− 2 and CH3SO− 3 were the least abundant ions, with average concentrations of 0.05 and 0.08 µg m−3. The two ions were only observed in a subset of samples; CH3SO− 3 was quantiﬁed in 67 % of samples and quantiﬁcation of NO− 2 was only achieved in 23 % of samples. The lower ratios recorded in Bachok suggest the majority of nssSO2− 4 at the site originates from anthropogenic sources. 4 As previously discussed, a study by Dunmore et al. (2016) revealed that levels of NOx and anthropogenic VOCs at the Bachok measurement site were signiﬁcantly higher when the air being sampled had passed over nearby land and dropped signiﬁcantly at around 11:00 LT when a sea breeze devel- oped. This indicated that air quality in Bachok is inﬂuenced by local sources of pollution, such as vehicle emissions and burning domestic waste. Pollution rose plots (Fig. 5) show the relationship between wind direction and speed and SO2 concentration at the Bachok measurement site. The left panel shows recorded SO2 concentrations ≥5 ppb and for compar- ison, the right panel shows all SO2 data recorded during the measurement period. SO2 concentrations < 5 ppb have been excluded from the left panel plot to further investigate the meteorological conditions when the spikes in SO2 concen- tration occur. In the lower right corner of each pollution rose, “mean” represents the mean SO2 concentration and “calm” represents the fraction of data that cannot be attributed to a speciﬁc wind direction. N. J. Farren et al.: Chemical characterisation of water-soluble ions During the elevated SO2 periods (≥5 ppb), weak southwesterlies dominated and the average wind speed was 1.1 m s−1. The mean SO2 concentration was 24.7 ppb during these events. Over the entire measurement campaign, however, average wind speed was considerably higher, 2.8 m s−1, and the air arrived predominantly from the east. In summary, the majority of higher SO2 events were ob- served in calmer conditions when the air arriving at the site had passed over nearby land; these observations provide fur- ther evidence for the inﬂuence of local sources of pollution. 3.2.1 Aerosol composition and determination of non-sea-salt and sea salt components The total concentration of measured water-soluble ions in the PM2.5 during the campaign ranged from 8.06 to 27.0 µg m−3, with an average concentration of 16.2 µg m−3. Table 1 shows the mean and maximum water-soluble ion concentrations measured throughout the campaign, and Fig. 3 shows time series for all the ions measured in the aerosol. (5) (6) (6) Figure 4 shows a series of stacked bar charts to summarise the average mass composition of water-soluble ions in at- mospheric aerosol and the distribution of non-sea-salt and sea salt components. The results show that the water-soluble ion fraction of the aerosol is dominated by SO2− 4 , which on average made up 65.6 % of the total ion content by mass. NH+ 4 and NO− 3 concentrations were signiﬁcantly lower, with mean concentrations of 1.69 and 0.61 µg m−3 respectively. As the composition of the water-soluble ions present in aerosol collected at the Bachok site was inﬂuenced by both marine and continental sources, it is useful to make an es- timation of non-sea-salt (nss) and sea salt (ss) components, using Eqs. (1)–(4). Total Na+ and Ca2+ concentrations have been measured in this study, and the mean Ca2+/Na+ ratio Atmos. Chem. Phys., 19, 1537–1553, 2019 www.atmos-chem-phys.net/19/1537/2019/ 1543 N. J. Farren et al.: Chemical characterisation of water-soluble ions 3.2.2 Sources and formation of sulfate (SO2− 4 ) The average SO2− 4 concentration during the measurement campaign was 10.7 µg m−3, with a maximum concentration of 20.8 µg m−3 recorded. The formation of SO2− 4 in the parti- cle phase occurs when emitted SO2 is oxidised by OH in the gas phase, or by O3 or H2O2 in the aqueous phase (Fisher et al., 2011). The most dominant anthropogenic sources of SO2 include fuel and industrial emissions, as well as open biomass burning. By using ssNa+ as a sea spray marker to determine non- sea-salt and sea salt components of the aerosol, it was found that on average 96 % of the measured SO2− 4 was nssSO2− 4 , and only 4 % of the SO2− 4 was from sea salt. As a potential biogenic source of nssSO2− 4 is DMS emissions from marine biota, and the main atmospheric source of MSA is the ox- idation of DMS, it is possible to use the MSA−/nssSO2− 4 ratio as a tracer to assess the contribution of biogenic sources to nssSO2− 4 in the atmosphere (Legrand and Pasteur, 1998). In this study, MSA−concentrations ranged from 0.02 to 0.22 µg m−3, with an average concentration of 0.08 µg m−3. As a result, the MSA−/nssSO2− 4 ratio ranged from 1.6 × 10−3 to 2.2 × 10−2, with an average value of 8.0 × 10−3. These values are low compared to MSA−/nssSO2− 4 ratios recorded at remote sites; for example, a MSA−/nssSO2− 4 mean mass ratio of 7 × 10−2 has been measured on Fanning Island and American Samoa (Savoie and Prospero, 1989). At the Bachok measurement site, no obvious relationship was observed between SO2 and particulate SO2− 4 concen- tration or between SO2− 4 concentration and wind direction. This is likely to be because it takes time for SO2 to oxidise to SO2− 4 and that the SO2− 4 fraction of the aerosol is more heavily inﬂuenced by long-range transport of aged emissions from East Asia. To investigate this further, the backward air mass trajectories were coloured by the concentration of N. J. Farren et al.: Chemical characterisation of water-soluble ions Figure 5. Pollution rose plots to show the relationship between wind direction and SO2 concentration at the Bachok measurement site. Panel (a) represents recorded SO2 concentrations ≥5 ppb and panel (b) represents all SO2 data recorded during the measurement period. Plot constructed using the openair package in RStudio (Carslaw and Ropkins, 2012; Carslaw, 2015). Figure 5. Pollution rose plots to show the relationship between win Panel (a) represents recorded SO2 concentrations ≥5 ppb and panel Figure 5. Pollution rose plots to show the relationship between wind direction and SO2 concentration at the Bachok measurement site. Panel (a) represents recorded SO2 concentrations ≥5 ppb and panel (b) represents all SO2 data recorded during the measurement period. Plot constructed using the openair package in RStudio (Carslaw and Ropkins, 2012; Carslaw, 2015). SO2− 4 , as shown in Fig. 6. To allow for comparison with Fig. 1, trajectories with a mean height of 80 m or below are shown on the left panel and trajectories with a mean height greater than 80 m are shown on the right panel. At the lower altitudes (≤80 m), it is clear that the SO2− 4 content of the aerosol is highest (ca. 15–20 µg m−3) when the site is inﬂu- enced by continental air masses from industrialised regions of East Asia and lower when the air masses have a more sig- niﬁcant marine inﬂuence. This trend is less clear at higher altitudes (right panel, mean height > 80 m), but nevertheless some of the lowest SO2− 4 levels are observed (∼5 µg m−3) when marine air that has passed over the East and South China seas arrives at the site. With this information available, it is useful to perform cluster analysis on the back trajecto- ries; this type of analysis groups air masses of similar geo- graphic origin together, which provides more information on pollutant species with similar chemical histories. A distance matrix is used to create a speciﬁed number of clusters with the most different air mass pathways. Figure 7 shows the ﬁve- cluster solution to back trajectories calculated for the Bachok site during the measurement campaign. ing altitude and increasing inﬂuence from industrialised re- gions. The mean altitude of air masses within cluster 5 was 1027 m, i.e. signiﬁcantly higher than air masses in clusters 1–4. N. J. Farren et al.: Chemical characterisation of water-soluble ions Furthermore, the air masses in cluster 5 arriving at the site were not a result of northeasterly monsoon winds from East Asian countries travelling across the South China Sea. Instead, there is evidence of a cyclonic weather system off the coast of the island of Borneo over the South China Sea. The air masses in cluster 5 arrived during the ﬁnal 24 h pe- riod of the measurement campaign, and only one SO2− 4 mea- surement is available (18.1 µg m−3). It is not possible to fully understand the nature of air masses from this different region without further measurements. Despite potential differences in atmospheric lifetimes and behaviour, other pollutants aside from SO2− 4 can provide fur- ther evidence that substantial amounts of industrial pollu- tion from East Asia are undergoing atmospheric transport to tropical regions of the Western Paciﬁc. For example, Oram et al. (2017) measured chlorine-containing, very short-lived substances (Cl-VSLSs) at the Bachok research station dur- ing the winter monsoon season in late January–early Febru- ary 2014. Cl-VSLSs are ozone-depleting species with short atmospheric lifetimes, typically less than 6 months. Species include dichloromethane (CH2Cl2) and 1,2-dichloroethane (CH2ClCH2Cl). A 7-day pollution or cold-surge event was reported between 19 and 26 January 2014, when signiﬁcantly enhanced concentrations of Cl-VSLSs were observed. Dur- ing this pollution episode, the measured samples were heav- ily impacted by emissions from the East Asian mainland, whilst this inﬂuence was less signiﬁcant during the cleaner, non-polluted periods. In fact, the total median concentra- tion of the four measured Cl-VSLSs was 546 ppt between 20 and 26 January and 243 ppt during the less polluted pe- riod (27 January to 5 February 2014). Oram et al. (2017) noted that even after the cold-surge event, the levels of Cl- VSLSs were still signiﬁcantly higher than expected, indicat- ing that this region of the South China Sea is widely im- Clusters 1 and 3 were associated with mean SO2− 4 con- centrations of 14.4 and 13.8 µg m−3 respectively. The mean altitude of cluster 1 was 72 m and the mean altitude of cluster 3 was 501 m. Full details of the air mass trajectories within each cluster are detailed in the Supplement (Table S3). These clusters contained air masses that had passed over several highly industrialised regions en route to Bachok, including cities such as Guangzhou (China) and Ho Chi Minh City (Vietnam). www.atmos-chem-phys.net/19/1537/2019/ www.atmos-chem-phys.net/19/1537/2019/ Atmos. Chem. Phys., 19, 1537–1553, 2019 1544 3.2.3 Correlation of SO2− 4 with NH+ 4 and implications for aerosol acidity Ammonium (NH+ 4 ) was the second most abundant ion in the aerosol; on average it made up 10.4 % of the total ion content, and the mean and maximum concentrations were 1.69 and 4.73 µg m−3 respectively. The strong positive correlation be- tween SO2− 4 and NH+ 4 was observed (R = 0.77, p < 0.001). A similar observation was reported by Keywood et al. (2003) during an investigation of the sources of particles contribut- ing to haze in the Klang Valley, Malaysia. The strong re- lationship between these species is due to neutralisation of SO2− 4 by NH+ 4 . Local sources of NH3 in the rural Bachok region are likely to come from agricultural practices such as animal husbandry, fertiliser use and agricultural waste burn- ing. The average NH+ 4 /SO2− 4 molar ratio was 0.81, which in- dicated that there was insufﬁcient gaseous NH3 in the atmo- sphere to neutralise SO2− 4 . Under an ammonia-poor regime, the uptake of SO2− 4 is preferential to the uptake of NO− 3 be- cause sulfuric acid has a lower vapour pressure than nitric acid (Seinfeld and Pandis, 2006). Although measurements of the total amounts of ammonia and sulfate in the gas, aque- ous and solid phases would provide a better prediction of the aerosol acidity, the results presented in this study indicated that an ammonia-poor regime exists and that the aerosol is likely to be acidic. In these scenarios, the NH3 partial pres- sure is low, and therefore the NH3-HNO3 partial pressure product is also low, meaning that the concentrations of am- monium nitrate are low or zero (Seinfeld and Pandis, 2006). This hypothesis can be supported by the fact that NO− 3 con- centrations in this study were very low, ranging from 0.005 to 1.52 µg m−3. As a result, NO− 3 made up a signiﬁcantly smaller fraction of the total ion content compared to SO2− 4 ; the average percentage mass composition of the total ion con- tent was 3.8 % for NO− 3 and 65.6 % for SO2− 4 . Figure 7. Five-cluster solution to backward air mass trajectories centred on the Bachok research station between 18 January and 7 February 2014. pacted by emissions from East Asia. N. J. Farren et al.: Chemical characterisation of water-soluble ions N. J. Farren et al.: Chemical characterisation of water-soluble ions Figure 6. The 7-day HYSPLIT backward air mass trajectories centred on the Bachok research station between 18 January and 7 Febru- ary 2014. The back trajectories are coloured by the concentration of SO2− 4 (µg m−3). Trajectories with a mean height of 80 m or below are shown in panel (a) and trajectories with a mean height greater than 80 m are shown in panel (b). Figure 6. The 7-day HYSPLIT backward air mass trajectories centred on the Bachok research station between 18 January and 7 Febru- ary 2014. The back trajectories are coloured by the concentration of SO2− 4 (µg m−3). Trajectories with a mean height of 80 m or below are shown in panel (a) and trajectories with a mean height greater than 80 m are shown in panel (b). Figure 7. Five-cluster solution to backward air mass trajectories centred on the Bachok research station between 18 January and 7 February 2014. N. J. Farren et al.: Chemical characterisation of water-soluble ions Average SO2− 4 concentrations for clusters 2 and 4 were 8.4 and 8.3 µg m−3, with mean altitudes of 169 and 37 m respectively. These air masses are likely to have ex- perienced some continental inﬂuence from the east coast of China and the island of Taiwan but have a much more sig- niﬁcant marine inﬂuence from passing over the South China Sea before arriving at Bachok. For air masses within clus- ters 1–4, the concentration of SO2− 4 increases with decreas- www.atmos-chem-phys.net/19/1537/2019/ Atmos. Chem. Phys., 19, 1537–1553, 2019 1545 N. J. Farren et al.: Chemical characterisation of water-soluble ions The positive strong acidity values provide an initial indica- tion that the aerosol is acidic and allows an estimate of the proton loading to be made (average H+ = 0.11 µg m−3). The ISORROPIA II thermodynamic model predictions of PM2.5 pH are shown in Fig. 9. Particle pH was estimated with ISORROPIA II run in the reverse mode without gas phase species input and ranged from −0.97 to 1.12 during the mea- surement period, implying that the aerosol was highly acidic. The pH prediction for the aerosol collected between midday on 3 and 4 February 2014 was 7.06 and has been excluded from Fig. 9. The particle concentrations input on this day correspond to negative values of strong acidity, and there- fore the model balances charge by assuming [OH−] > [H+]; this leads to a calculated pH of greater than 7. Murphy et al. (2017) reported that pH prediction is sensitive to strong acidity in the limit of strong acidity approaching zero and that the model can be drastically improved if gas phase NH3 and HNO3 measurements are included. The gas-to-particle parti- tioning of these species is sensitive to pH under conditions commonly encountered in the atmosphere; therefore gaseous NH3 and HNO3 measurements provide better constraint on the thermodynamic model. To obtain error bars for the strong acidity predictions (Fig. 8), H+ max and H+ min were calculated according to Eqs. (10) and (11) respectively. For H+ max the anions were ad- justed up to within their uncertainties (i.e. +%RSDtot), and the cations were adjusted down to within their uncertainties (i.e. −%RSDtot). For H+ min the anions were adjusted down and the cations were adjusted up (Murphy et al., 2017). The %RSDtot value was estimated by combining the error of the recovery process for each ion (%RSDrec, Table S1) and the error of the instrument for each ion (%RSDins, Table S2). H+ max = X max anion equivalents − X min cation equivalents (10) H+ min = X min anion equivalents − X max cation equivalents (11) (10) (11) Acidic particles can have detrimental effects on human health, air quality, and the health of aquatic and terrestrial ecosystems (Hennigan et al., 2015). For example, Gwynn et al. N. J. Farren et al.: Chemical characterisation of water-soluble ions To estimate proton loading in atmospheric particles, the strong acidity approach can be used, as shown in Eq. (7). This approach assumes that any deﬁcit in measured cation charge compared to measured anion charge can be attributed to H+. Total anion and total cation equivalents can be esti- mated using Eqs. (8) and (9). other atmospheric species such as HCO− 3 and basic amines, will lead to inaccuracies in the calculated H+. Thermodynamic equilibrium models such as ISOR- ROPIA II can also be used to predict aerosol pH (Fountoukis and Nenes, 2007). The ambient temperature and relative hu- midity data recorded nearby at the Sultan Ismail Petra Air- port were assumed to be representative of the Bachok re- search station, as further investigation of data from two other meteorological stations revealed that temperature and rela- tive humidity remained consistent. A map to show the loca- tion of the three meteorological stations along the east coast, as well as the Bachok research station, can be found in the Supplement (Fig. S1). The two other stations are Narathi- wat Airport (lat 6.520, long. 101.743) and Sultan Mahmud Airport (lat 5.383, long. 103.103). Average relative humid- ity between 18 January and 7 February 2014 was 77.7 % at Sultan Petra Ismail Airport, 75.4 % at Narathiwat Airport and 77.0 % at Sultan Mahmud Airport. Average temperatures recorded at the stations during this time were 24.8, 25.6 and 25.4 ◦C for Sultan Petra Ismail, Sultan Mahmud and Narathi- wat airports respectively. strong acidity (µeq.m−3) = X anion equivs. (µeq.m−3) − X cation equivs. (µeq.m−3) ( strong acidity (µeq.m−3) = X anion equivs. (µeq.m−3) g y (µ q ) X q (µ q ) − X cation equivs. (µeq.m−3) (7) X anion equivs. (µeq.m−3) = SO2− 4 48 + NO− 3 62 + Cl− 35.5 + PO3− 4 31.6 q + C2O2− 4 44 + NO− 2 46 + CH3SO− 3 95 (8) X cation equivs. (µeq.m−3) = Na+ 23 + NH+ 4 18 + K+ 39 − X cation equivs. (µeq.m−3) (7) (8) 23 18 39 + Mg2+ 12 + Ca2+ 20 (9) + Mg2+ 12 + Ca2+ 20 (9) As shown in Fig. 8, strong acidity values ranged from 0.03 to 0.19 µeq. m−3, with an average value of 0.11 µeq. m−3. 3.2.3 Correlation of SO2− 4 with NH+ 4 and implications for aerosol acidity Many other chemical pollutants, aside from short-lived chlorinated gases, will be present in these air masses from East Asia and will have a large impact on regional air quality. Oram et al. (2017) per- formed a Numerical Atmospheric dispersion Modelling En- vironment (NAME) trajectory analysis using carbon monox- ide (CO) as a tracer of industrial emissions from regions north of 20◦N for six winter seasons (2009–2010 to 2014– 2015). A strong correlation between CO and CH2Cl2 (a mea- sured Cl-VSLS) was observed during the pollution episode in late January 2014. Analysis of CO time series over the six winter seasons revealed that cold-surge events are likely to be repeated regularly each winter, demonstrating that pollu- tion rapidly undergoes long-range transport across the South China Sea on a regular basis during the northeast monsoon. Atmos. Chem. Phys., 19, 1537–1553, 2019 www.atmos-chem-phys.net/19/1537/2019/ www.atmos-chem-phys.net/19/1537/2019/ 1546 N. J. Farren et al.: Chemical characterisation of water-soluble ions N. J. Farren et al.: Chemical characterisation of water-soluble ions N. J. Farren et al.: Chemical characterisation of water-soluble ions 1547 Figure 8. Predictions of particle strong acidity for the aerosol collected during the Bachok measurement campaign. Dashed lines represent predictions of minimum and maximum limits of strong acidity. Figure 8. Predictions of particle strong acidity for the aerosol collected during the Bachok measurement campaign. Dashed lines represent predictions of minimum and maximum limits of strong acidity. Figure 8. Predictions of particle strong acidity for the aerosol collected during the Bachok measurement campaign. Dashed lines represent predictions of minimum and maximum limits of strong acidity. Figure 9. Predicted PM2.5 pH at the Bachok measurement site using ISORROPIA II (Fountoukis and Nenes, 2007). Dashed lines represent predictions of minimum and maximum limits of pH. Figure 9. Predicted PM2.5 pH at the Bachok measurement site using ISORROPIA II (Fountoukis and Nenes, 2007). Dashed lines represent predictions of minimum and maximum limits of pH. nate from local sources of pollution, it is possible that people living in these rural areas are also exposed to an additional risk, as the region appears to be sensitive to the effects of industrialisation further aﬁeld in East Asia. centrations in urban TSPs (total suspended particles) include measurements of 0.10–0.48 µg m−3 in Shanghai (Jiang et al., 2011) and 0.27 µg m−3 in Tokyo (Sempere and Kawamura, 1994). To investigate possible oxalate sources and formation pathways, it is necessary to consider the correlation of ox- alate with different atmospheric species. Jiang et al. (2011) report using NO− 2 as an indicator for vehicle emissions, nssSO2− 4 and NO− 3 for secondary formation through differ- ent pathways, and K+ for biomass burning. A study carried out by Huang et al. (2006) in the urban area of Shenzhen (Southern China) reported that, whilst good correlation of droplet oxalate with K+ was observed (R2 = 0.75, average diameter: 1.0 µm), there was poor correlation between ox- alate and K+ in the condensation mode (R2 = 0.10, average diameter: 0.4 µm). This implied that whilst biomass burning was probably not an important primary source of conden- sation mode oxalate, it is likely that biomass burning parti- cles act as effective CCN, promoting in-cloud sulfate and ox- alate formation. N. J. Farren et al.: Chemical characterisation of water-soluble ions In this study, nssK+ levels remained below 0.7 µg m−3; this likely represents background biomass burn- ing emissions over the region during the measurement pe- riod, as evidenced by ﬁre hotspots from satellite data shown in Fig. S4 (Giglio et al., 2003). A strong correlation was ob- N. J. Farren et al.: Chemical characterisation of water-soluble ions (2000) performed a time-series analysis of acidic PM and daily mortality and morbidity in the Buffalo, New York, re- gion; several signiﬁcant pollutant–health effect associations were identiﬁed, the strongest being the correlation between atmospheric SO2− 4 concentration and respiratory hospital ad- missions. Furthermore, particle acidity can inﬂuence vari- ous atmospheric chemical processes, including SO2 oxida- tion, halogen chemistry, and the partitioning of ammonia, ni- tric acid, organic acids and isomeric epoxy diols from iso- prene photooxidation (IEPOX) (Hennigan et al., 2015; Sur- ratt et al., 2010; Lin et al., 2012). In summary, whilst some of the risks associated with aerosol acidity in Bachok origi- In most cases, H+ min remains above zero. However, between 2 and 6 February 2014, slightly negative H+ min values between −1 × 10−3 and −2 × 10−2 µeq. m−3 were calculated, which are physically implausible (Murphy et al., 2017). These re- sults highlight the possible sources of error associated with the strong acidity approach for estimating aerosol acidity. For example, Hennigan et al. (2015) report that organic acids (which are mostly excluded from this study, except for MSA and oxalic acid) can have an important inﬂuence on aerosol acidity, especially at relatively low acidities where organic acids dissociate and contribute to the ion balance. Further- more, they can form salt complexes with inorganic species e.g. ammonium oxalate. Neglecting organic acids, as well as In most cases, H+ min remains above zero. However, between 2 and 6 February 2014, slightly negative H+ min values between −1 × 10−3 and −2 × 10−2 µeq. m−3 were calculated, which are physically implausible (Murphy et al., 2017). These re- sults highlight the possible sources of error associated with the strong acidity approach for estimating aerosol acidity. For example, Hennigan et al. (2015) report that organic acids (which are mostly excluded from this study, except for MSA and oxalic acid) can have an important inﬂuence on aerosol acidity, especially at relatively low acidities where organic acids dissociate and contribute to the ion balance. Further- more, they can form salt complexes with inorganic species e.g. ammonium oxalate. Neglecting organic acids, as well as www.atmos-chem-phys.net/19/1537/2019/ Atmos. Chem. Phys., 19, 1537–1553, 2019 N. J. Farren et al.: Chemical characterisation of water-soluble ions 1547 3.2.5 Sea salt aerosol and factors affecting chloride depletion There was no correlation between nssNa+ and Cl−(R = −0.01), but a strong positive correlation between ssNa+ and Cl−was observed (R = 0.84, p < 0.001). During the mea- surement period, the Cl−/ssNa+ molar ratio ranged from 0.003 to 1.10 with an average value of 0.40. A time series of Cl−/ssNa+ molar ratio can be found in the Supplement (Fig. S5). Signiﬁcant Cl−depletion was observed, as all of the ratios recorded were lower than that of bulk seawater, 1.18 (Boreddy and Kawamura, 2015). Figure 10 shows back- ward air mass trajectories arriving at the Bachok research sta- tion, coloured by the Cl−/ssNa+ molar ratio. There was no signiﬁcant correlation observed between ox- alate and NO− 3 , or between oxalate and NO− 2 . The lack of correlation with NO− 3 suggests that the two species do not have similar formation pathways and that vehicle emissions are not an important secondary source of oxalate. It is also unlikely that vehicle emissions contribute to the primary sources of oxalate, due to the lack of correlation between ox- alate and NO− 2 . However, NO− 2 was only detected in 7 out of the 30 samples collected, so it is difﬁcult to ascertain whether there is a relationship between these two species or not. A strong correlation between oxalate and nssSO2− 4 (R = 0.69, p < 0.001) was observed, suggesting a common formation pathway of the two species. It is well known that SO2− 4 forms via aqueous oxidation (Seinfeld and Pandis, 2006), and modelling studies also suggest that aqueous chemistry is a large contributor of oxalate formation globally (Myrioke- falitakis et al., 2011). Furthermore, Carlton et al. (2006) re- port that whilst there are likely to be many sources of ox- alate, oxidation of pyruvate in the aqueous phase is known to form oxalate at dilute (cloud-relevant) concentrations. Tan et al. (2012) also state that aqueous acetate oxidation is a key source of oxalate. A positive correlation between ox- alate and NH+ 4 was also observed (R = 0.73, p < 0.001). A similar observation was reported by Jiang et al. (2011) in a study of aerosol oxalate in Shanghai. Using size distribution data, they were able to propose that the correlation was due to the presence of ammonium oxalate in the aerosol. 3.2.5 Sea salt aerosol and factors affecting chloride depletion served between oxalate and nssK+ (R = 0.78, p < 0.001), and the oxalate / nssK+ ratio ranged from 0.68 to 4.64 dur- ing the Bachok measurement campaign. These ratios are sig- niﬁcantly higher than those found in aerosol directly emit- ted from vegetation ﬁres in the Amazon Basin (Yamasoe et al., 2000), suggesting that biomass burning in the wider re- gion inﬂuenced the secondary formation of oxalate, rather than acting as a primary source. A similar hypothesis was proposed by Huang and Yu (2007), who measured ambient PM2.5 in an urban environment in the Pearl River Delta re- gion of China, and reported oxalate / K+ ratios of 0.57 and 0.33 in summer and winter respectively. It is worth noting that whilst the higher ratios reported in both studies indi- cate that biomass burning is not a major primary source of oxalate, measurements of oxalate and K+ in local biomass burning aerosols (rather than aerosol in the Amazon Basin) would provide a better indication of the source contribution by biomass burning. served between oxalate and nssK+ (R = 0.78, p < 0.001), and the oxalate / nssK+ ratio ranged from 0.68 to 4.64 dur- ing the Bachok measurement campaign. These ratios are sig- niﬁcantly higher than those found in aerosol directly emit- ted from vegetation ﬁres in the Amazon Basin (Yamasoe et al., 2000), suggesting that biomass burning in the wider re- gion inﬂuenced the secondary formation of oxalate, rather than acting as a primary source. A similar hypothesis was proposed by Huang and Yu (2007), who measured ambient PM2.5 in an urban environment in the Pearl River Delta re- gion of China, and reported oxalate / K+ ratios of 0.57 and 0.33 in summer and winter respectively. It is worth noting that whilst the higher ratios reported in both studies indi- cate that biomass burning is not a major primary source of oxalate, measurements of oxalate and K+ in local biomass burning aerosols (rather than aerosol in the Amazon Basin) would provide a better indication of the source contribution by biomass burning. On average, Na+ and Cl−contributed 7.0 % and 4.1 % to the total measured water-soluble ion content respectively, and 72 % of the measured Na+ was attributed to ssNa+. The con- centration of ssNa+ ranged from 0.24 to 2.35 µg m−3, whilst the concentration of Cl−ranged from 0.003 to 2.38 µg m−3. 3.2.5 Sea salt aerosol and factors affecting chloride depletion In this study there are no size distribution data available, and so it is important to consider the fact that the correlation may be linked to the inﬂuence of sulfate on both NH+ 4 and oxalate in aerosol; NH+ 4 partitions to the aerosol from gaseous NH3 in an attempt to neutralise acidic sulfate particles, whilst oxalate exhibits similar formation pathways to SO2− 4 . , y / A study of marine aerosol at the remote Chichi-jima is- land in the western North Paciﬁc during 2001 and 2002 re- ported that mean Cl−/Na+ molar ratios were highest (1.34) in September 2001 and lowest (0.30) in May 2002, and the mean ratio across the 2-year period was 1.10 (Boreddy et al., 2014). Boreddy et al. (2014) reported that the observed chlo- ride depletion was likely due to acid displacement occurring as a result of atmospheric mixing of anthropogenic pollutants such as SOx and NOx. Acid displacement can occur when sea salt particles react with acids such as H2SO4, HNO3, ox- alic acid (C2H2O4) and methanesulfonic acid (CH3SO3H) in the atmosphere. Such processes are of atmospheric impor- tance as they lead to the formation of gaseous HCl and po- tentially affect acid deposition conditions in the region. It is widely accepted that Cl−depletion through the volatilisation of HCl via acid displacement occurs particularly in polluted marine air masses (Newberg et al., 2005; Sturges and Shaw, 1993). In this study, the lowest Cl−/ssNa+ ratios were ob- served when air masses arriving at the site had previously passed over highly industrialised countries such as China and Vietnam (Fig. 10). Higher Cl−/ssNa+ ratios are found in ma- rine air masses from less polluted regions. It is possible that aerosol, particularly in northern China, may be rich in chlo- ride in the winter due to coal burning. However, this study suggests that the inﬂuence of other anthropogenic pollutants, such as HNO3 and H2SO4, means that as the air is trans- ported over the South China Sea to the Bachok region, sig- niﬁcant Cl−depletion can occur prior to arrival at the mea- surement site. In a more recent study, Boreddy and Kawa- mura performed regression analysis between the Cl−/Na+ mass ratio and various acidic species, including nssSO2− 4 , NO− 3 , MSA−and oxalic acid. Atmos. Chem. Phys., 19, 1537–1553, 2019 3.2.4 Sources and formation of oxalate, C2O2− 4 Oxalic acid is the most abundant dicarboxylic acid in tro- pospheric aerosol (Sareen et al., 2016). This major water- soluble organic component can alter the hygroscopicity of aerosols and can either act as cloud condensation nuclei (CCN), or reduce the surface tension of particles to form CCN (Saxena and Hildemann, 1996; Novakov and Penner, 1993; Facchini et al., 1999; Kerminen, 2001). In this study, oxalate made up 2.6 % of the total measured water-soluble ion content. The average concentration was 0.42 µg m−3, and throughout the measurement period the concentration ranged from 0.15 to 0.65 µg m−3. Interestingly, such levels of ox- alate in atmospheric aerosol are typical of urban environ- ments, despite the fact that the Bachok research station is located in a rural coastal region. For example, Freitas et al. (2012) report average oxalate concentrations in PM10 at an urban site and a rural site in Londrina city, Brazil, of 0.57 and 0.03 µg m−3 respectively. Other reported oxalate con- www.atmos-chem-phys.net/19/1537/2019/ Atmos. Chem. Phys., 19, 1537–1553, 2019 1548 N. J. Farren et al.: Chemical characterisation of water-soluble ions N. J. Farren et al.: Chemical characterisation of water-soluble ions 1549 Figure 10. The 7-day HYSPLIT back trajectories centred on the Bachok research station, between 18 January and 7 February 2014. The back trajectories are coloured by the Cl−/ssNa+ molar ratio. Trajectories with a mean height of 80 m or below are shown in panel (a) and trajectories with a mean height greater than 80 m are shown in panel (b). Figure 10. The 7-day HYSPLIT back trajectories centred on the Bachok research station, between 18 January and 7 February 2014. The back trajectories are coloured by the Cl−/ssNa+ molar ratio. Trajectories with a mean height of 80 m or below are shown in panel (a) and trajectories with a mean height greater than 80 m are shown in panel (b). measurement period was 0.34 µg m−3, but during this 12 h episode the PO3− 4 concentration increased to 1.88 µg m−3. The increase in ion concentration was less pronounced for nssNa+ and Mg2+, but the concentrations were still 0.26 and 0.88 µg m−3 above the average for the whole measurement period respectively. Time-series plots for nssCa2+, PO3− 4 , Mg2+ and nssNa+ can be found in the Supplement (Fig. S6). correlated with the Cl−/Na+ mass ratio in the summer, there was a signiﬁcant negative correlation between oxalic acid and the mass ratio in the other three seasons, providing con- ﬁrmation that oxalic acid plays an important role in chloride loss at Chichi-jima island. A similar regression analysis was carried out on the measurements obtained during the Bachok measurement campaign, and whilst a weak negative corre- lation was found between SO2− 4 and the Cl−/ssNa+ mass ratio (R = −0.41, p < 0.001), there was no correlation of the Cl−/ssNa+ mass ratio with oxalate or MSA−. These re- sults imply that whilst H2SO4 played an important role in chloride depletion, methanesulfonic acid and oxalic acid may not have. Interestingly, a strong positive correlation was ob- served between NO− 3 and Cl−/ssNa+ mass ratio (R = 0.82, p < 0.001). These measurements may be linked to each other through the important role of H2SO4 in the atmosphere. Acid displacement, when sea salt reacts with H2SO4, leads to the removal of Cl−from the aerosol as gaseous HCl, and a par- titioning of SO2− 4 to the aerosol as Na2SO4. N. J. Farren et al.: Chemical characterisation of water-soluble ions Furthermore, under an ammonia-poor regime (as observed in this study), H2SO4 has a lower vapour pressure than HNO3, leading to the preferential formation of ammonium sulfate over ammo- nium nitrate when there is insufﬁcient NH3 available to fully neutralise sulfate and nitrate (Seinfeld and Pandis, 2006). Between midday on 30 January 2014 and midnight on 31 January 2014, the contribution of nssCa2+ to the to- tal measured water-soluble ion content was more signiﬁcant (1.26 %) than the average contribution during the remain- der of the measurement period (0.35 %). The Bachok re- search station is located in the outﬂow region of Asian dusts, and these measurements suggest that long-range transport of Asian dusts over the measurement site has occurred during this time. In fact, during this episode, the back trajectories ar- riving at the site can be traced back to the North China Plains and the Horqin Desert in eastern China; source attribution studies by Ginoux et al. (2012) have revealed that large an- thropogenic dust sources are found in these regions. In the future, it is important that longer term measurements are car- ried out at the Bachok research station to provide conﬁrma- tion that dust episodes in these regions of Asia are respon- sible for the elevated levels of nssCa2+ and other associated ions. Dust is one of the most abundant types of aerosol in the atmosphere and can have important impacts on both air qual- ity and climate; therefore it is important that seasonal and annual trends in water-soluble ions are studied in more detail at the Bachok research station. 3.2.6 Using nssCa2+ as a potential tracer for dust episodes NssCa2+ can be used as a tracer for atmospheric dust (Boreddy and Kawamura, 2015). The average concentra- tion of nssCa2+ during the measurement campaign was 0.07 µg m−3, and the maximum concentration recorded was 0.28 µg m−3. On most days, the nssCa2+ concentration was below 0.10 µg m−3, but between midday on 30 January 2014 and midnight on 31 January 2014 (local time), elevated nssCa2+ levels were observed, and the average concentration during this period was 0.22 µg m−3. The same trend was ob- served for PO3− 4 ; the average concentration across the entire 4 Conclusions An accurate and reliable technique relying on ion chromatog- raphy has been used to make time-resolved measurements of water-soluble ions in atmospheric aerosol at the Bachok Ma- rine Research Station. Using meteorological data from the nearby airport and HYSPLIT backward air mass trajectories www.atmos-chem-phys.net/19/1537/2019/ 3.2.5 Sea salt aerosol and factors affecting chloride depletion They found a moderate neg- ative correlation between Cl−/Na+ mass ratio and nssSO2− 4 , and a moderate to weak negative correlation with NO− 3 ; this suggested that sulfate had a higher inﬂuence on chloride de- pletion than nitrate. Furthermore, whilst MSA−moderately Atmos. Chem. Phys., 19, 1537–1553, 2019 www.atmos-chem-phys.net/19/1537/2019/ 1549 N. J. Farren et al.: Chemical characterisation of water-soluble ions N. J. Farren et al.: Chemical characterisation of water-soluble ions 1550 centred on the Bachok research station, it was possible to ob- serve the diurnal wind pattern behaviour and assess where the air masses arriving at the site originated from. Air quality at this remote location is inﬂuenced by local anthropogenic and biogenic emissions, as well as marine air masses from the South China Sea and aged emissions transported from highly polluted East Asian regions during the winter mon- soon season. In general, the site was inﬂuenced by south- westerlies coming from the land from the early hours of the morning until approximately 11:00 LT, and then a dramatic shift in wind direction occurred and a sea breeze was present for the remainder of the day. This shift was accompanied by a drop in the concentrations of NOx and anthropogenic VOCs (Dunmore et al., 2016). a result of air masses arriving at the site from the North China Plains and Horqin Desert. Longer term measurements are re- quired to fully investigate the inﬂuence of Asian dusts at this remote coastal location. To our knowledge, time-resolved measurements of water- soluble ions in PM2.5 are virtually non-existent in rural lo- cations on the east coast of Peninsular Malaysia. The data presented in this study have demonstrated the capabilities of the new atmospheric tower at the Bachok research sta- tion and has provided an initial insight into factors affecting aerosol composition on this coastline. In the future, it is im- portant that longer term measurements are carried out, with increased time-resolved sampling and particle size fractiona- tion, to provide a better understanding of the factors affecting aerosol composition at this measurement site. This remote lo- cation is susceptible to the effects of local, regional and inter- national air pollution, and rapid industrialisation in East Asia is inﬂuencing air quality along the east coast of Peninsular Malaysia. Twelve atmospheric water-soluble ions were measured in this study, and SO2− 4 was found to be the most dominant ion present, making up 66 % of the total measured ion content on average. The non-sea-salt and sea salt components of SO2− 4 , Na+, K+ and Ca2+ were determined, and it was found that 96 % of the measured SO2− 4 was non-sea-salt SO2− 4 . N. J. Farren et al.: Chemical characterisation of water-soluble ions Predic- tions of aerosol pH were made using the ISORROPIA II ther- modynamic model and it was estimated that the aerosol was highly acidic, with pH values ranging from −0.97 to 1.12; such levels of acidity are likely to have a detrimental impact on human health and the health of ecosystems at this remote coastal location. A clear difference in aerosol composition was found between continental air masses originating from industrialised regions of East Asia and marine air masses pre- dominantly inﬂuenced by the South China Sea. For example, elevated SO2− 4 concentrations were observed when continen- tal air masses that had passed over highly industrialised re- gions of East Asia arrived at the measurement site. Data availability. Raw data are available on PURE (https://doi.org/10.15124/bd4a9045-832b-4ff8-aecc-ef1653603f1d; Farren et al., 2018). Supplement. The supplement related to this article is available online at: https://doi.org/10.5194/acp-19-1537-2019-supplement. Supplement. The supplement related to this article is available online at: https://doi.org/10.5194/acp-19-1537-2019-supplement. Author contributions. All authors contributed to the ﬁnal version of this article. NJF analysed the aerosol samples and wrote the paper under the supervision of JFH. RED collected the aerosol samples, and BJB made the SO2 measurements. MIM, MSMN, AAS and SMP coordinate and manage the University of Malaya BMRS. WTS coordinated the Bachok demonstration “International Opportunities Fund” campaign. g Correlation analyses amongst ionic species and assess- ment of ratios between different ions provided an insight into common sources and formation pathways of key at- mospheric ions. Oxalate concentrations were recorded and found to be more comparable to measurements made at ur- ban locations rather than rural ones. A strong correlation of C2O2− 4 with SO2− 4 suggested a common aqueous oxidation formation pathway. A strong correlation between C2O2− 4 and K+ coupled with high C2O2− 4 /nssK+ ratios indicated that biomass burning was an important secondary source of ox- alate in the Bachok region, whereas a lack of correlation with NO− 2 and NO− 3 suggested that vehicular emissions were not an important source. The average Cl−/ssNa+ molar ra- tio during the measurement campaign was 0.40, signiﬁcantly lower than that of bulk seawater (1.18). Analysis of back tra- jectories revealed that chloride depletion was greater when the aerosol was more inﬂuenced by anthropogenic sources of pollution. Elevated levels of nssCa2+ and other ions such as PO3− 4 , Mg2+ and nssNa+ were observed between midday on 30 January 2014 and midnight on 31 January 2014. www.atmos-chem-phys.net/19/1537/2019/ Atmos. Chem. Phys., 19, 1537–1553, 2019 N. J. Farren et al.: Chemical characterisation of water-soluble ions References S.: Hygroscopic properties of particles nebulized from water extracts of aerosols collected at Chichijima Island in the western North Paciﬁc: An outﬂow region of Asian dust, J. Geophys. Res.-Atmos., 119, 167–178, https://doi.org/10.1002/2013JD020626, 2014. EPA: 40 Protection of Environment Part 136 Guidelines Establish- ing Test Procedures for the Analysis of Pollutants, in: Electronic Code of Federal Regulations, U.S. Government Publishing Of- ﬁce, Washington, DC, available at: https://www.ecfr.gov/cgi-bin/ text-idx?SID=57d5771415f8a36cb8c0e2ca2a1583a2&mc= true&node=pt40.25.136&rgn=div5#ap40.25.136_17.b, last access: 20 July 2017. Bowen, H. J. M.: Environmental chemistry of the elements, Aca- demic Press, London, 1979. Carlton, A. G., Turpin, B. J., Lim, H. J., Altieri, K. E., and Seitzinger, S.: Link between isoprene and secondary organic aerosol (SOA): Pyruvic acid oxidation yields low volatility organic acids in clouds, Geophys. Res. Lett., 33, L06822, https://doi.org/10.1029/2005gl025374, 2006. Facchini, M. C., Mircea, M., Fuzzi, S., and Charlson, R. J.: Cloud albedo enhancement by surface-active organic solutes in growing droplets, Nature, 401, 257–259, https://doi.org/10.1038/45758, 1999. Farren, N. J., Dunmore, R. E., Mead, M. I., Mohd Nadzir, M. S., Samah, A. A., Phang, S.-M., Bandy, B. J., Sturges, W. T., and Hamilton, J. F.: Chemical Characterisation of Water-soluble Ions in Atmospheric Particulate Matter on the East Coast of Peninsular Malaysia, https://doi.org/10.15124/bd4a9045-832b- 4ff8-aecc-ef1653603f1d, 2018. Carpenter, L. J., Fleming, Z. L., Read, K. A., Lee, J. D., Moller, S. J., Hopkins, J. R., Purvis, R. M., Lewis, A. C., Muller, K., Heinold, B., Herrmann, H., Fomba, K. W., van Pinxteren, D., Muller, C., Tegen, I., Wiedensohler, A., Muller, T., Niedermeier, N., Achterberg, E. P., Patey, M. D., Kozlova, E. A., Heimann, M., Heard, D. E., Plane, J. M. C., Mahajan, A., Oetjen, H., Ingham, T., Stone, D., Whalley, L. K., Evans, M. J., Pilling, M. J., Leigh, R. J., Monks, P. S., Karunaharan, A., Vaughan, S., Arnold, S. R., Tschritter, J., Pohler, D., Friess, U., Holla, R., Mendes, L. M., Lopez, H., Faria, B., Manning, A. J., and Wallace, D. W. R.: Seasonal characteristics of tropical marine boundary layer air measured at the Cape Verde Atmospheric Observatory, J. Atmos. Chem., 67, 87–140, https://doi.org/10.1007/s10874-011-9206-1, 2010. Fisher, J. A., Jacob, D. J., Wang, Q. Q., Bahreini, R., Carouge, C. C., Cubison, M. J., Dibb, J. E., Diehl, T., Jimenez, J. L., Leibensperger, E. M., Lu, Z. F., Meinders, M. B. J., Pye, H. O. T., Quinn, P. K., Sharma, S., Streets, D. G., van Donkelaar, A., and Yantosca, R. References Ashfold, M. J., Sturges, W. T., Harris, N. R. P., Robinson, A. D., and Pyle, J. A.: Characterisation of particle mass and num- ber concentration on the east coast of the Malaysian Peninsula during the northeast monsoon, Atmos. Environ., 117, 187–199, https://doi.org/10.1016/j.atmosenv.2015.07.018, 2015. Ashfold, M. J., Sturges, W. T., Harris, N. R. P., Robinson, A. D., and Pyle, J. A.: Characterisation of particle mass and num- ber concentration on the east coast of the Malaysian Peninsula during the northeast monsoon, Atmos. Environ., 117, 187–199, https://doi.org/10.1016/j.atmosenv.2015.07.018, 2015. Ashfold, M. J., Latif, M. T., Samah, A. A., Mead, M. I., and Har- ris, N. R. P.: Inﬂuence of Northeast Monsoon cold surges on air quality in Southeast Asia, Atmos. Environ., 166, 498–509, https://doi.org/10.1016/j.atmosenv.2017.07.047, 2017. Draxler, R. R. and Hess, G. D.: Description of the HYSPLIT_4 modeling system, NOAA Air Resources Laboratory, Silver Spring, MD, NOAA Tech. Memo. ERL ARL-224, 24, 1997. Behrenfeld, M. J., O’Malley, R. T., Siegel, D. A., McClain, C. R., Sarmiento, J. L., Feldman, G. C., Milligan, A. J., Falkowski, P. G., Letelier, R. M., and Boss, E. S.: Climate-driven trends in contemporary ocean productivity, Nature, 444, 752–755, https://doi.org/10.1038/nature05317, 2006. Draxler, R. R. and Hess, G. D.: An overview of the HYSPLIT_4 modeling system of trajectories, dispersion, and deposition, Aust. Meteorol. Mag., 47, 295–308, 1998. Bloss, W. J., Evans, M. J., Lee, J. D., Sommariva, R., Heard, D. E., and Pilling, M. J.: The oxidative capacity of the tropo- sphere: Coupling of ﬁeld measurements of OH and a global chemistry transport model, Faraday Discuss., 130, 425–436, https://doi.org/10.1039/b419090d, 2005. Draxler, R. R.: HYSPLIT4 user’s guide, NOAA Air Resources Lab- oratory, Silver Spring, MD, NOAA Tech. Memo. ERL ARL-230, 1999. Dunmore, R. E., Hopkins, J. R., Lidster, R. T., Mead, M. I., Bandy, B. J., Forster, G., Oram, D. E., Sturges, W. T., Phang, S. M., Abu Samah, A., and Hamilton, J. F.: Development of a Combined Heart-Cut and Comprehensive Two-Dimensional Gas Chromatography System to Extend the Carbon Range of Volatile Organic Compounds Analysis in a Single Instrument, Separa- tions, 3, 21, https://doi.org/10.3390/separations3030021, 2016. Boreddy, S. K. R. and Kawamura, K.: A 12-year observation of water-soluble ions in TSP aerosols collected at a remote marine location in the western North Paciﬁc: an outﬂow re- gion of Asian dust, Atmos. Chem. Phys., 15, 6437–6453, https://doi.org/10.5194/acp-15-6437-2015, 2015. Boreddy, S. K. R., Kawamura, K., and Jung, J. N. J. Farren et al.: Chemical characterisation of water-soluble ions As- suming that nssCa2+ can be used as a tracer for atmospheric dust, it was proposed that the increased concentrations were Competing interests. The authors declare that they have no conﬂict of interest. Acknowledgements. The ﬁnancial support of the Natural Envi- ronment Research Council (Naomi J. Farren, PhD studentship NE/L501751/1) is gratefully acknowledged. Naomi J. Farren would like to thank David Carslaw and Will Drysdale for their assistance using R. All authors would like to acknowledge NERC (NE/J016012/1 and NE/J016047/1) for funding the Bachok demonstration “International Opportunities Fund” campaign and HICoE-MoHE IOES-2014 (Air-Ocean-Land Interactions) for supporting the Bachok Marine Research Station facilities. Edited by: Hang Su Reviewed by: two anonymous referees www.atmos-chem-phys.net/19/1537/2019/ Atmos. Chem. Phys., 19, 1537–1553, 2019 N. J. Farren et al.: Chemical characterisation of water-soluble ions 1551 N. J. Farren et al.: Chemical characterisation of water-soluble ions 1552 GAW-SIS (Global Atmospheric Watch Station Information System): available at: https://gawsis.meteoswiss.ch/GAWSIS/ /index.html#/, last access: 12 April 2017. Murphy, J. G., Gregoire, P. K., Tevlin, A. G., Wentworth, G. R., Ellis, R. A., Markovic, M. Z., and Vandenboer, T. C.: Obser- vational constraints on particle acidity using measurements and modelling of particles and gases, Faraday Discuss., 200, 379– 395, 2017. Giglio, L., Descloitres, J., Justice, C. O., and Kaufman, Y. J.: An en- hanced contextual ﬁre detection algorithm for MODIS, Remote Sens. Environ., 87, 273–282, https://doi.org/10.1016/S0034- 4257(03)00184-6, 2003. Myriokefalitakis, S., Tsigaridis, K., Mihalopoulos, N., Sciare, J., Nenes, A., Kawamura, K., Segers, A., and Kanakidou, M.: In-cloud oxalate formation in the global troposphere: a 3-D modeling study, Atmos. Chem. Phys., 11, 5761–5782, https://doi.org/10.5194/acp-11-5761-2011, 2011. Ginoux, P., Prospero, J. M., Gill, T. E., Hsu, N. C., and Zhao, M.: Global-Scale Attribution of Anthropogenic and Natural Dust Sources and Their Emission Rates Based on Modis Deep Blue Aerosol Products, Rev. Geophys., 50, RG3005, https://doi.org/10.1029/2012rg000388, 2012. Neale, R. and Slingo, J.: The maritime continent and its role in the global climate: A GCM study, J. Climate, 16, 834–848, https://doi.org/10.1175/1520- 0442(2003)016<0834:Tmcair>2.0.Co;2, 2003. Gwynn, R. C., Burnett, R. T., and Thurston, G. D.: A time-series analysis of acidic particulate matter and daily mortality and mor- bidity in the Buffalo, New York, region, Environ. Health Persp., 108, 125–133, https://doi.org/10.2307/3454510, 2000. Newberg, J. T., Matthew, B. M., and Anastasio, C.: Chloride and bromide depletions in sea-salt particles over the north- eastern Paciﬁc Ocean, J. Geophys. Res.-Atmos., 110, D06209, https://doi.org/10.1029/2004jd005446, 2005. Hennigan, C. J., Izumi, J., Sullivan, A. P., Weber, R. J., and Nenes, A.: A critical evaluation of proxy methods used to estimate the acidity of atmospheric particles, Atmos. Chem. Phys., 15, 2775– 2790, https://doi.org/10.5194/acp-15-2775-2015, 2015. NOAA (National Oceanic and Atmospheric Administration Inte- grated Surface Database): available at: https://www.ncdc.noaa. gov/isd (last access: 20 July 2017), 2003. Huang, X. F. and Yu, J. Z.: Is vehicle exhaust a signiﬁcant primary source of oxalic acid in ambient aerosols?, Geophys. Res. Lett., 34, L02808, https://doi.org/10.1029/2006gl028457, 2007. Novakov, T. and Penner, J. E.: Large Contribution of Organic Aerosols to Cloud-Condensation-Nuclei Concentrations, Nature, 365, 823–826, https://doi.org/10.1038/365823a0, 1993. Huang, X. F., Yu, J. Z., He, L. Y., and Yuan, Z. B.: Water- soluble organic carbon and oxalate in aerosols at a coastal ur- ban site in China: Size distribution characteristics, sources, and formation mechanisms, J. Geophys. Res.-Atmos., 111, D22212, https://doi.org/10.1029/2006jd007408, 2006. Oram, D. N. J. Farren et al.: Chemical characterisation of water-soluble ions E., Ashfold, M. J., Laube, J. C., Gooch, L. J., Humphrey, S., Sturges, W. T., Leedham-Elvidge, E., Forster, G. L., Harris, N. R. P., Mead, M. I., Samah, A. A., Phang, S. M., Ou-Yang, C.-F., Lin, N.-H., Wang, J.-L., Baker, A. K., Brenninkmeijer, C. A. M., and Sherry, D.: A growing threat to the ozone layer from short-lived anthropogenic chlorocarbons, Atmos. Chem. Phys., 17, 11929–11941, https://doi.org/10.5194/acp-17-11929- 2017, 2017. Ismail, M., Yuen, F. S., and Abdullah, S. S.: Particulate Matter Sta- tus and its Relationship with Meteorological Factors in the East Coast of Peninsular Malaysia, J. Eng. Appl. Sci., 11, 2588–2593, 2016. Jiang, Y., Zhuang, G., Wang, Q., Liu, T., Huang, K., Fu, J. S., Li, J., Lin, Y., Zhang, R., and Deng, C.: Characteristics, sources and formation of aerosol oxalate in an Eastern Asia megacity and its implication to haze pollution, Atmos. Chem. Phys. Discuss., 11, 22075–22112, https://doi.org/10.5194/acpd-11-22075-2011, 2011. Pyle, J. A., Ashfold, M. J., Harris, N. R. P., Robinson, A. D., War- wick, N. J., Carver, G. D., Gostlow, B., O’Brien, L. M., Manning, A. J., Phang, S. M., Yong, S. E., Leong, K. P., Ung, E. H., and Ong, S.: Bromoform in the tropical boundary layer of the Mar- itime Continent during OP3, Atmos. Chem. Phys., 11, 529–542, https://doi.org/10.5194/acp-11-529-2011, 2011. Kerminen, V. M.: Relative roles of secondary sulfate and organics in atmospheric cloud condensation nuclei pro- duction, J. Geophys. Res.-Atmos., 106, 17321–17333, https://doi.org/10.1029/2001jd900204, 2001. Ripp, J.: Analytical Detection Limit Guidance and Laboratory Guide for Determining Method Detection Limits, Wisconsin De- partment of Natural Resources, Madison, WI, PUBL-TS-056-96, 1996. Robinson, A. D., Harris, N. R. P., Ashfold, M. J., Gostlow, B., War- wick, N. J., O’Brien, L. M., Beardmore, E. J., Nadzir, M. S. M., Phang, S. M., Samah, A. A., Ong, S., Ung, H. E., Peng, L. K., Yong, S. E., Mohamad, M., and Pyle, J. A.: Long-term halocarbon observations from a coastal and an inland site in Sabah, Malaysian Borneo, Atmos. Chem. Phys., 14, 8369–8388, https://doi.org/10.5194/acp-14-8369-2014, 2014. Keywood, M. D., Ayers, G. P., Gras, J. L., Boers, C. P., and Leong: Haze in the Klang Valley of Malaysia, Atmos. Chem. Phys., 3, 591–605, https://doi.org/10.5194/acp-3-591-2003, 2003. Lawrence, M. G., Jöckel, P., and von Kuhlmann, R.: What does the global mean OH concentration tell us?, Atmos. Chem. Phys., 1, 37–49, https://doi.org/10.5194/acp-1-37-2001, 2001. Legrand, M. and Pasteur, E. References M.: Sources, dis- tribution, and acidity of sulfate-ammonium aerosol in the Arctic in winter-spring, Atmos. Environ., 45, 7301–7318, https://doi.org/10.1016/j.atmosenv.2011.08.030, 2011. R. J., Monks, P. S., Karunaharan, A., Vaughan, S., Arnold, S. R., Tschritter, J., Pohler, D., Friess, U., Holla, R., Mendes, L. M., Lopez, H., Faria, B., Manning, A. J., and Wallace, D. W. R.: Seasonal characteristics of tropical marine boundary layer air measured at the Cape Verde Atmospheric Observatory, J. Atmos. Chem., 67, 87–140, https://doi.org/10.1007/s10874-011-9206-1, 2010. Fountoukis, C. and Nenes, A.: ISORROPIA II: a computa- tionally efﬁcient thermodynamic equilibrium model for K+- Ca2+-Mg2+-NH+ 4 -Na+-SO2− 4 -NO− 3 -Cl−-H2O aerosols, At- mos. Chem. Phys., 7, 4639–4659, https://doi.org/10.5194/acp-7- 4639-2007, 2007. Carslaw, D. C.: The openair manual – open-source tools for analysing air pollution data, Manual for version 1.1-4., King’s College London, 2015. Carslaw, D. C. and Ropkins, K.: openair – An R package for air quality data analysis, Environ. Modell. Softw., 27–28, 52–61, https://doi.org/10.1016/j.envsoft.2011.09.008, 2012. Freitas, A. D., Martins, L. D., and Solci, M. C.: Size-Segregated Particulate Matter and Carboxylic Acids over Urban and Rural Sites in Londrina City, Brazil, J. Brazil Chem. Soc., 23, 921–930, 2012. Charlson, R. J., Langner, J., Rodhe, H., Leovy, C. B., and War- ren, S. G.: Perturbation of the Northern-Hemisphere Radia- tive Balance by Backscattering from Anthropogenic Sulfate Aerosols, Tellus A, 43, 152–163, https://doi.org/10.1034/j.1600- 0870.1991.00013.x, 1991. Garreaud, R. D.: Subtropical cold surges: Regional aspects and global distribution, Int. J. Climatol., 21, 1181–1197, https://doi.org/10.1002/Joc.687, 2001. Dominick, D., Latif, M. T., Juneng, L., Khan, M. F., Amil, N., Mead, M. I., Nadzir, M. S. M., Moi, P. S., Abu Samah, A., Atmos. Chem. Phys., 19, 1537–1553, 2019 www.atmos-chem-phys.net/19/1537/2019/ www.atmos-chem-phys.net/19/1537/2019/ N. J. Farren et al.: Chemical characterisation of water-soluble ions C.: Methane sulfonic acid to non-sea-salt sulfate ratio in coastal Antarctic aerosol and surface snow, J. Geophys. Res.-Atmos., 103, 10991–11006, https://doi.org/10.1029/98jd00929, 1998. Sareen, N., Carlton, A. G., Surratt, J. D., Gold, A., Lee, B., Lopez- Hilﬁker, F. D., Mohr, C., Thornton, J. A., Zhang, Z., Lim, Y. B., and Turpin, B. J.: Identifying precursors and aqueous organic aerosol formation pathways during the SOAS campaign, Atmos. Chem. Phys., 16, 14409–14420, https://doi.org/10.5194/acp-16- 14409-2016, 2016. Lin, Y. H., Zhang, Z. F., Docherty, K. S., Zhang, H. F., Bud- isulistiorini, S. H., Rubitschun, C. L., Shaw, S. L., Knipping, E. M., Edgerton, E. S., Kleindienst, T. E., Gold, A., and Sur- ratt, J. D.: Isoprene Epoxydiols as Precursors to Secondary Or- ganic Aerosol Formation: Acid-Catalyzed Reactive Uptake Stud- ies with Authentic Compounds, Environ. Sci. Technol., 46, 250– 258, https://doi.org/10.1021/es202554c, 2012. Savoie, D. L. and Prospero, J. M.: Comparison of Oceanic and Con- tinental Sources of Non-Sea-Salt Sulfate over the Paciﬁc-Ocean, Nature, 339, 685–687, https://doi.org/10.1038/339685a0, 1989. www.atmos-chem-phys.net/19/1537/2019/ Atmos. Chem. Phys., 19, 1537–1553, 2019 N. J. Farren et al.: Chemical characterisation of water-soluble ions 1553 Saxena, P. and Hildemann, L. M.: Water-soluble organics in atmo- spheric particles: A critical review of the literature and appli- cation of thermodynamics to identify candidate compounds, J. Atmos. Chem., 24, 57–109, https://doi.org/10.1007/Bf00053823, 1996. Tahir, N. M., Suratman, S., Fong, F. T., Hamzah, M. S., and Latif, M. T.: Temporal Distribution and Chemical Characteri- zation of Atmospheric Particulate Matter in the Eastern Coast of Peninsular Malaysia, Aerosol Air Qual. Res., 13, 584–595, https://doi.org/10.4209/aaqr.2012.08.0216, 2013. Seinfeld, J. H. and Pandis, S. N.: Atmospheric chemistry and physics: from air pollution to climate change, 1–1203, 2nd Edn., J. Wiley, Hoboken, NJ, 2006. Tan, Y., Lim, Y. B., Altieri, K. E., Seitzinger, S. P., and Turpin, B. J.: Mechanisms leading to oligomers and SOA through aque- ous photooxidation: insights from OH radical oxidation of acetic acid and methylglyoxal, Atmos. Chem. Phys., 12, 801–813, https://doi.org/10.5194/acp-12-801-2012, 2012. Sempere, R. and Kawamura, K.: Comparative Distributions of Dicarboxylic-Acids and Related Polar Compounds in Snow Rain and Aerosols from Urban Atmosphere, Atmos. Environ., 28, 449–459, https://doi.org/10.1016/1352-2310(94)90123-6, 1994. Yamasoe, M. A., Artaxo, P., Miguel, A. H., and Allen, A. G.: Chemical composition of aerosol particles from direct emis- sions of vegetation ﬁres in the Amazon Basin: water-soluble species and trace elements, Atmos. Environ., 34, 1641–1653, https://doi.org/10.1016/S1352-2310(99)00329-5, 2000. Stein, A. F., Draxler, R. R., Rolph, G. D., Stunder, B. J. B., Cohen, M. D., and Ngan, F.: Noaa’s Hysplit Atmospheric Transport and Dispersion Modeling System, B. Am. Meteorol. Soc., 96, 2059– 2077, https://doi.org/10.1175/Bams-D-14-00110.1, 2015. Zhang, Y., Sperber, K. R., and Boyle, J. S.: Climatology and interannual variation of the East Asian winter mon- soon: Results from the 1979-95 NCEP/NCAR reanalysis, Mon. Weather Rev., 125, 2605–2619, https://doi.org/10.1175/1520- 0493(1997)125<2605:Caivot>2.0.Co;2, 1997. Sturges, W. T. and Shaw, G. E.: Halogens in Aerosols in Central Alaska, Atmos. Environ. A-Gen., 27, 2969–2977, https://doi.org/10.1016/0960-1686(93)90329-W, 1993. Surratt, J. D., Chan, A. W. H., Eddingsaas, N. C., Chan, M. N., Loza, C. L., Kwan, A. J., Hersey, S. P., Flagan, R. C., Wennberg, P. O., and Seinfeld, J. H.: Reactive in- termediates revealed in secondary organic aerosol formation from isoprene, P. Natl. Acad. Sci. USA, 107, 6640–6645, https://doi.org/10.1073/pnas.0911114107, 2010. Atmos. Chem. Phys., 19, 1537–1553, 2019 Atmos. Chem. Phys., 19, 1537–1553, 2019 www.atmos-chem-phys.net/19/1537/2019/ www.atmos-chem-phys.net/19/1537/2019/
https://openalex.org/W3015213037	https://europepmc.org/articles/pmc7155754?pdf=render	English	null	Comparison of Laparoscopy and Laparotomy for Para-Aortic Lymphadenectomy in Women With Presumed Stage I–II High-Risk Endometrial Cancer	Frontiers in oncology	2,020	cc-by	4,861	ORIGINAL RESEARCH published: 07 April 2020 doi: 10.3389/fonc.2020.00451 Comparison of Laparoscopy and Laparotomy for Para-Aortic Lymphadenectomy in Women With Presumed Stage I–II High-Risk Endometrial Cancer E Sun Paik 1†, Seung Hun Baek 2†, Jun Hyeok Kang 2, Soo Young Jeong 2, Myeong Seon Kim 3, Woo Young Kim 1, Yoo-Young Lee 2, Chel Hun Choi 2, Jeong-Won Lee 2, Byoung-Gie Kim 2, Duk-Soo Bae 2 and Tae-Joong Kim 2* Keywords: endometrial cancer, laparoscopy, lymphadenectomy, para-aortic lymph node, postoperative complications Edited by: Diana English, University of South Florida, United States 1 Department of Obstetrics and Gynecology, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, South Korea, 2 Department of Obstetrics and Gynecology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, South Korea, 3 Department of Obstetrics and Gynecology, Saint Vincent’s Hospital, The Catholic University of Korea, Suwon, South Korea Reviewed by: Emanuele Perrone, A. Gemelli University Hospital Foundation, Italy Lucia Tortorella, A. Gemelli University Hospital Foundation, Italy Objective: To compare laparoscopic surgery to laparotomy for harvesting para-aortic lymph nodes in presumed stage I–II, high-risk endometrial cancer patients. Methods: Patients with histologically proven endometrial cancer, presumed stage I-II with high-risk tumor features who had undergone hysterectomy, bilateral salpingoophorectomy, or pelvic and para-aortic lymphadenectomy by either laparoscopy or laparotomy in Samsung Medical Center from 2005 to 2017 were retrospectively investigated. The primary outcome was para-aortic lymph node count. Secondary outcomes were pelvic lymph node count, perioperative events, and postoperative complications. *Correspondence: Tae-Joong Kim tj28.kim@gmail.com †These authors have contributed equally to this work Specialty section: This article was submitted to Women’s Cancer, a section of the journal Frontiers in Oncology Results: A total of 90 patients was included (35 for laparotomy, 55 for laparoscopy) for analysis. The mean (±SD) para-aortic lymph node count was 10.66 (±7.596) for laparotomy and 10.35 (±5.848) for laparoscopy (p = 0.827). Mean pelvic node count was 16.8 (±6.310) in the laparotomy group and 16.13 (±7.626) in the laparoscopy group (p = 0.664). Lower estimated blood loss was shown in the laparoscopy group. There was no difference in perioperative outcome between the groups. Additional multivariate analysis showed that survival outcome was not affected by surgical methods in presumed stage I-II, high-risk endometrial cancer patients. Received: 05 December 2019 Accepted: 13 March 2020 Published: 07 April 2020 Surgical Management g g Surgical staging for endometrial cancer was performed by laparoscopy (four-port conventional) or laparotomy (midline incision). In our institution, a staging operation was routinely performed according to the Korean Gynecologic Oncology Group (KGOG) surgical manual (11). Regarding lymphadenectomy, pelvic lymph nodes removal was performed from the distal one-half of the common iliac artery to the circumﬂex iliac vein, and nodal tissue was removed at anterior of the obturator nerve and around the iliac vessels. The para-aortic lymph nodes indicated those covering the vena cava (right para-aortic), middle part between the vena cava and aorta, and to the left of the aorta (left para-aortic). The cephalad border of the para-aortic lymph node was commonly, but not limited to, the inferior mesenteric artery (IMA). The distal border was the midpoint of the common iliac artery (8). Laparoscopic PALND was performed after elevating the IMA for identiﬁcation of the left ureter. PALND proceeded to the left of the aorta and down the lateral aspect of the left common iliac artery to the midpoint. Six surgeons who specialize in gynecologic oncology in our institution performed surgical staging. All six surgeons were all skilled for laparoscopy and laparotomy procedures in gynecologic surgery, and they were all trained as gynecologic oncologists in major institution in Korea. The lymph node specimens were sent to the pathology department for histological evaluation which was performed by gynecologic expert pathologists. Surgery for endometrial cancer was traditionally performed with laparotomy, but minimally invasive surgery has increased gradually due to lower perioperative and postoperative complication rates and shorter hospital stays without diﬀerences in survival outcomes (6, 7). Previously, large-scale randomized trials were conducted for comparison between laparoscopic staging operation for endometrial cancer and laparotomy, but there were no detailed results regarding utility of laparoscopic procedure for systematic lymph node dissection, speciﬁcally for the PALND which can be more complex procedure (8, 9). In this study, we aimed to compare laparoscopy with laparotomy for PALND in patients with presumed stage I- II, high-risk endometrial cancer. The primary outcome of this study was number of harvested para-aortic lymph nodes, and secondary outcomes for study were pelvic node count and perioperative/postoperative surgical outcomes. Additionally, we compared overall survival and recurrence for two patient groups. Citation: Paik ES, Baek SH, Kang JH, Jeong SY, Kim MS, Kim WY, Lee Y-Y, Choi CH, Lee J-W, Kim B-G, Bae D-S and Kim T-J (2020) Comparison of Laparoscopy and Laparotomy for Para-Aortic Lymphadenectomy in Women With Presumed Stage I–II High-Risk Endometrial Cancer. Front. Oncol. 10:451. doi: 10.3389/fonc.2020.00451 Conclusions: Study results demonstrate comparable para-aortic lymph node count with less blood loss in laparoscopy over laparotomy. In women with presumed stage I-II, high-risk endometrial cancer, laparoscopy is a valid treatment modality. April 2020 \| Volume 10 \| Article 451 Frontiers in Oncology \| www.frontiersin.org 1 Laparoscopic Para-Aortic Lymphadenectomy Paik et al. INTRODUCTION features; non-endometrioid type/endometrioid type FIGO grade 3/endometrioid type FIGO grade 2 with >50% myometrial invasion or invasion in the cervical stroma, evaluated by trans-vaginal ultrasound and/or magnetic resonance imaging (MRI) of pelvis] (10). Criteria were based on concept of using clinically predicted stage before surgery as in actual practice by preoperative evaluation, and pathologic ﬁndings after surgical staging (lymphovascular space invasion, lymph node metastasis), were not included in inclusion criteria. Exclusion criteria were ongoing anti-tumor treatment, pre-operative imaging indicating extrauterine spread, and disseminated disease diagnosed during surgery. After diagnostic endometrial biopsy by curettage, all patients had undergone pre-operative evaluation, blood tests, and imaging studies (computed tomography (CT) of the chest and abdomen/pelvis and trans-vaginal ultrasound and/or MRI of the pelvis). Endometrial carcinoma is the most rapidly increasing female genital tract malignancy. In 2018, it was the sixth most common malignancy in females, accounting for 382,069 new cases world- wide (1). In Korea, the incidence of newly diagnosed endometrial cancer consistently increased in all age groups from 1999 to 2015 (2). Endometrial cancer usually occurs in women after menopause, and most cases present with early-stage disease due to frequent symptoms of abnormal vaginal bleeding, which leads to early detection. The European Society for Medical Oncology (ESMO), European Society of Gynaecological Oncology (ESGO), and European Society for Radiotherapy and Oncology (ESTRO) guidelines recommend that treatments of high-intermediate risk, high risk, and/or advanced endometrial cancer should include surgical treatment with hysterectomy, bilateral salpingo- oophorectomy with additional pelvic lymph node dissection (PLND), and para-aortic lymph node dissection (PALND) (3, 4). The survival eﬀect of para-aortic lymphadenectomy in endometrial cancer (SEPAL) study demonstrated a signiﬁcant beneﬁt of pelvic and para-aortic lymphadenectomy on survival in endometrial cancer patients with intermediate or high risk of recurrence (5). The results of lymphadenectomy provide physicians relevant information for adjuvant treatment in the clinical ﬁeld. Frontiers in Oncology \| www.frontiersin.org Patients The primary outcome measure was the number of harvested para-aortic lymph nodes. Secondary outcomes were number of pelvic nodes and perioperative/postoperative outcomes. Operation time was designated as the time (minutes) from ﬁrst incision of the operation to last closure of the skin. Postoperative complications were investigated until 30 days after surgery, and readmissions within 30 days from surgery were investigated. Length of hospital stay was measured (in days) for initial operation, and in the case of readmission, hospital days after readmission were not included. Additionally, survival outcomes, recurrence-free survival (RFS) and overall survival (OS), were After obtaining the Institutional Review Board (IRB) approval (No. 2019-07-133-002), we retrospectively reviewed data from all patients with primary endometrial cancer who were surgically treated (laparoscopy and laparotomy) at Samsung Medical Center from January 2005 to December 2017. We investigated data from electronic medical records and included patients based on the following criteria: (1) histologically conﬁrmation of endometrial cancer, (2) presumed stage I or II [International Federation of Gynecology and Obstetrics (FIGO)], (3) pre- operatively evaluated as high-risk tumor [one of following April 2020 \| Volume 10 \| Article 451 Frontiers in Oncology \| www.frontiersin.org 2 Laparoscopic Para-Aortic Lymphadenectomy Paik et al. Categorical variables were shown as frequency (percentage). After conﬁrmation of normal distributions with the Shapiro– Wilk test, the Mann–Whitney test was performed to compare median values, and Student’s t-test was performed to compare mean values. Fisher’s exact test or χ2 test was performed for analysis of the distribution of characteristics. The Cox proportional hazards model was used for univariate and multivariate analyses to evaluate the prognostic signiﬁcance of clinicopathologic features for RFS and OS. For multivariate analysis, a stepwise backward method was used. Variables associated with RFS and OS with a signiﬁcance level of P < 0.05 in univariate analysis were selected as possible variables for multivariate logistic regression analysis. A 95% conﬁdence interval (CI) was used to quantify the correlation between RFS or OS and each independent feature. All P-values were two-sided, and P < 0.05 were regarded as statistically signiﬁcant. All statistical analyses were accomplished using IBM SPSS Statistics software Version 21.0 (IBM, Armonk, New York, USA). investigated for two surgical methods. RFS was described as the time (months) from surgery to the recurrence or last follow-up. OS was deﬁned as the time (months) from surgery to date of death or last follow-up. Statistical Analysis For purposes of this study, summary statistics were used for description of the data. Median (range) or mean (standard deviation) were used to describe continuous variables. TABLE 1 \| Baseline demographics and clinicopathologic characteristics of endometrial cancer patients by surgical technique (n = 90). Characteristic Laparotomy (n = 35) Laparoscopy (n = 55) p-value Age, years Median (range) 59 (39–77) 53 (35–70) 0.002 BMI, kg/m2 Median (range) 23.4 (18.3–38.0) 23.8 (17.4–39.3) 0.699 Menopause 0.072 Yes 27 (77.1) 33 (60.0) No 8 (22.9) 22 (40.0) Stage (presumed by preoperative evaluation) 0.322 IA 16 (45.7) 33 (60.0) IB 13 (37.1) 18 (32.7) II 6 (17.2) 4 (7.3) Histology 0.558 Endometrioid 29 (82.9) 41 (74.5) Papillary serous 1 (2.9) 2 (3.6) Clear cell 3 (8.6) 2 (3.6) Others 2 (5.8) 10 (18.2) Grade 0.261 2 4 (11.4) 3 (5.5) 3 31 (88.6) 52 (94.5) Pre-operative serum CA-125 level, U/mL 0.135 Median (range) 11.8 (2.0–408) 10.4 (3.0–145) Myometrial invasion 0.084 No 7 (20.0) 18 (32.7) Superﬁcial 3 (8.6) 0 Inner half 11 (31.4) 20 (36.4) Outer half 11 (31.4) 16 (29.1) Full 3 (8.6) 1 (1.8) Lymphovascular space invasion 0.298 No 20 (57.1) 38 (69.1) Yes 15 (42.9) 17 (30.9) Adjuvant treatment >0.999 Chemotherapy 2 (5.7) 3 (5.5) Radiotherapy 21 (60.0) 34 (61.8) Concurrent chemo-radiotherapy 1 (2.9) 2 (3.6) No adjuvant therapy 11 (31.4) 16 (29.1) BMI, body mass index; CA-125, cancer antigen 125. TABLE 1 \| Baseline demographics and clinicopathologic characteristics of endometrial cancer patients by surgical technique (n = 90). Survival Outcomes In addition to comparison of number of lymph nodes and perioperative complications, survival comparison was performed. After a median follow-up of 61.6 months (range 3.0–136.9 months), there were 17 patients with recurrences, TABLE 3 \| Perioperative and postoperative outcomes per surgical modality. Outcomes Laparotomy (n = 35) Laparoscopy (n = 55) p-value Operation time, minutes 0.095 Median (range) 245 (180–502) 267 (144–727) Estimated total blood loss, mL 0.001 Median (range) 350 (50–800) 200 (50–500) Blood transfusion, no. (%) 0.007 No 30 (85.7) 55 (100) 1 pint 2 (5.7) 0 2 pint 2 (5.7) 0 4 pint 1 (2.9) 0 Length of hospital stay, days 0.104 Median (range) 10 (6–86) 9 (6–20) Intraoperative adverse events, n (%) 1 (2.9) 3 (5.5) 0.456 Ureter injury 1 (2.9) 0 Bladder injury 0 2 (3.6) IVC injury 0 1 (1.8) Postoperative complications within 30 days, n (%) 12 (34.3) 15 (27.3) 0.196 Fever 4 (11.4) 3 (5.5) Ileus 4 (11.4) 10 (18.2) Surgical wound infection 4(11.4) 1 (5.6) Leg pain 0 1 (1.8) IVC, inferior vena cava. discrepancies between preoperative clinical stage and result of pathologic evaluation (lymph node metastasis). Multivariate analysis revealed that survival outcome was not aﬀected by surgical method in presumed stage I–II, high-risk endometrial cancer patients. The eﬀect of PALND in endometrial cancer was shown in a previous retrospective cohort study (SEPAL study) (5). In SEPAL study, result of multivariate analysis had shown that PLND and PALND reduced the risk of death compared with that in the PLND only group. Thus, PALND and PLND are recommended for patients with endometrial cancer with intermediate or high risk of recurrence. In a large-scale, randomized trial comparing laparoscopy and laparotomy in endometrial cancer (Gynecologic Oncology Group Study LAP2 trial) (8), results showed that laparoscopic surgical staging for TABLE 4 \| Survival comparison by patients group with laparoscopy and laparotomy. Characteristic Laparotomy (n = 35) Laparoscopy (n = 55) p-value Recurrence event 9 (25.7) 8 (14.5) 0.269 Recurrence-free survival, months Median (range) 51.5 (3.0–136.9) 60.5 (3.0–135.4) 0.646 Death event 7 (20.0) 3 (5.5) 0.043 Overall-survival, months Median (range) 56.0 (3.0–136.9) 61.6 (3.0–136.9) 0.734 TABLE 3 \| Perioperative and postoperative outcomes per surgical modality. Outcomes Laparotomy (n = 35) Laparoscopy (n = 55) p-value Operation time, minutes 0.095 Median (range) 245 (180–502) 267 (144–727) Estimated total blood loss, mL 0.001 Median (range) 350 (50–800) 200 (50–500) Blood transfusion, no. DISCUSSION In this study, we concluded that laparoscopy is an acceptable surgical method for harvesting para-aortic lymph nodes in presumed stage I–II high-risk endometrial cancer. Between laparoscopy and laparotomy groups, number of harvested para-aortic lymph nodes was comparable, as well as number of pelvic lymph nodes. Less blood loss was shown in the laparoscopy group, and there were no signiﬁcant diﬀerences in operation time, length of hospital stay, and intraoperative and postoperative complications between the two groups. For survival comparison, no meaningful results were drawn, possibly due to unevenly distributed characteristics and discrepancies between preoperative clinical stage and result of pathologic evaluation (lymph node metastasis). Multivariate analysis revealed that survival outcome was not aﬀected by surgical method in presumed stage I–II, high-risk endometrial cancer patients. Perioperative and Postoperative Outcomes Perioperative and Postoperative Outcomes Comparison of perioperative and postoperative outcomes between two groups are shown in Table 3. There was signiﬁcantly less estimated total blood loss in the laparoscopy group than in the laparotomy group (200 vs. 350 mL, p = 0.001). Notably, there was need for blood transfusion in 14.3% of the laparotomy cases compared to none in the laparoscopy group. No conversion to laparotomy was shown in the laparoscopy group. Regarding operation time and length of hospital stay, no diﬀerences between two groups were shown. There were no diﬀerences in intraoperative complication events (5.5 vs. 2.9%, p = 0.456) or postoperative 30 day complications (27.3 vs. 34.3%, p = 0.196) between the two groups. No patient in either group experienced readmission within 30 days, and no mortality was shown within 30 days after surgery. Patients Characteristics This study analyzed 90 presumed stage I–II, high-risk endometrial cancer patients treated in Samsung Medical Center from 2005 to 2017. Within this patient group, 35 patients underwent laparotomy, and 55 patients had laparoscopic surgical staging. Basal characteristics are shown in Table 1. There were signiﬁcant diﬀerences in age among patients undergoing the two surgical methods, but there were no statistically signiﬁcant diﬀerences for other variables between the two groups. Analysis of PALND and PLND Regarding number of harvested para-aortic lymph nodes [mean (±SD)], no diﬀerence between laparoscopy [10.35 (±5.848)], and laparotomy [10.66 (±7.596)] (p = 0.827) was shown (Table 2). Also, pelvic lymph node count was not diﬀerent between the two groups [laparoscopy16.13 (±7.626) vs. laparotomy 16.8 (±6.310), p = 0.664]. Lymph node metastasis was shown only in laparotomy group (pelvic 14.3%, para-aortic 8.65%). In this April 2020 \| Volume 10 \| Article 451 Frontiers in Oncology \| www.frontiersin.org 3 Laparoscopic Para-Aortic Lymphadenectomy Paik et al. study, patients were included according to clinical FIGO stage I– II, which was presumed by preoperative evaluation, and patients with advanced stage disease due to lymph node metastasis in pathologic staging were inevitably included. and 10 deaths (Table 4). Regarding recurrence events, diﬀerence between the two groups was not signiﬁcant (p = 0.269), but there was a signiﬁcant diﬀerence regarding death events (p = 0.043). Kaplan–Meier plots for RFS and OS are shown in Supplementary Material 1. Unevenly distributed ages between groups and discrepancy between preoperative clinical stage and result of pathologic evaluation (lymph node metastasis) may be reasons for the survival diﬀerence. In multivariate analysis for OS, type of surgical procedure was not a signiﬁcant factor for OS [HR (95% CI) 0.444 (0.097–2.030), p = 0.295, Table 5]. Results of multivariate analysis demonstrated that survival outcome was not aﬀected by surgical method in presumed stage I–II, high-risk endometrial cancer patients. April 2020 \| Volume 10 \| Article 451 Survival Outcomes (%) 0.007 No 30 (85.7) 55 (100) 1 pint 2 (5.7) 0 2 pint 2 (5.7) 0 4 pint 1 (2.9) 0 Length of hospital stay, days 0.104 Median (range) 10 (6–86) 9 (6–20) Intraoperative adverse events, n (%) 1 (2.9) 3 (5.5) 0.456 Ureter injury 1 (2.9) 0 Bladder injury 0 2 (3.6) IVC injury 0 1 (1.8) Postoperative complications within 30 days, n (%) 12 (34.3) 15 (27.3) 0.196 Fever 4 (11.4) 3 (5.5) Ileus 4 (11.4) 10 (18.2) Surgical wound infection 4(11.4) 1 (5.6) Leg pain 0 1 (1.8) IVC, inferior vena cava. TABLE 3 \| Perioperative and postoperative outcomes per surgical modality. The eﬀect of PALND in endometrial cancer was shown in a previous retrospective cohort study (SEPAL study) (5). In SEPAL study, result of multivariate analysis had shown that PLND and PALND reduced the risk of death compared with that in the PLND only group. Thus, PALND and PLND are recommended for patients with endometrial cancer with intermediate or high risk of recurrence. In a large-scale, randomized trial comparing laparoscopy and laparotomy in endometrial cancer (Gynecologic Oncology Group Study LAP2 trial) (8), results showed that laparoscopic surgical staging for Frontiers in Oncology \| www.frontiersin.org 4 Laparoscopic Para-Aortic Lymphadenectomy Paik et al. TABLE 5 \| Univariate and multivariate Cox proportional hazards analysis for recurrence-free survival (RFS), and overall survival (OS) to adjust risk associated with prognostic features. Variables Univariate Multivariate HR (95% CI) p-value HR (95% CI) p-value RFS Age 1.009 (0.953–1.068) 0.769 N/A Clinical stage (I vs. II) 3.587 (1.160–11.090) 0.027 2.303 (0.565–9.381) 0.244 Histology (Endometrioid vs. non-endometrioid) 0.791 (0.227–2.753) 0.712 N/A Grade (2 vs. 3) 0.634 (0.145–2.773) 0.545 N/A Myometrial invasion (<1/2 vs. >1/2) 4.167 (1.537–11.300) 0.005 3.679 (1.337–10.122) 0.012 Lymphovascular space invasion (no vs. yes) 3.898 (1.440–10.552) 0.007 1.988 (0.659–5.997) 0.222 Cervical stromal invasion (no vs. yes) 3.184 (1.171–8.662) 0.023 0.534 (0.118–2.412) 0.415 LN metastasis (no vs. yes) 8.341 (2.656–26.201) <0.001 6.677 (2.073–21.509) 0.001 Type of surgical procedure (Laparotomy vs. Laparoscopy) 0.532 (0.205–1.380) 0.195 N/A OS Age 1.076 (0.998–1.077) 0.058 N/A Clinical stage (I vs. II) 5.528 (1.412–21.513) 0.014 1.791 (0.325–9.853) 0.503 Histology (Endometrioid vs. non-endometrioid) 2.648 (0.747–9.392) 0.132 N/A Grade (2 vs. 3) 0.719 (0.091–5.677) 0.719 N/A Myometrial invasion (<1/2 vs. >1/2) 2.341 (0.677–8.103) 0.179 N/A Lymphovascular space invasion (no vs. yes) 3.244 (0.914–11.520) 0.069 N/A Cervical stromal invasion (no vs. yes) 6.053 (1.744–21.004) 0.005 3.794 (0.903–15.949) 0.069 LN metastasis (no vs. Survival Outcomes yes) 9.684 (2.474–37.907) 0.001 4.587 (0.946–22.244) 0.059 Type of surgical procedure (Laparotomy vs. laparoscopy) 0.250 (0.064–0.966) 0.044 0.444 (0.097–2.030) 0.295 n = 90. RFS, recurrence-free survival; OS, overall-survival; HR, hazard ratio; CI, conﬁdential interval; LN, lymph node; N/A, not available. TABLE 5 \| Univariate and multivariate Cox proportional hazards analysis for recurrence-free survival (RFS), and overall survival (OS) to adjust risk associated with prognostic features TABLE 5 \| Univariate and multivariate Cox proportional hazards analysis for recurrence-free survival (RFS), and overall survival (OS) to adjust risk associated with prognostic features. Variables Univariate Multivariate that had not been thoroughly investigated so far. Using speciﬁc patient groups for comparison, outcomes of procedures could be more clearly determined. However, a number of potential limitations exist in the current study. The study had used data that were retrospectively collected for analysis. Therefore, it may be biased by patients not-randomly assigned, and incomplete data collection. Inclusion criteria were based on clinically presumed stage as in actual practice by preoperative evaluation and pathologic ﬁndings after surgical staging (lymphovascular space invasion, lymph node metastasis), which could be risk factor for adjuvant treatment, were not included. This could be potential bias of patients’ selection. Also, this may have led to unevenly distributed characteristics between two groups, and this made it diﬃcult to obtain a meaningful result in comparing the survival of the two groups. Kaplan–Meier plot for OS showed a signiﬁcant diﬀerence between two surgical groups, and we assumed that discrepancy between preoperative clinical stage and result of pathologic evaluation (lymph node metastasis) may be reasons for the survival diﬀerence. However, since the primary outcome of this study was a comparison of surgical outcomes, the diﬀerence in characteristics between the two groups may not be a problem, and multivariate analysis results demonstrated that OS was not aﬀected by surgical method in this patients group. In addition to limitations and biases associated with its retrospective nature, this study resulted from data of a small number of patients for comparisons. A further large-scale study will provide more meaningful results. endometrial cancer is feasible and safe in regards to postoperative outcomes which resulted in fewer complications and shorter hospital stays. The LAP2 trial included all stages of endometrial cancer compared to high-risk, presumed stage I–II in our study. In LAP2 trial, pelvic and para-aortic lymph nodes were not removed in 8% of patients of laparoscopy and 4% of laparotomy. Frontiers in Oncology \| www.frontiersin.org REFERENCES 10. Pecorelli S. Revised FIGO staging for carcinoma of the vulva, cervix, and endometrium. Int J Gynaecol Obstet. (2009) 105:103–4. doi: 10.1016/j.ijgo.2009.02.012 1. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. (2018) 68:394–424. doi: 10.3322/caac.21492 1. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. (2018) 68:394–424. doi: 10.3322/caac.21492 11. Lee M, Choi CH, Chun YK, Kim YH, Lee KB, Lee SW, et al. Surgical manual of the Korean Gynecologic Oncology Group: classiﬁcation of hysterectomy and lymphadenectomy. J Gynecol Oncol. (2017) 28:e5. doi: 10.3802/jgo.201 7.28.e5 2. Lim MC, Won YJ, Ko MJ, Kim M, Shim SH, Suh DH, et al. Incidence of cervical, endometrial, and ovarian cancer in Korea during 1999-2015. J Gynecol Oncol. (2019) 30:e38. doi: 10.3802/jgo.2019.30.e38 12. Gallotta V, Nero C, Chiantera V, Scambia G. Minilaparoscopic aortic lymphadenectomy. J Minim Invasive Gynecol. (2015) 22:546–7. doi: 10.1016/j.jmig.2015.02.011 3. Colombo N, Creutzberg C, Amant F, Bosse T, Gonzalez-Martin A, Ledermann J, et al. ESMO-ESGO-ESTRO Consensus Conference on Endometrial Cancer: diagnosis, treatment and follow-up. Ann Oncol. (2016) 27:16–41. doi: 10.1093/annonc/mdv484 13. Gallotta V, Giudice MT, Conte C, Sarandeses AV, D’Indinosante M, Federico A, et al. Minimally invasive salvage lymphadenectomy in gynecological cancer patients: a single institution series. Eur J Surg Oncol. (2018) 44:1568–72. doi: 10.1016/j.ejso.2018.08.006 4. Burke WM, Orr J, Leitao M, Salom E, Gehrig P, Olawaiye AB, et al. Endometrial cancer: a review and current management strategies: part I. Gynecol Oncol. (2014) 134:385–92. doi: 10.1016/j.ygyno.2014.05.018 14. Gallotta V, Federico A, Gaballa K, D’Indinosante M, Conte C, Giudice MT, et al. The role of robotic aortic lymphadenectomy in gynecological cancer: surgical and oncological outcome in a single institution experience. J Surg Oncol. (2019) 119:355–60. doi: 10.1002/jso. 25335 5. Todo Y, Kato H, Kaneuchi M, Watari H, Takeda M, Sakuragi N. Survival eﬀect of para-aortic lymphadenectomy in endometrial cancer (SEPAL study): a retrospective cohort analysis. Lancet. (2010) 375:1165–72. doi: 10.1016/s0140-6736(09)62002-x 15. Salehi S, Avall-Lundqvist E, Legerstam B, Carlson JW, Falconer H. Robot-assisted laparoscopy versus laparotomy for infrarenal paraaortic lymphadenectomy in women with high-risk endometrial cancer: a randomised controlled trial. Eur J Cancer. (2017) 79:81–9. doi: 10.1016/j.ejca.2017.03.038 6. ETHICS STATEMENT The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fonc. 2020.00451/full#supplementary-material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fonc. 2020.00451/full#supplementary-material This study was approved by the institutional review boards of Samsung Medical Center in accordance with the Declaration of Helsinki and the International Conference on Harmonization Good Clinical Practice guidelines. Supplementary Material 1 \| Kaplan–Meier curves for (A) recurrence-free survival and (B) overall survival by laparoscopy and laparotomy in presumed early-stage, high-risk endometrial cancer patients. AUTHOR CONTRIBUTIONS Despite these limitations, we demonstrated that laparoscopy is an acceptable surgical method for harvesting para-aortic lymph nodes in presumed stage I–II, high-risk endometrial cancer patients. Study results showed a similar number of harvested lymph nodes, no signiﬁcant diﬀerences in intraoperative and postoperative complications between laparoscopy and laparotomy, and less blood loss in the laparoscopy group. EP, SB, and T-JK contributed to the conception and design of the study. SB, JK, SJ, and MK organized the database. EP and T-JK performed the statistical analysis. EP and SB wrote the ﬁrst draft of the manuscript. WK, Y-YL, CC, J-WL, B- GK, and D-SB wrote sections of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version. EP, SB, and T-JK contributed to the conception and design of the study. SB, JK, SJ, and MK organized the database. EP and T-JK performed the statistical analysis. EP and SB wrote the ﬁrst draft of the manuscript. WK, Y-YL, CC, J-WL, B- GK, and D-SB wrote sections of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version. EP, SB, and T-JK contributed to the conception and design of the study. SB, JK, SJ, and MK organized the database. EP and T-JK performed the statistical analysis. EP and SB wrote the ﬁrst draft of the manuscript. WK, Y-YL, CC, J-WL, B- GK, and D-SB wrote sections of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version. ACKNOWLEDGMENTS The datasets for this article are not publicly available because IRB of Samsung Medical Center restricted providing raw data of this data to public for privacy issues. Requests to access the datasets should be directed to Review committee of Research Data under IRB in Samsung Medical Center. This topic was presented as oral presentation at AAGL 2018 Global Congress, Las Vegas in November 12, 2018. Survival Outcomes Lymph node metastases were found in 9% of patients and were similar in both groups. In present study, lymph node metastasis was present in only the laparotomy group, possibly due to uneven randomization between groups. This may have resulted in survival diﬀerences in the current study. Previously, number of studies had shown feasibility of minimally invasive surgery in lymphadenectomy of gynecologic cancers (12–14). A recent randomized trial in Sweden that compared robotic- assisted laparoscopy and laparotomy for infrarenal PALND in early stage high-risk endometrial cancer showed non-inferiority in para-aortic lymph node count, similar complication rates, shorter hospital length, and lower total cost for robotic-assisted laparoscopy over laparotomy (15). Inclusion criteria and primary outcomes were similar to those of the current study, but our study compared a broader ﬁeld of laparoscopy. We did not compare health care costs, and the results of the previous study may not be universal because medical expenses may vary depending on insurance system of each country. The strength of the current study is that analysis was performed for data of the procedure (PALND) in a speciﬁc group of patients (stage I–II, high-risk endometrial cancer patients) April 2020 \| Volume 10 \| Article 451 5 Laparoscopic Para-Aortic Lymphadenectomy Paik et al. REFERENCES Mourits MJ, Bijen CB, Arts HJ, ter Brugge HG, van der Sijde R, Paulsen L, et al. Safety of laparoscopy versus laparotomy in early-stage endometrial cancer: a randomised trial. Lancet Oncol. (2010) 11:763–71. doi: 10.1016/s1470-2045(10)70143-1 7. Wright JD, Burke WM, Tergas AI, Hou JY, Huang Y, Hu JC, et al. Comparative eﬀectiveness of minimally invasive hysterectomy for endometrial cancer. J Clin Oncol. (2016) 34:1087–96. doi: 10.1200/jco.2015.65.3212 Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. 8. Walker JL, Piedmonte MR, Spirtos NM, Eisenkop SM, Schlaerth JB, Mannel RS, et al. Laparoscopy compared with laparotomy for comprehensive surgical staging of uterine cancer: gynecologic oncology group study LAP2. J Clin Oncol. (2009) 27:5331–6. doi: 10.1200/jco.2009.22.3248 Copyright © 2020 Paik, Baek, Kang, Jeong, Kim, Kim, Lee, Choi, Lee, Kim, Bae and Kim. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 9. Obermair A, Janda M, Baker J, Kondalsamy-Chennakesavan S, Brand A, Hogg R, et al. Improved surgical safety after laparoscopic compared to open surgery for apparent early stage endometrial cancer: results from a randomised controlled trial. Eur J Cancer. (2012) 48:1147–53. doi: 10.1016/j.ejca.2012.02.055 April 2020 \| Volume 10 \| Article 451 Frontiers in Oncology \| www.frontiersin.org 6
https://openalex.org/W3129775810	https://www.crimejusticejournal.com/article/download/852/615	English	null	Confronting the North’s South: On Race and Violence in the United States	Springer eBooks	2,018	cc-by	8,397	Abstract More than any other ‘Northern’ country, the United States is distinctive in the degree to which its social, economic, and cultural development has been entwined with the global South from the beginning: and we cannot adequately understand the state of crime and punishment in the US without taking that uniquely ‘Southern’ history into account. In this paper, I sketch some of the dimensions of one crucial reflection of that Southern legacy: the extraordinary racial disparities in the experience of violent death between African‐ Americans and Whites. These disparities contribute substantially to radically different patterns of life and death between the races, and constitute a genuine social and public health emergency. But their structural roots remain largely unaddressed; and in some respects, the prospects for seriously confronting these fundamental inequalities may be receding. Please cite this article as: Currie E (2017) Confronting the North’s South: On race and violence in the United States. International Journal for Crime, Justice and Social Democracy 6(1): 23‐34. DOI: 10.5204/ijcjsd.v6i1.382. This work is licensed under a Creative Commons Attribution 4.0 Licence. As an open access journal, articles are free to use, with proper attribution, in educational and other non‐ commercial settings. ISSN: 2202‐8005 www.crimejusticejournal.com IJCJ&SD 2017 6(1): 23‐34 www.crimejusticejournal.com IJCJ&SD 2017 6(1): 23‐34 ISSN 2202–8005 Confronting the North’s South: On Race and Violence in the United States Confronting the North’s South: On Race and Violence in the United States Elliott Currie University of California, USA; Queensland University of Technology, Australia Introduction The United States occupies a remarkable and complex place within the ‘North/South’ dichotomy. As one of the richer, and certainly most globally powerful, of the advanced industrial societies, it is undeniably part of the ‘North’. And the reach of American economic (and cultural) influence arguably shapes the world social order more than any other ‘Northern’ country. But the United States is also distinctive in the degree to which its development has been intricately entwined with the global ‘South’ from the beginning, in ways that are more intimate and fundamental than in other advanced industrial societies. As Kerry Carrington, Russell Hogg, and Maximo Sozzo (2016: 5) remind us, ‘There is no global North that is not also the product of centuries‐old interactions between regions and cultures spanning the globe’. But the American experience stands out. Many Northern countries developed colonial regimes: the United States incorporated a full‐fledged plantation‐based social order within its borders from the start, one that shared more in common with the slavery‐based societies of the wider Caribbean‐Latin American region than it did with the socio‐economic systems of other Northern countries. Migration from colonies and former colonies has shaped the experience of crime and punishment significantly in most industrial societies of the North: in the United States, the most important colony was internal, and massive migration from that internal South to the country’s North was one of the defining social dynamics of the last hundred years (Wilkerson 2011). We cannot begin to grasp either the nature or the origins of America’s outsized problem of violent crime (or of punishment) without placing that ‘Southern’ legacy in the foreground. In a sense, it is the ‘elephant in the room’ in discussions of violence in America: a looming presence that everyone is aware of, but that is rarely confronted directly. Both in its dimensions and in many of its central characteristics, the American problem of violent crime looks more like that of many Third World countries than that of other advanced industrial nations. In this paper, I explore some of those dimensions, focusing on the stunning disparities in levels of violent death and injury between African‐American and white populations. I emphasize that this focus reflects only one facet of the ‘Southern’ effect on crime and punishment in the United States. Keywords African Americans; causes of death; homicide rates; racial disparities; violence. This work is licensed under a Creative Commons Attribution 4.0 Licence. As an open access journal, articles are free to use, with proper attribution, in educational and other non‐ commercial settings. ISSN: 2202‐8005 © The Author(s) 2017 Elliott Currie: Confronting the North’s South: On Race and Violence in the United States Online version via www.crimejusticejournal.com Introduction The country shares with some others – notably Australia, New Zealand and Canada – a history of systematic and violent subordination of indigenous populations, and an ever‐ evolving pattern of migration from a wide variety of ‘Southern’ regions. It also, unlike most other ‘Northern’ countries, has a long and historically porous geographical border with a ‘Southern’ country: Mexico. Both of these realities, and others, help to shape the lineaments of crime and punishment in the United States. But here I narrow the lens to the present state of violence within African‐American communities, in part because that is where the impact of America’s unique ‘Southern‐ness’ appears in its deadliest and most catastrophic form. Let me acknowledge at the outset that some might find my characterization of this legacy as ‘catastrophic’ to be both somewhat overwrought and perhaps a little dangerous. The dominant view of the state of violent crime in the United States in the mass media, and even among many social scientists, has lately been a generally celebratory one. The ‘story’ about crime in America today is that we are experiencing ‘historic lows’ following the unprecedented ‘crime drop’ that began in the early 1990s. In contrast to the 1960s, 1970s and 1980s, ‘crime in the streets’ receded as a national political issue in the early years of the twenty‐first century (cf. Currie 2010), only re‐emerging, to some extent, in the Presidential contest of 2016. But though the dialing‐down of overheated political rhetoric about crime is surely a welcome development, the recent national complacency about crime masks the continuing concentration of staggeringly high levels of death, injury and suffering in many African‐American communities, and whitewashes stark racial disparities in the experience of violence. As I will show, these differences translate into radically divergent patterns of life and death between the races. Our tendency to gloss over these realities helps to render invisible a degree of collective trauma and needless suffering that represents not only a demonstrably avoidable injustice but also a violation of fundamental human rights. 24 IJCJ&SD © 2017 6(1) Online version via www.crimejusticejournal.com Elliott Currie: Confronting the North’s South: On Race and Violence in the United States Our tendency to accept those conditions as part of the normal backdrop of American life stands in sharp contrast to the widespread outrage that has erupted in recent years over a spate of police killings of black Americans, especially young African‐American men. Introduction That outrage has galvanized serious protest across the country and sparked a movement built around the insistence that ‘black lives matter’. Yet no such national movement has emerged in protest of the black lives lost to routine, non‐official violence in neighborhoods throughout the United States. But those two kinds of violence – the unjustified killings of black Americans by police and the tragic killings of black Americans mostly by each other – are two sides of the same coin. Both reflect the same underlying reality: the continuing state of marginalization and structural oppression faced by many black communities today, which is deeply rooted in the peculiarly ‘Southern’ character of American social and economic development. We find it easier to comprehend that connection when it comes to one side of the coin – the police killings – and therefore easier to condemn. We find it harder to make the connection between the long legacy of systematic racial oppression and high rates of violence within the black community. When the hand that holds the gun that kills a young black man belongs to a white police officer, it is easy to see the connection. When that hand belongs to another young black man, the connection is complex, often indirect, and consequently harder to grasp. And because we find it harder to make that connection we too often tend to accept the ongoing social disaster of violence in many black communities as simply part of the American landscape. Others of us may be troubled by it but – for reasons I will explore later – find it too perilous or too sensitive a subject to talk about honestly. But confronting this toxic legacy of America’s hybrid history is essential if we want to put an end to a level of communal violence that has no counterpart anywhere else in the global North. Violent death in black and white For white Americans – more precisely, for what our Census Bureau calls ‘white non‐ Hispanic’ Americans – the rate is about 2.7 per 100,000. That means that the overall homicide death rate for white Americans over the past decade and a half has averaged about one‐seventh that of black Americans (Centers for Disease Control and Prevention 2016). whites. For white Americans – more precisely, for what our Census Bureau calls ‘white non‐ Hispanic’ Americans – the rate is about 2.7 per 100,000. That means that the overall homicide death rate for white Americans over the past decade and a half has averaged about one‐seventh that of black Americans (Centers for Disease Control and Prevention 2016). That disparity may seem shocking in itself. But it is only the tip of the iceberg. The connections between race and violence in America are sufficiently profound that they overturn some of our usual expectations about violence and whom it strikes most. One of those expectations has to do with gender; another has to do with age. Consider gender first. It is a truism in criminology that homicide usually strikes men far more often than women, and that is particularly true in societies that, like the United States, have high rates of homicide overall. And when we look at how violent death has played out in recent years within different racial groups, that truism holds, up to a point. Among those 124,000 black Americans who died by homicide during this century so far, about 105,000 – 85 per cent – were male. But so strong is the effect of race that a black woman has more than half again the chance of dying by homicide as a white man. For younger black women, the situation is worse: at age 21 years, the homicide death rate for black women is double the rate for white men. Now consider age. In the United States, as in most other societies that suffer high levels of violence, it is the young who suffer it the most. And again, that relationship holds true within every racial group in America. More than half of all homicide deaths among black men in America take place among those aged 18 to 30 years. The person with the greatest likelihood of dying by violence in the United States today is a 21 year old black man. Violent death in black and white Among whites, it is a 26 year old man. So age matters, and it matters a lot. But when we put age together with race, this usually very predictable relationship is disrupted. First of all, even though it is young men who are most likely to die by violence, just how likely they are to die depends overwhelmingly on their race. In 2014, the homicide death rate among male African‐American 21 year olds was roughly 83 per 100,000, which is a rate of violent death higher than that of any country in the world except Honduras. The rate for white, non‐Hispanic 21 year olds was under five per 100,000 – roughly one‐seventeenth the black rate. Not half the black rate: not one‐fourth the black rate: one‐seventeenth the black rate (Centers for Disease Control and Prevention 2016; United Nations 2014). But it gets worse. As I said, the chance that a man will die of homicide generally declines with age. It starts to rise in the teens, peaks in the twenties for both races, and then begins a long slide downward as they go into middle age and beyond. But once again, race radically complicates the picture. And that trajectory can be seen, to be sure, within both races. An older black man is much less likely to be a homicide victim than a black youth in his twenties. But a 60 year old black man remains more than twice as likely to suffer a violent death as a white man of 26 – less than half his age – the age of highest risk for white men in America. What makes these disparities even more sobering is that the rates of violent death for white men in the United States are themselves quite high by comparison with those of men in other advanced industrial societies around the world. Put up against men in, say, Japan or Hong Kong or Germany, young white men in the United States constitute a distinctly vulnerable group. They are several times more likely to meet a violent death than men of all races together in many countries at comparable levels of economic development. So when we say that a young black man of 21 is 17 times more likely than his white counterpart to be murdered, we are comparing him with people whose risk of being murdered is itself unusually high in the developed world. Violent death in black and white Let me begin with some numbers. Everyone knows that several countries in West Africa were recently struck by an epidemic of Ebola, one of the most frightening and deadliest of diseases; an epidemic that, unsurprisingly, drew an enormous amount of attention and concern around the world. And the world drew a deep collective breath of relief when the epidemic was officially declared to be over, or nearly over, in 2015. According to the World Health Organization, something over 11,000 people died in this epidemic in the hardest‐hit countries: Sierra Leone, Guinea and Liberia. That is a lot of people. It is also less than 18 months’ worth of homicide deaths among African‐Americans in the United States. From the start of the twenty‐first century up through the year 2014 – that is, in the first 15 years of the new century – almost 124,000 black Americans lost their lives to homicide (Centers for Disease Control and Prevention 2016). That is a population the size of a substantial city, larger by several thousand than the total population of the city of Berkeley, California, where I live. If we go back just a little farther in time to look at the picture over the last 20 years, the numbers become even more astonishing. From 1995 through 2014, roughly 169,000 black Americans lost their lives to homicide. To put that very abstract number into some more tangible frame, imagine that we had lined up the entire population of a medium‐sized city and mowed them down mercilessly, killing every man, woman, and child: that is the magnitude of the slow massacre of African‐ Americans in the past two decades. Keep in mind that those 20 years were years of generally declining crime in the United States; years when it has often been said that violent crime is not the problem it used to be. The figure of about 124,000 deaths since the start of the twenty‐first century translates into a homicide death rate for the African‐American population as a whole of about 20 per 100,000 people. Again, that number, by itself, doesn’t mean much – until we compare it with the rate for 25 IJCJ&SD © 2017 6(1) Online version via www.crimejusticejournal.com Elliott Currie: Confronting the North’s South: On Race and Violence in the United States whites. Violent death in black and white And, to add to the mix, the numbers I have outlined so far are averages, for the United States as a whole. In many American states, the presence of violent death looms especially large. 26 IJCJ&SD © 2017 6(1) Online version via www.crimejusticejournal.com Elliott Currie: Confronting the North’s South: On Race and Violence in the United States In the state of Illinois, the homicide death rate for black males aged 15 to 29 was 132 per 100,000 in 2014. In Missouri, it was 142 per 100,000. Those numbers, again, may sound technical and bloodless. But the take‐away point is that homicide death rates that high are unknown anywhere on the planet outside of some of the most violent countries of the global South. You can find them in some parts of Central America. You can find them in some cities in Brazil. You can find them in some places in South Africa. You will not find them in any other advanced industrial society: not even close. What this tells us is that the experience of violent death in the black community in the United States is very, very different from the experience among whites. That difference is not subtle, and it is not just a matter of degree. We are talking about a fundamental division that profoundly affects the quality of life that people of different races can enjoy. And, again, the difference does not only apply to young black men. We focus on them the most, not surprisingly, since their level of victimization is highest. But the special experience of violent death reaches out to affect groups we may not immediately think of as being uniquely vulnerable: black women as well as black men; older black men as well as younger black men. We can see those divergent realities even more clearly if we look at the racial disparities in violent death through some other lenses. For example, the huge differences I have described in homicide death rates mean that the ranking of causes of death between the two races looks remarkably different. People die, of course, for all kinds of reasons. But the way in which those reasons line up tells us a lot about the kinds of lives that different groups lead and the kinds of risks that they have to face routinely. Violent death in black and white Looking at the pattern of causes of death – what people most often die of – provides a sort of dark window on some very basic things about our society, and about the human meaning of racial inequality in particular. In the United States, homicide becomes the leading cause of death for black males by age 15, and continues as their number one cause of death through age 35. In other words, during every one of those years from 15 to 35, the single thing black men are most likely to die of is violence. Homicide is never the leading cause of death for non‐Hispanic white men at any age. It rises to fourth place for white males for a while during early childhood, and again at ages 15 to 24. On average, for white men of all ages, homicide ranks nineteenth among causes of death. Considerably fewer white men die of homicide than die, for example, from aortic aneurisms or benign cancers. For black men overall, it is fifth, exceeded only by heart disease, cancer, unintentional injuries, and stroke. And, again, it beats out all other causes of death for younger black men between ages 15 and 35 (Centers for Disease Control and Prevention 2016). And when we say that homicide is the leading cause of death among those black young men, we are not saying that it edges out other causes: we are saying that it overwhelms them. At age 21, homicide accounts for almost half of all deaths among black men. More 21 year old black men die of homicide than die of the next 19 biggest causes of death combined. By contrast, just one in 24 white men that age dies of homicide. And once again, these racial disparities are so large and so pervasive that they bend the usual expectations about gender and homicide. Again, only about one in 24 white males who die at age 21 dies by violence, but almost one in five black women who die at that age suffer a violent death. But even these numbers understate the impact of violent death in many black communities. One way to better appreciate that impact is to look at it through still another lens: what public health researchers call ‘years of potential life lost’, or YPLL. YPLL is a measure of how much premature death results from some particular cause. Online version via www.crimejusticejournal.com Violent death in black and white You choose an endpoint – say, age 65 years – and ask how many years are lost before that age from some given cause of death, or from all causes put together. So YPLL puts together two important factors: how widespread the cause of death is – 27 IJCJ&SD © 2017 6(1) Online version via www.crimejusticejournal.com Elliott Currie: Confronting the North’s South: On Race and Violence in the United States how many people die of it – and how old they are when they die. Putting the two together provides another, especially revealing window on how death plays out in different communities. In the United States today, heart disease is the leading single cause of death: that is, ultimately, more people die of it than of any other cause. But it is not the leading cause of years of potential life lost, because people tend to die of heart disease at relatively older ages, thus losing fewer ‘potential’ years of life when they do die. Instead, the biggest culprit for YPLL is what public health statisticians call ‘unintentional injuries’, a broad category that includes accidents – notably motor vehicle accidents – and ‘poisoning’, which includes drug overdose deaths. And that is because those deaths typically happen earlier in life. If I die at 63 of a heart attack, then I have lost two years of potential life before age 65. If I die at age 15 years in a car accident, I have lost 50 years of potential life. And that explains why, for the population as a whole, unintentional injuries are the biggest single source of years of potential life lost before age 65. And this measure becomes very important in understanding the social and personal burden of violence in the United States because, like accidents, homicide strikes hardest at younger people. Again, the highest number of homicide deaths in the US is among people in their twenties. And because it hits people younger, the impact of homicide on YPLL is much greater than the simple numbers or rates of death by homicide – bad enough in themselves – would suggest. As I said, homicide is the seventeenth leading cause of death in the United States today. But it is the sixth biggest cause of years of life lost before age 65. So how does this measure look when we compare whites and blacks in the United States? Violent death in black and white For black men overall, homicide is the leading cause of years of life lost before age 65. For whites, it is ninth. One out of every six years of life that black men lose prematurely before age 65 is lost to violence. For white men, that figure is less than one in 50. During the twenty‐first century, homicide has taken more years of life from African‐American men than cancer, diabetes and stroke combined. The difference between black and white men in this respect is so great that, even though black men make up a far smaller proportion of the male population in the United States than white men do, they collectively lose far more years of life to homicide than white men. There are roughly five times as many white non‐Hispanic men in the American population as black men. But black men as a group lose almost three times as many years of life to homicide as white men do (Centers for Disease Control and Prevention 2016). And once again, the racial effect on years of life lost bends the usual expectations about gender. As I have said, only about one in 50 years of life lost prematurely among white men is lost to violence. For black women, the proportion is just about double that: one in 25. Gender does indeed make a big difference, up to a point. Black women lose far fewer years of life to violence than black men do. But race upends the relationship. Measured this way – by how many years of their lives it steals – violence looms much larger in the lives of black women than in the lives of white men. Let’s step back from the numbers and ask what all of this is telling us. The numbers are a cold and, in many ways, inadequate tool when we are trying to understand something as complex and humanly significant as the impact of violence on the lives of individuals and the fabric of communities. But what we are seeing here through the lens of YPLL is nothing less than the massive eradication of human potential: the elimination of human possibilities. It tells us something more than the simple fact that violent death strikes black Americans more often. It tells us that, because it strikes so disproportionately at the young, it erases a substantial part of the future of an entire community. Why not? Part of the reason may be that, as I said at the start, it is genuinely harder for most people to grasp the connection between this kind of communal violence and the impact of generations of racial oppression. It is a complicated idea to grasp that those horrific levels of deadly violence are not just the result of bad choices by a handful of antisocial individuals; or of bad family upbringing or the noxious influence of violent movies or rap lyrics; or the leniency of the courts. It is harder to wrap our minds around the idea that these individual acts are the complex and often indirect result of large social forces that we often cannot see: the result of generations of blasted opportunities, of stunted chances, of bad alternatives, of shrunken possibilities for meaning as well as for work and a decent living. But those connections are very real. There is, of course, a unique story behind every one of those violent deaths. But those stories do not exist in a vacuum. They are set in particular social contexts. And those contexts are often quite predictable. But another part of the reason why our society tends to accept high levels of violent death in the black community, I suspect, is less conceptual and more political. It is that the great majority of the people whose deaths are reflected in these statistics do not come from places with any significant political influence, visibility or resources. Outside of their own neighborhoods, they are literally not ‘seen’ in any meaningful sense. Within those neighborhoods, there is often outrage and anger over the losses they have had to live with for so long. Outside them, not so much. The places that suffer the worst rates of violent death in America – as is true around the world – are by now depressingly predicable. They are invariably places that have been especially devastated by the harsh underside of America’s partly ‘Southern’ pattern of development. The highest urban homicide rate in 2013 in the United States among cities of reasonable size was suffered by the city of Gary, Indiana, where the murder rate greatly exceeds that of Kingston, Jamaica, the most violent city in one of the world’s most perennially violent countries. Gary is also a stunningly poor and heavily African‐American city. A neglected emergency I believe that these numbers constitute the metrics of an ongoing national disaster. But, as I’ve said, that disaster and the deep racial divide in the very right to life that it represents remain, to a remarkable degree, mostly a taken‐for‐granted backdrop of American life. If black Americans had the same risk of violent death as white Americans, we would have saved almost seven out of every eight black lives lost to homicide in 2014, or close to 7,000 people. Seen from the other direction, if white youth had the same risk of homicide as their black counterparts, there would have been roughly 13,000 white, non‐Hispanic, young male homicide victims that year, instead of the 787 who actually died. I think I can guarantee you that if that many young white American men had been murdered, there would have been a huge outcry and a vocal demand that something be done. Yet the violent deaths of more than 80,000 young black men over the past 20 years brought no such outrage, much less an urgent and serious search for solutions. Violent death in black and white 28 IJCJ&SD © 2017 6(1) Online version via www.crimejusticejournal.com Elliott Currie: Confronting the North’s South: On Race and Violence in the United States Online version via www.crimejusticejournal.com Elliott Currie: Confronting the North’s South: On Race and Violence in the United States Elliott Currie: Confronting the North’s South: On Race and Violence in the United States A recent survey of the neighborhood experiences of people of different races in Chicago, where homicides and gun violence generally are rising sharply as I write, drives this point home. Eighty‐ six per cent of blacks, versus 50 per cent of whites, said that it was ‘very likely’ or ‘somewhat likely’ that a young person in their neighborhood would be a victim of violent crime; and 81 per cent of blacks, versus 41 per cent of whites, believed that it was either ‘very’ or ‘somewhat’ likely that a young person in their neighborhood would go to jail (Davey and Russonello 2016). And the survey also highlights black residents’ deep sense of political marginalization in the city: only 8 per cent believed that the city’s mayor ‘cared a lot’ about people like them, while more than two‐ thirds believed that he cared either ‘not much’ or ‘not at all’. On another level, I think that, for a variety of reasons, it is often hard for people outside those communities to face up to these issues, even people who are generally concerned with social justice, and even people whose job it is to study violent crime. Part of the problem is that some people are paralyzed by the worry that, if we call attention to the scope of violent death in black communities, we will be ‘pathologizing’ those communities and feeding negative stereotypes about people of color; in particular, demonizing the young men who are both the most frequent perpetrators of violence and its most frequent victims. The concern is understandable: the negative stereotypes about blacks and crime in America (as in many other countries) are very real, and highlighting the troubles of those communities without simultaneously emphasizing their deeper structural roots is indeed both misleading and destructive. But so is ignoring tough but inescapable realities. You cannot do anything to alleviate a human tragedy if you are unwilling even to acknowledge that it exists. And in the absence of action based on being willing to look those realities in the face, people keep on dying. There is a racism of exaggeration, victim‐blaming and malevolent stereotyping, but there is also a racism of silence and a racism of denial. This look‐the‐other‐way attitude, unfortunately, is disturbingly common among social scientists. Elliott Currie: Confronting the North’s South: On Race and Violence in the United States Some of the people whose job it is to understand this problem actively deny that it exists. There has been a resurgence in the last few years, perhaps more in the United States than elsewhere, of a kind of scholarship and media commentary that downplays, or dismisses, the significance of violence as a social problem, and sometimes suggests that the plague of violence in many African‐ American communities is largely a social construction (Murakawa 2014). Some of this scholarship has been fairly good at describing the racial inequalities in our criminal justice system, and that is certainly necessary and important. But in largely dismissing the idea that violence is a real problem in a country that loses roughly 8,000 African American lives to violence every year, this perspective comes perilously close to dismissing the people who suffer from it the most: perilously close to suggesting that in fact those lives don’t matter. One result of the failure to acknowledge this emergency is that our attention is shifted away from thinking about strategies that could confront head‐on the historical and structural roots of endemic violence, toward relatively peripheral, if well‐intentioned, responses. In the face of the recent highly visible police killings of black Americans, much of the discussion has focused on narrow issues of police reform: whether police should be required to wear body cameras more often; or take sensitivity training to become more aware of their implicit biases against people of color; or develop better internal mechanisms for disciplining officers who behave badly. I am not suggesting that these measures might not be useful. But they are a long way from addressing the larger structural context of violent death in the black community. And if that is all we think about doing, we are not being serious. That lack of seriousness represents an abdication; a failure of the moral as well as the sociological imagination. Why not? Sixty per cent of its children are in poverty: it has recently averaged an official unemployment rate of just under 20 per cent – about four times the national average – which surely greatly underestimates the actual number of people without legitimate work. And it is 83 per cent black (United States Census Bureau 2016). A roughly similar picture holds for all the other major cities that consistently experience the highest homicide rates in America: cities like Newark, New Jersey; New Orleans, Louisiana; Baltimore, Maryland; St. Louis, Missouri; or Detroit, Michigan. Poor people in poor cities – just like poor people in poor countries – have very little voice and very little visibility. And so a lot of what happens in their lives – including the worst as well as the best – flies under the radar. 29 IJCJ&SD © 2017 6(1) Online version via www.crimejusticejournal.com Confronting the ‘Southern’ legacy If we really want to show that we believe black lives matter, we have to fold measures that are specifically aimed at changing police practices into a much wider strategy of tackling the sources 30 IJCJ&SD © 2017 6(1) Online version via www.crimejusticejournal.com Elliott Currie: Confronting the North’s South: On Race and Violence in the United States of black marginality in the United States. Yet to an important degree, the kind of structural and historical analysis that would make such a strategy possible has receded from the current discussion. The tendency to minimize these realities, or to accept them as a taken–for‐granted part of the American landscape, obscures the depth and historical embeddedness of the forces that have created the ‘Third World’ levels of violence that continue to plague African‐American communities in the age of the ‘great American crime decline’. And it obscures the degree to which those forces are the product of specific choices: choices deliberately made or heedlessly not made; choices that could have been made differently; choices that, at best, have failed to address the special needs created by the legacy of our ‘Southern’ pattern of development and, at worst, have actively perpetuated that legacy. I won’t go into detail about those choices here, but the historical record is unambiguous. The context of post‐Civil War communal violence in the black communities of the United States begins with the retreat of the federal government from its embryonic efforts at racially equitable economic development and land reform in the era of Reconstruction. That enormously fateful choice left Southern blacks landless and without stable sources of livelihood, and rendered them dependent on the shifting labor needs of white employers, who rapidly restored a system of labor discipline and social subordination that was remarkably akin to pre‐Civil War slavery (Foner 2015). That retrogression was enforced both through escalating private violence by whites and through the development of a criminal justice system geared explicitly toward mass intimidation and exploitation of the African‐American population as a whole. That system notably included the practice of convict leasing, through which local authorities routinely rounded up tens of thousands of blacks on vagrancy charges and other minor and/or vaguely defined offenses and sold them to private employers to work off the resulting fines, a practice that, in turn, was passively enabled by the persistent hands‐off response of federal authorities (Blackmon 2009; Litwack 1999). Confronting the ‘Southern’ legacy The degree to which this decades‐long pattern of systemic state and private violence is implicated in the current crisis of violent crime within black communities is rarely discussed by criminologists, but it had two major effects that have resonated to this day.1 This specifically ‘Southern’ reign of public and private violence and repression was a central pillar of the successful social and economic subordination of African‐Americans after the formal end of slavery, brutally enforcing harsh racial disparities in poverty and economic insecurity that, in the absence of national action on a scale sufficient to alter it, have been a defining feature of the African‐ American experience, North and South, ever since. At the same time, it arguably provided the template for the vast expansion of mass incarceration in America after the early 1970s, by modeling the deployment of the justice system as a key mechanism of control of entire populations rather than a response to individual criminality. As Douglas Blackmon (2009: 6) puts it in his stunning history of the ‘re‐enslavement’ of African Americans from the Civil War to World War II, it is essential to acknowledge the ‘effects of cycle upon cycle of malevolent defeat, of the injury of seeing one generation rise above poverty only to be indignantly crushed, of the impact of repeating tsunamis of violence and obliterated opportunities on each new generation’. ‘Southern’ resistance also limited the beneficial impact of key progressive legislation in the 1930s that established much of America’s still fairly rudimentary system of public social supports for working people and the deeply disadvantaged. The resistance of Southern employers kept most agricultural and domestic workers – categories that at the time included a majority of employed African‐Americans – out of the Social Security program for retirees and the unemployment insurance system for the jobless for decades. Similar pressure helped to keep other public assistance benefits abysmally low, because of the fear that decent benefits would cause blacks to spurn work at rock‐bottom wages (Brown 1999; Fox 2012), again helping to cement enduring racial differences in poverty and economic insecurity that continue to shape the conditions of life in black communities. Confronting the ‘Southern’ legacy 31 IJCJ&SD © 2017 6(1) Online version via www.crimejusticejournal.com Elliott Currie: Confronting the North’s South: On Race and Violence in the United States Still another crucial point of choice came at the end of World War II when, with the experience of the Great Depression still fresh, progressive legislators put forward a visionary proposal for full employment, which among other things committed the federal government to guaranteeing steady work to everyone who wanted a job (Currie 1976). That legislation faced fierce and largely successful resistance from employers and their representatives, including Southern congressmen whose constituents often relied on the presence of a large pool of precarious and intimidated black workers. Had the bill passed in anything like its original form, post‐war economic development in the United States would have been radically transformed, as the gains from rapidly growing productivity would have been far more widely and equitably shared. Instead, the tacit encouragement of high levels of joblessness, coupled with the precarious condition of black workers in a harshly segregated job structure, ensured that even a period of post‐war prosperity was marked by the growing economic marginalization of many black Americans, especially the young, and by the ‘hardening’ of a pattern of economic expendability and social abandonment. All of these processes were compounded by a somewhat more subtle and broadly political effect of America’s ‘Southerness’. The relative weakness of organized labor in the United States is widely understood to be an important part of the explanation for the distinctive underdevelopment of ‘social‐democratic’ policies in America versus most other advanced Northern countries. That difference affects nearly every aspect of American life, from the lack of accessible and affordable health care to elevated rates of child poverty and, not least, levels of violent crime. And though the sources of the relative weakness of the American labor movement are complex, one part of the explanation surely involves the historical ability of American employers and white elites generally to weaken class solidarity by pitting white workers against black, a practice that was a particularly prominent feature of the post‐Civil War industrialization of the American South. Online version via www.crimejusticejournal.com Confronting the ‘Southern’ legacy There is a growing and remarkably nonpartisan sentiment in favor of reining in the worst excesses of mass incarceration and racially 32 IJCJ&SD © 2017 6(1) Online version via www.crimejusticejournal.com Elliott Currie: Confronting the North’s South: On Race and Violence in the United States targeted policing. And there is a much broader willingness, especially among the young, to confront the meaning and consequences of the country’s long legacy of racial oppression. None of this makes the task easy, because there are equally powerful forces working to perpetuate the conditions that undermine opportunity, dignity and security in many black communities. I think we have to acknowledge the hard reality that, from the perspective of too many of the people now running the economy and the polity in the United States, black lives actually don’t matter, at least, not all black lives, and not very much. They don’t matter very much because from the standpoint of the imperatives of our heedless global economy they are relatively expendable. That is why, for example, we are quietly willing to tolerate levels of black youth joblessness that virtually guarantee that places like Gary or Baltimore or Detroit will continue to have the homicide numbers they do. If we want to enduringly end the routine and massive loss of black lives to violence – and for that matter to preventable diseases and substance abuse and all the other well‐known and exhaustively catalogued destroyers of black lives – we will have to challenge those imperatives. Whether we will rise to that challenge is, of course, an open question. The crisis of black economic exclusion and expendability is, after all, hardly a new problem. More than a century ago the great African‐American scholar and activist WEB DuBois (1967) wrote about crime and race in Philadelphia, the city where he conducted one of the first really wide‐ranging empirical sociological studies ever done in the United States. DuBois fully acknowledged that the problem of crime in Philadelphia’s black community was real and serious. But he also insisted that it was fixable, and that our failure to fix it when we had the resources to do so amounted to a historic default, one that we would be judged for in the future. Confronting the ‘Southern’ legacy This has meant that the kinds of deep structural measures that might have alleviated the marginalization and dispossession of black Americans – adequate income support, universal medical care, an inclusive housing policy, a commitment to full employment – perennially lacked a strong enough constituency to make real headway, even in relatively progressive periods in recent American history. The trajectory of marginality and violence in African‐American communities reflects a history of crucial opportunities not taken. Reversing that trajectory requires facing up to the magnitude of the consequences of this legacy, and advocating for social policies on the scale and depth that this history necessitates. Obviously, that is a formidable task and one that has been made more difficult by several adverse developments in the United States in recent years. One is the continuing shrinkage of opportunities for steady and sustaining work – the product of global wage competition, technological change, the ongoing decimation of public sector employment and, especially since the recession of the early twenty‐first century, the reluctance of private employers to invest in job creation. Another is the devastating impact of mass incarceration – again, in a real sense a ‘Southern’ invention – which has compounded the problem of mass joblessness and sealed the economic redundancy of great numbers of black Americans. And a third is the startling rise of the political power of the extreme Right, which, not coincidentally, draws much of its greatest support from the regional heartland of the traditional South. It would be difficult to overstate how much the national shift to the Right has contributed to the continued precariousness of African‐American lives, both by imposing harsh austerity measures on already reeling communities and, as importantly, by pushing crucially needed social and economic policies off the political agenda of both major political parties. But there are also strongly positive developments in American political culture that could shift the political balance in ways that bode well for the prospects of finally challenging the deeply entrenched forces that contribute to endemic violence in black communities. There is a budding, if still fragmented, movement against the extremes of economic inequality and heedless austerity that we have encouraged in the twenty‐first century. possible in the twentieth century, then our civilization is vain and the republic is a mockery and a farce. (DuBois 1967: 388) possible in the twentieth century, then our civilization is vain and the republic is a mockery and a farce. (DuBois 1967: 388) Correspondence: Professor Elliott Currie, Criminology, Law and Society, University of California, Irvine, California 92697, USA; Adjunct Professor, School of Justice, Faculty of Law, Queensland University of Technology, 2 George Street, Brisbane 4000 QLD, Australia. Email: ecurrie@uci.edu 1 The distinctive pattern of private and official violence in the American South also probably had a shaping influence on the overall American pattern of gun ownership and gun violence, which also sharply distinguishes the United States from other ‘Northern’ nations. That connection is sufficiently complex to require a treatment of its own. Confronting the ‘Southern’ legacy He said: Other centuries looking back upon the culture of the nineteenth would have a right to suppose that, if in a land of freemen, eight millions of human beings were found to be dying of disease, the nation would cry with one voice, ‘heal them!’ if they were staggering on in ignorance, it would cry, ‘train them!’ If they were harming themselves and others by crime, it would cry, ‘guide them!’. He went on to say that ‘such cries are heard and have been heard in the land; but it was not one voice’. And ‘its volume has ever been broken by counter‐cries and echoes; “let them die!” “train them like slaves!” “let them stagger downward!”’. DuBois wrote those words in 1899, and his point surely rings even more true more than one hundred years later. We now have far more resources at our disposal to train and to heal, and we have learned a lot more, since DuBois wrote, about what has gone wrong and what needs to be done. We can and do argue about specifics: about what kinds of social investments would give us the most return; about which toxic policies that now contribute to the ongoing social and economic impoverishment of black communities most need changing. But no one who has been paying attention can seriously claim that we don’t broadly understand some things we can do that would make a very substantial difference in the prospects of those communities and hence in the level of violence, fear and grief that they now suffer. DuBois put it this way back at the close of the nineteenth century: he acknowledged that the problems were ‘difficult, extremely difficult’. But he insisted that ‘they are such as the world has conquered before and can conquer again’. And taking them on was not just a matter of altruism, but a ‘battle for humanity and human culture’. He said: If in the heyday of the greatest of the world’s civilizations, it is possible for one people … [to] slowly murder [another] by economic and social exclusion until they disappear from the face of the earth – if the consummation of such a crime be 33 IJCJ&SD © 2017 6(1) Online version via www.crimejusticejournal.com Elliott Currie: Confronting the North’s South: On Race and Violence in the United States References Blackmon D (2008) Slavery by Another Name: The Re‐Enslavement of Black Americans from the Civil War to World War II. New York: Anchor Books. Brown M (1999) Race, Money, and the American Welfare State. Ithaca, New York: Cornell University Press. Carrington C, Hogg R and Sozzo M (2016) Southern riminology. British Journal of Criminology 56(1): 1‐20. DOI 10.1093/bjc/azv083. Centers for Disease Control and Prevention (2016) Injury Prevention and Control: Data and Statistics: Fatal Injury Data. Available at http://www.cdc.gov/injury/wisqars/fatal.html (accessed 1 February 2016). Currie E (1977) The politics of jobs: Humphrey‐Hawkins and the dilemmas of full employment. Socialist Review 7: 93‐114. Currie E (2009) An unchallenged crisis: The curious disappearance of crime as a public issue in the United States. Criminal Justice Matters 75(1):24‐26. DOI: 10.1080/09627250802699723. Davey M and Russonello D (2016) Chicago survey finds many see city gone awry. New York Times, 30 May. Available at http://www.nytimes.com/2016/05/07/us/chicago‐racial‐ divisions‐survey.html?_r=0 (accessed 30 June 2016). Du Bois WEB (1967) [1899] The Philadelphia Negro: A Social Study. New York: Shocken B Du Bois WEB (1967) [1899] The Philadelphia Negro: A Social Study. New York: Shocken Books. Foner E (2015) A Short History of Reconstruction Updated Edition New York: Harper Foner E (2015) A Short History of Reconstruction, Updated Edition. New York: Harper. Fox C (2012) Three Worlds of Relief. Princeton, New Jersey: Princeton University Press. Litwack L (1999) Trouble in Mind: Black Southerners in the Age of Jim Crow. New York: Vintage Books. Murakawa N (2014) The First Civil Right: How Liberals Built Prison America. New York: Oxford University Press. United Nations (2014) Global Study on Homicide. Available at https://www.unodc.org/gsh (accessed 15 May 2015). United States Census Bureau (2016) ACS Data Tables on American FactFinder. Washington, DC: US Department of Commerce. Available at http://www.census.gov/acs/www/data/data‐ tables‐and‐tools/american‐factfinder (accessed 30 May 2016). Wilkerson I (2011) The Warmth of Other Suns: The Epic Story of America’s Great Migration. New York: Vintage. 34 IJCJ&SD © 2017 6(1) Online version via www.crimejusticejournal.com
https://openalex.org/W2968178677	https://www.zora.uzh.ch/id/eprint/175230/1/Catani2019_Article_Soft-gluonEffectiveCouplingAnd.pdf	English	null	Soft-gluon effective coupling and cusp anomalous dimension	European physical journal. C, Particles and fields	2,019	cc-by	10,538	Zurich Open Repository and Archive University of Zurich University Library Strickhofstrasse 39 CH-8057 Zurich www.zora.uzh.ch Year: 2019 g y p Catani, Stefano; de Florian, Daniel; Grazzini, Massimiliano (2019). Soft-gluon effective coupling and cusp anoma- lous dimension. European Physical Journal C - Particles and Fields, 79(8):685. DOI: https://doi.org/10.1140/epjc/s10052-019-7174-9 Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 Abstract We consider the extension of the CMW soft- gluon effective coupling [1] in the context of soft-gluon resummation for QCD hard-scattering observables beyond the next-to-leading logarithmic accuracy. We present two proposals of a soft-gluon effective coupling that extend the CMW coupling to all perturbative orders in the MS cou- pling αS. Although both effective couplings are well-deﬁned in the physical four-dimensional space time, we examine their behaviour in d = 4 −2ǫ space time dimensions. We uncover an all-order perturbative relation with the cusp anomalous dimension: the (four dimensional) cusp anoma- lous dimension is equal to the d-dimensional soft-gluon effective coupling at the conformal point ǫ = β(αS), where the d-dimensional QCD β-function, β(αS)−ǫ, vanishes. We present the explicit expressions of the two soft-gluon cou- plings up to O(α2 S) in d dimensions. In the four-dimensional case we compute the two soft couplings up to O(α3 S). For one of the two couplings, we conﬁrm the O(α3 S) result previously presented by other authors. For the other coupling, we obtain the explicit relation with the cusp anomalous dimension up to O(α4 S). We comment on Casimir scaling at O(α4 S). Real emission has a logarithmic spectrum for radiation that is soft and/or collinear to the direction of the observed hard jets (partons). This spectrum produces perturbative con- tributions that have at most two powers of logarithms for each power of the QCD coupling αS. In the case of many observ- ables, the double-logarithmic (DL) terms can be resummed to all orders in αS in exponentiated form. For most of these observables resummation can be extended in exponentiated form to higher (or arbitrary) logarithmic orders. This feature depends on the hard-scattering process and, especially, on the speciﬁc kinematical properties of the observable under consideration. In the case of observables that fulﬁl exponen- tiation, it is natural to device a resummed perturbative expan- sion by systematically organizing the exponent in classes of subsequent logarithmic accuracy: leading logarithmic (LL) terms, next-to-leading logarithmic (NLL) terms, next-to- next-to-leading logarithmic (NNLL) terms and so forth. The explicit computation and resummation of Sudakov- type logarithms can be performed by using both traditional perturbative QCD methods and techniques based on Soft CollinearEffectiveTheory.Verymanyobservablesarenowa- days known up to NLL accuracy (see, e.g., the reviews in Refs. a e-mail: grazzini@physik.uzh.ch Soft-gluon effective coupling and cusp anomalous dimension Stefano Catani1, Daniel de Florian2, Massimiliano Grazzini3,a 1 INFN, Sezione di Firenze and Dipartimento di Fisica e Astronomia, Università di Firenze, Sesto Fiorentino, 50019 Florence, Italy 2 International Center for Advanced Studies (ICAS), ECyT-UNSAM, Campus Miguelete, 25 de Mayo y Francia, 1650 Buenos Aires, Argentina 3 Physik Institut, Universität Zürich, 8057 Zurich, Switzerland Catani, Stefano ; de Florian, Daniel ; Grazzini, Massimiliano Posted at the Zurich Open Repository and Archive, University of Zurich ZORA URL: https://doi.org/10.5167/uzh-175230 Journal Article Published Version The following work is licensed under a Creative Commons: Attribution 4 Posted at the Zurich Open Repository and Archive, University of Zurich ZORA URL: https://doi.org/10.5167/uzh-175230 Journal Article Published Version Originally published at: Catani, Stefano; de Florian, Daniel; Grazzini, Massimiliano (2019). Soft-gluon effective coupling and cusp anoma- lous dimension. European Physical Journal C - Particles and Fields, 79(8):685. DOI: https://doi.org/10.1140/epjc/s10052-019-7174-9 Eur. Phys. J. C (2019) 79:685 https://doi.org/10.1140/epjc/s10052-019-7174-9 Letter Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 (1) over the observable-dependent phase space produces large DL terms (in the vicinity of the exclusive boundary) and infrared diver- gent contributions that are cancelled by one-loop virtual- radiation effects. In the case of Sudakov sensitive observ- ables that fulﬁl exponentiation, DL resummation is achieved by simply using dwDL i as integration kernel in the exponent of the observable-dependent Sudakov form factor. The intensity of soft-gluon radiation in Eq. (1) is CiαS/π. The NLL resummation of the contributions from soft and collinear radiation is obtained (see, for instance, Eqs. (10), (12) and (26) in Ref. [6], or Eqs. (2.16) and (2.29) in Ref. [7]) by using the DL kernel of Eq. (1) and simply replacing the intensity of the soft-gluon coupling as follows We recall that the CMW coupling plays a role in two other contexts directly or indirectly related to Sudakov resumma- tion. The coupling αC MW S can be used in Monte Carlo event generators (see, e.g., Ref. [10]) to improve the logarithmic accuracy of corresponding parton shower algorithms [1]. The dispersive approach to power-behaved terms in QCD hard processes [11,12] uses αC MW S to combine contributions from the low-momentum (non-perturbative) region with perturba- tive contributions at next-to-leading order (NLO) in the high- momentum region. The extension of αC MW S beyond NLL accuracy can be useful also for higher-order studies within these two contexts. Ci αS π →AC MW i (αS(q2 T )) = Ci αC MW S (q2 T ) π = Ci αS(q2 T ) π 1 + αS(q2 T ) 2π K , (2) (2) where αC MW S is the CMW coupling [1] and αS(μ2) is the QCD running coupling at the renormalization scale μ in the MS renormalization scheme. The value of the coefﬁcient K in Eq. (2) is (nF is the number of massless-quark ﬂavours) where αC MW S is the CMW coupling [1] and αS(μ2) is the QCD running coupling at the renormalization scale μ in the MS renormalization scheme. The value of the coefﬁcient K in Eq. (2) is (nF is the number of massless-quark ﬂavours) The outline of the paper is as follows. We ﬁrst recall (in a sketchy way) the role of the CMW coupling in NLL resummed calculations. Then we introduce two deﬁnitions of the soft-gluon effective coupling at arbitrary perturba- tive orders. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 At the lowest perturbative order, the probability of radia- tion of a single soft gluon that is soft (ω ≪E) and collinear (θ ≪1) to the direction of a massless hard parton is given by the well-known DL spectrum dwDL i = Ci αS π dω ω dθ2 θ2 ≃Ci αS π dz 1 −z dq2 T q2 T , (1) (1) A relevant feature of the NLL results that we have brieﬂy recalled is that the ‘dominant’ (soft and collinear) part of the NLL kernel of the generalized Sudakov form factor is obtained simply and in complete form through the use of the QCD coupling αC MW S [1] in the Catani–Marchesini–Webber (CMW)scheme(orbremsstrahlungscheme).TheCMWcou- pling αC MW S has the meaning of an effective (physical) cou- pling for inclusive radiation of soft and collinear gluons. The purpose of the present paper is to extend the deﬁnition of αC MW S beyond NLL accuracy. A deﬁnition of such exten- sion has been proposed in Ref. [4]. Moreover, the authors of Ref. [4] present the relation between the effective coupling and the customary MS renormalized coupling up to O(α3 S), and they embody the effective coupling in the context of an explicit formulation of NNLL resummation for generic two- jet observables in e+e−annihilation [8,9]. We think that there is no unique extension of αC MW S beyond NLL accu- racy. By unique extension, we mean an extension with the sameuniversalityfeaturesasthoseofαC MW S atNLLaccuracy (we postpone additional comments on this). Therefore, in the present paper, besides considering the deﬁnition of Ref. [4], we propose a different deﬁnition of the soft-gluon effective coupling, and we present some results for both deﬁnitions. where ω is the energy of the soft gluon, E is the energy of the radiating hard parton and θ is the gluon emission angle. To DLaccuracy,dwDL i canbeequivalentlyexpressedintermsof the longitudinal-momentum fraction 1−z (1−z ≃ω/E) and transverse momentum qT (qT ≃ωθ) of the soft gluon. The subscript i denotes the type of radiating parton (i = q, ¯q, g), and Ci is the corresponding quadratic Casimir coefﬁcient. We have Ci = CF if i = q (or i = ¯q)) and Ci = CA if i = g, with CF = (N 2 c −1)/(2Nc) and CA = Nc in SU(Nc) QCD with Nc colours. The integration of the spectrum in Eq. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 [2,3] and references therein for a large, though still incomplete, list of NLL results), and several observables are known to NNLL or higher logarithmic accuracy (a list of results can be found in Ref. [4]). A well known feature of QCD is that perturbative computa- tions of hard-scattering processes are sensitive to soft-gluon effects. These effects manifest themselves in hard-scattering observables that are evaluated close to the exclusive bound- ary of the phase space. In such kinematical conﬁgurations, real-radiation contributions in the inclusive ﬁnal state are stronglysuppressedandtheycannotbalancevirtual-radiation effects (which are always kinematically allowed). The unbal- ance leads to large logarithmic radiative corrections (the argument of the logarithms is the distance from the exclu- sive boundary). These contributions are often refereed to as logarithmically-enhanced terms of Sudakov type. Direct inspection of NLL resummed results shows that they have a high degree of universality, with general struc- turesandingredientsthathavea‘minimal’dependenceonthe hard-scattering process and on the speciﬁc observable to be treated. Roughly speaking, the resummed logarithmic contri- butions are embodied in a ‘generalized’ Sudakov form factor and they are combined with ‘hard’ (non-logarithmic) factors, which are computable at ﬁxed perturbative orders. The uni- versality structure of NLL resummation is evident in the con- text of process-independent and observable-independent for- mulations of resummation that have been explicitly worked 12 3 3 685 Page 2 of 11 Eur. Phys. J. C (2019) 79 :685 685 Page 2 of 11 out [5–7] for large classes of (properly speciﬁed) observ- ables. In particular, within such formulations, the Sudakov form factor is obtained by integration (over an observable- dependent phase space) of a universal kernel that is explicitly evaluated up to NLL accuracy. In view of these NNL results, progress is being carried out [8,9] to extend such observable- independent resummation program to NNLL accuracy. Obvi- ously, an improved understanding of NNLL contributions and their possible universality features is also relevant in the context of resummed calculations for speciﬁc observables, independently of any observable-independent treatment. will appear in a separate publication). The results include an all-order relation with the cusp anomalous dimension and perturbative expressions up to O(α4 S). Finally, we conclude with a summary and some general comments. 123 Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 (2) and (3). One may be tempted to con- clude that the cusp anomalous dimension provides a sensible deﬁnition of a physical (though effective) soft-gluon cou- pling beyond O(α2 S). The equivalence between Ai(αS) and soft-gluon coupling, however, cannot hold in general. Indeed Ai(αS) depends on the factorisation scheme of collinear sin- gularities, while the physical coupling should not. Pii(αS; z) = 1 1 −z Ai(αS) + · · · , (z < 1), (4) (4) g p y p g We add more comments on this point, since there are conceptual analogies (and differences) between the soft- collinear part of the Sudakov kernel and the soft limit of the DGLAP kernel. The DGLAP kernel is related to the proba- bility of correlated emission of collinear partons with com- parable values of transverse momenta (independent collinear emission is instead taken into account through the perturba- tive iteration of the kernel). To obtain the DGLAP kernel, the transverse momenta are integrated up to some value of the evolution (or factorization) scale. The transverse-momentum integral is collinear divergent in the low-momentum region: within the MS factorization scheme, the divergences are han- dled by using dimensional regularization in d = 4−2ǫ space time dimensions, and the DGLAP kernel is deﬁned as the coefﬁcient of the ensuing 1/ǫ pole [see related comments after Eq. (22)]. This is an unphysical procedure, although it is perfectly well deﬁned for factorization purposes (a differ- ent factorization procedure would lead to a different DGLAP kernel). In contrast, the qT integration of the Sudakov ker- nel does not lead to collinear divergences since the low-qT region is ‘physically’ regularized by the deﬁnition of the measured observables. Nonetheless, the equality between the where the dots on the right-hand side denote terms that are less singular than (1 −z)−1 (we have also neglected contact terms, proportional to δ(1 −z), of virtual origin). The soft behaviour in Eq. (4) also applies to the collinear evolution of the parton fragmentation functions. The perturbative function Ai(αS) in Eq. (4) is usually called (light-like) cusp anomalous dimension, since it can also be related to the renormalization of cusp singularities of Wilson loops [22,23]. In the context of our discussion, Ai(αS) directly refers to the soft limit in Eq. (4), indepen- dently of any relations with Wilson loop renormalization. The perturbative expansion of Ai(αS) reads Ai(αS) = ∞ n=1 αS π n A(n) i . Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 The fourth-order coef- ﬁcient A(4) i is known in approximate numerical form [26,27] (the calculation in full analytic form is under completion), and a ﬁrst numerical estimate of A(5) q has been presented recently[28].BydirectinspectionofEqs.(6)–(8)wenotethat the dependence on i (the type of radiating parton) of the per- turbative function Ai(αS) is entirely speciﬁed up to O(α3 S) by the overall colour factor Ci. This overall dependence on Ci, which is customarily named as Casimir scaling relation, fol- lows from the soft-parton origin of Ai(αS) [29], and it is vio- lated at higher perturbative orders [30], starting from O(α4 S). From Eqs. (5) to (7) we see that, up to the second per- turbative order, Ai(αS) coincides with the CMW coupling AC MW i (αS) in Eqs. (2) and (3). One may be tempted to con- clude that the cusp anomalous dimension provides a sensible deﬁnition of a physical (though effective) soft-gluon cou- pling beyond O(α2 S). The equivalence between Ai(αS) and soft-gluon coupling, however, cannot hold in general. Indeed Ai(αS) depends on the factorisation scheme of collinear sin- gularities, while the physical coupling should not. where ζk is the Riemann ζ-function. The fourth-order coef- ﬁcient A(4) i is known in approximate numerical form [26,27] (the calculation in full analytic form is under completion), and a ﬁrst numerical estimate of A(5) q has been presented recently[28].BydirectinspectionofEqs.(6)–(8)wenotethat the dependence on i (the type of radiating parton) of the per- turbative function Ai(αS) is entirely speciﬁed up to O(α3 S) by the overall colour factor Ci. This overall dependence on Ci, which is customarily named as Casimir scaling relation, fol- lows from the soft-parton origin of Ai(αS) [29], and it is vio- lated at higher perturbative orders [30], starting from O(α4 S). Since the CMW Sudakov kernel refers to soft and collinear radiation, it can be viewed as obtained by considering the soft limit of multiple collinear radiation. In this respect it is natural to compare it with the DGLAP kernel [19,20] that controls the collinear evolution of the parton distribu- tion functions (PDFs). In the soft limit, z →1, the ﬂavour diagonal DGLAP kernel Pii(αS; z) (1−z is the longitudinal- momentum fraction that is radiated in the ﬁnal state) has the following behaviour [21]: S From Eqs. (5) to (7) we see that, up to the second per- turbative order, Ai(αS) coincides with the CMW coupling AC MW i (αS) in Eqs. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 We illustrate various perturbative results for both effective couplings, and we present some brief comments on their derivation (details about the derivation of the results K = 67 18 −π2 6 CA −5 9nF, (3) (3) as it turned out since early works on NLL resummation of several observables [13–16]. 123 123 Eur. Phys. J. C (2019) 79 :685 Page 3 of 11 685 e 3 of 11 685 685 A(1) i = Ci, (6) A(2) i = 1 2 K Ci, (7) A(3) i = Ci 245 96 −67 216π2 + 11 720π4 + 11 24ζ3 C2 A + −209 432 + 5 108π2 −7 12ζ3 CA nF + −55 96 + 1 2ζ3 CF nF − 1 108n2 F , (8) A(1) i = Ci, (6) A(2) i = 1 2 K Ci, (7) A(3) i = Ci 245 96 −67 216π2 + 11 720π4 + 11 24ζ3 C2 A + −209 432 + 5 108π2 −7 12ζ3 CA nF + −55 96 + 1 2ζ3 CF nF − 1 108n2 F , (8) Two effects are embodied in the DL kernel dwDL i through the replacement in Eq. (2). The QCD coupling αS is evalu- ated at the scale of the soft-gluon transverse momentum qT [17,18]: this accounts for the resummation of the LL terms. The gluon coupling acquires a correction of O(α2 S) (which is controlled by the coefﬁcient K in the MS renormalization scheme): this produces the resummation of NLL terms. Since the replacement takes place in the exponent of the Sudakov form factor, it is produced by the correlated radiation of soft partons (both two soft gluons and a soft q ¯q pair), whereas the independent emission of soft gluons is taken into account through the exponentiation. We also note that AC MW i (αS) is an effective coupling at the inclusive level, since it is obtained by integrating over the momenta of the ﬁnal-state correlated partons. The coupling AC MW i (αS) refers to radiation that is both soft and collinear. The Sudakov form factor includes other NLL terms due to soft wide-angle (i.e., non-collinear) radiationandhard(i.e.,non-soft)collinearradiation:wepost- pone some comments on these terms. (6) (8) where ζk is the Riemann ζ-function. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 The deﬁnitions are AT,i(αS(μ2); ǫ) = 1 2 μ2 ∞ 0 dm2 T dk2 T δ(μ2 −k2 T ) wi(k; ǫ), (13) A0,i(αS(μ2); ǫ) = 1 2 μ2 ∞ 0 dm2 T dk2 T δ(μ2 −m2 T ) wi(k; ǫ), (14) αu S μ2ǫ 0 Sǫ = αS(μ2 R) μ2ǫ R Z(αS(μ2 R); ǫ), Sǫ = (4π)ǫ e−ǫγE , (9) (9) (13) where μ0 is the dimensional regularization scale, μR is the renormalization scale and γE is the Euler number. The renor- malization function Z(αS; ǫ) is (14) Z(αS, ǫ) = 1 −αS β0 ǫ + α2 S β2 0 ǫ2 −β1 2ǫ + O(α3 S), (10) where AT,i(αS; ǫ = 0) corresponds2 to the soft coupling of Ref. [4]. (10) Thedeﬁnitions inEqs. (13) and(14) differ onlyinthekine- matical variable that is kept ﬁxed in the integration procedure over k: AT,i(αS(μ2); ǫ) is deﬁned at ﬁxed value kT = μ of the transverse momentum, while A0,i(αS(μ2); ǫ) is deﬁned at ﬁxed value mT = μ of the transverse mass. In the right- hand side of Eqs. (13) and (14), the factor μ2 is introduced for dimensional reasons (so that Ai is dimensionless) and the factor 1/2 takes into account the fact that the integration of wi(k; ǫ) includes the angular regions where the soft momen- tum k is collinear to the momentum of each of the two hard partons. In the deﬁnitions of Eqs. (13) and (14) the renor- malization scale μR is set to the value μR = μ. Obviously, the soft couplings AT,i(αS(μ2); ǫ) and A0,i(αS(μ2); ǫ) are renormalization group invariant quantities, so that, at the per- turbative level, they can equivalently be expressed in terms of the running coupling αS(μ2 R) and the ratio μ2/μ2 R. where β0 and β1 are the ﬁrst two perturbative coefﬁcients of the QCD β-function β(αS): β(αS) = −β0 αS −β1 α2 S + O(α3 S), (11) 12π β0 = 11CA −2nF 24π2β1 = 17C2 A −5CAnF −3CFnF. (12) (11) (12) As we have already stated, an all-order deﬁnition of soft- gluon effective coupling has been given in Ref. [4]. We use the same starting point as in Ref. [4]. We consider a generic hard-scattering process that involves only two massless hard partons, which can be either a q ¯q pair (i = q) or two gluons (i = g). 2 The function wi(k; ǫ) in Eqs. (13) and (14) and the web function in Eq. (2.25) of Ref. [4] are directly proportional, and the proportionality relation includes the overall factor (k2 T )−ǫ that makes AT,i and A0,i dimensionless in any number d of dimensions. 1 Note that this ‘probability’ is not positive deﬁnite since it refers to the correlation part of the total emission probability. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 (5) (5) where αS is the renormalized MS coupling. The perturbative coefﬁcients A(1) i , A(2) i [24,25] and A(3) i [19,20] are explicitly known. Using the MS factorization scheme for PDFs and fragmentation functions, these coefﬁcients are 12 3 Eur. Phys. J. C (2019) 79 :685 685 Page 4 of 11 685 Page 4 of 11 sion at ﬁxed total momentum k is a quantity that is infrared and collinear safe. Therefore, infrared singularities cancel in the computation of wi(k; ǫ) and, after renormalization of αS, the soft function wi(k; ǫ) is ﬁnite in the physical four- dimensional limit ǫ →0. For our subsequent purposes, we consider the general d-dimensional function wi(k; ǫ), although it is well deﬁned at ǫ = 0. cusp anomalous dimension and the CMW coupling at O(α2 S) is not completely accidental, since at this perturbative order the MS factorization procedure is equivalent to introduce a lower bound on the transverse momentum [1], which practi- cally acts as the regularization procedure that can be imple- mented through the use of a collinear safe observable. In the following we introduce the all-order deﬁnitions of two soft-gluon effective couplings, and we present some per- turbative results. The results are obtained by regularizing ultraviolet and infrared divergences (which are encountered at intermediate stages of the calculations) through analytic continuation in d = 4 −2ǫ space time dimensions. Speciﬁ- cally, we use the customary scheme of conventional dimen- sional regularization (CDR). The QCD bare coupling αu S and the renormalized running coupling αS(μ2 R) in the MS renor- malization scheme are related by the following standard def- inition: A relevant property of wi(k; ǫ) is its invariance under lon- gitudinal boosts along the direction of the momenta of the two hard partons in their centre-of-mass frame. It follows that wi(k; ǫ) actually depends only on two kinematical variables: the transverse-momentum component kT of k with respect to the direction of the radiating partons, and the transverse mass mT (m2 T = k2 T +k2). We propose the deﬁnition of two differ- ent effective couplings, AT,i(αS; ǫ) and A0,i(αS; ǫ), which measure the intensity of inclusive soft-parton radiation. 3 An equality between the soft anomalous dimension (which is related to the cusp anomalous dimension) and the d-dimensional rapid- ity anomalous dimension at the point ǫ = β(αS) is presented in Refs. [31,32]. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 (17) (17) where Ai is equivalently AT,i or A0,i, and Pǫ is the pro- jection operator [24] that extracts the ǫ poles (in MS form) of the function of αS(μ2 F) and ǫ in the curly bracket. The equivalence between Eqs. (21) and (22) can be proven by using some d-dimensional technicalities. We would like to point out that the relation between Ai and Ai as expressed in the form of Eq. (22) is in direct correspondence with our previous qualitative discussion about the relation between the Sudakov kernel and the DGLAP kernel. The soft cou- pling Ai((αS(q2 T ); ǫ) gives the intensity of the spectrum of correlated soft and collinear emission of partons with total transverse momentum qT . In the right-hand side of Eq. (22), the qT spectrum is integrated over the region from qT = 0 up to some value of the factorization scale μF. Following the MS factorization procedure, the ǫ poles that arise from the d-dimensional regularization of the collinear singularities in the region around qT ≃0 are then extracted to obtain (actu- ally, to deﬁne) the intensity Ai(αS(μ2 F)) of soft radiation in the DGLAP kernel [i.e., the cusp anomalous dimension in Eq. (4)]. Tomakeexplicitthedeﬁnitionoftheoverallnormalization of Ai(αS; ǫ) (and wi(k; ǫ)), we report the expression of the lowest-order contribution: A(1) T,i(ǫ) = A(1) 0,i (ǫ) = Ci c(ǫ), (18) (18) where c(ǫ) ≡ eǫγE Ŵ(1 −ǫ) = 1 −π2 12 ǫ2 −1 3 ζ3 ǫ3 + O(ǫ4), (19) (19) and Ŵ(z) is the Euler Ŵ-function. We note that the two soft couplings A(1) T,i and A(1) 0,i are exactly equal at the lowest per- turbative order. This equality simply follows from the fact that the lowest-order contribution to wi(k; ǫ) is proportional to δ(k2) = δ(m2 T −k2 T ). We also note [see Eq. (19)] that the ǫ dependence of A(1) i (ǫ) starts at O(ǫ2) (i.e., the coefﬁ- cient A(1;1) i at O(ǫ) vanishes). This mild ǫ dependence is of entirely‘kinematical’origin(itarisesfromthed-dimensional phase space), since (due to helicity conservation) the dynam- icsofsoft-gluonradiationdoesnotproduceanyǫ dependence at the lowest perturbative order. Equation (21) relates3 the cusp anomalous dimension to the d-dimensional soft-gluon coupling at the conformal point ǫ = β(αS), where the d-dimensional QCD β-function β(αS)−ǫ vanishes. The relation (21) is not speciﬁc of QCD, and it also applies to other gauge theories. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 We compute the probability for emitting a set of soft partons (soft gluons and soft q ¯q pairs), and we consider the function wi(k; ǫ) that gives the ‘probability’1 of correlated emission (including the corresponding virtual corrections) of an arbitrary number of soft partons with total momentum k. This function is formally deﬁned in Eq. (2.25) of Ref. [4], and it is called web function therein. R R The integration over k in Eqs. (13) and (14) is infrared and collinear safe, so that the limit ǫ →0 is ﬁnite and well deﬁned. Therefore, the soft-gluon effective couplings AT,i(αS) and A0,i(αS) in the physical four-dimensional space time are simply Contributions to wi(k; ǫ) from virtual and real radiative corrections separately lead to ultraviolet and infrared diver- gences. However, the probability of correlated soft emis- 12 3 Eur. Phys. J. C (2019) 79 :685 Page 5 of 11 685 685 One of the main results of this paper is the following all- orderrelationbetweenthecuspanomalousdimension Ai(αS) and the soft-gluon couplings: (15) AT,i(αS; ǫ = β(αS)) = A0,i(αS; ǫ = β(αS)) = Ai(αS). (21 AT,i(αS; ǫ = β(αS)) = A0,i(αS; ǫ = β(αS)) = Ai(αS). (21) (21) This relation can be derived in differents ways. A procedure that we have used consists in considering threshold resum- mation [16,33,34] for the production of high-mass systems in hadron collisions. The threshold resummed cross section is related to the evolution of the PDFs in the soft limit [see Eq. (4)]. We have applied both soft couplings in Eqs. (13) and (14) to the computation of the threshold resummed cross section and we have obtained the result in Eq. (21). The coefﬁcients of the perturbative expansion of Ai and Ai are deﬁned analogously to those in Eq. (5): Ai(αS; ǫ) = ∞ n=1 αS π n A(n) i (ǫ), Ai(αS) = ∞ n=1 αS π n A(n) i . (16) The relation in Eq. (21) can be rewritten in the following form: (16) Ai(αS(μ2 F)) = d d ln μ2 F Pǫ μ2 F 0 dq2 T q2 T Ai((αS(q2 T ); ǫ) , (22) The ǫ-expansion at the n-th perturbative order is denoted as follows The ǫ-expansion at the n-th perturbative order is denoted as follows (22) A(n) i (ǫ) = A(n) i + ∞ k=1 ǫk A(n;k) i . Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 In particular, in the We anticipate (see below) that, in the physical four- dimensional space time, both soft couplings in Eqs. (13) and (14) are equal to the CMW coupling AC MW i up to O(α2 S). Therefore, we have A(2) T,i = A(2) 0,i = A(2) i . (20) 12 3 685 Page 6 of 11 Eur. Phys. J. C (2019) 79 :685 685 Page 6 of 11 685 +CA c(2ǫ) h(ǫ) π 2 sin(πǫ) −CA c(2ǫ) π2 2 sin2(πǫ) 2 −sin2(πǫ) cos(πǫ) −2 sin(πǫ) πǫ , (23) case of N = 4 maximally supersymmetric Yang–Mills the- ory we have β(αS) = 0 and, therefore, the cusp anomalous dimension coincides with the physical (four-dimensional) soft-gluon coupling: AT (αS) = A0(αS) = A(αS). According to Eq. (21), there is a non-trivial interplay between the perturbative dependence of the cusp anoma- lous dimension and the d-dimensional dependence of the soft-gluon coupling. In particular, since the ǫ-dependence of A(1) i (ǫ) starts at O(ǫ2) [see Eqs. (18) and (19)], Eq. (21) directly implies the equivalence up to O(α2 S) [see Eq. (20)] between the cusp anomalous dimension and the four- dimensional soft-gluon coupling (or the CMW coupling). As we have already recalled, this equivalence is not com- pletely accidental [1] and, at the purely technical level, it can be viewed as a consequence of the mild O(ǫ2) dependence in Eqs. (18) and (19). (23) where where h(ǫ) = γE + ψ(1 −ǫ) + 2 ψ(1 + 2ǫ) −2 ψ(1 + ǫ), (24) and ψ(1+z) = d ln Ŵ(1+z) dz . In the case of A0,i(αS; ǫ) we ﬁnd h(ǫ) = γE + ψ(1 −ǫ) + 2 ψ(1 + 2ǫ) −2 ψ(1 + ǫ), (24) and ψ(1+z) = d ln Ŵ(1+z) dz . In the case of A0,i(αS; ǫ) we ﬁnd A(2) 0,i (ǫ) = Ci −c(ǫ) (11CA −2nF) 12 ǫ ) = Ci −c(ǫ) (11CA −2nF) 12 ǫ + c2(2ǫ) ǫ c2(ǫ) [CA(11 −7ǫ) −2 nF(1 −ǫ)] 4(3 −2ǫ)(1 −2ǫ) + CA c2(2ǫ)r(ǫ) 2(1 −2ǫ) c2(ǫ) −CA c(2ǫ) 2 ǫ2 (πǫ)2 cos(πǫ) sin2(πǫ) + π ǫ sin(πǫ) −2 c(2ǫ) c2(ǫ) , (25) The relation in Eq. (21) also states that the two d- dimensional soft couplings, AT,i(αS; ǫ) and A0,i(αS; ǫ), become equal by setting ǫ = β(αS). Starting from O(α2 S) [see Eqs. (23) and (25) below], the ǫ-dependence of the two soft couplings is very different. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 In view of this, we ﬁnd it remarkable that such a different ǫ-dependence conspires to make the coupling equal at ǫ = β(αS). Incidentally, such a different ǫ-dependence and the relation (21) imply that the two four-dimensional soft couplings, AT,i(αS) and A0,i(αS), inevitably differ starting from O(α3 S). Moreover, the differ- ence AT,i(αS) −A0,i(αS) is necessarily due to perturbative contributions that are proportional to the coefﬁcients, β0, β1 and so forth, of the QCD β-function. (25) where r(ǫ) = 2 1 + ǫ 3F2(1, 1, 1 −ǫ; 2 −2ǫ, 2 + ǫ; 1) − 1 1 −ǫ 3F2(1, 1, 1 −ǫ; 2 −2ǫ, 2 −ǫ; 1), (26) (26) and 3F2(α, β, γ ; δ, ρ; z) is the generalized hypergeometric function of the variable z. 2 The ǫ-expansion up to O(ǫ2) of the second-order expres- sions in Eqs. (23) and (25) gives In addition to be interesting for its intrinsic structure, the relation in Eq. (21) can be exploited for several different purposes. It can be used to crosscheck explicit perturbative computations of Ai(αS) and Ai(αS; ǫ). Once one the the three functions Ai, AT,i and A0,i is known at some pertur- bative order, Eq. (21) can exploited to extract information on the other two functions [in the following we explicitly make this use of Eq. (21)]. The relation (21) can also be used to obtain the cusp anomalous dimension Ai(αS) through the d- dimensional perturbative calculation of one of the two soft couplings Ai(αS; ǫ). 2 A(2) T,i(ǫ) = A(2) i + ǫ Ci CA 101 27 −11 π2 144 . Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 (29) (29) + CACFnF 17ζ3 9 + 11π4 1440 + 55π2 576 −7351 2304 + CAn2 F −179ζ3 432 + 13π4 4320 −695π2 3888 + 13819 15552 + CFn2 F −19ζ3 72 −π4 720 −5π2 288 + 215 384 + n3 F −ζ3 108 + 5π2 648 − 29 1458 . (32 This relation applies to both soft couplings AT,i and A0,i (we have omitted the corresponding subscripts T and 0), and we have also used A(1;1) i = 0 [see Eqs. (18) and (19)]. Since we have determined A(1) i (ǫ) and A(2) i (ǫ) to all orders in the ǫ- expansion, the explicit values of the coefﬁcients A(1;2) i and A(2;1) i can be directly read from Eqs. (18), (19), (27) and (28). Inserting these coefﬁcients in Eq. (29) we can explicitly relate A(3) i to the coefﬁcient A(3) i [see Eq. (8)] of the cusp anomalous dimension. We obtain the following results: (32) This fourth-order result can be used for applications to soft- gluon resummed calculations of infrared and collinear safe observables at the next-to-next-to-next-to-leading logarith- mic (N3LL) accuracy. This fourth-order result can be used for applications to soft- gluon resummed calculations of infrared and collinear safe observables at the next-to-next-to-next-to-leading logarith- mic (N3LL) accuracy. A(3) T,i = A(3) i + Ci (β0π)2 π2 12 + Ci (β0π) CA 101 27 −11 π2 144 −7ζ3 2 +nF π2 72 −14 27 , (30) A(3) 0,i = A(3) i + Ci (β0π)2 π2 12 + Ci (β0π) CA 101 27 −55 π2 144 −7ζ3 2 +nF 5 π2 72 −14 27 . (31) Knowing the result in Eq. (32) and exploiting the rela- tion in Eq. (21), we can also explicitly determine the third- order coefﬁcient A(3) 0,i (ǫ) of the d-dimensional soft coupling at O(ǫ). To illustrate the procedure, we perturbatively expand Eq. (32) in terms of the coefﬁcients A(n;k) i of Eq. (17), and we obtain (30) A(4) 0,i = A(4) i +(β0π) A(3;1) 0,i −(β0π)2 A(2;2) 0,i +(β1π2) A(2;1) 0,i +(β0π)3 A(1;3) 0,i −2(β1β0π3) A(1;2) 0,i , (33) (33) (31) where we have used A(1;1) 0,i = 0. The explicit coefﬁcients A(1;2) 0,i and A(1;3) 0,i at the ﬁrst order and A(2;1) 0,i and A(2;2) 0,i at the second order can be read from Eqs. (18), (19) and (28), respectively. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 −7ζ3 2 +nF π2 72 −14 27 +ǫ2 Ci CA 607 81 −67 π2 216 −77ζ3 36 −7 π4 120 +nF 5 π2 108 −82 81 + 7ζ3 18 + O(ǫ3), (27) A(2) 0,i (ǫ) = A(2) i + ǫ Ci CA 101 27 −55 π2 144 −7ζ3 2 +nF 5 π2 72 −14 27 +ǫ2 Ci CA 607 81 −67 π2 72 −143ζ3 36 −π4 36 +nF 5 π2 36 −82 81 + 13ζ3 18 + O(ǫ3), (28) +nF 5 π2 108 −82 81 + 7ζ3 18 + O(ǫ3), (27) A(2) 0,i (ǫ) = A(2) i + ǫ Ci CA 101 27 −55 π2 144 −7ζ3 2 +nF 5 π2 72 −14 27 +ǫ2 Ci CA 607 81 −67 π2 72 −143ζ3 36 −π4 36 +nF 5 π2 36 −82 81 + 13ζ3 18 + O(ǫ3), (28) (27) We have computed the soft function wi(k; ǫ) at O(α2 S) by combining the one-loop correction to single soft-gluon radiation [35] with the d-dimensional integration of dou- ble soft-parton radiation at the tree level [36]. Then, using Eqs. (13) and (14), we have computed the soft-gluon effec- tive couplings in d dimensions at O(α2 S), and we obtain the following results [37]. In the case of AT,i(αS; ǫ) we ﬁnd (28) A(2) T,i(ǫ) = Ci −c(ǫ) (11CA −2nF) 12 ǫ +c(2ǫ) π sin(πǫ) [CA(11 −7ǫ) −2 nF(1 −ǫ)] 4(3 −2ǫ)(1 −2ǫ) where A(2) i is given in Eq. (7). From these equations we see that the ǫ dependence of the two soft couplings A(2) T,i(ǫ) and A(2) 0,i (ǫ) is already different at O(ǫ). We also see that the 123 123 Eur. Phys. J. C (2019) 79 :685 Page 7 of 11 685 685 limit ǫ →0 of our explicit calculation at O(α2 S) leads to the equality in Eq. (20) between the two soft couplings and the CMW coupling. At the computational level the equality A(2) T,i = A(2) 0,i originates as follows. Since A(1) T,i(ǫ) = A(1) 0,i (ǫ), the value of A(2) i (ǫ) at ǫ = 0 is determined by the behaviour of the soft function wi(k; ǫ) in the region where k2 ≃0. In this region we have m2 T ≃k2 T and, therefore, the difference between the right-hand side of Eqs. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 (13) and (14) (and, hence, between the two soft couplings) is not effective. limit ǫ →0 of our explicit calculation at O(α2 S) leads to the equality in Eq. (20) between the two soft couplings and the CMW coupling. At the computational level the equality A(2) T,i = A(2) 0,i originates as follows. Since A(1) T,i(ǫ) = A(1) 0,i (ǫ), the value of A(2) i (ǫ) at ǫ = 0 is determined by the behaviour of the soft function wi(k; ǫ) in the region where k2 ≃0. In this region we have m2 T ≃k2 T and, therefore, the difference between the right-hand side of Eqs. (13) and (14) (and, hence, between the two soft couplings) is not effective. nel for soft-gluon resummation of infrared and collinear safe observables at NNLL accuracy [4]. In the case of the soft-gluon coupling A0,i(αS) we have also computed its relation with the MS coupling at O(α4 S). More precisely, we obtain an explicit relation between A(4) 0,i and the corresponding coefﬁcient A(4) i of the cusp anomalous dimension. We ﬁnd A(4) 0,i = A(4) i + Ci C3 A 121π2ζ3 288 −21755ζ3 864 + 33ζ5 4 +847π4 17280 −41525π2 15552 + 3761815 186624 + C2 AnF −11π2ζ3 144 + 6407ζ3 864 −3ζ5 2 −11π4 432 + 9605π2 7776 −15593 1944 + CACFnF 17ζ3 9 + 11π4 1440 + 55π2 576 −7351 2304 + CAn2 F −179ζ3 432 + 13π4 4320 −695π2 3888 + 13819 15552 + CFn2 F −19ζ3 72 −π4 720 −5π2 288 + 215 384 + n3 F −ζ3 108 + 5π2 648 − 29 1458 . (32 A(4) 0,i = A(4) i + Ci C3 A 121π2ζ3 288 −21755ζ3 864 + 33ζ5 4 +847π4 17280 −41525π2 15552 + 3761815 186624 + C2 AnF −11π2ζ3 144 + 6407ζ3 864 −3ζ5 2 −11π4 432 + 9605π2 7776 −15593 1944 We now present our computation of the third-order coefﬁ- cients A(3) T,i and A(3) 0,i of both four-dimensional soft couplings. To this purpose we use Eq. (11) and we perturbatively expand Eq. (21) in terms of the coefﬁcients A(n;k) i that are deﬁned in Eq. (17). We obtain A(3) i = A(3) i −(β0π)2 A(1;2) i + (β0π) A(2;1) i . Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 Therefore, by comparing Eqs. (32) and (33) we obtain Our result in Eq. (30) for the third-order coefﬁcient of the soft-gluon coupling AT,i(αS) agrees with the correspond- ing result presented in Ref. [4] [see Eqs. (3.9) and (3.11b) therein]. (3) A(3) 0,i (ǫ) = A(3) 0,i + ǫ A(3;1) 0,i + O(ǫ2), (34) Using the value of A(3) i in Eq. (8), the results in Eqs. (30) and (31) explicitly relate the four-dimensional (physical) soft-gluon effective couplings AT,i(αS) and A0,i(αS) with the MS renormalized coupling αS up to O(α3 S). This rela- tion generalizes the O(α2 S) CMW relation in Eq. (2) to the third order, and it can be used to construct the Sudakov ker- (34) with the explicit result with the explicit result A(3;1) 0,i = Ci C2 A 11π2ζ3 24 −225ζ3 8 + 9ζ5 + 121π4 4320 123 3 685 Page 8 of 11 Eur. Phys. J. C (2019) 79 :685 −4651π2 1296 + 403861 15552 + CAnF 289ζ3 72 −29π4 2160 + 2717π2 2592 −48241 7776 + CFnF 19ζ3 12 + π4 120 + 7π2 96 −1711 576 + n2 F −ζ3 18 −5π2 72 + 70 243 . (35) −4651π2 1296 + 403861 15552 + CAnF 289ζ3 72 −29π4 2160 + 2717π2 2592 −48241 7776 + CFnF 19ζ3 12 + π4 120 + 7π2 96 −1711 576 + n2 F −ζ3 18 −5π2 72 + 70 243 . (35) The presence in Eq. (36) of the quartic Casimir invari- ants violates Casimir scaling (i.e., the proportionality rela- tion Ai ∝Ci). Nonetheless A(4) i in Eq. (36) still fulﬁls a form of generalized Casimir scaling (in terms of three colour coefﬁcients that depend on i) since A(4) [2], A(4) [4A] and A(4) [4F] do not depend on the hard parton i. Setting ǫ = 0 in Eq. (36) and using Eq. (21), we obtain the colour structure of the four-dimensional soft coupling A(4) 0,i (or, analogously4, A(4) T,i) and of the cusp anomalous dimen- sion A(4) i : (35) We comment on our derivation of the result in Eq. (32). The soft-gluon effective coupling A0,i(αS; ǫ) is particularly suitable in the context of threshold resummation [16,33,34] for the production of colourless high-mass systems in hadron collisions. 4 The expression in Eq. (37) is equally valid for the soft coupling A(4) T,i through the replacement A(4) 0[2] →A(4) T [2]. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 The threshold resummed cross section for these processes is presently known in explicit form up to N3LL accuracy [38–47]. We have applied A0,i(αS; ǫ) to threshold resummation and, exploiting the known N3LL results [46], we obtain Eq. (32). A(4) 0,i = Ci A(4) 0[2] + d(4) Ai Ni A(4) [4A] + nF d(4) Fi Ni A(4) [4F], (37) A(4) i = Ci A(4) [2] + d(4) Ai Ni A(4) [4A] + nF d(4) Fi Ni A(4) [4F], (38) (37) (38) where, analogously to Eq. (36), the full dependence on the colour of the hard parton i is entirely controlled by the Casimir dependent coefﬁcients Ci, d(4) Ai /Ni and d(4) Fi /Ni. We note that, to obtain Eqs. (37) and (38) from Eq. (36), we have exploited Eq. (21) and the property that the dif- ference A(4) 0,i −A(4) i fulﬁls Casimir scaling [see Eq. (33)], since the perturbative terms A(n) 0,i (ǫ) with n = 1, 2, 3 ful- ﬁl Casimir scaling. In particular, in Eqs. (37) and (38) we have set A(4) 0[2](ǫ = 0) ≡A(4) 0[2], then we have related the ‘quadratic’ coefﬁcients of the soft coupling (A(4) 0[2])) and of the cusp anomalous dimension (A(4) [2]) through Casimir scal- ing: The result in Eq. (32) relates the fourth-order perturbative term A(4) 0,i of the soft coupling A0,i(αS) to the corresponding term A(4) i of the cusp anomalous dimension Ai(αS). Since A(4) i is not fully known in analytic form, we add some com- ments on the fourth-order results. We have examined the colour structure of soft multiparton radiation from two hard partons at O(α4 S) and, consequently, we can obtain the general colour structure of the soft func- tion wi(k; ǫ) or, equivalently [due to Eqs. (13) and (14)], the colour structure of the soft coupling. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 [4] the soft-gluon coupling AT,i has been explicitly applied to the resummation of a wide class of observables, by using the master formula in Eq. (2.45) therein. According to the notation in Eq. (2.45) of Ref. [4], NNLL terms are partly included in the exponentiated radiator R and partly assigned to the multiplicative factor δFNNLL. Within this NNLL formulation, the soft couplings AT,i and A0,i are equivalent at the practical level, since the replace- ment AT,i →A0,i in the exponentiated radiator can be compensated by a corresponding redeﬁnition of the factor δFNNLL. The equivalence of AT,i and A0,i does no longer hold if the NNLL terms are resummed in fully exponentiated form. Considering fully exponentiated logarithmic terms, we have already mentioned that the soft-gluon coupling A0,i is particularly suitable in the context of threshold resummation and related observables, and its application to other classes of observables can be investigated. By modifying the δ-function constraints in Eqs. (13) and (14), other deﬁnitions of soft- gluon effective couplings can be introduced. Such deﬁnitions can possibly be of interest for resummation purposes [i.e., independently of relations such as that in Eq. (21)] of certain class of observables. The quantitative effect on the soft coupling A(4) 0,i of the present numerical uncertainty of A(4) i is very small, since the quantitative value of A(4) 0,i turns out to be dominated by the contribution A(4) 0,i −A(4) i that we have explicitly computed in Eq. (32). To see this, we write A(4) 0,i = A(4) 0,i −A(4) i [ 1 + i ] , i ≡ A(4) i A(4) 0,i −A(4) i . (4 (41) The term i depends on nF. Using A(4) 0,i −A(4) i from Eq. (32) and A(4) i from Ref. [27] and setting nF = 5 (with Nc = 3) we obtain i(nF = 5) = −0.222(5) δiq + 4.05(4) δig × 10−2, (42) where the numbers in brackets indicate the numerical uncer- tainty (due to A(4) i [27]) of the preceding digit. Similar quan- titative results are obtained for nF = 3, 4. The term i turns out to contribute to A(4) 0,i at the level of few percents, so that a small uncertainty on A(4) i leads to a very small uncertainty on A(4) 0,i . As observed in Ref. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 C (2019) 79 :685 Page 9 of 11 685 1 685 685 A0,i(αS) = Ci αS π 1 + 0.54973 αS −1.7157 α2 S − 5.9803(3) δiq + 6.236(2) δig α3 S + O(α4 S) . (43) is absent in A(4) g of Eq. (38). Such contribution to P(4) gg (αS; z) vanishes in the soft limit, consistently with the approximate numerical result of Ref. [27]. (4) The fourth-order term A(4) i of the cusp anomalous dimen- sion is not yet known in full analytic form, although it is known with good numerical accuracy. The analytic results, which regard the coefﬁcients of various colour factors, have been obtained by using different methods: the computation of the soft limit of the DGLAP kernel [26,48,49], the fourth- order evaluation of form factors [50–55], the cusp renor- malization of Wilson loops [56–60] [as we have previously observed, the relation (21) leads to another method to com- pute Ai(αS) through the evaluation of the d-dimensional soft coupling Ai(αS; ǫ)]. In particular, the ‘quartic’ coefﬁcient A(4) [4F] in Eqs. (37) and (38) has been computed very recently [54,55]. The coefﬁcients of the remaining colour factor con- tributions to A(4) i have been evaluated in approximate numer- ical form [27]. (4) (43) The perturbative expansion in Eq. (43) can be compared with the corresponding perturbative expansion of the cusp anoma- lous dimension in Eq. (4.4) of Ref. [27]. From the comparison we can see that the third-order5 and fourth-order numeri- cal coefﬁcients in A0,i(αS) are sizeably larger than those in Ai(αS). Nonetheless the perturbative expansion of A0,i(αS) is still numerically well behaved. We also see that the viola- tion of Casimir scaling in the fourth-order term of A0,i(αS) is numerically at the 4% level. We add some general (though brief) comments on the soft- gluon effective coupling and Sudakov resummation. The resummation procedure of logarithmic contributions of Sudakov type requires proper kinematical approximations of the phase space for multiparton ﬁnal-state radiation. Such approximationsarespeciﬁcofthephysicalobservablesunder consideration. As a consequence, the use of one or the other of the two soft-gluon couplings AT,i and A0,i can be more appropriate depending on the observables. The two soft cou- plings can alternatively (or equivalently) be used for the resummation treatment of different classes of observables. Some observables can also require a combined use of both soft couplings. In Ref. 5 For comparison with the value 1.7157 in Eq. (43), we note that the numerical value of the third-order coefﬁcient [see Eq. (30)] of the soft coupling AT,i(αS) is 0.49121 . Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 We write this structure in the following form: ‘quadratic’ coefﬁcients of the soft coupling (A(4) 0[2])) and of the cusp anomalous dimension (A(4) [2]) through Casimir scal- ing: A(4) 0,i −A(4) i = Ci A(4) 0[2] −A(4) [2] , (39) (39) A(4) i (ǫ) = Ci A(4) [2](ǫ) + d(4) Ai Ni A(4) [4A](ǫ) + nF d(4) Fi Ni A(4) [4F](ǫ), (36) and, ﬁnally, we have derived and implemented the following equalities A(4) 0[4A](ǫ = 0) ≡A(4) 0[4A] = A(4) [4A], A(4) 0[4F](ǫ = 0) ≡A(4) 0[4F] = A(4) [4F], (40) (36) (40) where Ni is the dimension of the colour representation of the hard parton i (Ni = NA = N 2 c −1 if i = g, and Ni = NF = Nc if i = q, ¯q), and d(4) xy are the quartic Casimir invariants(we use the normalization of d(4) xy as in Eqs. (2.6)– (2.10) of Ref. [27]). The entire dependence of A(4) i (ǫ) (for both couplings A(4) T,i(ǫ) and A(4) 0,i (ǫ)) on the colour of the hard parton i is embodied in the Casimir dependent factors that we have explicitly written in the right-hand side of Eq. (36). The ‘quartic’ ( A(4) [4A](ǫ) and A(4) [4F]) and ‘quadratic’ ( A(4) [2]) coefﬁcients do not depend on the type of radiating parton i. In particular, A(4) [4A](ǫ) and A(4) [4F] are colour blind (they do not depend on Nc and nF). The coefﬁcient A(4) [2] still depends on Nc and nF, and this dependence involves all the colour structures that appear in the curly bracket of Eq. (32) plus an additional term with colour factor C2 FnF. between the ‘quartic’ coefﬁcients of the soft coupling (A(4) 0[4A], A(4) 0[4F]) and of the cusp anomalous dimension (A(4) [4A], A(4) [4F]). Our result in Eq. (32) is fully consistent with the Casimir scaling relation in Eq. (39). We note that the generalized Casimir scaling of the soft coupling in Eq. (36) and the relation in Eq. (21) necessar- ily imply the same scaling for the cusp anomalous dimen- sion in Eq. (38). The generalized Casimir scaling of the cusp anomalous dimension has been conjectured and veriﬁed to good numerical accuracy in Ref. [27]. We also note that at the fourth order the DGLAP kernel Pgg(αS; z) includes a contribution with the quartic Casimir invariant d(4) F F, which 12 3 Eur. Phys. J. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 [27], due to the actual values of the ‘quartic’ coefﬁcients A(4) [4A] and A(4) [4F] in Eq. (38), numerical Casimir scaling is completely broken in the fourth-order term A(4) i of the cusp anomalous dimension. However, due to the smallness of i, the soft coupling A(4) 0,i still fulﬁls numerical Casimir scaling (A(4) 0,i ∝Ci) modulo corrections at the few percent level. We report the numerical value of the soft coupling A0,i(αS) with Nc = 3 up to O(α4 S). Using A(3) 0,i from Eq. (31), A(4) 0,i −A(4) i from Eq. (32), A(4) i (with its numerical uncer- tainty) from Ref. [27] and setting nF = 5, we have 12 3 685 Page 10 of 11 685 Page 10 of 11 Eur. Phys. J. C (2019) 79 :685 The soft-gluon coupling Ai controls the intensity of the spectrum of soft and collinear radiation in the Sudakov ker- nel. The Sudakov kernel has other dynamical components that, roughly speaking, are due to soft non-collinear (i.e., wide-angle) radiation and hard (i.e., non-soft) collinear radi- ation. Both components have to be included in a resummed calculation (see, e.g., Refs. [5–7] at NLL accuracy and Refs. [4,8,9] at NNLL accuracy), and their inclusion has to be properly performed (i.e., properly matched) according to the soft coupling (either AT,i or A0,i) that is speciﬁcally used in the soft-collinear component. However, we note that, at a given ﬁxed perturbative order (say, αn S) in the Sudakov ker- nel,thesoft-collinearcomponentislogarithmicallyenhanced (by at least one power of log) with respect to the two other components. Therefore, the Sudakov kernel at NkLL accu- racy requires the knowledge of the soft coupling Ai(αS) up to O(αk+1 S ) and the computation of the other components up to O(αk S) (i.e., one order lower than the soft coupling). For instance, to achieve NNLL accuracy in the Sudakov kernel, the third-order results in Eqs. (30) and (31) for the soft cou- pling have to be combined with the calculation at O(α2 S) of the other dynamical components. and A0,i(αS). The coupling AT,i(αS) has originally been deﬁned in Ref. [4], and its explicit relation with αS up to O(α3 S) has been presented therein. We have computed both couplings, AT,i(αS) and A0,i(αS), up to O(α3 S) and, in the case of AT,i(αS), our independent calculation conﬁrm the result in Ref. [4]. Received: 6 May 2019 / Accepted: 29 July 2019 / Published online: 14 August 2019 © The Author(s) 2019 In the case of A0,i(αS; ǫ) we are able to compute its third-order contribution up to O(ǫ) and, in the four-dimensional limit, we obtain an explicit relation at O(α4 S) between A0,i(αS) and the cusp anomalous dimen- sion Ai(αS). Moreover, we have presented the explicit d- dimensional results (e.g., to all orders in the ǫ expansion) for both soft couplings up to O(α2 S). Acknowledgements We would like to thank Bryan Webber for com- ments on the manuscript. This work is supported in part by the Swiss National Science Foundation (SNF) under contracts 200020_169041 and IZSAZ2_173357, by MINCyT under contract SUIZ/17/05, by Con- icet and by ANPCyT. Data Availability Statement This manuscript has no associated data or the data will not be deposited. [Authors’ comment: The research presented in this manuscript is of purely theoretical nature and does not have associated data which need to be deposited. All our results are provided in analytic form within the paper.] A ﬁnal comment regards the process dependence of the Sudakov kernel. The soft couplings in Eqs. (13) and (14) are computed by considering soft-parton radiation from two hard partons in a colour singlet conﬁguration. Soft-gluon radiation in processes that involve several hard partons is deﬁnitely more complex than in the case of two hard-parton processes. This complex structure of soft-gluon radiation has to be taken properly into account. However, this does not affect the soft coupling Ai, since Ai measures the intensity of radiation that is both soft and collinear to parton i. The complex structure of soft radiation in multiparton hard scattering only affects the soft wide-angle component of the Sudakov kernel (see, e.g., Refs. [5–7] at NLL accuracy). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecomm ons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Funded by SCOAP3. References 1. S. Catani, B.R. Webber, G. Marchesini, Nucl. Phys. B 349, 635 (1991) 2. G. Luisoni, S. Marzani, J. Phys. G 42(10), 103101 (2015). arXiv:1505.04084 [hep-ph] We conclude the paper with a brief summary of its con- tent. We have considered the all-order extension of the CMW effective coupling in the context of soft-gluon resumma- tion beyond NLL accuracy. We have argued that there is no unique all-order extension, namely, no extension that shares all the universality (i.e., observable-independent) features of the CMW coupling at O(α2 S). Starting from the emission probability of an arbitrary number of soft partons, we have introduced the deﬁnition in d = 4 −2ǫ space-time dimen- sionsoftwoeffectivecouplings, AT,i(αS; ǫ)and A0,i(αS; ǫ), which measure the intensity of the inclusive spectrum for soft and collinear radiation from a massless hard parton i (i = q, ¯q, g). We have found that, to all perturbative orders, the two soft couplings are equal if they are evalu- ated at the d-dimensional point ǫ = β(αS), and they coin- cide with the (four-dimensional) cusp anomalous dimension Ai(αS). The limit ǫ →0 is smooth and it can be used to deﬁne the four-dimensional (‘physical’) couplings AT,i(αS) 3. T. Becher, A. Broggio, A. Ferroglia, Lect. Notes Phys. 896, (2015). arXiv:1410.1892 [hep-ph] 4. A. Banﬁ, B.K. El-Menouﬁ, P.F. Monni, JHEP 1901, 083 (2019). arXiv:1807.11487 [hep-ph] 5. A. Banﬁ, G.P. Salam, G. Zanderighi, Phys. Lett. B 584, 298 (2004). arXiv:hep-ph/0304148 6. R. Bonciani, S. Catani, M.L. Mangano, P. Nason, Phys. Lett. B 575, 268 (2003). arXiv:hep-ph/0307035 7. A. Banﬁ, G.P. Salam, G. Zanderighi, JHEP 0503, 073 (2005). arXiv:hep-ph/0407286 8. A. Banﬁ, H. McAslan, P.F. Monni, G. Zanderighi, JHEP 1505, 102 (2015). arXiv:1412.2126 [hep-ph] 9. A. Banﬁ, H. McAslan, P.F. Monni, G. Zanderighi, Phys. Rev. 117, 172001 (2016). arXiv:1607.03111 [hep-ph] 10. A. Buckley et al., Phys. Rep. 504, 145 (2011). arXiv:1101.2599 [hep-ph] 11. Y.L. Dokshitzer, G. Marchesini, B.R. Webber, Nucl. Phys. B 469, 93 (1996). arXiv:hep-ph/9512336 12. Y.L. Dokshitzer, A. Lucenti, G. Marchesini, G.P. Salam, JHEP 9805, 003 (1998). arXiv:hep-ph/9802381 13. J. Kodaira, L. Trentadue, Phys. Lett. 112B (1982) 66, report SLAC- PUB-2934 (1982) 123 Eur. Phys. J. C (2019) 79 :685 Page 11 of 11 685 685 41. E. Laenen, L. Magnea, Phys. Lett. B 632, 270 (2006). arXiv:hep-ph/0508284 14. J. Kodaira, L. Trentadue, Phys. Lett. 123B, 335 (1983) 15. S. Catani, E. D’Emilio, L. Trentadue, Phys. Lett. B 211, 335 (1988) 42. T. References Becher, M. Neubert, G. Xu, JHEP 0807, 030 (2008). arXiv:0710.0680 [hep-ph] 16. S. Catani, L. Trentadue, Nucl. Phys. B 327, 323 (1989) 17. D. Amati, A. Bassetto, M. Ciafaloni, G. Marchesini, G. Veneziano, Nucl. Phys. B 173, 429 (1980) 43. C. Anastasiou, C. Duhr, F. Dulat, E. Furlan, T. Gehrmann, F. Herzog, B. Mistlberger, Phys. Lett. B 737, 325 (2014). arXiv:1403.4616 [hep-ph] 18. Y.L. Dokshitzer, D. Diakonov, S.I. Troian, Phys. Rep. 58, 269 (1980) 44. Y. Li, A. von Manteuffel, R.M. Schabinger, H.X. Zhu, Phys. Rev. D 91, 036008 (2015). arXiv:1412.2771 [hep-ph] 19. S. Moch, J.A.M. Vermaseren, A. Vogt, Nucl. Phys. B 688, 101 (2004). arXiv:hep-ph/0403192 20. S. Moch, J.A.M. Vermaseren, A. Vogt, Nucl. Phys. B 691, 129 (2004). arXiv:hep-ph/0404111 45. T. Ahmed, M. Mahakhud, N. Rana, V. Ravindran, Phys. Rev. Lett. 113(11), 112002 (2014). arXiv:1404.0366 [hep-ph] (2004). arXiv:hep-ph/0404111 21. G.P. Korchemsky, Mod. Phys. Lett. A 4, 1257 (1989) 46. S. Catani, L. Cieri, D. de Florian, G. Ferrera, M. Grazzini, Nucl. Phys. B 888, 75 (2014). arXiv:1405.4827 [hep-ph] 22. R.A. Brandt, F. Neri, M.A. Sato, Phys. Rev. D 24, 879 47. M. Bonvini, L. Rottoli, Phys. Rev. D 91(5), 051301 (2015). arXiv:1412.3791 [hep-ph] 23. G.P. Korchemsky, A.V. Radyushkin, Nucl. Phys. B 283, 342 (1987) 24. G. Curci, W. Furmanski, R. Petronzio, Nucl. Phys. B 175, 27 (1980) 48. J.A. Gracey, Phys. Lett. B 322, 141 (1994). [hep-ph/9401214] 25. W. Furmanski, R. Petronzio, Phys. Lett. 97B, 437 (1980) 26. S. Moch, B. Ruijl, T. Ueda, J.A.M. Vermaseren, A. Vogt, JHEP 1710, 041 (2017). arXiv:1707.08315 [hep-ph] 26. S. Moch, B. Ruijl, T. Ueda, J.A.M. Vermaseren, A. Vogt, JHEP 26. S. Moch, B. Ruijl, T. Ueda, J.A.M. Vermaseren, 49. J. Davies, A. Vogt, B. Ruijl, T. Ueda, J.A.M. Verm 49. J. Davies, A. Vogt, B. Ruijl, T. Ueda, J.A.M. Vermas Phys. B 915, 335 (2017). arXiv:1610.07477 [hep-ph] j 1710, 041 (2017). arXiv:1707.08315 [hep-ph] g j Phys. B 915, 335 (2017). arXiv:1610.07477 [hep-ph 27. S. Moch, B. Ruijl, T. Ueda, J.A.M. Vermaseren, A. Vogt, Phys. L tt B 782 627 (2018) Xi 1805 09638 [h h] 27. S. Moch, B. Ruijl, T. Ueda, J.A.M. Vermaseren, A. Vogt, Phys. Lett. B 782, 627 (2018). arXiv:1805.09638 [hep-ph] 50. J.M. Henn, A.V. Smirnov, V.A. Smirnov, M. Steinhauser, JHEP 1605, 066 (2016). arXiv:1604.03126 [hep-ph] j Lett. B 782, 627 (2018). arXiv:1805.09638 [hep-ph] ( ) p p 28. F. Herzog, S. Moch, B. Ruijl, T. Ueda, J.A.M. Vermaseren, A. Vogt,. arXiv:1812.11818 [hep-ph] p p 51. A. von Manteuffel, R.M. Schabinger, Phys. Rev. D 95, 034030 (2017). arXiv:1611.00795 [hep-ph], arXiv:1902.08208 [hep-ph] p p 28. F. Herzog, S. Moch, B. Ruijl, T. Ueda, J.A.M. Ver g j Vogt,. arXiv:1812.11818 [hep-ph] 52. J. Henn, A.V. Smirnov, V.A. Smirnov, M. Steinhauser, R.N. Lee, JHEP 1703, 139 (2017). arXiv:1612.04389 [hep-ph] 29. J.G.M. Gatheral, Phys. Lett. 133B, 90 (1983) 30. J. Frenkel, J.C. Taylor, Nucl. Phys. B 246, 231 (1984) 53. R.N. Lee, A.V. Smirnov, V.A. Smirnov, M. Steinhauser, Phys. Rev. 53. R.N. Lee, A.V. Smirnov, V.A. Smirnov, M. Steinhauser, D 96(1), 014008 (2017). arXiv:1705.06862 [hep-ph] 31. A.A. Vladimirov, Phys. Rev. Lett. 118(6), 062001 (2017). arXiv:1610.05791 [hep-ph] 32. A.A. Vladimirov, JHEP 1804, 045 (2018). arXiv:1707.07606 [hep- ph] 54. R.N. Lee, A.V. Smirnov, V.A. Smirnov, M. Steinhauser, JHEP 1902, 172 (2019). arXiv:1901.02898 [hep-ph] 33. G.F. Sterman, Nucl. Phys. B 281, 310 (1987) 55. J.M. Henn, T. Peraro, M. Stahlhofen, P. Wasser,. arXiv:1901.03693 [hep-ph] 34. S. Catani, L. Trentadue, Nucl. Phys. B 353, 183 (1991) 56. M. Beneke, V.M. Braun, Nucl. Phys. B 454, 253 (1995). arXiv:hep-ph/9506452 35. S. Catani, M. Grazzini, Nucl. Phys. B 591, 435 (2000). arXiv:hep-ph/0007142 36. S. Catani, M. Grazzini, Nucl. Phys. B 570, 287 (2000). arXiv:hep-ph/9908523 57. A. Grozin, J.M. Henn, G.P. Korchemsky, P. Marquard, JHEP 1601, 140 (2016). arXiv:1510.07803 [hep-ph] 58. A. Grozin, PoS LL 2016, 053 (2016). arXiv:1605.03886 [hep-ph] 37. S. Catani, D. de Florian, S. Devoto, M. Grazzini, J. Mazzitelli (in preparation) 59. A. Grozin, JHEP 1806 (2018) 073 Addendum: [JHEP 1901 (2019) 134] arXiv:1805.05050 [hep-ph] A. Vogt, Phys. Lett. B 497, 228 (2001). arXiv:hep-ph/00101 60. R. Brüser, A. Grozin, J.M. Henn, M. Stahlhofen,. arXiv:1902.05076 [hep-ph] 39. S. Catani, D. de Florian, M. Grazzini, JHEP 0105, 025 (2001). arXiv:hep-ph/0102227 40. S. Moch, A. Vogt, Phys. Lett. B 631, 48 (2005). arXiv:hep-ph/0508265 123 12 123
https://openalex.org/W2955858505	https://www.ssoar.info/ssoar/bitstream/document/63063/1/ssoar-socialinclusion-2019-2-mesic_et_al-Use-values_for_inclusion_mobilizing_resources.pdf	English	null	Use-Values for Inclusion: Mobilizing Resources in Popular Education for Newly Arrived Refugees in Sweden	Social inclusion	2,019	cc-by	9,738	www.ssoar.info Issue This article is part of the issue “The European Refugee Controversy: Civil Solidarity, Cultural Imaginaries and Political Change”, edited by Gert Verschraegen (University of Antwerp, Belgium) and Robin Vandevoordt (University of Oxford, UK/University of Antwerp, Belgium). © 2019 by the authors; licensee Cogitatio (Lisbon, Portugal). This article is licensed under a Creative Commons Attribu- tion 4.0 International License (CC BY). conditions with regard to formal assessment of identity and citizenship, living conditions, local community rela- tions, social networks, labour market, welfare and car- ing and curing institutions. Thus, the asylum process is a complex process of life transition and everyday learn- ing, adapting to new conditions. Educational and occupa- tional background, recognition of prior learning as well as language training are crucial components for the adap- tation to the Swedish society and labour market (see e.g., Delmi, 2015). Abstract In times of market reforms and international migration, the Swedish welfare model has been seriously challenged. In the context of the arrival of refugees in 2015–2017, the state turned to civil society in facing up to the challenges. In this article, we direct our attention to the Workers’ Educational Association’s (ABF) state-funded work with refugees, with a specific focus on the activities conducted, the resources making them possible and the use-value of the resources mobilised. The article is based on observations and interviews with study circle leaders, managers and asylum seekers. The analysis illus- trates that ABF, in line with its historical legacy, the broader workers’ movement, the strong notion of popular education as ‘free and voluntary’, has, with its well-established connections throughout the country, not solely taken on the task defined by the state. In solidarity, ABF has also responded to the needs of the refugees. As highlighted in the analysis, ABF has mobilized a wide range of resources, not least providing refugees with social networks and help in contacting the authorities. With such mobilization, opportunities were provided for the inclusion of refugees in Sweden. Submitted: 15 January 2019 \| Accepted: 10 April 2019 \| Published: 27 June 2019 Submitted: 15 January 2019 \| Accepted: 10 April 2019 \| Published: 27 June 2019 www.ssoar.info Empfohlene Zitierung / Suggested Citation: Mešić, N., Dahlstedt, M., Fejes, A., & Nyström, S. (2019). Use-values for inclusion: mobilizing resources in popular education for newly arrived refugees in Sweden. Social Inclusion, 7(2), 85-95. https://doi.org/10.17645/si.v7i2.1971 Empfohlene Zitierung / Suggested Citation: Mešić, N., Dahlstedt, M., Fejes, A., & Nyström, S. (2019). Use-values for inclusion: mobilizing resources in popular education for newly arrived refugees in Sweden. Social Inclusion, 7(2), 85-95. https://doi.org/10.17645/si.v7i2.1971 Empfohlene Zitierung / Suggested Citation: Mešić, N., Dahlstedt, M., Fejes, A., & Nyström, S. (2019). Use-values for inclusion: mobilizing resources in popular education for newly arrived refugees in Sweden. Social Inclusion, 7(2), 85-95. https://doi.org/10.17645/si.v7i2.1971 Nutzungsbedingungen: Dieser Text wird unter einer CC BY Lizenz (Namensnennung) zur Verfügung gestellt. Nähere Auskünfte zu den CC-Lizenzen finden Sie hier: https://creativecommons.org/licenses/by/4.0/deed.de Social Inclusion (ISSN: 2183–2803) 2019, Volume 7, Issue 2, Pages 85–95 DOI: 10.17645/si.v7i2.1971 Keywords Keywords asylum seeker; civil society; inclusion; migrants; popular education; refugee; Sweden model Use-Values for Inclusion: Mobilizing Resources in Popular Education for Newly Arrived Refugees in Sweden Nedžad Mešić 1,, Magnus Dahlstedt 1, Andreas Fejes 2 and Sofia Nyström 2 1 Department of Social and Welfare Studies, Linköping University, 601 74 Norrköping, Sweden; E-Mails: nedzad.mesic@liu.se (N.M.), magnus.dahlstedt@liu.se (M.D.) 2 Department of Behavioural Sciences and Learning, Linköping University, 581 83 Linköping, Swed E-Mails: andreas.fejes@liu.se (A.F.), sofia.nystrom@liu.se (S.N.) 1 Department of Social and Welfare Studies, Linköping University, 601 74 Norrköping, Sweden; E-Mails: nedzad.mesic@liu.se (N.M.), magnus.dahlstedt@liu.se (M.D.) 2 Department of Behavioural Sciences and Learning, Linköping University, 581 83 Linköping, Sweden; E-Mails: andreas.fejes@liu.se (A.F.), sofia.nystrom@liu.se (S.N.) 1 Department of Social and Welfare Studies, Linköping University, 601 74 Norrköping, Sweden; E-Mails: nedzad.mesic@liu.se (N.M.), magnus.dahlstedt@liu.se (M.D.) 2 Department of Behavioural Sciences and Learning, Linköping University, 581 83 Linköping, Sweden; E-Mails: andreas.fejes@liu.se (A.F.), sofia.nystrom@liu.se (S.N.) Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 Corresponding author * Corresponding author 1. Introduction Current migration patterns in Europe, and particularly the historically high number of refugees arriving in Sweden and other member states of the EU, are chal- lenging in terms of inclusion. How can refugees and mi- grants be supported in order for them to gain access to the labour market and Swedish society more widely? This question comprises a number of policy challenges and institutional innovations not only for the govern- ment, regional organizations and municipalities, but also for social partners and in a broader sense civil society. Asylum seekers are facing several problems and shifting In order to face these challenges, a wide range of institutions and actors have been mobilised, not least within civil society. Over the last century, there has been Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 85 a close relationship between the state and civil society organisations in Sweden, as part of corporatist arrange- ments of the Swedish welfare model—a model that, how- ever, is changing. More and more tasks, previously min- istered by the state, are now conducted by organisations in civil society. In this article, we focus on one such ac- tivity emerging in response to the refugee challenges of 2015—Swedish from day 1. Here, the state invited civil society organisations, among them study associations, to apply for funding in order to provide adult asylum seek- ers with introduction to Swedish language and society. The invitation was taken up by all ten study associations in Sweden. In this article, we investigate the resources mobilized in such activities, and the ways in which inclu- sion is facilitated through such mobilizations. By focusing on the use-value of resources mobilised by civil society, the aim of the article is to analyse how public funding in the wake of the recent refugee situation in Sweden be- comes transferred into public benefit in the form of in- clusion of newly arrived migrants. The article thus exam- ines the implications of collaboration between the state and civil society, and how added value of publicly funded civil society engagements becomes generated vis-à-vis resource conversion in the free and voluntary form of popular education. 1980s, several of the pillars of the Swedish model were challenged. In all, the centralized welfare state was now seen as an obstacle to the individual’s freedom and ac- tive responsibility (Boréus, 1994). 2.2. Integration Policy in Change The Swedish welfare model has been described as char- acterized by its ‘de-commodifying’ provision of welfare services, based on the pillars of industrial development, full employment, economic growth and income redis- tribution (Esping-Andersen, 1990; Weiss, 1998). One of the model’s main characteristics was symbolised by the institutionalised collaboration between employer and trade union organisations (Edling, 2019). In this model, relations between the state and civil society organisa- tions, referred to as ‘people’s movements’ (folkrörelser), were seen as important for developing a more equal and democratic society. Accordingly, a wide range of interest groups was involved in the decision-making process, thus becoming part of the social democratic welfare project (Rothstein & Bergström, 1999). The idea with such ar- rangement was to embed decisions in broad layers of the population, at the same time as high levels of par- ticipation in the activities of such organisations would contribute to a democratic fostering of the population (Dahlstedt, 2009a; Edquist, 2009). In a contemporary academic debate concerning integra- tion policies among EU member states, the state of na- tional integration models has been a contested issue. Here, Joppke (2007) has argued that multicultural inte- gration models have generally converged, in line with an assimilationist development, to such an extent that mak- ing distinctions between national model in terms of inte- gration becomes superfluous. Although adhering to the prevalence of such conversion, Jacobs and Rea (2007) contend that dominant integration policy discourses in different member states, by and large, are still rather intact. They argue that it is, nevertheless, both possi- ble and relevant to distinguish path dependencies in terms of national models and policy choices concern- ing integration. However, in the wake of the refugee situation in Europe since 2015, major policy changes have recently been made, not least in the case of Sweden. From an in- ternational perspective, Swedish integration policy was for a long time widely recognized, not least in research (cf. Borevi, 2014; Schierup, Hansen, & Castles, 2006), for being an inclusion pioneer and a symbol of the citizen- ship model described by Castles (1995) as multicultural— a model based on the principles of inclusion (mak- ing it relatively easy to obtain citizenship) and recogni- tion (guaranteeing minorities certain group rights). 1. Introduction Eventually, following the early 1990s economic recession and the election of a centre-right government, by the mid-1990s, the Social democrats had undertaken the development of a ‘new Swedish model’. It entailed a radical shift, from equality to freedom of choice, redistribution to activation, collec- tive rights to individual responsibilities as the main ideas guiding the development of welfare policy in Sweden (Dahlstedt, 2015; Edling, 2019). a close relationship between the state and civil society organisations in Sweden, as part of corporatist arrange- ments of the Swedish welfare model—a model that, how- ever, is changing. More and more tasks, previously min- istered by the state, are now conducted by organisations in civil society. In this article, we focus on one such ac- tivity emerging in response to the refugee challenges of 2015—Swedish from day 1. Here, the state invited civil society organisations, among them study associations, to apply for funding in order to provide adult asylum seek- ers with introduction to Swedish language and society. The invitation was taken up by all ten study associations in Sweden. In this article, we investigate the resources mobilized in such activities, and the ways in which inclu- sion is facilitated through such mobilizations. By focusing on the use-value of resources mobilised by civil society, the aim of the article is to analyse how public funding in the wake of the recent refugee situation in Sweden be- comes transferred into public benefit in the form of in- clusion of newly arrived migrants. The article thus exam- ines the implications of collaboration between the state and civil society, and how added value of publicly funded civil society engagements becomes generated vis-à-vis resource conversion in the free and voluntary form of popular education. The established corporatist arrangements gradually became disintegrated, as they were criticized for being too centralized, state-directed and, thus, paternalistic, co-opting the civil society organisations (cf. Rothstein & Bergström, 1999). However, in the 1990s the idea of collaboration between state and civil society reap- peared, in the form of the increasingly widespread no- tion of partnership. According to this notion, collabo- rations between state, civil society and private organi- sations were seen as being necessary in order to deal with societal challenges. To be successful, such partner- ship arrangements were seen as in need of being decen- tralized, short-term and based on the transgression of boundaries between public and non-public sectors (cf. Dahlstedt, 2009b). Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 2.3. Study Associations, Swedish from Day 1 and Material Turning to Sweden post-2015, and specifically to the mobilisation of popular education, an extensive engage- ment for the inclusion of newly arrivedrefugeeshas been identified, conducted largely outside the formal task set by the government. Such engagement has been aimed at providing refugees with a home in Sweden, and for the refugees to become part of Swedish society (cf. Fejes, 2019; Fejes & Dahlstedt, 2017; Håkansson & Nilsson Mohammadi, 2018; Nordvall, Pastuhov, & Osman, 2018). In this article, focus is specifically directed towards the mobilisation of popular education and particularly study associations, in relation to the inclusion of newly ar- rived refugees. Such discourse has induced suggestions that the bor- ders surrounding the national community should be closed and that specific community ‘values’ need to be protected from threats located both beyond and within the national community. Such development has not least taken place during and in the aftermath of the 2018 Swedish general election, which became a strong electoral success for the right-wing populist party the Sweden Democrats. It was followed by a right- left political interlacement and government proposals on restricted migration policies, combined with integra- tion measures extending workfare and sanctions specifi- cally targeting newly-arrived refugees (Social Democrats, 2019). Altogether, these policy changes are similar to those seen during the last decade in, not least, other EU member states (cf. Duyvendak, 2011; Houdt, Suvarierol, & Schinkel, 2011). g There is still a strong notion of popular education as ‘free and voluntary’, and in this sense, there is a continu- ity with the long corporatist tradition in the Swedish wel- fare model. In terms of popular education, the state pro- vides basic funding to folk high schools and study associ- ations, in order for these institutions to provide courses and study circles of their own design, in terms of con- tent as well as pedagogical forms. Furthermore, partici- pation in these activities is voluntary. Thus, popular ed- ucation is ‘free and voluntary’, in terms of organisation as well as participation. However, study associations and folk high schools need, at least in relation to the activ- ities conducted based on state funding, to follow the state aims for popular education. These aims are broadly defined in terms of giving ‘everyone the possibility, to- gether with others, to increase their knowledge and ‘bil- dung’ for personal development and participation in the society’ (Swedish Code of Statutes, 2015). 2.3. Study Associations, Swedish from Day 1 and Material Popular educa- tion should also, according to the aims of the state, sup- port activities that contribute to the strengthening and development of democracy, increase people’s influence on their life situation, create engagement to participate in societal development, and close the educational gaps between individuals and groups in society. Yet, what has been referred to as the long summer of migration in 2015 (cf. Hess et al., 2017) has induced the need for critical analyses of integration policies ex- ecuted in different EU states, the role of civil society and migrants’ mobilisation, and the emergent social rela- tions and cultural encounters (Youkhana & Sutter, 2017). In one study, Karakayali (2017) examines emotions in German social movements in relation to newly arrived refugees, illustrating how volunteers’ emotional manage- ment in the form of refraining from emotional involve- ment induce a narrow, rather than transnational form of solidarity—thus maintaining established boundaries of belonging, furthering the volunteers’ local belonging vis- à-vis an emotional regime of charity. In another study of the mobilisation of middle-class volunteers in Germany, Fleischmann and Steinhilper (2017) describe the com- plexities of such mobilisation: on the one hand, there is a constant risk that such mobilisation will reproduce al- ready existing hierarchies and inequalities. On the other hand, there is a potential in this engagement, for the development of new political subjectivities—facilitating contesting public discourse interventions. In the Austrian context, De Jong and Ataç (2017) have drawn attention to the engagement of social movement-oriented NGOs, identifying the creation of spaces where encounters be- tween refugees and engaged activists take place, offering possibilities for the development of new forms of solidar- ity and belonging. These spaces, in turn, make it possible to break refugees’ isolation and provide volunteers with both a sense of responsibility and opportunities for par- Over time, other means of income for popular edu- cation institutions than state funding have become im- portant, not least due to drastic decrease in funding for popular education on the local level (i.e., by county and municipal councils). Such other means of income could be commissioned tasks by the state, municipali- ties or state agencies. One such commissioned task is the focus of this article, namely Swedish from day 1. Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 2.2. Integration Policy in Change How- ever, with contemporary developments, such notions of ‘Swedish exceptionalism’ (Schierup & Ålund, 2011) have The state has for the last century, to varying de- grees, funded popular education as a part of the corpo- ratist Swedish model—at the same time as popular ed- ucation has been ‘free and voluntary’ (Micheletti, 1995; Premfors, 2000). However, in the 1980s and 1990s, these corporatist arrangements changed, with a wider trans- formation gradually influenced by neoliberal rationali- ties (Larsson, Letell, & Thörn, 2012). Starting in the late Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 86 ticipation in political struggles for extended citizenship rights of refugees. been seriously challenged due to major policy changes, towards a more repressive approach. In political dis- course, it has become more or less widely accepted that migration creates societal problems and tensions, that ‘integration has failed’ and that ‘parallel societies’ are emerging, not least in the suburbs where most of the mi- grants live (Sernhede, Thörn, & Thörn, 2016). Based on such discourse, there has been an increasingly sharp fo- cus on repressive measures with an emphasis on citizens’ duties rather than on their rights—primarily aimed at mi- grants (cf. Dahlstedt & Neergaard, 2016; Gardell, 2016). 2.3. Study Associations, Swedish from Day 1 and Material 3. Empirical Material and Methodological Considerations Manager interviews focused on the organisation of Swedish from day 1 and rationales for engagement. Ques- tions asked concerned engagement motives in these ac- tivities, their educational and occupational background, activity organisation, their thoughts about their par- ticipants and involvement in other tasks than manag- ing the activities. The circle leader interviews centred on motives for engagement, educational and occupa- tional background, their teaching and participants, as well their involvement in tasks other than teaching. As will be illustrated, most of the SCLs themselves had ex- periences of migration. In participant interviews, we fo- cused their ideas about their current studies and how these related to their past experiences of work and ed- ucation, as well as to their dreams of the future. Sur- prisingly, we did not need interpreters. Most interviews were conducted in Swedish, some in English, and one in Bosnian. All interviews were transcribed verbatim and edited for readability. Empirically, the largest study association, the Workers’ Educational Association (ABF), was selected for further study. ABF was the study association with the most par- ticipants registered in Swedish from day 1, and there was already well-developed collaboration between the study association and the researchers, i.e., access was easy. Three different ABF locations in Sweden where Swedish from day 1 was carried out were chosen for fur- ther study: one large city, one medium-sized city and one small town. In order to gain insight into and background informa- tion on what took place in these study circles, before con- ducting interviews we participated as observers in each location, taking field notes and conducting informal con- versations. We also collected documentation concern- ing the activities (course plans, teaching material, etc.). The observations also made it possible for the partici- pants to get acquainted with us and what we were do- ing, which thus made the planning and conduct of the interviews easier. At each site, a sample of participants, study circle leaders (SCLs), and managers were chosen for semi-structured interviews. All managers and SCLs working with the activities were interviewed. Among the managers, 3 were female and 6 male. Manager 2 and 3 were employed by ABF regional headquarters in Large City, while managers 5 and 6 were employed by ABF in Mid-City as well as Small Town (they were co-responsible for both locations). 2.3. Study Associations, Swedish from Day 1 and Material In the autumn 2015, the Swedish government, facing the largest migration flows since the Second World War, made funding available for study associations (and to some extent folk high schools) in order to set up study circles for asylum seekers providing an introduction to the Swedish language and society. Between the autumn of 2015 and the end of 2017, more than 120,000 unique Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 87 undergone ethical vetting and been approved by the re- gional ethical committee in Linköping (Dnr 2017/280-31). Each interviewee was informed about the research, the possibility of their withdrawing at any time, and the fact that the information would be securely stored as well as the identifying markers being deleted/changed in com- ing publications in order to safeguard their anonymity. participants (asylum seekers) had taken part in these ac- tivities, which makes up more than half of all asylum seekers arriving in Sweden during this period (Fejes & Dahlstedt, 2017; Fejes, Dahlstedt, Olson, & Sandberg, 2018; Swedish Government, 2017). 3. Empirical Material and Methodological Considerations Managers 1, 4, and 7–9 were em- ployed by three migrant organizations with whom ABF cooperated in Large City, which below will be denomi- nated as ABF Large City local offices. Among the SCLs, 4 were female and 4 male. SCL 2–4 and 6-8 worked in Large City, while SCL 1 and 5 worked in the other two locations. Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 5.1. Mobilizing Resources In our interviews, ABF has, by persons in all intervie- wee categories, been portrayed as a hub where study circle participants become acquainted with state bod- ies and their services and other civil society organisa- tions. The head of one local Large City office expounds on the matter: Resources may have different attributes. One such at- tribute is that of ‘use-value’, i.e., how easily resources are transferred between persons or organisations and thus converted into resources of other kinds—for instance, as in converting money by the purchase of equipment, the hiring of staff or the production of cultural products. Money is generally quite easy to convert into other kinds of resources. This means that organizations with access to large amounts of monetary resources enjoy greater flexibility in terms of the strategies available to them. Other kinds of resources, not least social-organizational and human, are comparably more context-dependent and, thus, difficult to convert into other resources. We have had 25 years of experience with education and we have vast contacts with the Migration Board, Public Employment Service, the National Society for Road Safety…and we have many presentations for our participants regarding road safety, employment and social security services. There are, besides [the circle], many matters regarding society…and they come and explain to our participants. (Manager 4) Apart from the actors mentioned here, prominent soci- etal actors that the participants, circle leaders and man- agers refer to include a dental office offering free care, the Swedish Church and local social services informing them about child services and related laws. As we can further see in a participant’s excerpt below, the list of or- ganisations involved in the activities of the study circles is quite long: Furthermore, there are various mechanisms of ac- cess and resource sources, such as self-production and patronage. Resources are not always easily accessible, and so to speak ‘ready’ to be mobilized, but resources may also be produced within the organisation itself, for instance by the creation of cultural frames and the train- ing of activists. Patronage refers to the existence of a substantial amount of financial support provided by ‘pa- trons’ external to the organisation, seeking to influence the way in which money is actually used. 4. Analytical Perspective and Analysis The project has Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 88 portation, the internet), social networks and organiza- tions, and more specifically access to the resources em- bedded in these, in terms of contacts, potential coali- tions as well as recruiting volunteers. Cultural resources include widely known and accessible ideas, traditions, conceptual tools, and tactical know-how, for instance concerning how to conduct collective action, run a meet- ing or make decisions. Such resources also include the use and production of cultural artefacts such as liter- ature, music, magazines or websites. Moral resources include legitimacy, solidarity and sympathetic support, most often located outside of the organization, and granted by external sources. As these external sources may also choose to retract such resources, moral re- sources may be less accessible than, for instance, cul- tural resources. 4. Analytical Perspective and Analysis This article departs from an approach to social move- ments focusing specifically on resource mobilization as a means of collective action (cf. McCarthy & Zald, 1973). In line with the broader tradition of popular movements in Sweden, ABF mobilizes a wide range of resources in order to provide popular education to all adults, as a social right. With the aim of combating social inequal- ities, not least in terms of social class, ABF promotes adult learning based on equal encounters among peo- ple with specific experiences, interests and living condi- tions. In order to further analyse the mobilization of re- sources in the case of ABFs activities targeting newly ar- rived refugees, we employ a typology of resources de- veloped by Edwards and McCarthy (2007). According to Edwards and McCarthy (2007, p. 142), ‘the availability of diverse kinds of resources to social actors and privileged access to them…enhances the likelihood of effective col- lective action’. Among the participants, we wanted to select an equal number of persons at each site, as well as securing representation in terms of age, gender, country of origin and educational and occupational background. 21 of the 46 interview persons were female and 25 male. There was an age span between 20 and 60 years, and the inter- viewees mainly originated from Afghanistan, Syria and Iran. ABF was for all the participants their first encounter with education in Sweden. Some of them had only been in Sweden for as little as three weeks, while others had been in Sweden waiting for a decision on their asylum application for up to three years. In the five-fold typology proposed, they distinguish between material, human, social-organizational, cultural and moral resources. Material resources include finan- cial and physical resources such as money, property, equipment, supplies and office space. Human resources consist of access to skills, experience and labour, within or outside the organization. These are commonly em- bodied in the individuals involved in the organization, in the form of leaders, staff and volunteers. Social- organizational resources may be divided into three differ- ent forms: infrastructures (such as postal services, trans- Each interview lasted between 20 and 75 minutes. In- terviews with participants were generally shorter as com- pared to those with managers and SCLs. 5. Result In this section, we introduce our results. These are di- vided into two parts. In the first part, we discuss the mobilization of resources in the activities arranged by ABF and in the second we discuss the use-value of such mobilizations. Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 5.1. Mobilizing Resources The headquarters’ con- fidence in the local offices derives from a decades long joint delivery of educational and cultural services to in- habitants with a migrant background, and not least, the local offices’ successes in mobilising circles with newly ar- rived migrants as participants. This socio-organisational resource in the form of access to social networks of potential participants has been accredited central use- value by the headquarter managers. One of them com- ments on the local offices’ capacity to attract asylum seekers to the numerous circle-courses offered: For most of the them [SCLs], it is some type of labour market subsidy programme…but it is still a cost to the association, even if a subsidy is received…and we come in and cover up…with at least one part of the cost of the difference. So, we finance the teach- ers…and, these associations [Large City local offices] have actually in principle no other assets whatsoever. They have no association subsidy from the municipal- ity as athletic clubs and the like have. (Manager 2) There has always been a waiting list…to take courses in Swedish and societal orientation. And, that is not something we advertise in some newspapers or on placards; rather, it spreads between friends and ac- quaintances, relatives and so. So, we have had a wait- ing list for the courses even before the arrival of this boom in 2015. (Manager 2) Another counteractive measure to redeem the ever- present burdens related to material resources that we encounter links in with the ideological and historical foundations of the ABF, i.e., its cultural resources. Activ- ities that were directed to asylum seekers were stated to be highly prioritised due to their clear accordance with the ABF’s founding mission and educational ambi- tions for workers’ educative development. According to the managers of ABF, this has contributed to a further strengthening of the Swedish from day 1 programme through local re-orientation of other available material resources. Two managers at the Large City headquarters expound on this type of self-production of material re- sources vis-à-vis the internal redistribution of capital: Both the managers and SCLs attest to the use-value of the long-standing relations both with and between local offices, as a means of facilitating the general groundwork. 5.1. Mobilizing Resources Where are you from?’ Everything, you dared to talk. (Participant 4) to support these people. It costs a lot of money be- cause it is free of charge to the participants. We are pretty much in accord, all of us working at ABF in Large City, that this is a highly prioritized activity…we also use money for this which we take from our regular state grants [for ordinary circles and cultural arrange- ments]. (Manager 2) In interviews with participants, SCLs, and managers, such an introduction to and interaction with influential soci- etal organisations is described as one of the cornerstones of the programme. In parallel to the extensive mobilisa- tion of socio-organisational resources, our empirical data has also revealed organisational strains in securing ma- terial resources such as supplies, staff wages and ade- quate premises. We must sponsor the weakest in the society. That is our ideology, because we are a people’s movement organisation, or workers’ organisation…We think that these groups (refugees) are important now…maybe 30–40 years ago, it was the workers at Volvo and oth- ers. Now, it is these groups who need our education. And, that means that we take a little from privileged groups who have money…and give to these groups that require it. (Manager 3) Counteracting strategies to such strains include both self-production and resource mobilisation arrangements such as the establishment of affordable course litera- ture lease programmes; and patronage-based arrange- ments in the form of the use of municipal registers for vol- unteers; the use of state wage-subsidised employments for circle leaders; and, in the case of one Large City lo- cal office, a successful application for a supplemental project grant. Despite such examples of successful forms of external mobilisation of material resources in terms of project grants, this local office has also been cornered in finding makeshift solutions such as alleviating a lack of classroom space by disposing of office-space for the course work. One of the managers at Large City ABF re- gional headquarters expounds further on the matter of accumulating material resources: It has accordingly, in the case of the Large City regional headquarters, been possible to re-dispose resources from the regular and larger state grants so as to fur- ther extend the financing of the Large City local offices’ Swedish from day 1 groundwork. 5.1. Mobilizing Resources Kaldor (2003) problematizes such influences on the part of the state in terms of ‘domestication’ by which the civil society would substitute state services in exchange for becoming institutionalized, and hence drift away from its activist- oriented agenda. Last week somebody came from the Swedish Pub- lic Employment Service. A couple of months ago a police officer came who spoke about laws and regu- lations…and recently…information about health care in Sweden. Somebody came and spoke about nature conservation. Folkuniversitetet [an adult educational association] also came and talked about their free of charge courses that one can attend, [and how to] val- idate earlier education. (Participant 46) Drawing on the typology of resource mobilization, the analysis is focused on the ways in which resources are mobilized in the activity Swedish from day 1, and how inclusion of newly arrived refugees might be made possible through such mobilizations. Specific attention is drawn to the ways in which the organisation and set- up of the activity is described in interviews with people involved in the activity, i.e., managers, SCLs and partici- pants. All interviews conducted were thematically anal- ysed, guided by the analytical approach outlined. Here, the focus was on identifying commonalities and differ- ences across interviews, as well as geographical settings, in terms of how resources are mobilized, and converted and by which means resources are accessed. Apart from the study circle, ABF also arranged for and created conditions for other activities that took place in relation to the study circles, such as language cafés, learning how to ride a bike and swim, cook, apply for job and make study visits. The language café is an activity that is highlighted in all contexts. The café was located at ABF’s premises, where the participants have the oppor- tunity to drink coffee, meet other participants as well as Swedes and thus practice speaking Swedish. One partici- pant said: There was a language café here. It helped us a lot. Swedes came here. We drank coffee and tried to talk Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 89 to everyone. We talked like this: ‘What’s your name? Where are you from?’ Everything, you dared to talk. (Participant 4) to everyone. We talked like this: ‘What’s your name? Where are you from?’ Everything, you dared to talk. (Participant 4) to everyone. We talked like this: ‘What’s your name? One manager argues that the circle has four functions: One manager argues that the circle has four functions: The first function is that you give them a tool, the lan- guage, so you can communicate. This is very impor- tant. Then you create a meeting place, since these in- dividuals are alone, with different backgrounds, trau- matized by their experiences…they have left a coun- try at war and all of that…the circle becomes a meet- ing point which has a social and psychological mean- ing for them, a support. The third is that you gain a network which is important in order to find differ- ent ways (ahead). To find an apartment, to find work and maybe find each other, a friend, and you can help each other in the tough situation…the fourth is the knowledge they acquire about the society. When they learn Swedish, they can…integrate into the soci- ety and get a job. They then know what ways there are. Where is the public employment office, the social service and where to get help? (Manager 3) The findings also show that the ABF vis-à-vis the various resources mobilised within the circles themselves jointly offers a social form of support through the relationships that are formed between the SCLs and the participants. The use-value of the material resources provided vis- à-vis state patronage through Swedish from day 1 in this way extends beyond the formal programme funding goals. Thus, rather than drifting away from its activist- oriented agenda (cf. Kaldor, 2003), our study illustrates that state patronage, on the contrary, has provided fuel for the activist orientation of ABF. Both the managers and SCLs describe in detail numerous ways how they con- duct voluntary and funding programme-supplementary social work in relation to the participants. These activi- ties constitute the study circle context as a kind of liai- son centre where both the managers and the SCLs assist the participants with social support in a broad sense— one that is akin to ‘bureaucratic support’ and described as indispensable in encounters with public and business services such as banks, doctors and migration officials. One manager, for example, states that her door is always open to those in need of help, whatever it might be: I can only recommend it to those that sit at home; they do not have to. They can come to ABF. Here you will find friends; you can learn Swedish and you can make more connections. 5.1. Mobilizing Resources Specific attention is paid to the local offices’ extensive human resources embodied in the individuals involved in the organization, in terms of their ability to communi- cate with participants in their mother tongue, which in turn makes it possible for the local offices to reach the participants. However, these conditionalities are rather disparate in the different contexts we have studied. In contrast to Large City, we have not found similar mo- bilisation on the basis of ethnic background in Mid-City and Small Town. Contemplating the potential for mobili- sation of not least moral resources vis-à-vis migrant or- ganisations, which may aid in attracting newly arrived participants by generating legitimation, signals the im- Our mission is the same today, but it’s a different tar- get group than 100 years ago. Therefore, it felt the whole time as a calling to be in on this. It is of huge importance for ABF to take part and do what we can Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 90 language, and by using the language, I can get to understand the Swedes better and then learn more Swedish. (Participant 46) portance of establishing and nurturing of such organisa- tional collaborations. 5.2. The Use-Value of Resources Mobilized Another participant further reflects on learning vis-à-vis the circle: In this section, we will scrutinize the use-value of a vari- ety of resources in the realm of study circles as a social support system for asylum seekers. One central aspect is how ABF vis-à-vis its socio-organizational, material and human resources manifested through the study circle it- self generates moral resource use-value in the form of participants’ attendance, simply by being a place for en- counters. Yet, it is not solely referred to as a place where participants simply hang out. Rather, it is described as a safe place, a place where the participants may escape the uncertainties of their everyday lives and thereby ‘break’ some of the isolation that they express. This also con- nects up with use-value related to the way in which the studies are organized and how they are regarded by the participants. The circle-form of education has through- out our observations and interviews distinguished itself from its convention where the classroom is isolated by walls and the learning closely follows the curriculum. In- stead, the idea of the classroom has assumed an ex- tended form, where activities, previously expounded on vis-à-vis socio-organisational resources, such as visits by external guests and city tours have been regarded by the participants as widely appreciated educational compo- nents. On the basis of such arrangements, along with the possibilities of encountering many new people and find- ing friends, the circles have been framed by the partici- pants in a rather soft-centred respect. One of the partic- ipants expounds: Much of the talk is about Sweden as a place, a country. We learn how to orient ourselves, north, south, how many citizens…and last time we had someone here that talked about how to understand traffic and the rules. (Participant 37) One manager argues that the circle has four functions: Another SCL describes his role: ‘You explain and give them some of your own experience about society. It is not only about learning Swedish, an introduction to this language; it is more about the society’ (SCL 6). All the SCLs that were interviewed showed a large social commitment, which also extends out into the lo- cal community, where they, for example, may run a local sports team or work extra hours at the library. Through these engagements, the SCLs often establish bonds with the participants outside their professional work at ABF, i.e., the SCLs also encounter participants in contexts out- side of the circle. One SCL states: ‘Well, the majority of the participants are my friends as well. I play football with them; I visit them and they come to me’ (SCL 5). As the quotes illustrate, the SCLs’ lived experiences of migration can be emphasised as human resources whose use-value translates into moral resources as the expe- riences may allow SCLs to act as guiding role models, i.e., be perceived as persons with whom the participants can identify. Thus, SCLs gain credibility and the oppor- tunity to share their experiences while having an under- standing of the participants’ situation. However, it is not solely the lived experiences of migration among the SCLs that can be used as a resource, but also their mother tongue, which in some study circles is used as a ‘spring- board’ for learning (in those cases where the SCLs share their mother tongue with the participants). As one of the participants describes it: ‘Those SCLs that speak Per- sian…they can explain it better for me and I understand’ (Participant 10). However, the close relationships and the social com- mitments that the SCLs offer the participants in the form of human resources, also come with certain trials. The findings indicate that the SCLs also take considerable re- sponsibility for the welfare of the participants. Many say that they have been or are struggling with how to handle the situation of always being ‘on call’. While underlining the importance of the relational aspects of their work, they also emphasize the costs that come with the form- ing of relationships with individuals that may be trans- ferred to another location or removed from the country at very short notice. One manager states that: ‘How they should handle this is very challenging for the participants and our leaders. One manager argues that the circle has four functions: For many of them, they are not psychol- ogists or social workers; they cannot take care of these individuals’ (Manager 3). One manager argues that the circle has four functions: ABF is not a school for me, ABF is a home. I feel at home when I’m here. (Participant 1) Many of the participants in the study express similar opinions about ABF and the study circle Swedish from day 1 as those in the quote above. The use of the metaphor ‘home’ indicates that there is more going on in the study circle apart from language learning, which indicates that several other forms of resource have been successfully converted into moral resources in the form of participants’ articulated satisfaction. The findings have also provided evidence that show how attending the ABF study circles denotes a feeling of belonging and inclusion, since the participants are not only learning about the Swedish language and about Sweden, but also about the society and how to get around. One participant emphasises the value of and strong will to be included into the Swedish society: I always have my door open, so everyone can come in if they want help in calling the authorities or interpret- It is by learning the language that I can get into the so- ciety. Here, I get to learn the language, the Swedish Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 91 going the asylum process, learning Swedish and building a future for themselves in Sweden. One SCL states: ing different papers. There is so much to do. Some- timesI stay untiltenin the evening, trying toreadtheir letters and maybe writing a ‘close relative application’ to the Migration Board. (Manager 1) I think they saw me as a role model and a leader. I could also see myself in them based on my own ex- perience of being a migrant. I worked; I applied for asylum and I got the chance to stay and this makes it easier to work with the participants. (SCL 5) The relationship that is formed between the participants and the SCL thus expands way beyond the professional responsibilities as SCLs at ABF. One reason for this is the precarious situation the asylum seekers are in, and this kind of support is described in terms of an obligation to fellow human beings in despair. Another SCL describes his role: ‘You explain and give them some of your own experience about society. It is not only about learning Swedish, an introduction to this language; it is more about the society’ (SCL 6). 6. Discussion This article has directed attention to one specific case of civil society mobilisation carried out across Europe in the wake of the long summer of 2015 (cf. Hess et al., 2017), taking place in Sweden and specifically within the realm of popular education. In the wake of the 2015 refugee situation, the Swedish state oc- casioned mobilisation of the civil society in the cause of language studies and introduction into the Swedish society, incorporating ten major study associations in Sweden (cf. Fejes & Dahlstedt, 2017; Håkansson & Nilsson Mohammadi, 2018). Among these, ABF was awarded the largest share of the targeted state funding programme for integration—Swedish from day 1. It is this specific activity that has been the focus of this ar- ticle. On the receiving end, such popular education di- rected to adult asylum seekers has been seen by the ABF leadership as directly contingent on the resources allocated by the government. However, apart from un- derlining the centrality of such patronage-based mate- rial resources for the operations of ABF, we have in this article identified the mobilisation of a wide range of resources as being crucial for the activities carried out within the organisation—material, human, socio- organisational, cultural as well as moral. In the article, we have analysed how different kinds of resources are mobilized within the Swedish from day 1 activities, as Another manager reasons on the investment made when forming a relationship: There is a risk; you form a close relationship with indi- viduals whose application [to stay in Sweden] are re- jected. Yes, it is tough. People that have been here ev- ery day and the next day you ask where they are and they have been deported. Then it is very tough for us. (Manager 6) Despite such hardships, the accounts by SCLs are indica- tive of a considerable use-value of human resources vis- à-vis the SCLs principal pursuit for social engagements that are anchored in their interest in benefitting the asy- lum seekers’ welfare. It is apparent that this commitment derives from a will to work with people, but also a will to give something back. A small group of SCLs were born and raised in Sweden, and have a great interest in work- ing relationally with people. 6. Discussion Despite the rather major shifts in terms of welfare and integration policy taking place in Sweden recently, with a greater focus on obligations, adaptability and individual responsibility (cf. Gardell, 2016; Schierup & Ålund, 2011), this kind of anal- ysis has made it possible to illustrate the complexities of such developments and is not as clear cut as first might seem to be the case. Popular education offers one ex- ample of the existing opportunities to work towards so- cial inclusion in Sweden, in a range of different ways, not the least by mobilising state resources in combination with other available resources. In order for such an en- deavour to be successful, moral resources such as legit- imacy and trust are needed, and these need to be care- fully managed. Concerning popular education in Sweden, such moral resources in terms of legitimacy and trust seem to be rather intact, despite the changes that have taken place. Thus, a focus on popular education might provide space for further debates on how to promote so- cial inclusion in times of migration. In addition to the patronage-based resource mobili- sation, ABF has also engaged in extensive self-production of resources, by relocating some of the general state funding provided for the Swedish from day 1 operations. This manoeuvre entails the ways in which cultural re- sources, in the form of ABF’s historical and ideological legacy of striving for workers’ rights in the society by means of providing adult education as a social right to the broad sections of the population, in turn may gener- ate additional material use-value. In the operations car- ried out by ABF, newly arrived refugees thus become per- ceived as the ‘new’ working class, or rather the working class of today (cf. Dahlstedt, 2009a; De Jong & Ataç, 2017; Osman, 2007). One of the central resources for the ground opera- tions found in our study is the one related to the gen- eral and extensive engagements on the part of the partic- ipants to attend the courses offered, by which the study circles commonly are denominated as a ‘home’ (as, for instance, illustrated by Participant 1), a place where the participants are seen as belonging and treated as equal fellow beings, even though they are newly arrived. 6. Discussion Another group, almost all of the SCLs interviewed, have themselves migrated to Sweden, and have corresponding experiences of under- Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 92 well as how various resources are converted from one form into another (cf. Edwards & McCarthy, 2007). engaged and on the call, which in turn raises questions concerning the ways in which boundaries between the professional and the private sphere are and should be drawn. Even though there is certainly great potential in the efforts and engagement of the SCLs, in terms of so- cial inclusion of newly arrived refugees, there are ob- vious risks in such a mobilisation of human resources within the organisation, not least considering the risk of SCLs hitting a wall, as there is not really a limit to the amount of work that would need to be carried out in order to meet the needs of newly arrived refugees (cf. Karakayali, 2017). Our data illustrates how state funding of the ‘free and voluntary’ format of the study circle, by extension, has contributed to the attainment of use-value in the form of socio-organisational resources within ABF. These re- sources have included the mobilization of an established broad social network of collaboration partners as well as the establishment of new alliances with an extended set of municipal, state and civil society actors through- out society. The material resources mobilised through the state-funded programme have, furthermore, vis-à- vis the accompanying moral resource in the form of a widely established legitimacy of the work carried out by popular education organisations more broadly, aided ABF in mobilising yet additional sets of patronage provid- ing material resources such as labour. y , ) Conclusively, in line with its historical legacy, and with its well-established connections throughout society, ABF not solely took on the task defined vis-à-vis state patronage. On the contrary, in converting the material resources provided by the state into other resources, and by mobilizing its traditional activist orientation (cf. Kaldor, 2003), ABF succeeded in both responding to the urgent needs of the newly arrived refugees and gaining legitimacy for such a response—from participants as well as from the state. So then, what is the point of focusing on popular education in terms of resource mobilisation in relation to newly arrived refugees? 6. Discussion The use-value of the moral resource that the participants’ at- tendance attests to has in turn been coupled with human resources embodied by the SCLs, their own experiences of migration, language skills and familiarity with Swedish society (cf. Fleischmann & Steinhilper, 2017; Karakayali, 2017). Additionally, both SCLs and participants refer to the study circles and the activities arranged by ABF as a safe place with the potential of breaking the social isola- tion related to living in refugee camps across the country, by offering the means to get acquainted with the new so- ciety and to learn the Swedish language in the ‘free and voluntary’ format provided by ABF (cf. De Jong & Ataç, 2017). Not least, the empirical material underscores the participants’ appreciation for the social and bureaucratic support provided in the activities conducted, for instance in terms of SCLs helping out with contacts with banks, doctors and migration officers. Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 References Boréus, K. (1994). Högervåg [Right wing wave]. Stock- holm: Tiden. Borevi, K. (2014). Multiculturalism and welfare state in- tegration. Identities, 21(6), 708–723. Castles, S. (1995). How nation-states respond to immi- gration and ethnic diversity. New Community, 21(3), 293–308. Castles, S. (1995). How nation-states respond to immi- gration and ethnic diversity. New Community, 21(3), 293–308. However, on the part of the SCLs, such engagements also come at the cost of persistently being emotionally Dahlstedt, M. (2009a). Democratic governmentality. Dahlstedt, M. (2009a). Democratic governmentality. Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 93 Journal for Critical Education Policy Studies, 7(2), 369–394. Jacobs, D., & Rea, A. (2007). The end of national models? International Journal on Multicultural Societies, 9(2), 264–283. Dahlstedt, M. (2009b). The partnering society. The Open Urban Studies Journal, 2(1), 18–27. Joppke, C. (2007). Beyond national models. West Euro- pean Politics, 30(1), 1–22. Dahlstedt, M. (2015). Discourses of employment and in- clusion in Sweden. In E. Righard, M. Johansson, & T. Salonen (Eds.), Transformation of Scandinavian cities (pp. 61–80). Lund: Nordic Academic Press. Kaldor, M. (2003). Global civil society. Cambridge: Polity Press. Karakayali, S. (2017). Feeling the scope of solidarity. So- cial Inclusion, 5(3), 7–16. Dahlstedt, M., & Neergaard, A. (2016). Crisis of solidar- ity? Critical Sociology, 45(1), 121–135. Larsson, B., Letell, M., & Thörn, H. (Eds.). (2012). Trans- formations of the Swedish welfare state. Basingstoke: Palgrave Macmillan. De Jong, S., & Ataç, I. (2017). Demand and deliver. Social Inclusion, 5(3), 28–37. McCarthy, J. D., & Zald, M. N. (1973). The trend of so- cial movements in America. Morristown, NJ: General Learning. Delmi. (2015). Integrationspolitik och arbetsmarknad [Policies of integration and labour market]. Stock- holm: Regeringskansliet. Micheletti, M. (1995). Civil society and state relations in Sweden. Aldershot: Avebury. Duyvendak, J. W. (2011). The politics of home. New York, NY: Palgrave Macmillan. Nordvall, H., Pastuhov, A., & Osman, A. (2018). Studiecirk- lar av och för migranter [Study circles by and for mi- grants]. Stockholm: ABF. Edling, N. (2019). The language of welfare in Sweden. In N. Edling (Ed.), The changing meanings of the welfare state. New York, NY: Berghahn. Edquist, S. (2009). En folklig historia [A folk story]. Umeå: Boréa. Osman, A. (2007). Invandrarföreningar [Immigrant asso- ciations]. In M. Dahlstedt et al. (Eds.), Utbildning, ar- bete, medborgarskap [Education, labour, citizenship] (pp. 341–361). Umeå: Boréa. Edwards, B., & McCarthy, J. D. (2007). Resources and social movement mobilization. In D. A. Snow, S. References A., Soule, & H. Kriesi (Eds.), The Blackwell companion to social movements (pp. 116–152). Oxford: Blackwell Publishers. Premfors, R. (2000). Den starka demokratin [The strong democracy]. Stockholm: Atlas. Rothstein, B., & Bergström, J. (1999). Korporatismens fall och den svenska modellens kris [The fall of cooper- atism and the crisis of the Swedish model]. Stock- holm: SNS. Esping-Andersen, G. (1990). The three worlds of welfare capitalism. Oxford: Polity Press. Fejes, A. (2019). Adult education and the fostering of asy- lum seekers into potential ‘full’ citizens. International Review of Education, 65(2), 233–250. Schierup, C. U., & Ålund, A. (2011). The end of Swedish exceptionalism? Race & Class, 53(1), 45–64. Fejes, A., & Dahlstedt, M. (2017). Popular education, mi- gration and a discourse of inclusion. Studies in the Ed- ucation of Adults, 49(2), 214–227. Schierup, C. U., Hansen, P., & Castles, S. (2006). Migra- tion, citizenship and the European welfare state. Ox- ford: Oxford University Press. Fejes, A., Dahlstedt, M., Olson, M., & Sandberg, F. (2018). Adult education and the formation of citizens. Lon- don: Routledge. Sernhede, O., Thörn, C., & Thörn, H. (2016). The Stock- holm uprising in context. In M. Mayer et al. (Eds.), Ur- ban uprisings. Basingstoke: Palgrave Macmillan. Fleischmann, L., & Steinhilper, E. (2017). The myth of apo- litical volunteering for refugees. Social Inclusion, 5(3), 17–27. Social Democrats. (2019). Sakpolitisk överenskom- melse mellan Socialdemokraterna, Centerpartiet, Liberalerna och Miljöpartiet de gröna [Political agree- ment between Social Democrats, Center party, Lib- erals and Swedish Green Party]. Socialdemokraterna. Retrieved from https://www.socialdemokraterna.se/ aktuellt/2019/Utkast Gardell, M. (2016). Raskrigaren [The race warrior]. Stock- holm: Pocketförlaget. Håkansson, P., & Nilsson Mohammadi, R. (2018). Slutrap- port: Folkhögskolornas verksamhet för nyanlända [Fi- nal report: The undertaking of folk high schools for newly arrived immigrants] (Dnr. UTB 3.7-2017/68). Malmö: Malmö University. Swedish Code of Statutes. (2015). Svensk förfat- tningssamling. Förordning om statsbidrag till folkbildningen (2015:218) [Swedish code of statutes. Regulation on subsidies for popular education]. Stockholm: Swedish government. Hess, S., Kasparek, B., Kron, S., Rodatz, M., Schwertl, M., & Sontowski, S. (2017). Der lange sommer der migra- tion [The long summer of migration]. In S. Hess, B. Kasparek, S. Kron, M. Rodatz, & M. Schwertl (Eds.), Der lange sommer der migration [The long summer of migration] (pp. 6–24). Hamburg: Assoziation A. Swedish Government. (2017). Government budget 2018 (Prop. 2017/18:1). Stockholm: Swedish Government. Weiss, L. (1998). The myth of the powerless state. Cam- bridge: Polity Press. Houdt, F. Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 References V., Suvarierol, S., & Schinkel, W. (2011). Neolib- eral communitarian citizenship. International Sociol- ogy, 26(3), 408–432. Youkhana, E., & Sutter, O. (2017). Perspectives on the Eu- ropean border regime. Social Inclusion, 5(3), 1–6. 94 Social Inclusion, 2019, Volume 7, Issue 2, Pages 85–95 About the Authors About the Authors Nedžad Mešić holds a PhD in Ethnicity and Migration and works as a Lecturer in education. His research interest concern migration and inclusion, with particular focus on the engagements of social movements for precarious migrants, newly arrived migrants and young people. (see e.g., Negotiating Solidarity: Collective Actions for Precarious Migrant Workers’ Rights in Sweden, Linköping University, 2017). Magnus Dahlstedt is a Professor in Social Work at Linköping University. His primary research inter- ests concern welfare changes, inclusion and exclusion in times of market-orientation and migration. Currently, he is conducting research on civil society organizations and the social inclusion of migrants in Sweden. Andeas Fejes is Professor and Chair of Adult Education research at Linköping University, Sweden. His research interests concern adult and popular education with a specific focus on issues pertaining to citizenship and migration. He is currently, together with M. Dahlstedt, directing a research programme on migration, learning and social inclusion, of which this article forms part. Fejes is one of the founding editors of the European Journal for Research on the Education and Learning of Adults. Sofia Nyström is an Associate Professor in Education at the Department of Behavior Sciences and Learning at Linköping University, Sweden. Her research interest focuses on the relation between edu- cation and work concerning arrangements for professional and vocational learning as well as identity formation using different theoretical perspectives. She is currently working on a research programme on migration, learning and social inclusion. 95
W2015772400.txt	https://zenodo.org/records/1522292/files/article.pdf	de		Zur F�llung des Magnesiums als Magnesium-Ammoniumphosphat	Analytical and bioanalytical chemistry/Analytical & bioanalytical chemistry	1,909	public-domain	758	Bericht: Chemische Analyse anorganischer KSrper. 479- 700 0 C.) geglfiht werden und das Filter sollte nach der Methode voa~ B u n s e n unterhalb Rotgltihhitze verbrannt werden. Bei der Bestimmung der Schwefelsiiure ist das in dem gewogenen :Niederschlag ent-haltene Chlorbaryum yon dem Gewicht des geglfihten Niederschlags in Abzug zu bringen, bei der Bestimmung des Baryums dagegen au[ Baryumsulfat umzurechnen und dem im Niederschlage gefundenen Baryumsulfat zuzuffigen. Die oben erwiihnten Erscheinungen beim Glfihen des chlorhaltigen, Baryumsulfats sind vielleicht dadurch zu erkliiren, dass sich etwas. B a C I . H S 0 4 bildet und yon dem l~iederschlage mitgerissen wird. der dann beim Glfihen Salzs~ure abgeben wfirde. Es w~re aber auch mOglich, dass die beiden Salzpaare (BaCI),2S0~ und (HS04).2Ba entstehen, die zusammen Salzs~ure bilden kSnnen. Ausgeschlossen ist es fibrigens nieht, dass Chlorbaryum in dem Baryumsulfatniederschlag zugegen seim kann, da Chlorbaryum in demselben Kristallsystem mit schwefelsaurem Baryum vorkommt, Wenn es richtig ist, dass die oben genannten Verbindungen yon dem Baryumsulfat niedergeschlagen werden, so rfihrem sie yon tier Anwesenheit der zwei Ionen BaCl" uud I-ISO4' her. Mare mfisste also reines Baryumsulfat erhalten, wenn diese zwei Ionen abwesend wiiren. Diese Bedingung wird getroffen, wenn man das Baryumsalz einer zweibasischen Siiure mit dem schwefelsauren Salz eines zweiwertigen Metalls in LSsung zusammenbringt. In der Tat erhielten die. Verfasser bei der Fi~llung yon reinem Baryumplatincyanfir mit einem. grossen Uberschuss einer MagnesiumsulfatlSsung reine ~iederschl~ge,_ die nichts mit sich zu Boden ftihrten und theoretisch richtige Werte. lieferten. L. L. de K o n i n c k ~) hat Versuche fiber das Glfihen der Baryumsulfatniederschliige ausgeffihrt und dabei festgestellt, dass eine Reduktiom des Baryumsulfats dureh Filterkohle nicht zu beffirchten ist, wenn man das Filter bei niederer Temperatur und bei Luftzutritt im offeneno Tiegel verascht. Zur ]?allung des Magnesiums als Magnesium-Ammoniumphosphat~ hat E. R a ffa2) Dinatriumammoniumphosphat in geringem Uberschusse benutzt und eine quantitative Abscheidung des Magnesiums erreicht._ Das Reagens wird bereitet, indem man eine LSsung yon Dinatrium1) Bulletin de la soc. chim. de Belgiqae 21, 116; yore Veffasser eingesandt.. 2) Gazz. Chim. Ital. 88. 556; The Analyst 84,, 73. 480 Bericht: Chemische Analyse anorganischer KSrper. phosphat mit genau so viel Ammoniak versetzt, als zur Bildung des Doppelsalzes erforderlich ist. Man bringt einen t)berschuss des Dinatriumammoniumphosphats (20 b i s 2 5 c c m einer 1/9-NormallOsung ) in ein Becherglas und gibt .10 c c m der Magnesiumchloridl(isung, welche 0,3 bis 0,5 °/o Magnesium enthalten soll, rasch hinzu. Beim Umrtihren der Fliissigkeit, wobei ein Bert~hren der Gefitsswandung zu vermeiden ist, scheidet sich der Niederschlag vollst~tndig ab. Er wird auf einem tarierten Filter oder in einem G o o c h ' s c h e n Tiegel gesammelt, mit 2,5-prozentigem Ammoniak ausgewaschen und in tiblicher Weise behandelt. Bei etwa 15 Minuten langem Glahen fiber dem T e c l u - B r e n n e r wird ,der Niedersehlag vollkommen weiss. ZurvolumetrischenBestimmung des M a g n e s i u m s b r i n g t man 10 ccn~ der Ltisung des Magnesiumsalzes rasch in 40 c c m des neutralen Reagens, das sich in einem mit Glasstopfen versehenen Zylinder befindet, schilttelt um und l~tsst bis zur volist~udigen Abscheidung des Niederschlags s t e h e n . Hierauf wird ein aliquoter Teil der klareu Fliissigkeit abgezogen, mit 1 bis 2 c c m starker Essigsgure versetzt und das iiberschassige Phosphat mi~ Uranylazetat titriert. Die hierbei erhaltenen Resultate schwankten bei einer LOsung, die 0,368 °/o Magnesium enthielt, yon 0,351 bis 0 , 3 7 8 ° / o , Bei Anwendung dieser Methode zur Analyse der Zemente etc. wird zunhehst das Kalzium entfernt, der nach dem Weggltihen der Ammonsalze verbleibende Riiekstand wieder aufgenommen und das Magnesium in vorstehender Weise bestimmt. Zur Bestimmung der Magnesia in Magnesiten benutzte J. M a y r .h o f er 1) ein yon Em. P o z zi - E s c o t'2) zur F~illung der Magnesia empfohlenes Verfahren, bei welchem als F~tllungsmittel eine PhosphorsalzlOsung verwendet wird, die in 250 ccm Wasser l t~ Phosphorsalz, 4 g Chlorammonium, 2 g Zitrouens~ure und 6 g konzentriertes Ammoniak enthi~lt. Es ergab sich jedoeb, dass die Methode far die direkte Bestimmung der Magnesia ungeeignet ist, weil die Konzentration an Zitronens~iure, sowie an Ammoniumsalzen zu gering ist, um die anderen Basen in LOsung zu h a l t e n ; dann wird auch durch das Chlorammonium das Ausfallen eines normalen phosphorsauren Salzes der Magnesia nicht gew~thrleistet. Das Ammoniumsulfat sicher~ dagegen eine gleiehmiffsige 1) Zeitschrift f. angew. Chemie 21, 592. ~) Ann. Chim. anal. apl)l. 7, 126.
https://openalex.org/W2953147631	https://journals.plos.org/plospathogens/article/file?id=10.1371/journal.ppat.1005379&type=printable	English	null	A conserved interaction between a C-terminal motif in Norovirus VPg and the HEAT-1 domain of eIF4G is essential for translation initiation	bioRxiv (Cold Spring Harbor Laboratory)	2,015	cc-by	22,632	OPEN ACCESS Citation: Leen EN, Sorgeloos F, Correia S, Chaudhry Y, Cannac F, Pastore C, et al. (2016) A Conserved Interaction between a C-Terminal Motif in Norovirus VPg and the HEAT-1 Domain of eIF4G Is Essential for Translation Initiation. PLoS Pathog 12(1): e1005379. doi:10.1371/journal.ppat.1005379 A Conserved Interaction between a C- Terminal Motif in Norovirus VPg and the HEAT-1 Domain of eIF4G Is Essential for Translation Initiation Eoin N. Leen1, Frédéric Sorgeloos2, Samantha Correia1¤a, Yasmin Chaudhry2, Fabien Cannac1¤b, Chiara Pastore1, Yingqi Xu1, Stephen C. Graham2, Stephen J. Matthews1, Ian G. Goodfellow2, Stephen Curry1 1 Department of Life Sciences, Imperial College London, London, United Kingdom, 2 Division of Virology, Department of Pathology, University of Cambridge, Addenbrooke's Hospital, Cambridge, United Kingdom ¤a Current address: Section of Virology, Faculty of Medicine, Imperial College London, St Mary's Campus, London, United Kingdom ¤b Current address: Department of Biology, Paul Scherrer Institut, ETH Zürich, Villigen PSI, Switzerland * eoin.leen07@imperial.ac.uk (ENL); s.curry@imperial.ac.uk (SC) ¤a Current address: Section of Virology, Faculty of Medicine, Imperial College London, St Mary's Campus, London, United Kingdom ¤b Current address: Department of Biology, Paul Scherrer Institut, ETH Zürich, Villigen PSI, Switzerland * eoin.leen07@imperial.ac.uk (ENL); s.curry@imperial.ac.uk (SC) RESEARCH ARTICLE Abstract This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Copyright: © 2016 Leen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Abstract Translation initiation is a critical early step in the replication cycle of the positive-sense, sin- gle-stranded RNA genome of noroviruses, a major cause of gastroenteritis in humans. Nor- ovirus RNA, which has neither a 5´ m7G cap nor an internal ribosome entry site (IRES), adopts an unusual mechanism to initiate protein synthesis that relies on interactions between the VPg protein covalently attached to the 5´-end of the viral RNA and eukaryotic initiation factors (eIFs) in the host cell. For murine norovirus (MNV) we previously showed that VPg binds to the middle fragment of eIF4G (4GM; residues 652–1132). Here we have used pull-down assays, fluorescence anisotropy, and isothermal titration calorimetry (ITC) to demonstrate that a stretch of ~20 amino acids at the C terminus of MNV VPg mediates direct and specific binding to the HEAT-1 domain within the 4GM fragment of eIF4G. Our analysis further reveals that the MNV C terminus binds to eIF4G HEAT-1 via a motif that is conserved in all known noroviruses. Fine mutagenic mapping suggests that the MNV VPg C terminus may interact with eIF4G in a helical conformation. NMR spectroscopy was used to define the VPg binding site on eIF4G HEAT-1, which was confirmed by mutagenesis and binding assays. We have found that this site is non-overlapping with the binding site for eIF4A on eIF4G HEAT-1 by demonstrating that norovirus VPg can form ternary VPg- eIF4G-eIF4A complexes. The functional significance of the VPg-eIF4G interaction was shown by the ability of fusion proteins containing the C-terminal peptide of MNV VPg to inhibit in vitro translation of norovirus RNA but not cap- or IRES-dependent translation. These observations define important structural details of a functional interaction between norovirus VPg and eIF4G and reveal a binding interface that might be exploited as a target for antiviral therapy. Editor: Bert L. Semler, University of California, Irvine, UNITED STATES Editor: Bert L. Semler, University of California, Irvine, UNITED STATES UNITED STATES Received: August 10, 2015 Accepted: December 10, 2015 Published: January 6, 2016 Copyright: © 2016 Leen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. UNITED STATES Received: August 10, 2015 Accepted: December 10, 2015 Published: January 6, 2016 Copyright: © 2016 Leen et al. Conserved MNV VPg Motif Binds eIF4G HEAT-1 SJM is funded by Senior Investigator Award from the Wellcome Trust (ref. 100280). CP was supported by intramural funding from Imperial College. BBSRC: http://www.bbsrc.ac.uk, Wellcome Trust: http://www. wellcome.ac.uk, Imperial College: http://www.imperial. ac.uk. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Author Summary Norovirus infections cause acute gastroenteritis and are a growing worldwide problem in human health. A critical early step in infection is translation of the viral RNA genome to produce the proteins needed to assemble new virus particles. In mouse noroviruses (MNV), which provide a useful model for studying human noroviruses, the VPg protein attached to the viral RNA is essential for this process because it interacts with a cellular protein, eIF4G, that is normally involved in initiating protein synthesis from the messen- ger RNA of host genes. We have used a variety of biochemical and biophysical experi- ments to measure how well MNV VPg binds to eIF4G and to identify the parts of both proteins that are involved in this interaction. We show that a sequence of about 20 amino acids at one end of MNV VPg–the C terminus– allows it to bind to a well-defined domain within eIF4G (called HEAT-1), and that it may adopt a helical structure when doing so. Our data suggest that this interaction is common to all noroviruses, including types that infect humans. We have also shown that the MNV VPg C-terminal peptide can inhibit norovirus protein synthesis, which raises the possibility that the VPg-eIF4G interaction could be targeted in the design of antiviral drugs. Competing Interests: The authors have declared that no competing interests exist. Competing Interests: The authors have declared that no competing interests exist. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: This work was funded by research grants from the Biotechnology and Biological Sciences Research Council (United Kingdom) to SCu (reference number BB/J001708/1) and to IGG (reference numbers BB/I012303/1 and BB/I01232X/ 1), and funding from the Wellcome Trust to IGG. IGG is a Wellcome Senior Fellow (reference number WT097997MA). SCG holds a Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society (Grant Number 098406/Z/12/Z). 1 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 Fig 1. MNV VPg interacts with eIF4GI via its HEAT-1 domain. (A) Schematic representation of eIF4GI (NCBI accession AAM69365.1), one of two eIF4G paralogues expressed in humans, and the paralog DAP5 (NCBI accession NP_001036024.3). Positions of domains that interact with other proteins of the translation initiation machinery are indicated, as are the cleavage sites of FMDV L protease and Rhinovirus 2A protease. The principal eIF4G fragments tha were sub-cloned for use in this study are also indicated. (B) SDS PAGE analysis of glutathione affinity pull-down assays that were performed to map the locus of MNV VPg binding. GST-fusions of various eIF4GI fragments (shown in panel A) were used as bait and His-tagged-MNV VPg(1–124) as prey. Left panel: protein mixtures applied to the glutathione-sepharose 4B beads (lanes 1–6). Right panel: proteins eluted with 10 mM glutathione (lanes 7–12). (C) SDS PAGE analysis of cobalt affinity pull-down assays to confirm that binding of MNV VPg occurs primarily through the eIF4G HEAT-1 domain. GST-fusion of various eIF4GI fragments with C-terminal His-tags were used as bait and untagged MNV VPg as prey. Lanes 1–4: protein mixtures applied to the cobalt resin; lanes 5–8: proteins eluted with 250 mM imidazole. (D) SDS PAGE analysis of cobalt affinity pull-down assays performed to confirm the eIF4G HEAT-1 domain as the locus of MNV VPg binding. His-tagged-MNV VPg(1–124) or His-tagged-FCV VPg(1–111) were used as bait proteins and GST-fusions of various eIF4GI fragments as prey. Left panel: protein mixtures applied to the cobalt resin (Lanes 1–6). Right panel: proteins eluted with 250 mM imidazole (lanes 7 12) Conserved MNV VPg Motif Binds eIF4G HEAT Fig 1. MNV VPg interacts with eIF4GI via its HEAT-1 domain. (A) Schematic representation of eIF4GI (NCBI accession AAM69365.1), one of two eIF4G paralogues expressed in humans, and the paralog DAP5 (NCBI accession NP_001036024.3). Positions of domains that interact with other proteins of the translation initiation machinery are indicated, as are the cleavage sites of FMDV L protease and Rhinovirus 2A protease. The principal eIF4G fragments that were sub-cloned for use in this study are also indicated. (B) SDS PAGE analysis of glutathione affinity pull-down assays that were performed to map the locus of MNV VPg binding. GST-fusions of various eIF4GI fragments (shown in panel A) were used as bait and His-tagged-MNV VPg(1–124) as prey. Left panel: protein mixtures applied to the glutathione-sepharose 4B beads (lanes 1–6). Introduction The Caliciviridae family of positive-sense, single-strand RNA viruses includes 5 genera that infect a variety of animals: Norovirus, Lagovirus, Nebovirus, Sapovirus and Vesivirus. Of these, the noroviruses constitute the greatest threat to human health. They cause around 18% of all cases of acute gastroenteritis worldwide and an estimated 200,000 deaths a year in children under 5 in developing nations [1, 2]. The Norovirus genus contains 6 genogroups (GI-GVI), three of which infect humans (GI, GII and GIV). The murine noroviruses (MNV) of the GV genogroup only infect rodents and have been studied extensively as a convenient model for human norovirus infections [3–6]. A critical step in calicivirus replication following delivery of the viral RNA to the cytoplasm of infected cells is the initiation of translation. As obligate intracellular parasites, viruses need to gain access to the protein synthesis machinery and, in common with many other viruses, caliciviruses do this by partially circumventing the host mechanisms of translation initiation [7]. The initiation of protein synthesis is a complex and highly regulated process by which the 43 S ribosomal initiation complex is first recruited to an mRNA molecule and then directed to the AUG start codon [8]. In eukaryotes this process involves a multitude of protein initiation factors along with initiator tRNA. In the normal or canonical mechanism of ribosomal recruitment an important early step is the binding of the heterotrimeric eukaryotic initiation factor 4F (eIF4F) to the m7G cap at the 5´-end of the mRNA. eIF4F mediates binding of the mRNA to the 43 S ribosomal initiation complex via its interaction with eIF3 and, in concert with a number of other initiation factors, promotes scanning of the mRNA until the initiation complex locates the AUG start codon [8, 9]. At this point ribosomal assembly can be completed and protein synthesis can commence. eIF4F consists of a large, multi-domain ‘platform’ protein, eIF4G, which has independent sites for recruitment of eIF4E, the cap-binding protein, and eIF4A, an ATP-dependent RNA heli- case that melts RNA secondary structure to facilitate ribosomal scanning (Fig 1A). eIF4G also serves as a platform for other proteins such as the polyA-binding protein (PABP), which aug- ments the efficiency of translation initiation, and the MAP kinase-interacting kinases 1 and 2, which are involved in regulating it (reviewed in [9]). In humans there are two closely-related PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 2 / 34 Conserved MNV VPg Motif Binds eIF4G HEAT-1 paralogs of eIF4G (I and II) which share 46% sequence identity [10]; there are also several alter- natively spliced isoforms for each paralog but the differentiation of their functions in transla- tion has yet to be fully determined [11]. DAP5 (Death-associated protein 5; also known as p97, NAT-1 and eIF4G2) is another eIF4G paralog [12]. It is missing the N-terminal portions of eIF4GI and eIF4GII which contain the motifs that bind PABP and eIF4E but it otherwise simi- lar in domain structure to the C-terminal two-thirds of these proteins (Fig 1A), sharing 39% and 43% amino acid sequence identity respectively. The adoption of non-canonical mechanisms of ribosomal recruitment is used by some RNA viruses to gain an advantage over host cell mRNA in competing for ribosome recruitment. A common strategy is the use of an internal ribosome entry site (IRES), a large RNA structure within the 5´-untranslated region (5´-UTR) of the viral genome, which can mediate ribosome binding independently of a 5´-cap and which typically has a reduced dependency on eIFs for translation initiation [13]. Viruses in possession of an IRES can therefore gain a competitive advantage by disabling redundant eIFs that are critical for cap-dependent translation initiation [7, 14, 15]. Some of the best studied examples of this strategy are to be found among the picor- naviruses (e.g. poliovirus (PV), human rhinovirus (HRV), and foot-and-mouth disease virus (FMDV)) [16–18]. Their RNA genomes lack a 5´-cap and encode proteases (PV 2Apro, HRV 2Apro, FMDV Lpro) that can specifically cleave eIF4G to separate the binding sites for eIF4E and eIF4A [19]. This disables cap-dependent translation initiation, leading to shut-off of host cell protein synthesis, but does not impair IRES-dependent translation initiation from the viral genome, which only requires the large C-terminal fragment of eIF4G retaining the binding site for eIF4A [9]. Consistent with its inability to bind the cap-binding protein eIF4E, DAP5 has been reported to preferentially support IRES-dependent translation initiation [20]. Although picornaviruses and caliciviruses share many similarities in their genome struc- tures and replication cycles, caliciviruses have acquired completely different mechanism for translation initiation. Right panel: proteins eluted with 10 mM glutathione (lanes 7–12). (C) SDS PAGE analysis of cobalt affinity pull-down assays to confirm that binding of MNV VPg occurs primarily through the eIF4G HEAT-1 domain. GST-fusions of various eIF4GI fragments with C-terminal His-tags were used as bait and untagged MNV VPg as prey. Lanes 1–4: protein mixtures applied to the cobalt resin; lanes 5–8: proteins eluted with 250 mM imidazole. (D) SDS PAGE analysis of cobalt affinity pull-down assays performed to confirm the eIF4G HEAT-1 domain as the locus of MNV VPg binding. His-tagged-MNV VPg(1–124) or His-tagged-FCV VPg(1–111) were used as bait proteins and GST-fusions of various eIF4GI fragments as prey. Left panel: protein mixtures applied to the cobalt resin (Lanes 1–6). Right panel: proteins eluted with 250 mM imidazole (lanes 7–12). Fig 1. MNV VPg interacts with eIF4GI via its HEAT-1 domain. (A) Schematic representation of eIF4GI (NCBI accession AAM69365.1), one of two eIF4G paralogues expressed in humans, and the paralog DAP5 (NCBI accession NP_001036024.3). Positions of domains that interact with other proteins of the translation initiation machinery are indicated, as are the cleavage sites of FMDV L protease and Rhinovirus 2A protease. The principal eIF4G fragments that were sub-cloned for use in this study are also indicated. (B) SDS PAGE analysis of glutathione affinity pull-down assays that were performed to map the locus of MNV VPg binding. GST-fusions of various eIF4GI fragments (shown in panel A) were used as bait and His-tagged-MNV VPg(1–124) as prey. Left panel: protein mixtures applied to the glutathione-sepharose 4B beads (lanes 1–6). Right panel: proteins eluted with 10 mM glutathione (lanes 7–12). (C) SDS PAGE analysis of cobalt affinity pull-down assays to confirm that binding of MNV VPg occurs primarily through the eIF4G HEAT-1 domain. GST-fusions of various eIF4GI fragments with C-terminal His-tags were used as bait and untagged MNV VPg as prey. Lanes 1–4: protein mixtures applied to the cobalt resin; lanes 5–8: proteins eluted with 250 mM imidazole. (D) SDS PAGE analysis of cobalt affinity pull-down assays performed to confirm the eIF4G HEAT-1 domain as the locus of MNV VPg binding. His-tagged-MNV VPg(1–124) or His-tagged-FCV VPg(1–111) were used as bait proteins and GST-fusions of various eIF4GI fragments as prey. Left panel: protein mixtures applied to the cobalt resin (Lanes 1–6). Right panel: proteins eluted with 250 mM imidazole (lanes 7–12). doi:10.1371/journal.ppat.1005379.g001 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 3 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 eIF4E-eIF4G interface by the translational regulator 4E-BP1 inhibits FCV translation, but has no effect on MNV translation [25]. Moreover, FCV translation is abrogated if FMDV Lpro is allowed to digest eIF4G, thereby separating its eIF4E and eIF4A binding domains, but MNV translation is unaffected by this proteolytic cleavage. These findings reinforce the idea that FCV VPg probably operates as a functional analogue of the m7G cap, by interacting directly with eIF4E, whereas MNV relies on a distinct mode of interaction with eIF4F (see below). More recent work suggests that PSaV VPg may direct translation initiation by a mechanism similar to FCV VPg since it also binds directly to eIF4E [36]; moreover, PSaV translation in vitro also exhibits the same sensitivity to inhibition by 4E-BP1 and Lpro digestion of eIF4G. The translation initiation directed by MNV VPg may not be completely independent of eIF4E in view of a recent study showing that modulation of the MAPK pathway to stimulate phosphorylation of eIF4E during MNV1 infection helps to promote virus replication [37]. Nevertheless it is clear that the interaction of MNV VPg with eIF4E is not its primary mode of engagement with the translational machinery. Recent work revealed that MNV VPg is instead capable of forming a specific interaction with eIF4G [38]. The interaction was identified ini- tially using tandem affinity purification of double-tagged MNV VPg to identify specific binding partners in eukaryotic cell lysates [38]. The direct nature of the interaction was confirmed using purified recombinant proteins. The site of VPg interaction was mapped to the central 4GM fragment of eIF4G (residues 652–1132) and alanine scanning mutagenesis of VPg sug- gested that residues in the C terminus made important contributions to binding affinity [38]. The work described here extends these initial observations by providing a detailed charac- terisation of the interaction of MNV VPg with eIF4G. We describe a biochemical and biophysi- cal dissection of the interaction using a combination of qualitative and quantitative binding assays, and NMR analyses to map precisely the particular regions of MNV VPg and eIF4G that mediate binding. We have discovered a C-terminal sequence motif in MNV VPg that is con- served in all noroviruses and is capable of recapitulating the interactions made by MNV VPg with the translation initiation machinery in living cells. We further find that this motif binds to the HEAT-1 domains of eIF4GI and eIF4GII with micromolar affinity. Our evidence suggests that this VPg peptide motif may adopt a helical conformation when bound to eIF4G HEAT-1 and can mediate the formation of a ternary VPg-eIF4G-eIF4A complex with 1:1:1 stoichiome- try in the presence of eIF4A. The functional significance of this interaction is evident from the observation that peptides corresponding to the eIF4G-binding motif of MNV VPg mediate an interaction with translation initiation complexes in cells that is indistinguishable from that made by the intact protein, and inhibit norovirus translation in vitro. Further our findings pro- vide proof of principle that the VPg-eIF4G interaction might provide a useful target for antivi- ral drug development. The observation that the VPg-NS6pro precursor is unable to bind eIF4G shows that a processing of the VPg-NS6 precursor is required for the interaction and suggests that proteolytic processing of the viral polyprotein precursor may be linked to the differential regulation of translation and RNA replication phases in the virus life cycle. Taken together our results provide significant new insights into the molecular mechanism of norovirus transla- tional initiation and identify an interaction interface that could be targeted for antiviral drug development. They have no IRES–indeed their 5´-UTRs are very small, in some cases as short as 4 nucleotides (nt) [5, 21]– but instead rely on a small protein called VPg (Virus Pro- tein, genome-linked) of ~110–140 amino acids that is covalently linked to the 5´-end of the viral genome for translation initiation. Picornaviruses are also known to have a VPg attached to the 5´-ends of their RNA genomes but the picornaviral protein, which is typically only around 22 amino acids and structurally unrelated to its caliciviral equivalent, is rapidly removed following cell entry [22] and plays no part in translation [23, 24]. In contrast, proteo- lytic digestion of caliciviral VPg in vitro has long been known to render its RNA non-infectious and to prevent translation in vitro [25–28]. The observation that infectivity of caliciviral RNA can be restored by covalent attachment of a 5´ m7G cap suggests that VPg serves as a proteina- ceous cap analogue [28, 29]. However, the particulars of the mechanism by which VPg per- forms this role have yet to be fully elucidated. Picornavirus VPg peptides appear to be largely unstructured in aqueous buffer [30], but recent work of the solution structures of caliciviral VPg proteins from three different genera has revealed that they all contain a helical core flanked by unstructured N and C termini. The feline calicivirus (FCV, a vesivirus) and porcine sapovirus (PSaV) VPg cores consist of very similar 3 helix bundles, while the core of MNV VPg is somewhat truncated and contains just the first two helices found in the other two proteins [31, 32]. These structural similarities belie the fact that, despite being critical for translation initia- tion, VPg proteins from different caliciviruses appear to perform this function in different ways. One of the earliest studies found that the VPg of noroviruses interact with eIF3 but the mechanistic implications of this finding remain unclear [33, 34]. Direct interactions with eIF4E have been reported for FCV VPg and MNV VPg [35] but, while this interaction appears to be crucial for FCV, its significance for MNV is more questionable. For example, disruption of the PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 4 / 34 Results MNV VPg binds to the eIF4G 4GM region via the eIF4G HEAT-1 domain We previously showed that the binding site on eIF4GI for MNV VPg resides within the central 4GM fragment (residues 652–1132; Fig 1A) [38]. This 4GM fragment contains a central HEAT-1 domain (residues 751–1010), a crescent-shaped module comprised of 10 α-helices Conserved MNV VPg Motif Binds eIF4G HEAT-1 [39], flanked by N- and C-terminal regions that are not known to be structured. To determine whether MNV VPg bound to the HEAT-1 domain we prepared GST-fusion proteins corre- sponding to sub-fragments of 4GM– the N- and C-terminal flanking regions (GST-4GM-N and GST-4GM-C respectively), and GST-eIF4G-HEAT-1 (Fig 1A)– and used them as bait pro- teins in glutathione pull-down assays to test their ability to interact with the viral protein (Fig 1B). These assays showed that GST-4GM and GST-eIF4G-HEAT-1 bound His-tagged MNV VPg with similar efficiency (lanes 7, 9) but that binding to GST-4GM-N was undetectable (lane 8), while binding to GST-4GM-C was at close to the background levels observed for GST (lanes 10, 11). To probe this interaction with greater sensitivity we compared the ability of GST-4GM con- structs with C-terminal His-tags to bind untagged MNV VPg in pull-down assays using cobalt affinity resin which exhibited less non-specific binding (Fig 1C). Further, using SYPRO-orange staining we were able to quantify binding by determining the ratio of the intensities of the VPg band to that of the band for the particular GST-4GM protein used in each assay. This con- firmed that GST-eIF4G-HEAT-1 bound VPg efficiently (lane 6) but that GST-4GM-N and GST-4GM-C both bound much more weakly (lanes 5, 7), each giving a band intensity ratio of only around 2% of that observed for the HEAT-1 domain. In a further set of pull-down assays, performed on cobalt affinity resin, this time using His- tagged MNV VPg as the bait (Fig 1D), we again demonstrated that GST-4GM and GST-eIF4- G-HEAT-1 bound His-tagged MNV VPg equally well (lanes 10, 12) but there was no observed interaction with GST alone (lane 8). This experiment also showed that FCV VPg, which is only 27% identical in amino acid sequence with MNV VPg, does not interact with either GST-4GM or GST-HEAT-1 (lanes 9, 11). Together these results suggest that MNV VPg interacts with the HEAT-1 domain within the 4GM fragment of eIF4GI. MNV VPg binds to the eIF4G 4GM region via the eIF4G HEAT-1 domain We previously showed that the binding site on eIF4GI for MNV VPg resides within the central 4GM fragment (residues 652–1132; Fig 1A) [38]. This 4GM fragment contains a central HEAT-1 domain (residues 751–1010), a crescent-shaped module comprised of 10 α-helices 5 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 A conserved sequence at the C terminus of MNV VPg is critical for HEAT-1 domain binding Previous work identified three mutations close to the C terminus of the 124-residue MNV VPg, V115A, D116A and F123A, which all reduced its interaction with eIF4F [38]. This obser- vation is likely due to the effect of the mutations on binding to eIF4G since the F123A mutation was also shown to substantially reduce binding of VPg to GST-4GM. The possible functional importance of the C terminus of MNV VPg was further underscored by alignment of VPg amino acid sequences from representative members of each of the norovirus genogroups, which revealed that residues 108–124 are highly conserved (S1 Fig). Since our NMR studies of MNV VPg have shown that the C-terminal half of MNV VPg (residues 63–124) is flexible and exhibits no interaction with the structured core of the protein [32], the sequence conservation in this region probably reflects the functional constraint imposed by the interaction with eIF4G. We therefore hypothesized that VPg is likely to bind to eIF4G HEAT-1 via a contiguous sequence close to the C terminus. As a first test of this hypothesis, we compared the ability of MNV VPg proteins with and without the unstructured C terminus to bind to eIF4G HEAT-1. Using GST-eIF4GI HEAT-1 containing a C-terminal His-tag as the bait protein in a cobalt affinity pull-down assay, we found that it could interact with untagged MNV VPg(1–124) but not with MNV VPg(1–85), a truncated version of the protein that retains the helical core but lacks the unstructured C termi- nus (Fig 2A). This confirms that binding to eIF4G requires the C terminus of VPg. To map the extent of the MNV VPg C terminus needed to bind eIF4G we introduced blocks of 3–4 amino acid substitutions within and beyond the conserved sequence motif. In addition 6 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 or a His-tagged GFP-VPg-NS6-4´ fusion (containing just the first 4 amino acids of NS6) as bait, and untagged eIF4GII HEAT-1 as prey. Lanes 1–5: protein mixes; lanes 6–10: proteins eluted with 250 mM imidazole. (E) Structure of the VPg-NS6 junction showing that the N-terminus of NS6 is tightly folded into the body of the protease. Structure shown is the crystal structure of VPg-NS6 from human Norwalk virus, which is very similar to the structure of MNV NS6 [40, 41]. Note that due to disorder in the crystal, only the C terminus of VPg is visible [41]. or a His-tagged GFP-VPg-NS6-4´ fusion (containing just the first 4 amino acids of NS6) as bait, and untagged eIF4GII HEAT-1 as prey. Lanes 1–5: protein mixes; lanes 6–10: proteins eluted with 250 mM imidazole. (E) Structure of the VPg-NS6 junction showing that the N-terminus of NS6 is tightly folded into the body of the protease. Structure shown is the crystal structure of VPg-NS6 from human Norwalk virus, which is very similar to the structure of MNV NS6 [40, 41]. Note that due to disorder in the crystal, only the C terminus of VPg is visible [41]. doi:10.1371/journal.ppat.1005379.g002 to the F123A mutation already known to reduce binding to eIF4G [38], we made the following non-conservative substitutions in His-tagged MNV VPg: 120KIN120ESA, 116DYGE116RAPK, 112DRQV112REAS, 108WADD108APRR, 104VGPS104KKAH, and an insertion of GSGSGS after residue 103 (Fig 2B). Wild-type and mutant MNV VPg proteins were used as bait proteins in cobalt affinity pull-down assays to determine their ability to bind GST-eIF4GI and GST-eIF4- GII HEAT-1 proteins, which are 83% identical in amino acid sequence within this domain (Fig 2C). All mutations made in the conserved region (residues 108–123) had drastically reduced binding capacity for eIF4GI HEAT-1. However, neither the 104VGPS104KKAH mutations nor the GSGSGS insertion at position 103, which introduced substitutions at the N-terminal end of the conserved motif, had any observable effect on MNV VPg binding to eIF4GI HEAT-1 (Fig 2C, upper panel). A very similar pattern of effects was observed when the MNV VPg mutants were tested for binding to GST-eIF4GII HEAT-1, with the exception that the 104VGPS104KKAH mutations were found to moderately reduce binding of the viral protein (Fig 2C, lower panel). These data are consistent with the hypothesis that the conserved motif within residues 108–124 of MNV VPg plays a critical role in eIF4G binding, and that the N-ter- minal boundary of the binding domain for eIF4GI and eIF4GII lies between residues 104 and 108. Finally, given that the eIF4G HEAT-1 binding site on MNV VPg is located very close to the C terminus of the protein, which is generated by proteolytic processing at the VPg-NS6 junc- tion within the virus polyprotein precursor, we asked whether the interaction with eIF4G HEAT-1 requires a free C terminus. To address this question we performed cobalt affinity pull down assays to compare the ability of eIF4GII HEAT-1 to bind wild-type His-GFP tagged MNV VPg and a His-GFP tagged MNV VPg-NS6 fusion protein which contained a C139A mutation in the NS6 protease to prevent auto-processing of the VPg-NS6 junction; (the GFP tags are not relevant to this assay–they were added in anticipation of using these reagents for other experiments that were not performed). We found that wild-type His-GFP-VPg was effec- tive at capturing eIF4GII HEAT-1 in these assays but His-GFP-VPg-NS6 had no observable binding activity (Fig 2D; lanes 6 and 8). The fusion protein exhibited the same lack of binding activity as the His-GFP-VPg (F123A) mutant, which was used as a negative control (lane 7). y g g Intriguingly, a His-GFP-VPg-NS6-4´ construct, which has just the first four amino acids of NS6 fused to the C-terminus of VPg, binds efficiently to HEAT-1 (lane 9)–in fact almost as well as His-GFP-VPg. The most plausible explanation for this result is structural. The folding of NS6, even when part of the VPg-NS6 fusion protein [41], tucks the N terminus of the NS6 protease into the globular fold of the molecule (Fig 2E). The proximity of this bulky folded pro- tease to the eIF4G-binding motif at the extreme C terminus of VPg (which is not present in His-GFP-VPg-NS6-4´) probably prevents binding to eIF4G HEAT-1 by steric hindrance. (We confirmed that the NS6 protease domain folds correctly within the context of our His- GFP-VPg-NS6 construct by showing that expression of an active form of the protein results in processing of the VPg-NS6 junction– see S2 Fig). These results suggest that processing of the VPg-NS6 junction is required for the MNV VPg-eIF4G interaction, but that a free VPg C-terminus is not a specific requirement for binding. Conserved MNV VPg Motif Binds eIF4G HEAT-1 Fig 2. MNV VPg interacts with eIF4GI and eIF4GII HEAT-1 domains via the C-terminal residues 104–124. (A) SDS PAGE analysis of cobalt affinity pull- down assay using a GST-eIF4GI HEAT-1 construct with a C-terminal His-tag as bait and either untagged MNV VPg(1–124) or MNV VPg(1–85) as prey. Protein mixtures are shown in lanes 1 and 2 (blue labels); bound proteins eluted with 250 mM imidazole are in lanes 3 and 4 (red labels). (B) Sequence alignment showing the location of amino acid substitutions introduced into the C terminus of His-tagged MNV VPg. (C) SDS PAGE analysis of cobalt affinity assays using His-tagged MNV mutants (panel B) as bait and either GST-eIF4GI-HEAT-1 (top panel) or GST-eIF4GII-HEAT-1 as prey. Lanes 1–2: input proteins; lanes 3–11: eluted proteins. (D) SDS PAGE analysis of cobalt affinity pull-down assay using either His-tagged GFP-VPg(1–124) wild-type, GFP-VPg(1–124) F123A, a His-tagged GFP-VPg-NS6 fusion (containing the inactivating C139A mutation of the protease active site Cys (NS6 numbering)) Fig 2. MNV VPg interacts with eIF4GI and eIF4GII HEAT-1 domains via the C-terminal residues 104–124. (A) SDS PAGE analysis of cobalt affinity pull- down assay using a GST-eIF4GI HEAT-1 construct with a C-terminal His-tag as bait and either untagged MNV VPg(1–124) or MNV VPg(1–85) as prey. Protein mixtures are shown in lanes 1 and 2 (blue labels); bound proteins eluted with 250 mM imidazole are in lanes 3 and 4 (red labels). (B) Sequence alignment showing the location of amino acid substitutions introduced into the C terminus of His-tagged MNV VPg. (C) SDS PAGE analysis of cobalt affinity assays using His-tagged MNV mutants (panel B) as bait and either GST-eIF4GI-HEAT-1 (top panel) or GST-eIF4GII-HEAT-1 as prey. Lanes 1–2: input proteins; lanes 3–11: eluted proteins. (D) SDS PAGE analysis of cobalt affinity pull-down assay using either His-tagged GFP-VPg(1–124) wild-type, GFP-VPg(1–124) F123A, a His-tagged GFP-VPg-NS6 fusion (containing the inactivating C139A mutation of the protease active site Cys (NS6 numbering)) PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 7 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 or a His-tagged GFP-VPg-NS6-4´ fusion (containing just the first 4 amino acids of NS6) as bait, and untagged eIF4GII HEAT-1 as prey. Lanes 1–5: protein mixes; lanes 6–10: proteins eluted with 250 mM imidazole. (E) Structure of the VPg-NS6 junction showing that the N-terminus of NS6 is tightly folded into the body of the protease. Structure shown is the crystal structure of VPg-NS6 from human Norwalk virus, which is very similar to the structure of MNV NS6 [40, 41]. Note that due to disorder in the crystal, only the C terminus of VPg is visible [41]. Conserved MNV VPg Motif Binds eIF4G HEAT-1 The C terminus of MNV VPg binds eIF4G HEAT-1 domains with micromolar affinity In these experiments, small volumes of untagged eIF4GII HEAT-1 were injected into a 200 μL sample of untagged MNV VPg(1–124); the heat changes arising from the interaction of the two proteins were measured and fitted to a single-site binding model (Fig 3C). The KD determined from three independent ITC experiments was 5.2 μM. This value is comparable to the KD obtained from fluorescent anisotropy measurements performed with the MNV VPg(104–124) peptide (2.8 μM) and strongly suggests that MNV VPg residues outside the C-terminal region do not contribute significantly to eIF4G binding. All subsequent binding experiments were performed with the tighter-binding MNV VPg 104–124 peptide. It was observed to bind with essentially identical affinity to the HEAT-1 domains of eIF4GII (KD = 2.8 μM) and eIF4GI (KD = 3.0 μM) (Fig 3A). However, MNV VPg (104–124) bound about six-fold less well to the DAP5 HEAT-1 domain (KD = 20 μM) (Fig 3B). This difference in affinity suggests that VPg is more likely to recruit eIF4GI and eIF4GII in infected cells than DAP5 – even though DAP5 appears to be about 2-fold more abundant [20, 43]–and may explain why DAP5 was not observed as a ligand for MNV VPg in the tandem-affin- ity purifications that were used originally to identify eIF4G as a direct ligand in cell lysates [38]. To determine the binding affinity of full-length MNV VPg for eIF4GII HEAT-1 we used iso- thermal titration calorimetry (ITC). In these experiments, small volumes of untagged eIF4GII HEAT-1 were injected into a 200 μL sample of untagged MNV VPg(1–124); the heat changes arising from the interaction of the two proteins were measured and fitted to a single-site binding model (Fig 3C). The KD determined from three independent ITC experiments was 5.2 μM. This value is comparable to the KD obtained from fluorescent anisotropy measurements performed with the MNV VPg(104–124) peptide (2.8 μM) and strongly suggests that MNV VPg residues outside the C-terminal region do not contribute significantly to eIF4G binding. The C terminus of MNV VPg binds eIF4G HEAT-1 domains with micromolar affinity The results described above indicate that a contiguous region at the C terminus of VPg medi- ates binding to the eIF4G HEAT-1 domain. To determine the affinity of the interaction, we used peptides corresponding to the MNV VPg C terminus in fluorescence anisotropy experi- ments. Peptides corresponding to MNV VPg 104–124 and 108–124 and incorporating an N- terminal FITC group were chemically synthesized and tested for binding to untagged versions of the HEAT-1 domains of eIF4GI (748–993) and eIF4GII (753–1003) (see Materials and Methods). In addition, we tested the binding of these peptide to the HEAT-1 domain of DAP5 (61–323), which shares approximately 40% amino acid sequence identity with the HEAT-1 domains of eIF4GI and eIF4GII and is very similar in structure [42]. In a first set of experiments, we compared binding of the two MNV VPg peptides to eIF4GII HEAT-1. These showed that MNV VPg 104–124 binds with a KD of 2.8 μM, while the shorter MNV VPg 108–124 peptide binds with about four-fold weaker affinity (KD = 10.6 μM) (Fig 3A). Very similar affinities for these two peptides were observed in experiments performed with the fusion protein MBP-eIF4GII HEAT-1 (S3 Fig). These results are consistent with the pull-down experiments, which found that mutations on the C-terminal side of position 104 in VPg (104VGPS104KKAH) reduced binding to eIF4GII, whereas the 103GSGSGS insertion imme- diately before position 104 did not (Fig 2C). All subsequent binding experiments were performed with the tighter-binding MNV VPg 104–124 peptide. It was observed to bind with essentially identical affinity to the HEAT-1 domains of eIF4GII (KD = 2.8 μM) and eIF4GI (KD = 3.0 μM) (Fig 3A). However, MNV VPg (104–124) bound about six-fold less well to the DAP5 HEAT-1 domain (KD = 20 μM) (Fig 3B). This difference in affinity suggests that VPg is more likely to recruit eIF4GI and eIF4GII in infected cells than DAP5 – even though DAP5 appears to be about 2-fold more abundant [20, 43]–and may explain why DAP5 was not observed as a ligand for MNV VPg in the tandem-affin- ity purifications that were used originally to identify eIF4G as a direct ligand in cell lysates [38]. To determine the binding affinity of full-length MNV VPg for eIF4GII HEAT-1 we used iso- thermal titration calorimetry (ITC). 8 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 Fig 3. MNV VPg 104–124 interacts with the HEAT-1 domains of eIF4GI, eIF4GII and DAP5 with low micromolar affinity. (A-B) FITC-labelled peptides– MNV VPg(104–124) and MNV VPg(108–124)–were used in fluorescence anisotropy binding assays with unlabelled HEAT-1 domains of eIF4GI (748–993), eIF4GII (745–1003) and DAP5 (61–323) in order to measure the affinity of the interaction. The normalised change in fluorescence anisotropy (relative to a no-protein control), ΔFA, is plotted against protein concentration. Error bars in ΔFP indicate the standard deviation of 5 (eIF4GI) or 10 (eIF4GII) independent measurements. The solid lines indicate the fit to a single-site binding model. (A) Comparison of the binding of MNV VPg(104–124) and MNV VPg(108–124) to eIF4GI and eIF4GII HEAT-1 domains. (B) Binding of MNV VPg(104–124) to the HEAT-1 domain of DAP5. The fit is calculated with the fluorescence anisotropy data from three independent experiments (all included in the graph). (C) ITC experiments in which unlabelled eIF4GII (745–1003) was titrated into MNV VPg(1–124). Top panel: raw data obtained for a representative experiment from 20 injections (firstly with a volume of 0.5 μL followed by 19 injections of 2 μL of eIF4GII HEAT-1). Bottom panel: the integrated data with a best-fit curve for the representative experiment generated for a single-site binding model using the Origin software package. Fig 3. MNV VPg 104–124 interacts with the HEAT-1 domains of eIF4GI MNV VPg(104–124) and MNV VPg(108–124)–were used in fluorescence a eIF4GII (745–1003) and DAP5 (61–323) in order to measure the affinity of no-protein control), ΔFA, is plotted against protein concentration. Error bar measurements. The solid lines indicate the fit to a single-site binding mode to eIF4GI and eIF4GII HEAT-1 domains. (B) Binding of MNV VPg(104–124 Fig 3. MNV VPg 104–124 interacts with the HEAT-1 domains of eIF4GI, eIF4GII and DAP5 with low micromolar affinity. (A-B) FITC-labelled peptides– MNV VPg(104–124) and MNV VPg(108–124)–were used in fluorescence anisotropy binding assays with unlabelled HEAT-1 domains of eIF4GI (748–993), eIF4GII (745–1003) and DAP5 (61–323) in order to measure the affinity of the interaction. The normalised change in fluorescence anisotropy (relative to a no-protein control), ΔFA, is plotted against protein concentration. Error bars in ΔFP indicate the standard deviation of 5 (eIF4GI) or 10 (eIF4GII) independent measurements. The solid lines indicate the fit to a single-site binding model. (A) Comparison of the binding of MNV VPg(104–124) and MNV VPg(108–124) to eIF4GI and eIF4GII HEAT-1 domains. The direct interaction between VPg and eIF4GI HEAT-1 is conserved in noroviruses Having established that the C terminus of MNV VPg could mediate binding of the protein to eIF4G HEAT-1 domains with micromolar affinity, we sought to test how well this function is conserved in related viruses. This seemed likely since 10 out of the 21 residues in the region defined by MNV VPg 104–124 are strictly conserved in all norovirus genogroups (S1 Fig). Intriguingly, there also appears to be significant conservation of the eIF4G-binding motif in the C terminus of human astrovirus 4 (11 out of 23 residues are identical), while there is very little indication of the occurrence of the motif in the C terminus of VPg from FCV, a member of the Vesivirus genera of the Caliciviridae family (Fig 4A). To investigate the eIF4G-binding properties of these various VPg proteins we fused C-terminal sequences corresponding to the sequences that aligned with residues 102–124 from MNV VPg to GST (Fig 4A). The resulting PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 9 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 Fig 4. The C-termini of Noroviral VPg proteins contain a conserved sequence motif that binds to eIF4G HEAT-1. (A) Top: Amino acid sequence alignments of representative sequences of the C-termini of all 6 genogroups of Norovirus (GI-GIV). Second from top: Alignment of the C-termini of Human Astrovirus 4 (HuAST4) VPg with MNV (GV) VPg. Third from top: Alignment of the C-termini of FCV VPg with MNV (GV) VPg. Bottom: Alignment of the C terminus of MNV (GV) VPg with Rice Yellow Mottle Virus (RYMV) VPg. The representative strains used in the alignments are GI—Hu/GI/Norwalk/1968/US, (NCBI accession AAC64602), GII—Lordsdale virus Hu/GII/Lordsdale/1993/UK (NCBI accession P54634), GIII—Bo/GIII/B309/2003/BEL (NCBI accession ACJ04905.1), GIV—Hu/GIV.1/LakeMacquarie/NSW268O (NCBI accession AFJ21375), GV—Mu/NoV/GV/MNV1/2002/USA (NCBI accession ABU55564.1), GVI—dog/GVI.1/HKU_Ca035F/2007/HKG (NCBI accession FJ692501), FCV—F9 strain (NCBI accession P27409.1), HuAst4—Human astrovirus 4 (NCBI accession Q3ZN07), Rice yellow mottle virus isolate CI4 (NCBI accession NC_001575). Sequence alignments were performed by ClustalW [6] and BioEdit (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). #Denotes sequences that were used in pull-down assays (see panel B). (B) The Conserved MNV VPg Motif Binds eIF4G HEAT oteins contain a conserved sequence motif that binds to eIF4G HEAT-1. (A) Top: Amino acid sequence the C-termini of all 6 genogroups of Norovirus (GI-GIV). Second from top: Alignment of the C-termini of Human ) VPg. Third from top: Alignment of the C-termini of FCV VPg with MNV (GV) VPg. Bottom: Alignment of the C w Mottle Virus (RYMV) VPg. The representative strains used in the alignments are GI—Hu/GI/Norwalk/1968/US, ale virus Hu/GII/Lordsdale/1993/UK (NCBI accession P54634), GIII—Bo/GIII/B309/2003/BEL (NCBI accession arie/NSW268O (NCBI accession AFJ21375), GV—Mu/NoV/GV/MNV1/2002/USA (NCBI accession 35F/2007/HKG (NCBI accession FJ692501), FCV—F9 strain (NCBI accession P27409.1), HuAst4—Human ce yellow mottle virus isolate CI4 (NCBI accession NC_001575). Sequence alignments were performed by ncsu edu/BioEdit/bioedit html) #Denotes sequences that were used in pull-down assays (see panel B) (B) The Fig 4. The C-termini of Noroviral VPg proteins contain a conserved sequence motif that binds to eIF4G HEAT-1. (A) Top: Amino acid sequence alignments of representative sequences of the C-termini of all 6 genogroups of Norovirus (GI-GIV). Second from top: Alignment of the C-termini of Human Astrovirus 4 (HuAST4) VPg with MNV (GV) VPg. Third from top: Alignment of the C-termini of FCV VPg with MNV (GV) VPg. Bottom: Alignment of the C terminus of MNV (GV) VPg with Rice Yellow Mottle Virus (RYMV) VPg. (B) Binding of MNV VPg(104–124) to the HEAT-1 domain of DAP5. The fit is calculated with the fluorescence anisotropy data from three independent experiments (all included in the graph). (C) ITC experiments in which unlabelled eIF4GII (745–1003) was titrated into MNV VPg(1–124). Top panel: raw data obtained for a representative experiment from 20 injections (firstly with a volume of 0.5 μL followed by 19 injections of 2 μL of eIF4GII HEAT-1). Bottom panel: the integrated data with a best-fit curve for the representative experiment generated for a single-site binding model using the Origin software package. doi:10.1371/journal.ppat.1005379.g003 GST-VPg fusion proteins included C-terminal VPg sequences from norovirus genogroups GI, GII, GIII and GV (MNV), from FCV F9 and from human astrovirus 4 (Fig 4A) and were used as prey in cobalt affinity pull-down assays to determine whether they bound to His-eIF4GI HEAT-1, which was used as bait. All of the norovirus VPg GST fusions were found to interact with eIF4GI HEAT-1 similarly effectively (lanes 4–7), while neither FCV nor human Astro- virus 4 GST-VPg fusions (lanes 8, 9) were observed to bind (Fig 4B). This strongly suggests that direct binding to eIF4G is a conserved function of the C terminus of norovirus VPg pro- teins. The interaction appears to be specific since partial conservation of the eIF4G-binding motif in human astrovirus 4 VPg was not sufficient to permit binding. The negative result for FCV VPg is consistent with previous work showing that it interacts directly with eIF4E and that, in contrast to MNV, FCV translation initiation is inhibited in the presence of the eIF4E regulatory protein, 4E-BP1 [25, 35]. 10 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 C terminus may adopt a helical conformation when bound to eIF4G -1 To begin to probe the structural basis of the interaction between MNV VPg and eIF4G HEAT- 1, we used GST-fusions with MNV VPg 102–124 to test the contribution to binding of specific amino acids in the C terminus of VPg. In an initial series of experiments we mutated conserved residues within the motif that is responsible for binding to eIF4G; these included W108A, V115A and the double-mutation D110A-D111A. The effects of the mutations were tested using His-eIF4GI HEAT-1 as the bait protein in a pull-down assay. In each case the mutation was found to abrogate binding (Fig 4C; lanes 5–7) (See Fig 4E for quantification). g g ( g ) ( g q ) The results of these mutagenesis experiments, along with the previous observation that the F123A point mutation also severely reduces binding (Fig 2C), were intriguing because they suggested that the binding activity is distributed throughout the conserved motif within resi- dues 108–124 of MNV VPg. Since this motif is 17 amino acids long and is found in an unstruc- tured C terminus in the free protein [32], the fact that single point mutations within it (W108A, V115A, F123A) all severely reduced binding suggested that the interaction with eIF4G is unlikely to involve the C terminus of VPg lying in an extended conformation across the surface of the HEAT-1 domain. For such a mode of binding, individual point mutations would be unlikely to completely abrogate binding, particularly for mutations at the extremities of the motif. However, the strong effects of point mutations might be rationalised if the MNV VPg C terminus adopts a rigid structure upon interaction with eIF4G. Consistent with this idea, a helical wheel representation of the MNV VPg C terminus suggests that the substitutions tested in the first series of experiments described above (and F123A) would all be on the same side of an α-helix (Fig 4F) (with the exception of D110A in the D110A/D111A double-substitu- tion). This model suggests that there would be an interacting flank (containing W108, V115 and F123) and a non-interacting flank. It also predicts that the reduction in binding observed for the D110A-D111A double mutation is likely to be attributable to the effects of the D111A substitution since it lies on the interacting flank of the helix. Conserved MNV VPg Motif Binds eIF4G HEAT-1 indicated VPg sequences in panel A were fused to the C terminus of GST for use as prey in cobalt affinity pull-down assays in which His-tagged eIF4GI HEAT-1 was used as bait, and analysed by SDS PAGE. Lanes 1–3: input samples of some of the proteins used in pull-down experiments (black labels); lanes 4–11: eluted proteins (red labels; bands at ~27 kDa indicate GST-VPg C-terminal constructs that bound to His-eIF4G HEAT-1); lanes 12–19 (blue labels)–protein mixes used in the pull-down experiments. (C, D) Mutational analysis of the C-terminal sequences of MNV VPg were performed using GST-MNV VPg C-terminal fusions in the same way as the experiment presented in panel B. Labelling is colour-coded as in panel B. (E) Quantification of the pull-down results shown in panels C and D. (F) Helical wheel representation of MNV VPg 102–124 generated using http://heliquest.ipmc.cnrs.fr/ [44]. Residues mutated for the pull-down assays are indicated by bold-face labels and colour-coded by the effect of the substitution on binding to eIF4G HEAT-1: red–no binding; black—binding at or near wild-type levels. doi:10.1371/journal.ppat.1005379.g004 The representative strains used in the alignments are GI—Hu/GI/Norwalk/1968/US, (NCBI accession AAC64602), GII—Lordsdale virus Hu/GII/Lordsdale/1993/UK (NCBI accession P54634), GIII—Bo/GIII/B309/2003/BEL (NCBI accession ACJ04905.1), GIV—Hu/GIV.1/LakeMacquarie/NSW268O (NCBI accession AFJ21375), GV—Mu/NoV/GV/MNV1/2002/USA (NCBI accession ABU55564.1), GVI—dog/GVI.1/HKU_Ca035F/2007/HKG (NCBI accession FJ692501), FCV—F9 strain (NCBI accession P27409.1), HuAst4—Human astrovirus 4 (NCBI accession Q3ZN07), Rice yellow mottle virus isolate CI4 (NCBI accession NC_001575). Sequence alignments were performed by ClustalW [6] and BioEdit (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). #Denotes sequences that were used in pull-down assays (see panel B). (B) The PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 11 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Mapping the MNV VPg binding site on eIF4G HEAT-1 To further probe the molecular details of the MNV VPg-eIF4G interaction we used nuclear magnetic resonance (NMR) spectroscopy to map the VPg binding site on the HEAT-1 domain. 1H-15N TROSY NMR experiments were used to monitor chemical shift perturbations in the spectra obtained from a solution of 15N labelled eIF4GI HEAT-1 domain as it was titrated with a synthetic unlabelled peptide corresponding to MNV VPg(104–124). We used the backbone resonance assignments of eIF4GI HEAT-1 deposited in the Biological Magnetic Resonance Data Bank [45] to assign changes that occurred in the 1H-15N TROSY NMR spectra on forma- tion of the VPg(104–124):eIF4GI HEAT-1 complex to particular residues in HEAT-1. Ideally in such experiments a limited number of amide (HN) resonances would shift (in fast or slow NMR time scales) in a dose-dependent manner. Since the spectra are dominated by sig- nals from backbone amides (though they also contain peaks from side-chain amides in Asn, Gln, Arg and Trp residues), they can be used to map ligand binding sites. Typically clusters of surface residues that are affected by titration of the ligand are presumed to define the likely binding surface. However, titration of the VPg peptide into 15N labelled eIF4GI HEAT-1 was found to cause significant perturbations of most of the HN resonances, which indicates that binding of VPg affects the entire HEAT-1 domain. The exact nature of these conformational changes in HEAT-1 cannot yet be determined precisely, though they are likely to be relatively subtle since the domain retains the capacity to interact with eIF4A, which requires a large bind- ing interface (see below and Fig 5C). As a result of the extensive perturbations, the spectrum observed at the end point of the titration (1.35 peptide:protein molar ratio) was very different to the reference spectrum obtained before the start of the titration (Figs 5A and S4). Moreover, many of the affected resonances appeared to enter a dynamics regime dominated by conforma- tional exchange on slow or intermediate time scales, since HN peaks disappeared from the HSQC as the titration proceeded. While some new peaks appeared later in the titration, it was not possible to relate these to the peaks that had disappeared earlier in the titration (S4 Fig). However, from closer inspection of HSQC spectra it is clear that a limited number of peaks are selectively broadened very early in the titration (Fig 5A). Conserved MNV VPg Motif Binds eIF4G HEAT-1 D111, V115 and F123. Confirmation of this model requires further structural analysis, but if it is confirmed the helical mode of binding would likely apply to all other norovirus VPg proteins since the sequence of the eIF4G-binding motif is highly conserved (Fig 4A). C terminus may adopt a helical conformation when bound to eIF4G -1 To test the possibility that the C terminus of VPg may adopt a helical conformation upon interacting with eIF4G HEAT-1, we prepared a second set of mutations in the context of our GST-MNV VPg (102–124) fusion protein for use in pull-down assays: D110A, R113A and K120A. Although these mutations are distributed throughout the conserved eIF4G-binding motif in MNV VPg, they were all predicted to be on the non-interacting flank of the putative helix. As can be seen in the pull-down experiments in Fig 4D (which are quantified in Fig 4E), GST-MNV VPg (102–124) fusions containing D110A, R113A, and K120A mutations are all capable of binding eIF4GI HEAT-1 (lanes 1–3). Of these, only the K120A mutant appears to bind as effectively as the wild-type sequence; although the binding of the D110A and R113A mutants appears to be slightly weakened (about 60–80% of wild-type; Fig 4E), they still bind much more effectively than the F123A mutant (lane 4). These results are consistent with the hypothesis that the C terminus of MNV VPg adopts a helical conformation upon interaction with eIF4G and suggest that the helical flank in contact with eIF4G contains residues W108, PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 12 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Mapping the MNV VPg binding site on eIF4G HEAT-1 Since these changes happen at low peptide concentrations, they probably indicate specific contacts made by the viral protein on the eIF4GI HEAT-1 domain. To obtain a quantitative understanding of how specific HN reso- nances were changing early in the titration, the volume of non-overlapped peaks in the HSQCs were determined for both the reference spectrum (eIF4GI HEAT-1 only) and the spectrum obtained after addition of 0.35 molar equivalents of peptide. Volume ratios of matching peaks in the two spectra were calculated for each assigned residue (Rv = \|Reference peak vol\|)/(\|0.35 eqv peak vol\|) (Fig 5B). Since the number of scans recorded for the reference and for the 0.35 molar equivalent HSQC spectra were 16 and 28 respectively and the total volumes are not sig- nificantly different, Rv should be p(16/28) (= 0.76) for unperturbed residues. The only HN res- onances for which this ratio is approximately 0.76 are at the N- and C-termini of the domain (residues 749–973 and 977–993) and within the loop formed by residues 869–880 of the mole- cule, which is in a disordered region in the eIF4GII HEAT-I crystal structure [39]; these fea- tures appear not to interact with the MNV VPg C-terminal peptide. Most other residues have higher values of Rv, as expected from the large number of perturbations caused by peptide binding. However, a small number of resonances deviate substantially from this expected value. In particular HN resonances from residues 807, 823, 825, 826, 851, 896–898, 904, 912, PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 13 / 34 Most of the other residues with very large Rv values cluster in two surface patches in the middle of the eIF4GI HEAT-1 domain, one formed by residues L897 and L939, which are close to the binding site for eIF4A, and one containing residues L912, E914, H918 and D919, which are on the opposite face of HEAT-1. Clearly, surface residues are more likely to be able to make direct contact with the bound MNV VPg peptide. To confirm whether resonance perturbations assigned to residues that cluster in the surface patches in the middle of the HEAT-1 domain are involved in direct contact with MNV VPg, we tested the effect of point mutants on binding of the viral protein. The following substitu- tions were made in eIF4GI HEAT-1: L897A and L939A in the cluster proximal to the eIF4A binding site, and L912A, H918A, D919R, and L939A in the distal cluster. (A fifth substitution in the distal cluster, E914R, was also prepared but found to have picked up an additional muta- tion– presumably due to a PCR error– resulting in the substitution K901M which is adjacent to the proximal cluster; this E914R/K109M double substitution therefore changes both putative VPg-binding sites). As a negative control we also made the eIF4GI HEAT-1 (K771A) mutant since this residue is in the N terminus of the HEAT-1 domain and was found in our NMR anal- ysis to be unaffected by binding of MNV VPg (Rv = 0.55). The VPg-binding ability of these eIF4G mutants was tested in a cobalt affinity pull-down assay in which His-MNV VPg(1–124) was used as bait. A first series of assays was performed using a binding buffer containing 150 mM NaCl (see Materials and Methods). This demon- strated that the H918A substitution in the distal cluster had the most severe defect in MNV VPg binding, reducing it to less than 10% of the control, while nearby substitutions D919R and L939A showed more modest decreases in binding (Fig 5D and 5E). Within the proximal clus- ter, the L897A mutation had no effect on binding, while the L939A mutation reduced binding to about 60% of the control value. To increase sensitivity, a second series of assays was per- formed in buffer containing 300 mM NaCl. Under these higher-salt conditions, binding defects due to the mutations were more pronounced. Conserved MNV VPg Motif Binds eIF4G HEAT-1 were taken from the Bio Magnetic Resonance Database (BMRB id 18738). (B) Plot of the ratio of the absolute values of non-overlapping peak volumes for assigned residues obtained at 0 and 0.35 molar equivalents of the MNV VPg(104–124) peptide. High values of the ratios indicate the residues in eIF4GI HEAT-1 most affected by MNV VPg binding. The red dashed line indicates the expected volume ratio (Rv) for unperturbed residues: 0.76, based on the relative number of scans performed in each HSQC experiment. Points above the green dashed line indicate residues for which the peak volume ratio was at least six times higher than the baseline for unperturbed amides. (C) Model of the complex of eIF4GI HEAT-1 (grey, surface representation) and eIF4A (cyan, cartoon representation) indicating the predicted location of the VPg binding site. Surface residues in eIF4GI HEAT-1 that exhibit the greatest changes in Rv in the presence of the MNV VPg(104–124) peptide are coloured yellow. The model of eIF4GI HEAT-1 was generated using SWISS-MODEL [46]; the complex was created by superposing this model on the eIF4G HEAT-1 component of the yeast eIF4G-eIF4A co-crystal structure [47]. (D) SDS PAGE analysis of pull- down assays to test the effect of mutations in the putative VPg-binding site on eIF4G HEAT-1 on binding to the viral protein. His-tagged MNV VPg(1–124) was used as bait; eIF4GI HEAT-1 wild-type or mutant proteins were used as prey, and the pull-down buffer contained 150 mM NaCl (top panel) or 300 mM NaCl (bottom panel). Lanes 1–8 (blue labels): input protein mixtures; lanes 9–16 (red labels): eluted proteins. (E) Graphical representation of the results of the cobalt affinity pull-down assays performed in the presence of 150 or 300 mM NaCl, plotted as the ratio of the optical densities of the eIF4GI HEAT-1 band to that of the His-MNV VPg band in eluted fractions (lanes 9–16 in panel D). Band densities were quantified using ImageJ (http://imagej.nih.gov/ij/) and the ratios were normalised to the wild-type control. doi:10.1371/journal.ppat.1005379.g005 914, 917–921, 935, 939 and 950 all have ratios that are more than 6 times greater than the expected value. The position of these residues in a homology model of the eIF4GI HEAT-1 domain is given in Fig 5C. Several of these residues (823, 917, 920, and 921) are internal and probably suffer chemical shift perturbations because of conformational adjustments induced by binding of MNV VPg. Conserved MNV VPg Motif Binds eIF4G HEAT-1 Fig 5. NMR chemical shift mapping reveals the binding site for MNV VPg on the eIF4GI HEAT-1 domain. (A) Superposition of the 1H15N TROSY HS spectra of 15N-labelled eIF4GI HEAT-1 (748–993) obtained in the absence (black) and presence (green) of 0.35 molar equivalents of unlabelled MNV VP (104–124) peptide. (1H15N TROSY NMR spectra for all of the points in the titration are given in S3 Fig). The residues for which the intensity of amide 1H15 signals exhibit the greatest reductions as the MNV VPg(104–124) peptide concentration is increased are indicated by their residue number. The assignm Fig 5. NMR chemical shift mapping reveals the binding site for MNV VPg on the eIF4GI HEAT-1 domain. (A) Superposition of the 1H15N TROSY HSQC spectra of 15N-labelled eIF4GI HEAT-1 (748–993) obtained in the absence (black) and presence (green) of 0.35 molar equivalents of unlabelled MNV VPg (104–124) peptide. (1H15N TROSY NMR spectra for all of the points in the titration are given in S3 Fig). The residues for which the intensity of amide 1H15N signals exhibit the greatest reductions as the MNV VPg(104–124) peptide concentration is increased are indicated by their residue number. The assignments Fig 5. NMR chemical shift mapping reveals the binding site for MNV VPg on the eIF4GI HEAT-1 domain. (A) Superposition of the 1H15N TROSY HSQC spectra of 15N-labelled eIF4GI HEAT-1 (748–993) obtained in the absence (black) and presence (green) of 0.35 molar equivalents of unlabelled MNV VPg (104–124) peptide. (1H15N TROSY NMR spectra for all of the points in the titration are given in S3 Fig). The residues for which the intensity of amide 1H15N signals exhibit the greatest reductions as the MNV VPg(104–124) peptide concentration is increased are indicated by their residue number. The assignments PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 14 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 MNV VPg forms a stable complex with eIF4A and eIF4G HEAT-1 at a stoichiometry of 1:1:1 The VPg-binding site on eIF4GI identified by our NMR and mutagenesis experiments lies on a surface of the HEAT-1 domain that is distinct from the eIF4A-binding site. This suggested to us that VPg binding would not interfere with the eIF4G-eIF4A and that MNV VPg might therefore be able to form a ternary complex with eIF4G and eIF4A. The presence of eIF4A in initiation complexes formed with MNV VPg is consistent with earlier observations that MNV VPg translation is sensitive to inhibition of the ATPase activity of eIF4A by hippuristanol [25]. We used size exclusion chromatography (SEC) to characterise the complexes formed by MNV VPg and eIF4A with eIF4GI HEAT-1. Stoichiometric mixes of MNV VPg and eIF4GI HEAT-1, and eIF4A and eIF4GI HEAT-1 both resulted in peaks that eluted at lower volumes (14.7 and 14.3 mL respectively) than observed for the individual component proteins, consis- tent with the formation of binary complexes (Fig 6A). The presence of these complexes was confirmed by SDS PAGE analysis (Fig 6A inset). An equimolar mix of 1:1:1 MNV VPg, eIF4GI HEAT-1, and eIF4A eluted at an even lower volume (13.6 mL), consistent with the formation of a complex involving all three proteins, an interpretation that was also confirmed by SDS PAGE analysis (Fig 6A). The VPg-binding site on eIF4GI identified by our NMR and mutagenesis experiments lies on a surface of the HEAT-1 domain that is distinct from the eIF4A-binding site. This suggested to us that VPg binding would not interfere with the eIF4G-eIF4A and that MNV VPg might therefore be able to form a ternary complex with eIF4G and eIF4A. The presence of eIF4A in initiation complexes formed with MNV VPg is consistent with earlier observations that MNV VPg translation is sensitive to inhibition of the ATPase activity of eIF4A by hippuristanol [25]. We used size exclusion chromatography (SEC) to characterise the complexes formed by MNV VPg and eIF4A with eIF4GI HEAT-1. Stoichiometric mixes of MNV VPg and eIF4GI HEAT-1, and eIF4A and eIF4GI HEAT-1 both resulted in peaks that eluted at lower volumes (14.7 and 14.3 mL respectively) than observed for the individual component proteins, consis- tent with the formation of binary complexes (Fig 6A). The presence of these complexes was confirmed by SDS PAGE analysis (Fig 6A inset). Conserved MNV VPg Motif Binds eIF4G HEAT-1 the negative control mutation K771A had no effect on binding at low or high salt (Fig 5D and 5E). These results are consistent with the location of the MNV VPg binding site on eIF4GI HEAT-1 that was mapped in NMR titration experiments. They also suggest that the binding site is centred on the surface patch (around residues 914–919) distal to the eIF4A binding site. The binding site may extend to residue L939 but residue L897, which did not affect binding of VPg when mutated and is closest to the binding site of eIF4A, may define one edge of the VPg binding site. It is worth noting that if residues 108–124 of MNV VPg were to form a helix on binding to eIF4G (see above), the helix would have five turns and therefore be of a similar length to the helices in the HEAT-1 domain. The H918A mutant was again found to exhibit the weakest binding of MNV VPg, but the deleterious effects of the E912R and D919R substitu- tions were more severe. In the proximal cluster, there was still no effect of the L897A substitu- tion but the binding defect due to the L939A substitution was exacerbated (Fig 5E). The interaction of VPg with the E914R/K901M double mutant, which was similar to wild-type at 150 mM NaCl, was reduced markedly at a salt concentration of 300 mM. Finally, as expected, PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 15 / 34 MNV VPg forms a stable complex with eIF4A and eIF4G HEAT-1 at a stoichiometry of 1:1:1 An equimolar mix of 1:1:1 MNV VPg, eIF4GI HEAT-1, and eIF4A eluted at an even lower volume (13.6 mL), consistent with the formation of a complex involving all three proteins, an interpretation that was also confirmed by SDS PAGE analysis (Fig 6A). The SEC analysis was coupled to multi-angle laser light scattering (SEC-MALLS) to mea- sure the molecular weights (MW) of these complexes and determine their stoichiometry (Fig 6B). For the VPg:eIF4G HEAT-1 complex the observed MW (45.9 kDa) was very close to that expected for a 1:1 complex (46.5 kDa). For the eIF4A:eIF4G HEAT-1 complex the observed 65.9 kDa MW was somewhat lower than the 79.7 kDa expected for a 1:1 complex. The reason for this small discrepancy is not clear– it may be due to partial dissociation during elution–but the formation of a 1:1 binary complex is still the most likely outcome. The average MW mea- sured for the ternary MNV VPg:eIF4A:eIF4GI HEAT-1 complex using SEC-MALLS was 94.5 kDa, which is very close to the 94.1 kDa expected for a complex of 1:1:1 stoichiometry. The for- mation of this ternary complex is consistent with the observation that MNV VPg and eIF4A have non-overlapping binding sites on the eIF4G HEAT-1 domain. This observation also helps to explain how MNV VPg is able to interact with functional eIF4E:eIF4G:eIF4A complexes, as shown by the ability of the protein to interact with these proteins in cell lysates [38]; the eIF4E binding site on eIF4G is upstream of the HEAT-1 domain within eIF4G [9], allowing eIF4E to bind independently of eIF4A and MNV VPg. MNV VPg C-terminal peptide fusions inhibited MNV VPg mediated translation in vitro but could not be shown to have any anti-viral activity Conserved MNV VPg Motif Binds eIF4G HEAT-1 Fig 6. MNV VPg forms a ternary complex with eIF4G and eIF4A. (A) SEC profiles monitored by differential refractive index (which is proportional to protein concentration) for eIF4A, eIF4G HEAT-1 and MNV VPg as well as for the binary complexes formed by mixing approximately equimolar quantities of eIF4A and eIF4G HEAT-1 (4A-4G), eIF4G HEAT-1 and MNV VPg (4G-VPg), and the ternary complex obtained from eIF4A, eIF4G HEAT-1, and MNV VPg (4A-4G-VPg). SDS PAGE analysis of peak fractions of the binary and ternary complexes obtained in the SEC experiments are also shown in A. (B) SEC-MALLS analysis of the molar mass distributions of the binary and ternary complexes plotted against the SEC profiles shown in A. doi:10.1371/journal.ppat.1005379.g006 Fig 6. MNV VPg forms a ternary complex with eIF4G and eIF4A. (A) SEC profiles monitored by differential Fig 6. MNV VPg forms a ternary complex with eIF4G and eIF4A. (A) SEC profiles monitored by differential refractive index (which is proportional to protein concentration) for eIF4A, eIF4G HEAT-1 and MNV VPg as well as for the binary complexes formed by mixing approximately equimolar quantities of eIF4A and eIF4G HEAT-1 (4A-4G), eIF4G HEAT-1 and MNV VPg (4G-VPg), and the ternary complex obtained from eIF4A, eIF4G HEAT-1, and MNV VPg (4A-4G-VPg). SDS PAGE analysis of peak fractions of the binary and ternary complexes obtained in the SEC experiments are also shown in A. (B) SEC-MALLS analysis of the molar mass distributions of the binary and ternary complexes plotted against the SEC profiles shown in A. doi:10.1371/journal.ppat.1005379.g006 Fig 6. MNV VPg forms a ternary complex with eIF4G and eIF4A. (A) SEC profiles monitored by differential refractive index (which is proportional to protein concentration) for eIF4A, eIF4G HEAT-1 and MNV VPg as well as for the binary complexes formed by mixing approximately equimolar quantities of eIF4A and eIF4G HEAT-1 (4A-4G), eIF4G HEAT-1 and MNV VPg (4G-VPg), and the ternary complex obtained from eIF4A, eIF4G HEAT-1, and MNV VPg (4A-4G-VPg). SDS PAGE analysis of peak fractions of the binary and ternary complexes obtained in the SEC experiments are also shown in A. (B) SEC-MALLS analysis of the molar mass distributions of the binary and ternary complexes plotted against the SEC profiles shown in A. doi:10.1371/journal.ppat.1005379.g006 doi:10.1371/journal.ppat.1005379.g006 possibly by adopting a helical conformation that interacts with a centrally-located binding site that does not overlap with the binding site for eIF4A. MNV VPg C-terminal peptide fusions inhibited MNV VPg mediated translation in vitro but could not be shown to have any anti-viral activity Our in vitro binding studies have shown that a conserved motif within the C terminus of MNV VPg mediate a micromolar affinity binding to the HEAT-1 domain of eIF4GI and eIF4GII, 16 / 34 16 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 Fig 7. GST-MNV VPg 102–124 inhibits MNV VPg-mediated translation in vitro but not cap-dependent or IRES-dependent translation. In vitro translation reactions programmed with VPg-linked MNV RNA were performed in the presence of increasing concentrations of GST-MNV VPg(102–124) WT protein or the GST-MNV VPg(102–124) F123A mutant that binds very poorly to eIF4G. Protein synthesis was monitored by autoradiography of SDS PAGE analysis of incorporation of 35S-methionine in translation reactions. (A) Top panel: Effect of exogenous GST-MNV VPg(102–124) proteins on translation from VPg-linked MNV RNA; bottom panel: quantitative analysis of the level of 35S-methionine incorporation. This result was confirmed by independent replication–see S5A Fig. (B) Top panel: Effect of exogenous GST-MNV VPg(102–124) proteins on translation from capped bi-cistronic mRNA constructs containing the FMDV IRES between the first (CAT) and second (Luc) cistrons; bottom panel: quantitative analysis of the level of 35S-methionine incorporation. doi:10.1371/journal.ppat.1005379.g007 Fig 7. GST-MNV VPg 102–124 inhibits MNV VPg-mediated translation in vitro but not cap-dependent g 102–124 inhibits MNV VPg-mediated translation in vitro but not cap-dependent Fig 7. GST-MNV VPg 102–124 inhibits MNV VPg-mediated translation in vitro but not cap-dependent or IRES-dependent translation. In vitro translation reactions programmed with VPg-linked MNV RNA were performed in the presence of increasing concentrations of GST-MNV VPg(102–124) WT protein or the GST-MNV VPg(102–124) F123A mutant that binds very poorly to eIF4G. Protein synthesis was monitored by autoradiography of SDS PAGE analysis of incorporation of 35S-methionine in translation reactions. (A) Top panel: Effect of exogenous GST-MNV VPg(102–124) proteins on translation from VPg-linked MNV RNA; bottom panel: quantitative analysis of the level of 35S-methionine incorporation. This result was confirmed by independent replication–see S5A Fig. (B) Top panel: Effect of exogenous GST-MNV VPg(102–124) proteins on translation from capped bi-cistronic mRNA constructs containing the FMDV IRES between the first (CAT) and second (Luc) cistrons; bottom panel: quantitative analysis of the level of 35S-methionine incorporation. doi:10.1371/journal.ppat.1005379.g007 Fig 7. GST-MNV VPg 102–124 inhibits MNV VPg-mediated translation in vitro but not cap-dependent or IRES-dependent translation. In vitro translation reactions programmed with VPg-linked MNV RNA were performed in the presence of increasing concentrations of GST-MNV VPg(102–124) WT protein or the GST-MNV VPg(102–124) F123A mutant that binds very poorly to eIF4G. Protein synthesis was monitored by autoradiography of SDS PAGE analysis of incorporation of 35S-methionine in translation reactions. To probe the physiological significance of this observation we tested whether GST-fusions containing the C terminus of MNV VPg could inhibit norovirus translation. To do so we performed in vitro translation assays in rabbit reticulo- cyte lysates (RRL) programmed with VPg-linked genomic RNA in the presence of increasing concentrations of GST-MNV VPg(102–124) or the GST-MNV VPg(102–124) F123A mutant, which binds eIF4G much more weakly. We have previously shown that MNV translation of VPg-linked MNV RNA in RRL is absolutely dependent on the presence of intact VPg [25]; the detection of protein products is therefore indicative of VPg-dependent translation. As shown in Fig 7A GST-MNV VPg(102–124) inhibits norovirus translation in a dose-dependent manner. At the highest concentration used (28 μM) the yield of translation product is reduced to 40% of the control value. In contrast, at all concentrations the GST-MNV VPg (102–124) F123A mutant, which binds very weakly to eIF4G (Fig 4D), causes no significant inhibition of translation. possibly by adopting a helical conformation that interacts with a centrally-located binding site that does not overlap with the binding site for eIF4A. To probe the physiological significance of this observation we tested whether GST-fusions containing the C terminus of MNV VPg could inhibit norovirus translation. To do so we performed in vitro translation assays in rabbit reticulo- cyte lysates (RRL) programmed with VPg-linked genomic RNA in the presence of increasing concentrations of GST-MNV VPg(102–124) or the GST-MNV VPg(102–124) F123A mutant, which binds eIF4G much more weakly. We have previously shown that MNV translation of VPg-linked MNV RNA in RRL is absolutely dependent on the presence of intact VPg [25]; the detection of protein products is therefore indicative of VPg-dependent translation. As shown in Fig 7A GST-MNV VPg(102–124) inhibits norovirus translation in a dose-dependent manner. At the highest concentration used (28 μM) the yield of translation product is reduced to 40% of the control value. In contrast, at all concentrations the GST-MNV VPg (102–124) F123A mutant, which binds very weakly to eIF4G (Fig 4D), causes no significant inhibition of translation. 17 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 Finally, since we had shown that MNV VPg(102–124) can inhibit noroviral translation in vitro and interacts specifically with initiation complexes in cells, we sought to determine Fig 8. mCherry-MNV VPg(102–124) co-immunoprecipitates with translation initiation proteins. (A) RFP-Trap immunoprecipitation of lysates of BV2 cells expressing mCherry, mCherry-MNV VPg(102–124) WT or mCherry-MNV VPg(102–124) F123A. Pre-purification lysates (Input), purified fractions (IP) and unbound fractions (Flow through) were analysed by SDS PAGE and western blotting. (B) Time course of BV2 infection (MOI 0.01 TCID50 units/cell) with MNV1 in the presence of mCherry-VPg proteins. Prior to infection BV2 cells were transduced with lentiviruses expressing mCherry, mCherry-MNV VPg(102 124) WT or mCherry-MNV VPg(102–124) F123A. The progress of infection was monitored by measuring the number of viral cDNA copies generated from quantitative RT-PCR analysis of whole-cell RNA. (C) Fluorescence microscopy analysis of cell penetration of the FITC-TAT-MNV VPg(102–124) peptides. The images shown are merged images of DAPI stained nuclear DNA (blue) and wild-type or F123A versions of the cell penetrating peptides (green). (D) Time course of BV2 infection (MOI 0.01 TCID50 units/cell) with MNV1 following pre-treatment for 150 minutes with 100 μM of wild-type or F123A versions of the cell penetrating peptides prior to infection. The progress of infection was monitored by RT-PCR analysis (as in panel B). doi:10.1371/journal.ppat.1005379.g008 Fig 8. mCherry-MNV VPg(102–124) co-immunoprecipitates with translation initiation proteins. (A) RFP-Trap immunoprecipitation of lysates of BV2 cells expressing mCherry, mCherry-MNV VPg(102–124) WT or mCherry-MNV VPg(102–124) F123A. Pre-purification lysates (Input), purified fractions (IP) and unbound fractions (Flow through) were analysed by SDS PAGE and western blotting. (B) Time course of BV2 infection (MOI 0.01 TCID50 units/cell) with MNV1 in the presence of mCherry-VPg proteins. Prior to infection BV2 cells were transduced with lentiviruses expressing mCherry, mCherry-MNV VPg(102– 124) WT or mCherry-MNV VPg(102–124) F123A. The progress of infection was monitored by measuring the number of viral cDNA copies generated from quantitative RT-PCR analysis of whole-cell RNA. (C) Fluorescence microscopy analysis of cell penetration of the FITC-TAT-MNV VPg(102–124) peptides. The images shown are merged images of DAPI stained nuclear DNA (blue) and wild-type or F123A versions of the cell penetrating peptides (green). (D) Time course of BV2 infection (MOI 0.01 TCID50 units/cell) with MNV1 following pre-treatment for 150 minutes with 100 μM of wild-type or F123A versions of the cell penetrating peptides prior to infection. (A) Top panel: Effect of exogenous GST-MNV VPg(102–124) proteins on translation from VPg-linked MNV RNA; bottom panel: quantitative analysis of the level of 35S-methionine incorporation. This result was confirmed by independent replication–see S5A Fig. (B) Top panel: Effect of exogenous GST-MNV VPg(102–124) proteins on translation from capped bi-cistronic mRNA constructs containing the FMDV IRES between the first (CAT) and second (Luc) cistrons; bottom panel: quantitative analysis of the level of 35S-methionine incorporation. doi:10.1371/journal.ppat.1005379.g007 doi:10.1371/journal.ppat.1005379.g007 As a further test of the specificity of the inhibition of VPg-mediated translation the same proteins were used at the same concentrations in alterative in vitro translation assays. These were programmed with a capped bicistronic construct containing an open reading frame (ORF) for chloramphenicol acetyl transferase (CAT) under the control of an m7G cap and a downstream luciferase (LUC) ORF under the control of an IRES. Two different constructs were tested, one with a FMDV IRES and the other with a PTV IRES. We found that neither the wild-type nor the F123A mutant form of GST MNV VPg (102–124) had any effect on transla- tion of either CAT or LUC at even the highest protein concentrations, irrespective of the type of IRES present between the cistrons (Figs 7B and S5B). This suggests that the inhibition of VPg-dependent translation by exogenous added GST-MNV VPg(102–124) is specific. It is likely mediated by competition for binding to eIF4G. As a second probe of the physiological relevance of our observation that the C terminus of MNV VPg mediates binding of the protein to eIF4G, we tested whether MNV VPg(102–124) could interact with the large ribosomal initiation complexes in eukaryotic cells, as has been pre- viously demonstrated for TAP-tagged MNV VPg(1–124) [38]. To do this plasmids encoding mCherry-MNV VPg(102–124) and mCherry-MNV VPg(102–124) F123A fusion proteins were expressed in BV2 macrophages using lentivirus vectors (see Materials and Methods). Cell lines expressing mCherry MNV VPg(102–124) wild-type or the F123A mutant were selected on the basis of equal fluorescence intensity. Subsequent α-mCherry co-immunoprecipitation and western blot analysis revealed that eIF4G, eIF4A, eIF4E, eIF3D and PABP all co-immu- noprecipitate with wild-type mCherry-MNV VPg(102–124) but not with the F123A version (Fig 8A). This suggests that the MNV VPg C-terminal peptide can form the same complexes in the cytosol as the full-length protein. 18 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Discussion It has been known for nearly 20 years that the VPg protein covalently attached to the 5´-end of calicivirus RNA genomes is critical for translation [26]. Since then, various interactions of VPg proteins with eukaryotic initiation factors, including eIF3, eIF4E and eIF4G, have been reported [25, 33–35, 38]. Of these, the interaction of FCV VPg with eIF4E and of MNV VPg with eIF4G appear to be the most functionally significant [25, 38] but mechanistic details beyond the idea that VPg serves to recruit the viral RNA to the translation machinery have been lacking. In this paper we report the discovery and characterisation of a conserved ~20 amino acid motif in the C terminus of norovirus VPg proteins that binds to the HEAT-1 domains of eIF4GI and eIF4GII with micromolar affinity. Our results show that the VPg interaction with the central 4GM fragment of eIF4G, which is crucial for the initiation of translation from the viral RNA genome [38], is focused on the HEAT-1 domain. Norovirus VPg therefore serves as a proteinaceous ‘5´-cap’ that does not require a functional interaction with the cap-binding protein, eIF4E [25]. Our data provide clear evidence for the mechanistic basis for an important and unusual paradigm of translation initiation that is independent of 5´-cap and IRES struc- tures, and places functional and geometric constraints on how it operates. We have shown that the equilibrium dissociation constant for binding of the C terminus of VPg for the eIF4G HEAT-1 domains is in the low micromolar range–around 3 μM (Fig 3). This is relatively weak when compared to the affinity of m7G capped mRNA for the eIF4F com- plex, which is in the low nanomolar range [49]. IRES RNA structures are also reported to bind to initiation factor complexes with affinities in the low nanomolar range [50–52]. At first sight it may therefore appear questionable whether the binding of MNV VPg to eIF4G HEAT-1 is of high enough affinity to recruit the ribosome, particularly if it is in competition with 5´-capped eukaryotic mRNAs. However, it seems likely that in infected cells the VPg:HEAT-1 interaction forms just one contact in the overall interaction between VPg-linked genomic RNA and the initiation machinery, though we suspect that these points of contact remain focused on eIF4G. Conserved MNV VPg Motif Binds eIF4G HEAT-1 virus. However, no difference was observed in MNV titre over a 24-hour time course between cells expressing mCherry MNV VPg (102–124), or the non-binding mCherry-MNV VPg (102–124) F123A mutant or mCherry alone (Fig 8B). The mCherry tag adds more than 25 kDa to the molecular weight of the peptide. The lack of inhibition of norovirus replication might therefore be due to the copy number of the molecule in cells being too low. As an alternative test of the inhibitory activity of the MNV VPg C termi- nus wild-type and F123A variants of the MNV VPg(102–124) that incorporated N-terminal cell penetrating peptide (CPP) sequences were synthesised. Two different CPP sequences were used, derived from HIV TAT and Bac7. In each case the peptides were also N-terminally tagged with FITC so that penetration into cells could be monitored with fluorescence microscopy. As shown in Figs 8C and S6A, though the staining pattern was somewhat punctate (perhaps due to peptide aggregation) and shows some nuclear localisation in the case of the FITC-TAT-VPg peptide [48], the peptides were clearly present throughout the cytoplasm. However, no differ- ence was observed in viral titre over a 24-hour time course between cells containing either wild-type or F123A versions of the cell penetrating VPg peptides when compared to controls that had only been exposed to the peptide solvent, DMSO (Figs 8D and S6B). The reason for the negative result is not clear although it may again be due to delivery of insufficient free pep- tide to the cell interior. The progress of infection was monitored by RT-PCR analysis (as in panel B). doi:10.1371/journal.ppat.1005379.g008 doi:10.1371/journal.ppat.1005379.g008 Finally, since we had shown that MNV VPg(102–124) can inhibit noroviral translation in vitro and interacts specifically with initiation complexes in cells, we sought to determine whether cells over-expressing mCherry MNV VPg(102–124) are resistant to infection by the PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 19 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 and eIF3 [33, 34], the functional significance of these interactions has yet to be demonstrated. Indeed, although there is residual binding of MNV VPg to eIF4E when the eIF4G binding site on the viral protein has been mutated, abrogation of eIF4E-dependent translation initiation, either through the addition of the 4E-binding protein or cleavage of eIF4G by the FMDV L protease, has no effect on translation of MNV VPg-linked RNA [25]. It seems more likely that RNA interactions play a significant role in the formation of initia- tion complexes with VPg-RNA, since capped and uncapped eukaryotic mRNA have been shown to bind to eIF4F with nanomolar affinity [49], mostly likely though interactions with RNA binding sites on eIF4G and eIF4A [39]. Although the cap adds little to the overall affinity, it may serve to specify the orientation of cellular mRNA binding to eIF4F since the interaction with RNA is not otherwise sequence-specific. It may be that VPg is acting in a similar mecha- nism to the m7G cap, providing an important topological constraint for a high-affinity interac- tion between noroviral RNA and the eIF4F complex to ensure efficient initiation. The precise geometry of the complex formed when noroviral RNA binds to eIF4F is difficult to determine. Although there is a crystal structure for the binary complex comprising eIF4A and the HEAT-1 domain of eIF4G [47] and we have at least partially mapped the binding site on eIF4G HEAT-1 for the conserved C-terminal peptide in MNV VPg (Fig 5), the C terminus of VPg is linked flexibly to the helical core of the protein (residues 11–62) which contains the covalent point of attachment (Tyr 26) of the viral RNA [32]. Since the AUG start codon in nor- oviral RNA is only 4 nucleotides from the 5´-end, the length of this linker may be required to permit positioning of the start codon in the ribosomal P-site without it being encumbered by the bulk of the eIF4F complex. Similar considerations apply to all noroviruses since the key functional features– proximity of the start codon to the 5´-end of the viral RNA, site of RNA attachment to VPg and length of the linker to the conserved eIF4G-binding motif– are con- served. However, high-resolution structural analyses will be required to establish the relative positions of norovirus VPg-RNA and eIF4F within the context of the 48S ribosomal initiation complex. Discussion Although initial interaction of the viral RNAs with eIF4F leads to recruitment of eIF3 and the 40S ribosomal subunit, and there is evidence that norovirus VPg can bind to both eIF4E [35] 20 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 precursor. This precursor is eventually processed by NS6pro into the functional proteins required for virus replication. The precise timings of the processing cascade are not known in detail, although evidence suggests that cleavage of VPg-NS6pro is significantly slower than the early cleavages at the NS2-3 and NS3-4 sites [54, 55]. It may well be that the timing of the pro- cessing of precursors containing VPg is tuned to permit an early burst of translation (unim- peded as long as VPg remains fused to NS6pro), followed by a switch to RNA replication that may be aided by the accumulation of free VPg (or precursors with VPg at the C terminus), which would bind to eIF4G and inhibit viral translation initiation. This might clear the viral RNA of ribosomes to facilitate initiation of RNA replication, which requires association of the viral NS7pol at the 3´-end, followed by movement along the RNA template in the 3´-5´ direc- tion. Consistent with this notion, we have shown that exogenous GST-VPg fusion proteins containing the eIF4G-binding peptide can inhibit translation from norovirus VPg-RNA (Fig 7A), but it remains to be seen if this hypothetical mechanism plays out in infected cells. Finally the discovery of a binding site on eIF4G that is critical for norovirus translation initi- ation raises the possibility that the VPg-eIF4G interaction might be targeted in the develop- ment of antiviral compounds. Our initial efforts to demonstrate the antiviral activity of VPg peptides failed to show any effect (Figs 8 and S5), but this could simply be because an insuffi- cient cytoplasmic concentration was achieved. In any case, small molecule inhibitors are likely to have improved pharmacokinetics and our results lay the necessary groundwork to enable high-throughput screening for compounds that may disrupt the VPg-eIF4G interaction and so block infection. Similar geometric constraints are likely to affect the translation initiation mechanisms of other genera within the Caliciviridae, even though their VPg proteins lack the eIF4G-binding motif found in norovirus VPg. For example, the best available evidence suggests that binding of FCV VPg to eIF4E is required for translation initiation [25, 35]. This positions the FCV VPg at a different locus within the eIF4F complex compared to MNV VPg, though their binding sites on eIF4F may not be far apart– the eIF4E binding site in eIF4G (residues 557–681) is close to the HEAT-1 domain (residues 751–1011) [9]. Intriguingly, the mechanism reported here for noroviruses may also operate in rice yellow mottle virus (RYMV) since there is evidence to sug- gest that its VPg protein binds to the eIF4G HEAT-1 domain [53] and appears to have a C-terminal sequence that is similar to the HEAT-1 binding motif identified in norovirus VPg (Fig 4A). The observation that VPg cannot bind to the eIF4G HEAT-1 domain while still fused to NS6pro (Fig 2D) suggests that the VPg-NS6pro junction has to be cleaved for a productive inter- action. This does not appear to reflect a strict requirement for a free VPg C terminus, since a fusion construct with just the first four residues of NS6 appended to VPg retained binding activity. Rather we suggest that the proximity of a relatively large, folded protease domain pre- vents the C terminus of VPg from becoming properly seated in the binding site on eIF4G (Fig 2E). The precise reasons for the inability of VPg-NS6pro fusion protein to binding eIF4G remain to be determined but the observation nevertheless raises interesting questions about the regulation of norovirus replication. As with all positive-sense, single-strand RNA viruses, the early stages of norovirus infection are dominated by translation, which allows the accumulation of copies of the polyprotein 21 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 fuse eGFP to MNV VPg(1–124) using a ENLYFQGAAA linker between the two proteins that incorporates contains a TEV NIa Cleavage site (underlined). The PCR product was ligated into pETM-11m1 (S2 Fig). Full length hexa-Histidine-tagged VPg constructs (MNV VPg(1–124) wild-type (WT) and the F123A mutant, and FCV VPg(1–111) WT), and GST-eIF4GI 4GM (751–1132) constructs are described elsewhere [32, 38]. pET15b-eIF4A was kindly supplied by Chris Hellen and Tat- yana Pestova (State University New York) [60, 61]. Almost all mutations were made by QuikChange site directed mutagenesis (Stratagene). The exception to this was the MNV VPg(102–124) K120A mutant, which was made by anneal- ing oligonucleotides and ligating the product as described above. Details of all of the constructs used in the study, including residue numbers, expression vector, purification tag details, prote- ase cleavage site to remove the tag and non-native residues in the sequence post-cleavage are summarised in S2 Table. All constructs were verified by DNA sequencing (MWG Operon). Plasmids for protein expression in BV2 cells Lentivirus vectors expressing mCherry-MNV1-VPg fusions were generated in two steps using the plasmid pTM900, a bicistronic lentiviral vector derived from pCCLsin.PPT.hPGK.GFP.pre that carries a multiple cloning site downstream of the human phosphoglycerate kinase pro- moter and coexpresses a hygromycin resistance gene under the control of the EMCV IRES [62, 63]. The mCherry coding sequence lacking an in-frame stop codon was first amplified by PCR (see S2 Table for primer details) and cloned into the lentiviral vector pTM900 between the BshTI and BsiWI restriction sites. In the second step, the coding sequence of wild-type and mutated VPg peptides were inserted in frame with the mCherry ORF by hybridizing comple- mentary single-stranded oligonucleotides encoding the last 23 amino acids of MNV1 VPg flanked by BsiWI and XbaI restriction sites and ligated into the corresponding sites of the vec- tor generated in the first step. The sequences of the oligonucleotides used to amplify MNV VPg wild-type and F123A mutants for cloning into the lentivirus vector are given in S2 Table. A bicistronic lentiviral vector co-expressing mCherry and a hygromycin resistance gene was used as negative control. All plasmid sequences were verified by DNA sequencing. Plasmids for protein expression in E. coli Six different expression vectors were used to make new expression constructs. Briefly these were pETM-11 which has been described elsewhere [56]. Two modified versions of pETM-11, one with a BamHI site replacing the NcoI site (pETM-11m1), and the other, a variant of pETM-11m1, in which a thrombin cleavage site downstream of the N-terminal His-tag replaces the Tobacco Etch Virus (TEV) NIa cleavage site (pETM-11m2) [57, 58]. pMALX(E) has been described elsewhere [59] and pGEX-2T is commercially available (GE healthcare). An additional modified pGEX-2T plasmid was used with a TEV NIa cleavage site replacing the thrombin site (pGEX-2Tm). eIF4GI, eIF4GII and DAP5 sequences were amplified by PCR (KOD polymerase hot start kit) from full length cDNA clones. cDNA for eIF4GI, eIF4GII and DAP5 were kindly supplied by Chris Hellen and Tatyana Pestova (State University New York), Mark Coldwell (University of Southampton) and Bhushan Nagar (McGill University, Montréal) respectively. The residue numbering used throughout this paper for these proteins is based on the following NCBI accession codes, eIF4GI (AAM69365.1), eIF4GII (NP_003751.2) and DAP5 (NP_001036024.3). The oligonu- cleotide primers used in amplification of cDNA are given in S1 Table. Ligation into expression vectors was performed using the T4 Quick Ligation Kit (New England Biolabs). Fusions of glutathione-S-transferase (GST) with VPg C-terminal peptide sequences were made by ligating annealed synthetic oligonucleotides into pGEX-2T (S1 Table). 5 μM of sense and anti-sense oligonucleotide were heated to 95°C for 12 minutes in NEB buffer 2 followed by slow cooling to room temperature over 2.5 hours. The oligonucleotides were designed so that annealing would generate sticky ends corresponding to the BamHI and EcoRI restriction sites that were used in ligation. Enhanced green-fluorescent protein (eGFP) fusions were made in the following way: cDNA for eGFP-VPg and eGFP-VPg-NS6 fusions were generated by using overlap extension PCR to 22 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 The purification of the various recombinant proteins from E. coli followed the general scheme presented below. The details of the buffers used for each protein purification are fully documented in S5 Table (and cross-referenced to the data presented in each figure); any signif- icant deviations from the protocol below are given in S1 Text. Cells were re-suspended in a purification buffer (spiked with 2 mg/mL final Chicken egg lysozyme, 0.1 mM phenylmethanesulfonyl fluoride (PMSF) and 0.1% Triton X-100) and lysed by sonication in 15-second pulses separated by 15-second rests on ice over a period of 5 min- utes. Lysates were clarified by centrifugation at 29,000 g for 20 minutes. DNA was precipitated by addition of protamine sulfate to a concentration of 1 mg/mL and the lysate centrifuged again at 29,000 g for 20 minutes. Clarified lysates were applied to their respective affinity resins: TALON (Clontech) or His-select (Sigma) for His-tagged protein, Glutathione 4B (GE Health- care) resin for GST tagged proteins, or amylose resin (New England Biolabs) for MBP-tagged proteins. In each case the lysate was slowly rotated with the resin for ~1 h at 4°C and then applied to a gravity flow column. The washing and elution strategy employed was dependent on the affinity strategy being employed. For His-tagged proteins the resin was washed sequen- tially with 25 mL batches of purification buffer with increasing concentrations of imidazole. Proteins typically eluted at an imidazole concentration of 100 mM. For GST-tagged proteins the resin was washed with four 25 mL batches of purification buffer and eluted with two washes of the same buffer containing 10 mM reduced glutathione, the first titrated to pH 8 the second to pH 9; MBP-tagged proteins bound to the amylose resin were washed in the same way but the protein was eluted using 15 mL of purification buffer containing 10 mM maltose. Pure frac- tions were dialysed overnight (14–17 hours) against 4 L of dialysis buffer (S5 Table). For pro- teins in which the affinity tag was removed this was done by addition of protease (typically 0.5 mg TEV NIa or 200 U Bovine thrombin for 2 L cultures) at the beginning of dialysis. Following dialysis, proteins which had not been processed to remove tags were concentrated using centrifugal concentrator, snap frozen in liquid nitrogen and stored at -80°C. Proteins that were proteolytically-processed during overnight dialysis were re-applied to the appropriate affinity resin to remove the cleaved tag. The flow-through fractions obtained at this stage, which contained the untagged protein, were concentrated in a centrifugal concentrator and further purified by size exclusion chromatography (SEC), typically using a Superdex 75 16/60 column connected to an ÄKTA FPLC (GE Healthcare). The SEC buffers are detailed in S5 Table. Pure fractions were pooled and concentrated using centrifugal concentrator and snap frozen in liquid N2 and stored at -80°C. Protein concentrations were determined from optical density measurements at 280 nm using calculated extinction coefficients; peptide concentra- tions were calculated from the dry weight of synthesised material. Protein expression and purification All proteins were expressed in E. coli BL21-CodonPlus(DE3) RIPL or BL21-Rosetta(DE3) cells. The cells type used for each construct is given in S3 Table. For non-isotopically labelled pro- teins 1 L of lysogeny broth (LB) was seeded using 50 mL of an overnight starter culture. The cells were grown at 220 RPM and 37°C until mid-log phase was reached (OD600 of ~0.6–0.8) at which point the cells were typically induced with 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) and allowed to grow at 220 RPM and a defined temperature and period of time. Cells were pelleted by centrifugation at ~5,000 g and stored at -80°C. The temperature and time regi- men used for each protein are listed in S3 Table. For expression of 15N labelled eIF4GI HEAT-1 four 1 L cultures of LB were seeded and grown to mid-log phase. At this point the cells were pelleted at ~3,000 g for 20 minutes. The supernatant was discarded and the cells re-suspended in 800 mL of sterile TBS (50 mM Tris pH 7.5, 150 mM NaCl). The centrifugation step was repeated, the supernatant discarded and the pellets finally re-suspended in a total of 1 L of minimal medium divided between two 1 L culture flasks. The minimal medium, which contained 15NH4Cl as a labelled nitrogen source, was prepared as detailed in S4 Table. Protein expression was induced after 2 hours at 30°C by addition of 1 mM final IPTG and continued for approximately 16 hours at 30°C and 220 RPM. The cells were harvested and stored as previously described. 23 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 buffer containing a low concentration of imidazole (typically 10 mM). Bound proteins were eluted by incubation of the resin for 2 minutes with 50 μL of binding buffer containing 250 mM imidazole, followed by centrifugation for 2 minutes at 2,000 g. In the case of Glutathione Sepharose 4B pull-downs, protocol was as per the TALON pull-downs, except that the wash solutions did not contain imidazole. In addition the elution buffer consisted of 100 mM Tris pH 8, 300 mM NaCl, and 10 mM reduced glutathione. Samples of eluates, protein mixes taken before application to the resin, and the prey and bait input proteins were analysed by SDS PAGE and stained with either Coomassie or SYPRO orange gel stains. All pull-down assays were independently repeated at least once. VPg-NS6 auto cleavage assay pETM-11m1 eGFP-VPg-NS6 WT or pETM-11m1 eGFP-VPg-NS6 C139A were expressed at 20°C in E. coli BL21-CodonPlus(DE3) RIPL cells, as described above (see Protein expression and purification). Samples were taken at 1, 2, 4 and 21 hours post induction. The equivalent of 200 μL of cells at OD600nm of 0.6 were taken at each time point and re-suspended in 80 μL SDS-loading buffer. Proteins were separated by SDS PAGE and transferred to a nitrocellulose membrane. Membrane blocking was performed with 5% skimmed milk powder in TBST (20 mM Tris pH 7.5, 150 mM NaCl, 0.1% tween-20). Primary antibodies were additionally sup- plemented with ~1 μg/mL non-expressing E. coli lysate. All antibodies were applied in blocking buffer. Membranes were first probed for VPg; they were subsequently stripped with a mild stripping buffer (0.15% Glycine, 0.1% SDS, 1% tween20, pH 2.2) and incubated overnight with ~200 nM recombinant VPg in blocking buffer. The membrane was subsequently probed for eGFP. Rabbit anti-VPg sera was raised in-house and used as a 1/1000 dilution. A HRP-conju- gated goat anti-rabbit secondary antibody was purchased from Santa Cruz Biotechnology (SC- 2004). A HRP conjugated goat anti-GFP primary antibody was purchased from Genetex Inc (HRP-GTX26663). HRP-conjugated antibodies were used as a 1/4000 dilution. Membranes were imaged by detection of chemiluminescence using a Fujifilm LAS3000 CCD detector. Pull-down assays Due to the number of pull-down assays presented in the manuscript, specific details of each pull-down, including the proteins used as bait and prey, the concentration of each protein, and the mixing buffer used, are given for each assay on a figure by figure basis in S6 Table. Pull- down assays were performed with TALON (Clontech Laboratories, Inc.) or Glutathione Sepharose 4B (GE healthcare) to select for His-tagged or GST tagged protein respectively. The bed volumes used were 25–50 μL of TALON and 75 μL of Glutathione Sepharose 4B. Typically the bait and proteins (combined volume < 50 μL) were mixed with 400–450 μL binding buffer in a SigmaPrep spin column. The resulting slurry was incubated for ~60 minutes in the cold room. Following incubation, the unbound fraction was collected by centrifugation at 100 g for 3–5 seconds. In the case of TALON pull-downs the resin was washed by centrifugation with 750 μL binding buffer. The resin was subsequently washed, at least once, with 750 μL binding PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 24 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Isothermal Titration Calorimetry (ITC) For ITC experiments MNV VPg 1–124 and eIF4GII HEAT-1 (745–1003)–both with their His- tags removed –were dialysed overnight into 1 L of MES binding buffer using Sigma-Pur-a- lyzer midi dialyser (3,500 kDa MWCO). Following dialysis the protein stock concentrations were determined by measurement of absorption at 280 nm to be 52 μM MNV VPg(1–124) and 544 μM eIF4GII HEAT-1. ITC experiments were performed at 29°C using a MicroCal iTC200 instrument by injecting small volumes of eIF4GII HEAT-1 in 200 μL MNV VPg(1–124) in the sample cell. A total of 20 injections were performed, the first of 0.5 μL followed by 19 injections of 2 μL. The first two injections were separated by 180 s; all subsequent injections were sepa- rated by 120 s. Integrated raw data analysis was performed using Origin9 software using the single-site binding model to determine the values of KD, ΔH, ΔS and stoichiometry (n). Binding measurements were performed in triplicate. In a control titration performed with buffer replac- ing eIF4GII HEAT-1, the heat evolved was constant and less than 0.4 kcal/mol. Conserved MNV VPg Motif Binds eIF4G HEAT-1 The average fluorescence anisotropy values for the no-protein controls were subtracted from the fluorescence anisotropy values for each point in the titration giving the difference (ΔFA) for each data point. The ΔFA values for each point in the titration were used to calculate the dissociation constant (KD) for each interaction tested. This was done by fitting all of the titra- tion data to a single-site binding model (ΔFA = FAmax[protein]/(KD + [protein]) where FAmax is the maximal change in fluorescence anisotropy) in Prism6 (GraphPad Software). Titration of up to 40 μM eIF4GI into an irrelevant peptide ([FITC]-GSSHRYFLERGLE- SATSL; [64]) did not result in significant increases in fluorescence anisotropy (ΔFA < 0.004), confirming the specificity of the interaction with the MNV VPg peptides. Nuclear Magnetic Resonance (NMR) experiments NMR chemical shift mapping was used to map interactions made by MNV VPg with the HEAT-1 domain of eIF4GI. Experiments were performed using a Bruker Avance II 800 MHz spectrometer at 298 K. Titration of MNV VPg 104–124 peptide into a solution containing 15N labelled untagged eIF4GI HEAT-1 truncated (748–993) was monitored by recording 1H-15N transverse-relaxation optimized spectroscopy (TROSY) NMR spectra. MNV VPg(104–124) peptide (ChinaPeptides Co., Ltd. Shanghai, China) was dissolved in 50 mM MES pH 6.5, 150 mM NaCl, 5 mM 2-mercaptoethanol and titrated to pH 6.5 using a very small volume of 10 M NaOH. The reference 1H-15N TROSY spectrum was recorded from a 450 μL sample of 15N eIF4GI HEAT-1 protein at 212 μM dissolved in the same buffer containing 5% D2O used as a lock signal. The effects of peptide titration were measured at peptide:protein molar ratios of 0.117, 0.352, 0.587, 0.936, and 1.404 by sequential addition of 1.83 μL of peptide stock solu- tion (6.1 mM). The number of scans recorded in each 1H-15N TROSY NMR spectra were as follows, reference (16), 0.117 (24), 0.352 (28), 0.587 (56), 0.936 (56), 1.404 (180). The data were transformed using NMRpipe and analysed using NMRview (One Moon Scientific) [65]. The deposited 1H-15N assignments for eIF4GI HEAT-1 truncated 748–993 (BMRB Entry 18738) was used to analyse the titration [45]. For peak volume analysis the peak list was adjusted to account for slight spectral changes. In addition, clearly overlapped peaks were deleted from the list. The eVolume function of NMRview was used to determine non-overlapped peak volumes. Fluorescence anisotropy binding assays Fluorescence anisotropy was used to measure the binding of synthetic VPg peptides (labelled at the N terminus with fluorescein isothiocyanate (FITC)) to HEAT-1 domains from eIF4G1, eIF4GII and DAP5. Measurements were performed at 29°C using a Spectramax i3 Spectrome- ter with excitation/emission wavelengths of 485 and 535 nm respectively, slit widths set to 20 nm (excitation) and 25 nm (emission) and an integration time of 400 ms. Samples were loaded into non-binding surface black half area plates (Corning) using a MES binding buffer composed of 25 mM MES pH 6.5, 150 mM NaCl, 3 mM DDT, 0.05% Tween-20. N-terminally FITC labelled peptides MNV VPg 104–124 (VGPSWADDDRQVDYGEKIN- FE-COOH) and MNV VPg 108–124 were purchased at >95% purity from ChinaPeptides Co., Ltd. (Shanghai, China). In both cases the FITC group was attached with via a 6-aminohexanoic acid group to the N terminus of the peptide. Peptides were dissolved in binding buffer and titrated to pH 6.5 using a small volume of concentrated NaOH. The final peptide concentration was kept at a constant 10 nM across all binding experiments. The proteins tested for binding– untagged eIF4GI HEAT-1 truncated (748–993), untagged eIF4GII HEAT-1 (745–1003), MBP tagged eIF4GII HEAT-1 (745–1003) and DAP5 HEAT-1 (61–323)– were all dissolved in MES binding buffer. For each experiment a two-fold dilution series of each HEAT-1 protein was performed (typically starting at 20 or 40 μM) using MES binding buffer as diluent. For each dilution series, a no-protein control was performed in which binding buffer replaced protein. Each measurement typically involved two independent dilution series performed in parallel. PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 25 / 34 In vitro translation assays MNV1 VPg-linked RNA was isolated from infected RAW 264.7 cells as described previously [25]. For these assays we used capped bicistronic constructs that contain an upstream chloram- phenicol acetyl transferase (CAT) open reading frame (ORF) separated from a downstream luciferase (LUC) ORF by the IRES from either porcine teschovirus (PTV) or FMDV, which have been described previously [25]. In vitro translation was performed using the Flexi rabbit reticulocyte lysate system (Promega) in the presence of 35S-labelled methionine (EasyTag, Per- kin Elmer). Translation reactions were performed essentially according to the manufacturer's protocol but scaled down to 12.5 μL and with the inclusion of 100 mM KCl. They were pre- incubated for 15 minutes with 3.5–28 μM purified protein GST-MNV VPg(102–124) (wild- type or F123A mutant) and then initiated by addition of either 25 ng/μL in vitro transcribed bicistronic RNA or 16 ng/μL MNV1 VPg-linked RNA to give a final reaction volumes of 12.5 μL. The reactions were incubated at 30°C for 90 min and terminated by addition of an equal volume of Tris buffer containing 10 mM ethylenediaminetetraacetic acid (EDTA) and 100 ng/mL RNase A. Translated proteins were re-suspended in SDS PAGE sample buffer and resolved on a 12.5% polyacrylamide gel. Protein synthesis levels were detected by autoradiogra- phy. The level of translation was quantified using ImageJ to integrate selected band intensities with reference to the value obtained in the absence of recombinant protein. Lentivirus vector particle production and transduction Lentivirus vectors were used to test intracellular effects of MNV VPg peptide sequences expressed as C-terminal fusions with the fluorescent mCherry protein tag (see above). Vesicu- lar stomatitis virus G-protein-pseudotyped lentiviral particles were generated by transient transfection of 293T cells grown in 6-wells plates using 1.25 μg lentiviral vector, 0.63 μg pMDLg/pRRE (Addgene #12251), 0.31 μg pRSV-Rev (Addgene#12253) and 0.38 μg pMD2.G (Addgene#12259) per well using published protocols [66]. Parental BV2 cells were transduced with lentiviral supernatants and incubated for 48 h. Transduced cells were then selected on the basis of their resistance to Hygromycin B at a concentration of 200 μg/mL and sub-cultured for three additional passages. Cell populations expressing equal level of mCherry fluorescence were then enriched by FACS prior to infection with MNV1 or immunoprecipitation of mCherry-VPg protein (see below). Peak protein fractions were also analysed by SDS PAGE and Coomassie staining. Conserved MNV VPg Motif Binds eIF4G HEAT-1 in a total volume of 210 μL with final concentrations of 117 μM, 110 μM, and 132 μM of eIF4A, eIF4G and MNV VPg respectively. The mix was rotated slowly at room temperature for 30 minutes and then concentrated in an Amicon Ultra 500 10,000 MWCO concentrator to 130 μL. This sample was injected onto a S200 10/300 GL column at 0.5 mL/min using a 1260 Infinity HPLC system (Agilent Technologies). Light scattering and refractive index were mea- sured using in-line miniDAWN TREOS and Optilab T-rEX detectors (Wyatt Technology). in a total volume of 210 μL with final concentrations of 117 μM, 110 μM, and 132 μM of eIF4A, eIF4G and MNV VPg respectively. The mix was rotated slowly at room temperature for 30 minutes and then concentrated in an Amicon Ultra 500 10,000 MWCO concentrator to 130 μL. This sample was injected onto a S200 10/300 GL column at 0.5 mL/min using a 1260 Infinity HPLC system (Agilent Technologies). Light scattering and refractive index were mea- sured using in-line miniDAWN TREOS and Optilab T-rEX detectors (Wyatt Technology). in a total volume of 210 μL with final concentrations of 117 μM, 110 μM, and 132 μM of eIF4A, eIF4G and MNV VPg respectively. The mix was rotated slowly at room temperature for 30 minutes and then concentrated in an Amicon Ultra 500 10,000 MWCO concentrator to 130 μL. This sample was injected onto a S200 10/300 GL column at 0.5 mL/min using a 1260 Infinity HPLC system (Agilent Technologies). Light scattering and refractive index were mea- sured using in-line miniDAWN TREOS and Optilab T-rEX detectors (Wyatt Technology). Binary complexes (eIF4A:eIF4GI HEAT-1 and MNV VPg:4GI HEAT-1) and individual protein samples were prepared the same way with the exception that matched buffer replaced one or more of the protein components in the mix. SEC MALLS data were analysed by Astra 6 software (Wyatt Technology) using a dn/dc value of 0.185 and weight-averaged molar masses are reported. Peak protein fractions were also analysed by SDS PAGE and Coomassie staining. Binary complexes (eIF4A:eIF4GI HEAT-1 and MNV VPg:4GI HEAT-1) and individual protein samples were prepared the same way with the exception that matched buffer replaced one or more of the protein components in the mix. SEC MALLS data were analysed by Astra 6 software (Wyatt Technology) using a dn/dc value of 0.185 and weight-averaged molar masses are reported. Size exclusion chromatography with multi-angle light scattering (SEC-MALLS) To test for the formation of a ternary complex, untagged eIF4A 1–406, untagged MNV VPg 1–124 and N-terminally His-tagged eIF4GI HEAT-1 truncated (748–993) were mixed together 26 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 disrupted by up and down pipetting and lysates were centrifuged at 12,000 g for 10 min at 4°C. Clarified supernatants were diluted to 500 μL with lysis buffer. RFP-Trap_A agarose beads (Chromotek GmbH, Cat No: RTA-20) were added and the lysates were incubated for 2 h at 4°C under gentle agitation. The beads were washed 5 times with lysis buffer and proteins were eluted with Laemmli sample buffer. Immunoprecipitated proteins were detected using SDS PAGE and immunoblot analysis. As a control of protein expression in cells used for co-immu- noprecipitation experiments, total cell lysate corresponding to 5% of the input used for immu- noprecipitation was also analysed by immunoblot. Antibodies used were obtained from a variety of sources. Mouse monoclonal anti-eIF4E (A- 10) (Cat No: sc-271480), rabbit polyclonal anti-eIF4G (H-300) (4GI) (cat No: sc-11373) and goat polyclonal anti-eIF4AI (N-19) (Cat No: sc-14211) were obtained from Santa Cruz Bio- technology Inc. Rabbit polyclonal anti-eIF3D was obtained from ProteinTech (Cat No: 10219- 1-AP). Mouse monoclonal anti-GAPDH was obtained from Ambion (Cat No: AM4300). Rat monoclonal anti-RFP (clone 5F8) was obtained from Chromotek GmbH. Rabbit polyclonal anti-PABP1 was obtained from Cell Signaling Technology (Cat No: #4992). Measurement of viral replication The effect of excess added MNV VPg peptide sequences on MNV replication was tested using cell-penetrating peptides and ectopic expression of mCherry VPg. We used synthetic FITC-labelled peptides containing cell-penetrating sequences from Bac7 or human Immunodeficiency Virus (HIV) TAT fused to residues 102–124 of MNV VPg (sourced from ChinaPeptides Co. Ltd., Shanghai, China). F123A mutant variants of both pep- tides (which bind much less well to eIF4G) were also used. The peptide sequences were as follows: Bac-WT: FITC-PRPLPFPRPGNVVGPSWADDDRQVDYGEKINFE; Bac-F123A: FITC-PRPLPFPRPGNVVGPSWADDDRQVDYGEKINAE; TAT-WT: FITC-GRKKRRQRRR PQNVVGPSWADDDRQVDYGEKINFE; TAT-F123A: FITC-GRKKRRQRRRPQNVVGPSW ADDDRQVDYGEKINAE. The identity and purity (>95%) of the labelled peptides were assessed by mass spectrometry and reverse-phase liquid chromatography. All peptides were dissolved in dimethyl sulfoxide (DMSO) at a concentration of 10 mM and stored at -80°C. Naïve BV2 cells were treated with peptides at a concentration of 100 μM in culture complete medium for 150 min before being infected with MNV1 at low MOI (0.01 TCID50 units/cell) in the presence of peptides. Untreated cells were cultured in the same conditions using DMSO. After the indicated period of time, total cell RNA was extracted using a GenElute Mammalian Total RNA Miniprep kit (Sigma) and normalised to a standard concentration before being reverse transcribed using random hexamers and MuMLV RT enzyme (Promega). SYBR green- based quantitative PCR was performed using MNV1-specific primers (S2 Table). Each sample was measured in biological triplicate and compared to a standard curve. Additional non-tem- plate and non-reverse transcriptase samples were analysed as negative controls. Data were col- lected using a ViiA 7 Real-Time PCR System (Applied Biosystems). Measurements of viral replication in BV2 cells ectopically expressing mCherry-VPg fusion proteins or mCherry alone (prepared by lentiviral transduction as described above) were per- formed using a similar experimental setup. Co-immunoprecipitation experiments with mCherry-VPg Approximately 107 BV2 transduced cells were washed with cold phosphate-buffered saline (PBS) and harvested in 200 μL lysis buffer (Tris-HCl 10 mM pH 7.5, NaCl 150 mM, NP40 0.5%, EDTA 0.5 mM, PMSF 1 mM and 1x protease inhibitors (Calbiochem)). Cells were 27 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 ACJ04905.1),Hu/GIV.1/LakeMacquarie/NSW268O (NCBI accession number AFJ21375), Mu/ NoV/GV/MNV1/2002/USA (NCBI accession ABU55564.1), GVI dog/GVI.1/HKU_Ca035F/ 2007/HKG (NCBI accession FJ692501). Sequence alignment was performed by ClustalW [67] and BioEdit (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). The conserved eIF4G-binding motif is boxed. Selected amino acids from this motif are indicated (with MNV VPg number- ing). (PDF) S2 Fig. Autoproteolysis of VPg-NS6 confirms that NS6 folds into an active protease within the fusion protein. (A) Schematic of the His-eGFP-VPg\|NS6 fusion protein, including the position of TEV NIa cleavage sites and the molecular weights of each of the fused proteins. (B) Auto-cleavage of the eGFP-VPg-NS6 fusion (described in A) in E. coli was monitored by west- ern blotting of SDS PAGE separated E. coli lysate. The membrane was first probed for VPg and subsequently stripped and probed for eGFP. The molecular weight of eGFP-VPg-NS6, eGFP-VPg and VPg are shown on the right in black and are based on the purified recombinant protein run in lanes 6, 7, and 8. Possible cleavage intermediates are shown in red. (PDF) S3 Fig. Comparison of binding of eIF4GII HEAT-1 and MBP-eIF4GII HEAT-1 to MNV VPg C-terminal peptides. FITC-labelled MNV VPg(104–124) and MNV VPg(108–124) pep- tides were used in fluorescence anisotropy binding assays with the HEAT-1 domain of eIF4GII (745–1003) in order to determine the KD of the interaction. ΔFP, the normalised change in fluoresce anisotropy (relative to a no protein control) is plotted against protein concentration. Where appropriate (N>1), error bars indicate the standard deviation in the mean ΔFP value observed. The assays were performed with untagged eIF4GII HEAT-1 (black) and an MBP- tagged version (purple). In each case the data were fit using GraphPad Prism to a single-site binding model. S4 Fig. NMR analysis of binding of MNV VPg(104–124) to eIF4GI. 1H-15N TROSY HSQC spectra obtained on titration of unlabelled MNV VPg(104–124) into 212 μM 15N-labelled eIF4GI HEAT-1 (748–993). (A) Reference spectrum obtained in the absence of MNV VPg (104–124). (B-F) Spectra obtained in the presence of (B) 0.12, (C) 0.35, (D) 0.57, (D) 0.94 and (F) 1.35 molar equivalents of MNV VPg(104–124) peptide superposed on the reference spec- trum. The molar equivalents of MNV VPg(104–124) peptide and the number of scans used to obtain the spectrum (which was increased as the average signal intensity decreased) each point in the titration. ( ) S5 Fig. Supporting Information S1 Fig. Amino acid sequence alignments of representative sequences for VPg of all 6 gen- ogroups of Norovirus (GI-GIV). The representative strains used in the alignments are GI Hu/ GI/Norwalk/1968/US, (NCBI accession AAC64602), GII Lordsdale virus Hu/GII/Lordsdale/ 1993/UK (NCBI accession P54634), GIII Bo/GIII/B309/2003/BEL (NCBI accession 28 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Tabulated data for all plots presented in the paper (organised by Figure and Supplementary Figure number). (XLSX) Acknowledgments We thank Xulin Liu and Stephen Hare for technical assistance, and Trevor Sweeney for critical reading of the manuscript. Conserved MNV VPg Motif Binds eIF4G HEAT-1 translation from capped bi-cistronic mRNA constructs containing the PTV IRES between the first (CAT) and second (Luc) cistrons; bottom panel: quantitative analysis of the level of 35S-methionine incorporation. (PDF) S6 Fig. Cell penetrating FITC-Bac7-MNV VPg(102–124) peptides do not inhibit MNV infection of BV2 cells. (A) Fluorescence microscopy analysis of cell penetration of the FITC- Bac7-MNV VPg(102–124) peptides. The images shown are merged images of DAPI stained nuclear DNA (blue) and wild-type or F123A versions of the cell penetrating peptides tagged with FITC (green). (B) Time course of BV2 infection (MOI 0.01 TCID50 units/cell) with MNV1 following pre-treatment for 150 minutes with 100 μM of wild-type or F123A versions of the cell penetrating peptides prior to infection. The progress of infection was monitored by RT-PCR analysis (as in Fig 8). (PDF) S1 Table. Plasmid construct details. (XLSX) S2 Table. Oligonucleotide primer sequences. (XLSX) S3 Table. Details of protein expression. (XLSX) S4 Table. Solutions use in NMR experiments. (XLSX) S5 Table. Buffer details (cross-referenced by Figure number in the paper). (XLSX) S6 Table. Details of pull-down experiments (cross-referenced by Figure number in the paper). (XLSX) S1 Text. Supporting Methods. Specific adjustments to purification protocols. (DOCX) S1 Data. Supporting data. Tabulated data for all plots presented in the paper (organised by Figure and Supplementary Figure number). (XLSX) Acknowledgments We thank Xulin Liu and Stephen Hare for technical assistance, and Trevor Sweeney for critica reading of the manuscript. translation from capped bi-cistronic mRNA constructs containing the PTV IRES between the first (CAT) and second (Luc) cistrons; bottom panel: quantitative analysis of the level of 35S-methionine incorporation. (PDF) S6 Fig. Cell penetrating FITC-Bac7-MNV VPg(102–124) peptides do not inhibit MNV infection of BV2 cells. (A) Fluorescence microscopy analysis of cell penetration of the FITC- Bac7-MNV VPg(102–124) peptides. The images shown are merged images of DAPI stained nuclear DNA (blue) and wild-type or F123A versions of the cell penetrating peptides tagged with FITC (green). (B) Time course of BV2 infection (MOI 0.01 TCID50 units/cell) with MNV1 following pre-treatment for 150 minutes with 100 μM of wild-type or F123A versions of the cell penetrating peptides prior to infection. The progress of infection was monitored by RT-PCR analysis (as in Fig 8). (PDF) S6 Table. Details of pull-down experiments (cross-referenced by Figure number in the paper). (XLSX) S1 Text. Supporting Methods. Specific adjustments to purification protocols. (DOCX) S1 Data. Supporting data. Specific inhibition of translation initiation from MNV RNA by VPg (A) Combined quantitative analysis of two independent experiments to assay the inhibition of translation in RRL from MNV RNA by GST-MNV VPg(102–124)–the first of which was presented in Fig 7. In each case the level of translation was quantified using ImageJ to integrate two selected band intensities; values are normalised to the value obtained in the absence of added recombinant protein. The pink diamond indicated the averaged normalised signal for both experiments. (B) GST-MNV VPg(102–124) does not inhibit cap-dependent or PTV IRES-dependent translation. In vitro translation reactions were performed in the pres- ence of increasing concentrations of GST-MNV VPg(102–124) WT protein or the GST-MNV VPg(102–124) F123A mutant that binds much less well eIF4G. Protein synthesis was moni- tored by autoradiography of SDS PAGE analysis of incorporation of 35S-methionine in transla- tion reactions. Top panel: Effect of exogenous GST-MNV VPg(102–124) proteins on 29 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 References 1. Patel MM, Widdowson MA, Glass RI, Akazawa K, Vinje J, Parashar UD. Systematic literature review of role of noroviruses in sporadic gastroenteritis. Emerging Infect Dis. 2008; 14(8):1224–31. doi: 10.3201/ eid1408.071114 PMID: 18680645 2. Ahmed SM, Hall AJ, Robinson AE, Verhoef L, Premkumar P, Parashar UD, et al. Global prevalence of norovirus in cases of gastroenteritis: a systematic review and meta-analysis. Lancet Infect Dis. 2014; 14(8):725–30. doi: 10.1016/S1473-3099(14)70767-4 PMID: 24981041 3. Jones MK, Watanabe M, Zhu S, Graves CL, Keyes LR, Grau KR, et al. Enteric bacteria promote human and mouse norovirus infection of B cells. Science. 2014; 346(6210):755–9. doi: 10.1126/science. 1257147 PMID: 25378626 4. Mesquita JR, Barclay L, Nascimento MSJ, Vinje J. Novel Norovirus in Dogs with Diarrhea. Emerg Infect Dis. 2010; 16(6):980–2. doi: 10.3201/eid1606.091861 PMID: 20507751 5. Karst SM, Wobus CE, Lay M, Davidson J, Virgin HW. STAT1-dependent innate immunity to a Norwalk- like virus. Science. 2003; 299(5612):1575–8. doi: 10.1126/science.1077905 PMID: 12624267 6. Siebenga JJ, Duizer E, Koopmans M.P.G. Norovirus Epidemiology. In: Hansman GS, Jiang J.X., Green K.Y., editor. Caliciviruses: Molecular and Cellular Virology: Caister Academic Press; 2010. p. 1–24. 7. Walsh D, Mathews MB, Mohr I. Tinkering with translation: protein synthesis in virus-infected cells. Cold Spring Harb Perspect Biol. 2013; 5(1):a012351. doi: 10.1101/cshperspect.a012351 PMID: 23209131 8. Hinnebusch AG, Lorsch JR. The mechanism of eukaryotic translation initiation: new insights and chal- lenges. Cold Spring Harb Perspect Biol. 2012; 4(10):a011544. doi: 10.1101/cshperspect.a011544 PMID: 22815232 9. Jackson RJ, Hellen CU, Pestova TV. The mechanism of eukaryotic translation initiation and principles of its regulation. Nat Rev Mol Cell Biol. 2010; 11(2):113–27. doi: 10.1038/nrm2838 PMID: 20094052 10. Gradi A, Imataka H, Svitkin YV, Rom E, Raught B, Morino S, et al. A novel functional human eukaryotic translation initiation factor 4G. Mol Cell Biol. 1998; 18(1):334–42. doi: 10.1128/MCB.18.1.334 PMID: 9418880 11. Coldwell MJ, Sack U, Cowan JL, Barrett RM, Vlasak M, Sivakumaran K, et al. Multiple isoforms of the translation initiation factor eIF4GII are generated via use of alternative promoters, splice sites and a non-canonical initiation codon. Biochem J. 2012; 448(1):1–11. doi: 10.1042/BJ20111765 PMID: 22909319 12. Imataka H, Olsen HS, Sonenberg N. A new translational regulator with homology to eukaryotic transla- tion initiation factor 4G. EMBO J. 1997; 16(4):817–25. doi: 10.1093/emboj/16.4.817 PMID: 9049310 13. Jackson RJ. Alternative mechanisms of initiating translation of mammalian mRNAs. Biochem Soc Trans. 2005; 33(Pt 6):1231–41. doi: 10.1042/BST20051231 PMID: 16246087 14. Conserved MNV VPg Motif Binds eIF4G HEAT-1 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Author Contributions Conceived and designed the experiments: ENL FS SCo YC SCG SJM IGG SCu. Performed the experiments: ENL FS SCo YC FC CP YX SCG. Analyzed the data: ENL FS SCo YC FC CP YX SCG SJM IGG SCu. Wrote the paper: ENL FS YC CP SCG SJM IGG SCu. 30 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 22. Virgen-Slane R, Rozovics JM, Fitzgerald KD, Ngo T, Chou W, van der Heden van Noort GJ, et al. An RNA virus hijacks an incognito function of a DNA repair enzyme. Proc Natl Acad Sci U S A. 2012; 109 (36):14634–9. doi: 10.1073/pnas.1208096109 PMID: 22908287 23. Lee YF, Nomoto A, Detjen BM, Wimmer E. A protein covalently linked to poliovirus genome RNA. Proc Natl Acad Sci U S A. 1977; 74(1):59–63. PMID: 189316 24. Nomoto A, Kitamura N, Golini F, Wimmer E. The 5'-terminal structures of poliovirion RNA and poliovirus mRNA differ only in the genome-linked protein VPg. Proc Natl Acad Sci U S A. 1977; 74(12):5345–9. PMID: 202952 25. Chaudhry Y, Nayak A, Bordeleau ME, Tanaka J, Pelletier J, Belsham GJ, et al. Caliciviruses differ in their functional requirements for eIF4F components. J Biol Chem. 2006; 281(35):25315–25. doi: 10. 1074/jbc.M602230200 PMID: 16835235 26. Herbert TP, Brierley I, Brown TD. Identification of a protein linked to the genomic and subgenomic mRNAs of feline calicivirus and its role in translation. J Gen Virol. 1997; 78 (Pt 5):1033–40. PMID: 9152420 27. Burroughs JN, Brown F. Presence of a covalently linked protein on calicivirus RNA. J Gen Virol. 1978; 41(2):443–6. PMID: 569187 28. Guix S, Asanaka M, Katayama K, Crawford SE, Neill FH, Atmar RL, et al. Norwalk virus RNA is infec- tious in mammalian cells. J Virol. 2007; 81(22):12238–48. doi: 10.1128/JVI.01489-07 PMID: 17855551 29. Sosnovtsev S, Green KY. RNA transcripts derived from a cloned full-length copy of the feline calicivirus genome do not require VPg for infectivity. Virology. 1995; 210(2):383–90. doi: 10.1006/viro.1995.1354 PMID: 7618275 30. Schein CH, Oezguen N, Volk DE, Garimella R, Paul A, Braun W. NMR structure of the viral peptide linked to the genome (VPg) of poliovirus. Peptides. 2006; 27(7):1676–84. doi: 10.1016/j.peptides.2006. 01.018 PMID: 16540201 31. Hwang HJ, Min HJ, Yun H, Pelton JG, Wemmer DE, Cho KO, et al. Solution structure of the porcine sapovirus VPg core reveals a stable three-helical bundle with a conserved surface patch. Biochem Bio- phys Res Commun. 2015; 459(4):610–6. doi: 10.1016/j.bbrc.2015.02.156 PMID: 25753201 32. Leen EN, Kwok KYR, Birtley JR, Simpson PJ, Subba-Reddy CV, Chaudhry Y, et al. Structures of the Compact Helical Core Domains of Feline Calicivirus and Murine Norovirus VPg Proteins. J Virol. 2013; 87(10):5318–30. doi: 10.1128/Jvi.03151-12 PMID: 23487472 33. Daughenbaugh KF, Fraser CS, Hershey JWB, Hardy ME. References Etchison D, Milburn SC, Edery I, Sonenberg N, Hershey JW. Inhibition of HeLa cell protein synthesis following poliovirus infection correlates with the proteolysis of a 220,000-dalton polypeptide associated with eucaryotic initiation factor 3 and a cap binding protein complex. J Biol Chem. 1982; 257 (24):14806–10. PMID: 6294080 15. Gradi A, Svitkin YV, Imataka H, Sonenberg N. Proteolysis of human eukaryotic translation initiation fac- tor eIF4GII, but not eIF4GI, coincides with the shutoff of host protein synthesis after poliovirus infection. Proc Natl Acad Sci U S A. 1998; 95(19):11089–94. doi: 10.1073/pnas.95.19.11089 PMID: 9736694 16. Kirchweger R, Ziegler E, Lamphear BJ, Waters D, Liebig HD, Sommergruber W, et al. Foot-and-mouth disease virus leader proteinase: purification of the Lb form and determination of its cleavage site on eIF-4 gamma. J Virol. 1994; 68(9):5677–84. PMID: 8057448 17. Krausslich HG, Nicklin MJ, Toyoda H, Etchison D, Wimmer E. Poliovirus proteinase 2A induces cleav- age of eucaryotic initiation factor 4F polypeptide p220. J Virol. 1987; 61(9):2711–8. PMID: 3039165 18. Haghighat A, Svitkin Y, Novoa I, Kuechler E, Skern T, Sonenberg N. The eIF4G-eIF4E complex is the target for direct cleavage by the rhinovirus 2A proteinase. J Virol. 1996; 70(12):8444–50. PMID: 8970966 19. Foeger N, Glaser W, Skern T. Recognition of eukaryotic initiation factor 4G isoforms by picornaviral pro- teinases. J Biol Chem. 2002; 277(46):44300–9. doi: 10.1074/jbc.M208006200 PMID: 12228254 20. Liberman N, Gandin V, Svitkin YV, David M, Virgili G, Jaramillo M, et al. DAP5 associates with eIF2β and eIF4AI to promote Internal Ribosome Entry Site driven translation. Nucleic Acids Res. 2015; 43 (7):3764–75. doi: 10.1093/nar/gkv205 PMID: 25779044 21. Hansman GS, Jiang XJ, Green KY. Caliciviruses: molecular and cellular virology. Norfolk, UK: Caister Academic Press; 2010. 31 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Conserved MNV VPg Motif Binds eIF4G HEAT-1 42. Virgili G, Frank F, Feoktistova K, Sawicki M, Sonenberg N, Fraser CS, et al. Structural analysis of the DAP5 MIF4G domain and its interaction with eIF4A. Structure. 2013; 21(4):517–27. doi: 10.1016/j.str. 2013.01.015 PMID: 23478064 43. Lee SH, McCormick F. p97/DAP5 is a ribosome-associated factor that facilitates protein synthesis and cell proliferation by modulating the synthesis of cell cycle proteins. EMBO J. 2006; 25(17):4008–19. doi: 10.1038/sj.emboj.7601268 PMID: 16932749 44. Gautier R, Douguet D, Antonny B, Drin G. HELIQUEST: a web server to screen sequences with specific alpha-helical properties. Bioinformatics. 2008; 24(18):2101–2. doi: 10.1093/bioinformatics/btn392 PMID: 18662927 45. Akabayov SR, Wagner G. Backbone resonance assignment of the HEAT1-domain of the human eukaryotic translation initiation factor 4GI. Biomol NMR Assign. 2014; 8(1):89–91. doi: 10.1007/ s12104-013-9459-5 PMID: 23325513 46. Arnold K, Bordoli L, Kopp J, Schwede T. The SWISS-MODEL workspace: a web-based environment for protein structure homology modelling. Bioinformatics. 2006; 22(2):195–201. doi: 10.1093/ bioinformatics/bti770 PMID: 16301204 47. Schutz P, Bumann M, Oberholzer AE, Bieniossek C, Trachsel H, Altmann M, et al. Crystal structure of the yeast eIF4A-eIF4G complex: an RNA-helicase controlled by protein-protein interactions. Proc Natl Acad Sci U S A. 2008; 105(28):9564–9. doi: 10.1073/pnas.0800418105 PMID: 18606994 48. Dang CV, Lee WM. Nuclear and nucleolar targeting sequences of c-erb-A, c-myb, N-myc, p53, HSP70, and HIV tat proteins. J Biol Chem. 1989; 264(30):18019–23. PMID: 2553699 49. Kaye NM, Emmett KJ, Merrick WC, Jankowsky E. Intrinsic RNA Binding by the Eukaryotic Initiation Factor 4F Depends on a Minimal RNA Length but Not on the m7G Cap. J Biol Chem. 2009; 284 (26):17742–50. doi: 10.1074/jbc.M109.009001 PMID: 19414591 50. Jan E, Sarnow P. Factorless ribosome assembly on the internal ribosome entry site of cricket paralysis virus. Journal of Molecular Biology. 2002; 324(5):889–902. doi: 10.1016/S0022-2836(02)01099-9 PMID: 12470947 51. Kieft JS, Zhou KH, Jubin R, Doudna JA. Mechanism of ribosome recruitment by hepatitis C IRES RNA. RNA. 2001; 7(2):194–206. doi: 10.1017/S1355838201001790 PMID: 11233977 52. Lomakin IB, Hellen CUT, Pestova TV. Physical association of eukaryotic initiation factor 4G (eIF4G) with eIF4A strongly enhances binding of elF4G to the internal ribosomal entry site of encephalomyocar- ditis virus and is required for internal initiation of translation. Mol Cell Biol. 2000; 20(16):6019–29. doi: 10.1128/Mcb.20.16.6019–6029.2000 PMID: 10913184 53. Albar L, Bangratz-Reyser M, Hebrard E, Ndjiondjop MN, Jones M, Ghesquiere A. Mutations in the eIF (iso)4G translation initiation factor confer high resistance of rice to Rice yellow mottle virus. The genome-linked protein VPg of the Nor- walk virus binds eIF3, suggesting its role in translation initiation complex recruitment. EMBO Journal. 2003; 22(11):2852–9. doi: 10.1093/Emboj/Cdg251 PMID: 12773399 34. Daughenbaugh KF, Wobus CE, Hardy ME. VPg of murine norovirus binds translation initiation factors in infected cells. Virol J. 2006; 3:33. doi: 10.1186/1743-422X-3-33 PMID: 16719923 35. Goodfellow I, Chaudhry Y, Gioldasi I, Gerondopoulos A, Natoni A, Labrie L, et al. Calicivirus translation initiation requires an interaction between VPg and eIF 4 E. EMBO Rep. 2005; 6(10):968–72. doi: 10. 1038/sj.embor.7400510 PMID: 16142217 36. Hosmillo M, Chaudhry Y, Kim DS, Goodfellow I, Cho KO. Sapovirus Translation Requires an Interaction Between VPg and the Cap Binding Protein eIF4E. J Virol. 2014; 88(21):12212–21. doi: 10.1128/JVI. 01650-14 37. Royall E, Doyle N, Abdul-Wahab A, Emmott E, Morley SJ, Goodfellow I, et al. Murine Norovirus 1 (MNV1) Replication Induces Translational Control of the Host by Regulating eIF4E Activity during Infec- tion. J Biol Chem. 2015; 290(8):4748–58. doi: 10.1074/jbc.M114.602649 PMID: 25561727 38. Chung L, Bailey D, Leen EN, Emmott EP, Chaudhry Y, Roberts LO, et al. Norovirus Translation Requires an Interaction between the C Terminus of the Genome-linked Viral Protein VPg and Eukary- otic Translation Initiation Factor 4G. J Biol Chem. 2014; 289(31):21738–50. doi: 10.1074/jbc.M114. 550657 PMID: 24928504 39. Marcotrigiano J, Lomakin IB, Sonenberg N, Pestova TV, Hellen CU, Burley SK. A conserved HEAT domain within eIF4G directs assembly of the translation initiation machinery. Mol Cell. 2001; 7(1):193– 203. PMID: 11172724 40. Leen EN, Baeza G, Curry S. Structure of a murine norovirus NS6 protease-product complex revealed by adventitious crystallisation. PLoS One. 2012; 7(6):e38723. doi: 10.1371/journal.pone.0038723 PMID: 22685603 41. Muhaxhiri Z, Deng L, Shanker S, Sankaran B, Estes MK, Palzkill T, et al. Structural basis of substrate specificity and protease inhibition in Norwalk virus. J Virol. 2013; 87(8):4281–92. doi: 10.1128/JVI. 02869-12 PMID: 23365454 32 / 34 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 Plant J. 2006; 47(3):417–26. doi: 10.1111/j.1365-313X.2006.02792.x PMID: 16774645 54. Belliot G, Sosnovtsev SV, Mitra T, Hammer C, Garfield M, Green KY. In vitro proteolytic processing of the MD145 norovirus ORF1 nonstructural polyprotein yields stable precursors and products similar to those detected in calicivirus-infected cells. J Virol. 2003; 77(20):10957–74. doi: 10.1128/Jvi.77.20. 10957–10974.2003 PMID: 14512545 55. May J, Korba B, Medvedev A, Viswanathan P. Enzyme kinetics of the human norovirus protease con- trol virus polyprotein processing order. Virology. 2013; 444(1–2):218–24. doi: 10.1016/j.virol.2013.06. 013 PMID: 23850457 56. Zou P, Gautel M, Geerlof A, Wilmanns M, Koch MH, Svergun DI. Solution scattering suggests cross- linking function of telethonin in the complex with titin. J Biol Chem. 2003; 278(4):2636–44. doi: 10.1074/ jbc.M210217200 PMID: 12446666 57. Sweeney TR, Cisnetto V, Bose D, Bailey M, Wilson JR, Zhang X, et al. Foot-and-mouth disease virus 2C is a hexameric AAA+ protein with a coordinated ATP hydrolysis mechanism. J Biol Chem. 2010; 285(32):24347–59. doi: 10.1074/jbc.M110.129940 PMID: 20507978 58. Birtley JR, Curry S. Crystallization of foot-and-mouth disease virus 3C protease: surface mutagenesis and a novel crystal-optimization strategy. Acta Crystallogr D Biol Crystallogr. 2005; 61(Pt 5):646–50. doi: 10.1107/S0907444905007924 PMID: 15858279 59. Moon AF, Mueller GA, Zhong X, Pedersen LC. A synergistic approach to protein crystallization: combi- nation of a fixed-arm carrier with surface entropy reduction. Protein Sci. 2010; 19(5):901–13. doi: 10. 1002/pro.368 PMID: 20196072 60. Pisarev AV, Unbehaun A, Hellen CU, Pestova TV. Assembly and analysis of eukaryotic translation initi- ation complexes. Methods Enzymol. 2007; 430:147–77. doi: 10.1016/S0076-6879(07)30007-4 PMID: 17913638 61. Pestova TV, Hellen CU, Shatsky IN. Canonical eukaryotic initiation factors determine initiation of trans- lation by internal ribosomal entry. Mol Cell Biol. 1996; 16(12):6859–69. PMID: 8943341 PLOS Pathogens \| DOI:10.1371/journal.ppat.1005379 January 6, 2016 33 / 34 Conserved MNV VPg Motif Binds eIF4G HEAT-1 62. Follenzi A, Ailles LE, Bakovic S, Geuna M, Naldini L. Gene transfer by lentiviral vectors is limited by nuclear translocation and rescued by HIV-1 pol sequences. Nature Genetics. 2000; 25(2):217–22. doi: 10.1038/76095 PMID: 10835641 62. Follenzi A, Ailles LE, Bakovic S, Geuna M, Naldini L. Gene transfer by lentiviral vectors is limited by nuclear translocation and rescued by HIV-1 pol sequences. Nature Genetics. 2000; 25(2):217–22. doi: 10.1038/76095 PMID: 10835641 63. Sorgeloos F, Jha BK, Silverman RH, Michiels T. Evasion of Antiviral Innate Immunity by Theiler's Virus L* Protein through Direct Inhibition of RNase L. PLOS Pathogens. 2013; 9(6):e1003474. doi: 10.1371/ journal.ppat.1003474 PMID: 23825954 64. Reijns MA, Bubeck D, Gibson LC, Graham SC, Baillie GS, Jones EY, et al. The structure of the human RNase H2 complex defines key interaction interfaces relevant to enzyme function and human disease. J Biol Chem. 2011; 286(12):10530–9. doi: 10.1074/jbc.M110.177394 PMID: 21177854 65. Delaglio F, Grzesiek S, Vuister GW, Zhu G, Pfeifer J, Bax A. NMRPipe: a multidimensional spectral pro- cessing system based on UNIX pipes. J Biomol NMR. 1995; 6(3):277–93. doi: 10.1007/BF00197809 PMID: 8520220 66. Dull T, Zufferey R, Kelly M, Mandel RJ, Nguyen M, Trono D, et al. A third-generation lentivirus vector with a conditional packaging system. J Virol. 1998; 72(11):8463–71. PMID: 9765382 67. Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, et al. Clustal W and Clustal X version 2.0. Bioinformatics. 2007; 23(21):2947–8. doi: 10.1093/bioinformatics/btm404 PMID: 17846036 34 / 34
https://openalex.org/W1534698081	https://link.springer.com/content/pdf/10.1007/s00213-015-3977-z.pdf	English	null	Negative mood reverses devaluation of goal-directed drug-seeking favouring an incentive learning account of drug dependence	Psychopharmacology/Psychopharmacologia	2,015	cc-by	12,194	Abstract Background Two theories explain how negative mood primes smoking behaviour. The stimulus–response (S-R) account ar- gues that in the negative mood state, smoking is experienced as more reinforcing, establishing a direct (automatic) associa- tion between the negative mood state and smoking behaviour. By contrast, the incentive learning account argues that in the negative mood state smoking is expected to be more reinforc- ing, which integrates with instrumental knowledge of the re- sponse required to produce that outcome. Methods To test this, overnight-deprived daily smokers (n= 44) acquired two instrumental responses for tobacco and chocolate points, respectively, before smoking to satiety. Half then received negative mood induction to raise the ex- pected value of tobacco, opposing satiety, whilst the remain- der received positive mood induction. Finally, a choice be- tween tobacco and chocolate was measured in extinction to test whether negative mood could augment tobacco choice, opposing satiety, in the absence of direct experience of tobac- co reinforcement. Objectives One differential prediction is that whereas the in- centive learning account anticipates that negative mood induc- tion could augment a novel tobacco-seeking response in an extinction test, the S-R account could not explain this effect Results Negative mood induction not only abolished the de- valuation of tobacco choice, but participants with a significant increase in negative mood increased their tobacco choice in extinction, despite satiety. Conclusions These findings suggest that negative mood aug- ments drug-seeking by raising the expected value of the drug through incentive learning, rather than through automatic S-R control. * Lee Hogarth l.hogarth@exeter.ac.uk 1 School of Psychology, University of Exeter, Washington Singer Building, Perry Road, Exeter EX4 4QG, UK Keywords Incentive learning . Goal-directed learning . Drug-seeking . Negative mood . Depression . Allostasis . Discriminative stimuli . Motivating operations . Negative reinforcement 2 Institute of Psychiatry, King’s College London, London, UK 3 Department of Psychiatry, University of Pittsburgh, Pittsburgh, PA, USA 4 School of Psychology, University of Plymouth, Plymouth, UK 5 Department of Psychology, Saint Anselm College, Manchester, NH, USA Psychopharmacology (2015) 232:3235–3247 DOI 10.1007/s00213-015-3977-z Psychopharmacology (2015) 232:3235–3247 DOI 10.1007/s00213-015-3977-z ORIGINAL INVESTIGATION Lee Hogarth1 & Zhimin He2 & Henry W. Chase3 & Andy J. Wills4 & Joseph Troisi II5 & Adam M. Leventhal6 & Amanda R. Mathew7 & Brian Hitsman8 Lee Hogarth1 & Zhimin He2 & Henry W. Chase3 & Andy J. Wills4 & Joseph Troisi II5 & Adam M. Leventhal6 & Amanda R. Mathew7 & Brian Hitsman8 Received: 16 January 2015 /Accepted: 20 May 2015 /Published online: 5 June 2015 # The Author(s) 2015. This article is published with open access at Springerlink.com because the extinction test prevents S-R learning by omitting experience of the reinforcer. Introduction The study should con- tribute to our understanding of how depression promotes drug dependence (Breslau et al. 1998; Hitsman et al. 2013; Hughes 1999; Kassel et al. 2007) and provide insight into the mecha- nisms underpinning the transition to drug dependence. Most accounts of how negative mood promotes tobacco- seeking draw inspiration from the self-medication hypothesis (Khantzian 1997). The core idea is that smoking acutely alle- viates negative mood, and this improvement in state back towards homeostasis (Brody et al. 2009; Ramsay and Woods 2014) increases the reinforcement value of smoking (Hursh and Silberberg 2008). Crucially, it is proposed that because negative mood signals that smoking has greater rein- forcement value, negative mood acquires the capacity to prime smoking behaviour (Audrain-McGovern et al. 2014; Leventhal et al. 2014b). Incentive learning theory has taken this type of modulatory account further (Balleine et al. 1994; Dickinson and Balleine 2010; Heyes and Dickinson 1990; see also, Pittenger and Bevins 2013). On this view, negative mood functions as an internal motivational state, which is much like a MO in reli- ably predicting that smoking has greater reinforcement value. However, rather than automatically modulating the efficacy of external SDs, incentive learning theory argues that smokers learn from experience that in the negative mood state, smoking is more reinforcing. This incentive learning experi- ence enables the motivational state, negative mood, to subse- quently retrieve an expectation that smoking currently has a higher value, which manifests as subjective desire for that outcome. This outcome expectancy is integrated with goal- directed instrumental knowledge of which response produces that outcome in the external discriminative context (Bradfield and Balleine 2013; Trask and Bouton 2014), and this conjunc- tion determines the selection and performance of the appro- priate smoking response. Thus, incentive learning argues that negative mood primes smoking behaviour via a conjunction of explicit desire and instrumental belief, whereas the S-R account proposes that negative mood primes smoking behav- iour directly without intervening decision processes. Theories differ in their description of the mechanisms by which negative mood comes to acquire control over smoking behaviour. Early accounts described negative mood as an in- ternal instrumental discriminative stimulus (SD) which ‘sets the occasion’ (Skinner 1938) in which smoking behaviour is more reinforcing, and thus comes to prime smoking behav- iour, although exactly how was not well defined (Carmody 1989; Pomerleau and Pomerleau 1984). Introduction 6 Departments of Preventive Medicine and Psychology, University of Southern California Keck School of Medicine, Los Angeles, CA, USA A key debate in contemporary addiction theory is whether the transition from recreational drug use to clinical drug depen- dence is driven by the emergence of automatic control over drug-seeking (Everitt and Robbins 2013; Koob 2013) or su- pernormal reinforcement value of the drug increasing inten- tional choice of this commodity (Bickel et al. 2014; Heyman 7 Department of Psychiatry and Behavioral Sciences, Medical University of South Carolina, Charleston, SC, USA 8 Department of Preventive Medicine, Northwestern University Feinberg School of Medicine, Chicago, IL, USA Psychopharmacology (2015) 232:3235–3247 3236 different smoking behaviours are required to produce rein- forcement in different contexts (for instance, obtaining ciga- rettes from a shop versus a machine requires different se- quences). Negative mood is not sufficiently discriminating on its own, they argue, to elicit the appropriate response se- quence in each context via S-R mapping, because negative mood is largely the same irrespective of context (Dougher and Hackbert 2000; Laraway et al. 2014; Michael 1993; Troisi 2014). For this reason, the behaviour analysts have argued that negative mood is more accurately described as a motivating operation (MO), rather than a SD, because it pre- dicts the greater reinforcement value of smoking irrespective of the specific response that is required to produce that rein- forcer in the context. SDs by contrast, are typically external cues which are scheduled to predict that a particular response will produce an outcome. Consequently, they argue, the mo- tivating operation, negative mood, should come to lower the threshold enabling external discriminative stimuli to evoke the specific response that produces the valued outcome in the context. On this view, negative mood does not elicit smoking directly (as predicted by S-R theory), but rather, modulates the ability of external discriminative stimuli to evoke smoking. 2013). As clinical drug dependence is comorbid with psychi- atric illness (SAMHSA 2012; Swendsen et al. 2010), one might recast this question as whether psychiatric states in- crease automatic or intentional drug choice. To explore this question, the current study used an outcome-devaluation assay to test whether negative mood would increase goal-directed drug-seeking in an extinction test, consistent with negative mood exerting its effect on drug-seeking through intentional rather than automatic processes (Hogarth 2012; Hutcheson et al. 2001; Willner and Jones 1996). Introduction This type of account was developed by negative reinforce- ment theories (Baker et al. 2004), allostasis theories (Koob 2009; Koob and Volkow 2010) and incentive habit theories (Belin et al. 2013). These accounts argue that experience of the greater reinforcement value of smoking in the negative mood state establishes a strong direct link between the nega- tive mood state and the motor sequence of smoking. This direct link has been variously described as automatic, uncon- scious, preconscious, habitual and compulsive. In essence, these theories have adopted Hull’s (Hull 1943) stimulus–re- sponse/reinforcement (S-R) account of instrumental discrimi- nation learning, wherein the negative mood state (S) acquires capacity to directly elicit the motor sequence of smoking (R) without retrieving an expectation of the reinforcer produced by that response. Such S-R accounts are attractive because they can explain how negative mood could prime smoking behaviour automatically, bypassing the individual’s intentions or beliefs about the nature of the drug and its current value. The incentive learning, MO and S-R accounts fall on a continuum with the incentive learning and S-R accounts at each extreme, so only the differential predictions of these latter two positions will be explored in the remainder of the paper. The incentive learning and S-R accounts make a differential prediction as to whether the increase in self-report desire to smoke associated with negative mood is causal or epiphenomenal in driving smoking behaviour. Existing Behaviour analysts have questioned S-R accounts of neg- ative mood on the logical grounds that topographically Psychopharmacology (2015) 232:3235–3247 3237 responses (Corbit et al. 2012; Dickinson et al. 2002; Hutcheson et al. 2001; Miles et al. 2003) are governed by incentive or S-R learning in animals. In one key study, Hutcheson et al. (2001) found that in animals that had experi- enced the greater value of heroin in withdrawal, being shifted to a state of heroin withdrawal augmented performance of a heroin-seeking response in extinction, even though that re- sponse had never before been reinforced in the withdrawal state. Thus, withdrawal must have augmented heroin- seeking by retrieving an expectation of the greater reinforce- ment value of heroin, which integrated with goal-directed knowledge of the response-heroin contingency, to augment performance of that response. Introduction This increase in heroin- seeking could not have been driven by the strengthening of any S-R association controlling the response because the ex- tinction test prevented direct experience of the greater rein- forcement value of heroin in the withdrawal state modulating the propensity to make the response through S-R learning. experimental data cannot readily distinguish these two posi- tions because although desire and behaviour are commonly correlated, causal control by desire over behaviour has not been empirically confirmed. To be specific, studies have shown that smokers who are prone to negative mood or are depressed verbally report intentionally smoking in order to alleviate negative mood (Lerman et al. 1996), and more often report smoking as their most preferred activity (Audrain- McGovern et al. 2014; Spring et al. 2003), show quicker la- tency to smoke and greater willingness to pay for cigarettes (Leventhal et al. 2014b) and show preferential selection of smoking over money reinforcement in a concurrent choice progressive ratio schedule (Audrain-McGovern et al. 2014). Negative mood induction is associated with similar enhance- ments of smoking motivation. Negative mood induction in- creases self-reported desire to smoke (Brandon et al. 1996; Payne et al. 1991; Perkins et al. 2013; Vinci et al. 2012), increases smoking behaviour (Conklin and Perkins 2005; Fucito et al. 2010; Payne et al. 1991) and increases response rate in an instrumental progressive ratio task reinforced by cigarette puffs (Willner and Jones 1996). Collectively, these studies show that negative mood (trait or state) produces cor- responding increases in a self-reported desire to smoke, instru- mental tobacco-seeking and smoking behaviour. However, it remains unclear whether self-reported desire is causal or epi- phenomenal in driving smoking-related behaviour. Thus, the empirical co-occurrence of desires and smoking behaviours under negative mood, whilst suggestive, does not completely resolve the issues of whether negative mood controls smoking behaviour effect via incentive learning or S-R mechanisms. p p y p g g Similar to Hutcheson et al. (2001), the current human pro- cedure tested whether negative mood would function as a motivational state raising the expected value of the drug in the extinction test. In previous versions of the current human design (Hogarth 2012; Hogarth and Chase 2011; Hogarth et al. Introduction 2013), smokers first underwent concurrent choice train- ing in which two key press responses earned points notionally exchangeable for tobacco and chocolate rewards, respectively, and participants reported explicit knowledge of these instru- mental response-outcome (R-O) contingencies. Smoking to satiety was then used to decrease the expected value of the tobacco outcome. Finally, choice between the two responses was tested in nominal extinction, where participants believed that the responses continued to earn their outcomes, but these outcomes were not displayed until the end (to prevent direct experience of the outcomes modifying the tendency to make each response in the extinction test through S-R learning). These studies found that the devaluation treatment (smoking satiety) reduced tobacco choice in the extinction test relative to concurrent training. Thus, the tobacco-seeking response in the extinction test was demonstrably governed by knowledge of the current expected low value of the tobacco outcome in the sated state, integrated with knowledge of the response- outcome contingencies learned in the concurrent training phase. This effect could not be explained by S-R learning, because the extinction test prevented direct experience of the current low value of the tobacco outcome weakening the S-R association controlling the response. g Fortunately, the incentive learning and S-R accounts make a second differential prediction that is more tractable to exper- imental dissection—whether or not negative mood induction could augment a novel tobacco-seeking response in an extinc- tion test. The importance of the extinction test is that it pre- vents the opportunity for direct experience of the greater value of tobacco outcome in the negative mood state strengthening the S-R association between the mood state and the tobacco- seeking response. Thus, S-R theory predicts that negative mood induction should not enhance a novel tobacco-seeking response in an extinction test and could not readily explain this effect if it was found. By contrast, the extinction test does allow an expectation about the current high value of tobacco evoked by negative mood to integrate with knowledge of the response that produces that outcome, to augment performance of the response. Consequently, the finding that negative mood induction enhances a novel tobacco-seeking response in ex- tinction would support the incentive learning account and dis- confirm the S-R account of how negative mood primes smoking behaviour. Introduction The question at stake in the present experiment was wheth- er negative mood induction (compared with positive mood induction) administered immediately prior to the extinction test would reverse/oppose the devaluation of goal-directed tobacco-seeking produced by satiety, consistent with negative mood functioning as a motivational state raising the expected value of the drug. Numerous studies have shown that negative The extinction test has become a standard procedure for testing whether naturally reinforced responses (Balleine et al. 1994; Kosaki and Dickinson 2010) or drug-seeking Psychopharmacology (2015) 232:3235–3247 3238 mood induction increases drug motivation, but it is not clear if this effect can compete with primary motivational states. By testing whether negative mood can counter drug satiety, we can ascertain whether negative mood is a strong motivational state, capable of competing with primary motivational states. An important precedent for this ‘oppositional’ design comes from Willner and Jones (1996). They found that smoking to satiety reduced response rate on the progressive ratio schedule in which cigarette puffs served as the reinforcer (also consis- tent with Perkins et al. 1994; Perkins et al. 1997). Crucially, negative mood induction partially reversed the devaluation effect produced by satiety, raising response rates back towards the abstinent baseline. Although negative mood opposed sati- ety, it is not clear how, given that tobacco-seeking was rein- forced with puffs at test. Negative mood may have augmented tobacco-seeking by raising the expected value of tobacco or by enhancing the experience of puffing reinforcement which strengthened the S-R association controlling the response. The key innovation in the present study was to test whether nega- tive mood induction would reverse the devaluation of tobacco-seeking in an extinction test. This effect would favour the incentive learning over the S-R account of how negative mood primes drug-seeking behaviour. More broadly, if mood functions as a motivational state, opposing satiety, this would extend the scope of incentive learning theory to include not just primary states like hunger, thirst, cold, sexual arousal etc. but to also include ‘higher-level’ psychiatric emotional states, like sadness, regret, confusion, anxiety etc., providing princi- pled insight into how these emotional states modulate goal- directed behaviour (Dickinson and Balleine 2010). was a deception. At the end, during debrief, participants were asked if they would be happy to receive £5 extra (£15 in total), ‘to save the experimenter having to restock the items’. All participants accepted this. Method Participants were then shown the rewards on the table: A packet of ten cigarettes of their preferred brand and a 200-g bar of Cadbury dairy milk (both sealed) above the keyboard on the same side as the key that earned that outcome in the task that followed. Their baseline desire for these rewards was recorded with two questions: ‘To what extent do you agree with the following statements? I would like to (smoke a cig- arette right now/eat chocolate right now)’ with a 7-point Likert scale underneath ranging from ‘strongly disagree’ to ‘strongly agree’. Participants Participants were daily smokers (n=48; 50 % male) recruited from students and staff at the University of Nottingham. All were asked to abstain from smoking for at least 3 h prior to the study, which amounted to overnight abstinence for the major- ity. In order to increase compliance with abstinence proce- dures, participants were informed that abstinence would be confirmed with a breath CO measurement at the start of the experiment (although this was not used as an exclusion crite- rion). Exclusion criteria employed during initial e-mail vetting were current use of illicit drugs except cannabis; aged less than 18 or over 40 years of age; significant current or past medical or psychiatric illness (including mood disorder); poor physical health; allergies; or dietary sensitivities to chocolate. Testing was between 10 a.m. and 5 p.m., lasting 1 h/subject. During initial contact, participants were told that they would be paid £10 for participation plus the additional chocolate and cigarettes won in a computer task up to the value of £5. This Introduction This arrangement ensures that par- ticipants believe they are earning tobacco and chocolate dur- ing the instrumental task, and mitigates the ethical problem of paying participants in cigarettes. Informed consent was ob- tained in accordance with the declaration of Helsinki, and ethical approval was granted by the University of Nottingham, School of Psychology Ethics Committee. Procedure Breath CO—baseline After informed consent was obtained, participants were pre- sented with a Bedfont Smokerlyzer and told that this device measures the time of their last cigarette. Participants self- reported their time of last cigarette (none reported smoking within 3 h, consistent with instructions) before providing their ‘baseline’ exhaled CO reading. Questionnaires A questionnaire pack followed which included smoking his- tory (age, cigarettes per day, years smoking, age of onset), the Cigarette Dependence Scale (CDS-5; Etter et al. 2003) and the Questionnaire of Smoking Urges (QSU; Tiffany and Drobes 1991). Reward desire—baseline Contingency knowledge test Immediately following concurrent training, participants were tested for knowledge of the instrumental contingencies through the on-screen questions: ‘Which key earned tobac- co/chocolate, the D or the H key? Please choose carefully’. The order of the two questions was randomised. Participants who got either question wrong were excluded (n=4). Concurrent choice—baseline The aim of concurrent training was to establish two instru- mental responses (key presses), which earn distinct reward. Participants were faced with the physical rewards (a pack of ten cigarettes of their preferred brand and a 200-g bar of Cadbury dairy milk) and on-screen instructions which stated: ‘This is a game in which you can win the cigarettes and choc- olate in front of you. In each trial, press the D or H key to see if you have won a point for these rewards. You will only win on Psychopharmacology (2015) 232:3235–3247 3239 with a 9-point Likert scale ranging from ‘Happy’ at 1, ‘Neutral’ at 5 and ‘Sad’ at 9. with a 9-point Likert scale ranging from ‘Happy’ at 1, ‘Neutral’ at 5 and ‘Sad’ at 9. some trials. Press the space bar to begin’. Each trial began with the centrally presented fixation cross, which remained until either the D or H key was pressed. Pressing one key immedi- ately presented the outcome text ‘tobacco point’, whereas pressing the other key produced the outcome text ‘chocolate point’, for 1 s. The response-outcome assignment was counterbalanced between participants. Each key had only a 50 % chance of yielding its respective outcome. On non- rewarded trials, the outcome text ‘nothing’ was presented. A random inter-trial interval between 750 and 1250 ms interced- ed between outcome offset and the fixation cross of the next trial. There were 40 trials of concurrent training in total. Per cent choice of the tobacco over the chocolate response was the main measure, where 50 %=indifference, >50 %=tobacco preference and<50 %=chocolate preference. Devaluation by smoking satiety All participants were then told: ‘We want to test how much you like smoking. There will be a 10-min break in which you can smoke as much or as little as you wish of a cigarette. Please report the pleasantness of each puff you take on this smoking satiety questionnaire’. On this questionnaire, partic- ipants reported the pleasantness of each puff they consumed on visual analogue scale, enabling quantification of the num- ber of puffs consumed and the decline in pleasure rating from the first to the last puff. Participants were escorted outside the building to a smoking area and left alone during the smoking period to minimise disturbance. The experimenter stated that they would return when the time was up to facilitate ad libitum consumption. The experimenter returned in 10 min. Mood induction procedure The nega- tive statements were: I feel a little down today; My work is harder than I expected; Sometimes I feel so guilty that I can’t sleep; I wish I could be myself, but nobody likes me when I am; Today is one of those days when everything I do is wrong; I doubt that I’ll ever make a contribution in the world; I feel like my life is in a rut that I’m never going to get out; My mistakes haunt me, I’ve made too many; Life is such a heavy burden; I’m tired of trying; Even when I give my best effort, it just doesn’t seem to be good enough; I don’t think things are ever going to get better; I feel worthless; What’s the point of trying; I feel cheated by life; Every time I turn around, something else has gone wrong. The positive statements were: I feel cheerful and lively; On the whole, I have very little difficulty in thinking clearly; I’m pleased that most people are so friendly to me; I can make friends extremely easily; I feel enthusiastic and con- fident now; There should be a lot of good times coming along; I’m able to do things accurately and efficiently; I know that I can achieve the goals I set; I have a sense of power and vigour; I’m feeling amazingly good today; I feel highly perceptive and refreshed; I can concentrate hard on anything I do; My thinking is clear and rapid; Life is so much fun; It seems to offer so many sources of fulfilment; Life is firmly in my control; I’m really feeling sharp now. Both groups listened to music during these statements played through noise cancelling headphones. The negative group received Barber’s Adagio for Strings, whereas the positive group received Mozart’s Einekleine Nachtmusik (Morrison and O’Connor 2008). A positive and negative mood group were contrasted in an attempt to match the arousing effects of the two mood conditions, whilst manipulating the valence. A neutral group was not included because the contrast Contingency knowledge test Mood induction procedure The positive statements were: I feel cheerful and lively; On the whole, I have very little difficulty in thinking clearly; I’m pleased that most people are so friendly to me; I can make friends extremely easily; I feel enthusiastic and con- fident now; There should be a lot of good times coming along; I’m able to do things accurately and efficiently; I know that I can achieve the goals I set; I have a sense of power and vigour; I’m feeling amazingly good today; I feel highly perceptive and refreshed; I can concentrate hard on anything I do; My thinking is clear and rapid; Life is so much fun; It seems to offer so many sources of fulfilment; Life is firmly in my control; I’m really feeling sharp now. Both groups listened to music during these statements played through noise cancelling headphones. The negative group received Barber’s Adagio for Strings, whereas the positive group received Mozart’s Einekleine Nachtmusik (Morrison and O’Connor 2008). A positive and negative mood group were contrasted in an attempt to match the arousing effects of the two mood conditions, whilst manipulating the valence. A neutral group was not included because the contrast AVelten mood induction procedure was then employed derived from previous studies (Berna et al. 2010; Lenton and Martin 1991; Richell and Anderson 2004; Velten 1968). Participants were presented with the instructions ‘You will now be shown a series of statements that represent a particular type of mood. Read each of the statements to yourself and focus your attention on it. Your success at coming to experience this mood will largely depend on your willingness to accept and respond to the idea in each statement and to allow each statement to act upon you. Attempt to respond to the feeling suggested by each statement. Then try to think of yourself as moving into that state. If it is natural for you to do so, try to visualise a scene in which you have had such a feeling. Press the space bar to begin’. Instructions were followed by 16 statements presented for 10 s each in random order, separated by a random ITI of 2500 to 3500 ms. The two randomly assigned groups (positive and negative) were exposed to different statements. Mood induction procedure AVelten mood induction procedure was then employed derived from previous studies (Berna et al. 2010; Lenton and Martin 1991; Richell and Anderson 2004; Velten 1968). Participants were presented with the instructions ‘You will now be shown a series of statements that represent a particular type of mood. Read each of the statements to yourself and focus your attention on it. Your success at coming to experience this mood will largely depend on your willingness to accept and respond to the idea in each statement and to allow each statement to act upon you. Attempt to respond to the feeling suggested by each statement. Then try to think of yourself as moving into that state. If it is natural for you to do so, try to visualise a scene in which you have had such a feeling. Press the space bar to begin’. Instructions were followed by 16 statements presented for 10 s each in random order, separated by a random ITI of 2500 to 3500 ms. The two randomly assigned groups (positive and negative) were exposed to different statements. The nega- tive statements were: I feel a little down today; My work is harder than I expected; Sometimes I feel so guilty that I can’t sleep; I wish I could be myself, but nobody likes me when I am; Today is one of those days when everything I do is wrong; I doubt that I’ll ever make a contribution in the world; I feel like my life is in a rut that I’m never going to get out; My mistakes haunt me, I’ve made too many; Life is such a heavy burden; I’m tired of trying; Even when I give my best effort, it just doesn’t seem to be good enough; I don’t think things are ever going to get better; I feel worthless; What’s the point of trying; I feel cheated by life; Every time I turn around, something else has gone wrong. Reward desire and breath CO - post devaluation The impact of devaluation on tobacco and chocolate desire was tested immediately after smoking satiety by participants completing the desire questions again (as above). Breath CO was recorded for a second time to quantify smoke exposure. Mood state—baseline Mood state—baseline To determine participants’ baseline mood, they were present- ed with the on-screen question ‘How do you currently feel?’ Psychopharmacology (2015) 232:3235–3247 3240 characteristics, ts(42)<−1.28, ps>0.20, and there was also no group difference in gender ratio, X2(1)=0.1, p=0.74. of the two more extreme groups offered the best strategy to detect an effect. Mood state—post-mood induction Table 1 shows the effect of smoking satiety and mood induc- tion on self-report measures and breath CO. The groups did not differ in their experience of smoking satiety but did di- verge with respect to mood following mood induction. Specifically, analysis of variance (ANOVA) with the breath CO data yielded a main effect of time (baseline, post-devalu- ation), F(1.42)=45.42, p<0.001, ŋp 2=0.52, demonstrating that smoking took place during the satiety manipulation, and there was no main effect of group or interaction between time and group, F<1. The groups also showed no difference in the number of puffs consumed, F(1.42)=1.3, p=0.26. ANOVA with the reward desire data yielded no significant main effects of interactions involving group, F(1.42)<2.62, p>0.11, but revealed a significant interaction between reward (tobacco, chocolate) and time (baseline, post-satiety), F(1.42)=89.98, p<0.001, ŋp 2=0.68. This interaction was due to desire for tobacco significantly decreasing after satiety, F(1.43)= 145.29, p<0.001, ŋp 2=0.77, whereas desire for chocolate showed no significant change, F(1.43)=3.15, p=0.08, ŋp 2= 0.07. Thus, smoking satiety decreased smoking pleasure and tobacco desire, comparably across groups. To measure participants’ mood following mood induction, they were again presented with the on-screen question ‘How do you currently feel?’ with a 9-point Likert scale ranging from ‘Happy’ at 1, ‘Neutral’ at 5 and ‘Sad’ at 9. The two groups were expected to diverge in their mood compared with their pre-induction baseline. Extinction test Finally, participants completed the concurrent choice proce- dure again, in extinction, to evaluate the combined impact of satiety and mood induction on tobacco-seeking. Participants were presented with the on-screen instructions: ‘You can now earn cigarettes and chocolate by pressing the D or H keys as before. You will only be told how many of each reward you have earned at the end of the experiment. You will also be asked to examine the mood statements. Press the space bar to begin’. The extinction test was identical to concurrent training except that outcomes are omitted from trials. Thus, the initial fixation cross appeared until a response choice was made, which immediately launched the ITI before to the next trial. In addition however, to ensure that the mood induction con- tinued throughout this test phase, one of the mood statements (randomly selected from the set of 16 for the group) was presented for 3 s prior to the fixation cross of each choice trial, and remained on until a response choice was made. The per cent choice of the tobacco over the chocolate key was the main measure. The question at stake was whether satiety and mood induction would modulate the proportion of tobacco choices in this extinction test relative to the baseline concurrent choice phase. ANOVAwith self-reported mood data yielded a significant interaction between time (pre-induction, post-induction) and group, F(1.42)=29.95, p<0.001, ŋp 2=0.42, where the nega- tive group reported a significant increase in negative mood, F(1.19)=21.33, p<0.001, ŋp 2=0.53, and the positive group Table 1 The effect of smoking satiety (top) and mood induction (bottom two rows) on adjunct measures Measure Group Negative Positive Breath CO—baseline 3.9 (0.8) 3.8 (0.6) Breath CO—post-devaluation 5.7 (0.9) 5.7 (0.6) Number of puffs 17.1 (1.4) 15.3 (1.0) Pleasure first puff 79.5 (5.1) 83.9 (3.8) Pleasure last puff 61.0 (5.9) 49.8 (5.9) Desire tobacco—baseline 5.5 (0.3) 6.0 (0.2) Desire tobacco—post-satiety 3.0 (0.4) 2.5 (0.3) Desire chocolate—baseline 3.1 (0.4) 3.2 (0.3) Desire chocolate—post-satiety 3.6 (0.4) 3.5 (0.4) Mood state—pre-induction 3.2 (0.4) 3.3 (0.3) Mood state—post-induction 5.6 (0.4) 2.8 (0.3) Breath CO is in parts per million. Pleasure of puffs is per cent of visual analogue scale. Desire scores are on a 7-point scale (positive numbers equal greater desire). Mood state scores are on a 9-point scale (positive numbers equal greater sadness) Numbers are mean and sem in brackets Effects of satiety and mood on tobacco choice 3: high positive (filled squares), low positive (empty squares), low negative (empty circles) and high negative (filled circles) who did and did not show sig- nificant changes in mood following induction. The scatterplot shows that increased negative mood was associated with increased tobacco choice at test, opposing satiety Effects of satiety and mood on tobacco choice Baseline concurrent training contained 40 choice trials, and there was a significant linear increase in tobacco choice across successive quarters (62, 69, 65 and 71 %), F(1.43)=6.20, p<0.02, ŋp 2=0.13. For this reason, the final quarter (10 trials) of the baseline concurrent choice task were employed as the baseline. Figure 1 shows per cent tobacco choice at this base- line and in the extinction test following satiety and mood induction for the two groups. ANOVA with these data pro- duced a significant interaction between group (negative, pos- itive) and block (baseline, extinction), F(1.42)=5.4, p=0.02, ŋp 2=0.11, a significant main effect of block, F(1.42)=5.5, p= 0.02, ŋp 2=0.12, and no main effect of group, F<1. Furthermore, the main effect of block was significant in the positive group, F(1.23)=12.28, p=0.002, ŋp 2=0.35, but not in the negative group, F<1. These data indicate that satiety de- creased tobacco choice in the extinction test relative to base- line, consistent with previous findings (Hogarth 2012; Hogarth and Chase 2011; Hogarth et al. 2013). However, this devaluation effect was completely abolished by negative mood induction, suggesting that negative mood functioned as a motivational state raising the expected value of tobacco in opposition to satiety. A standard median split procedure was used to group par- ticipants into whether they showed a high or low change in mood consistent with their mood induction procedure, shown in Fig. 3a. ANOVA on these mood change scores (pre- to post- induction) with the factor sub-group (4) yielded a significant main effect, F(3.40) = 48.06, p < 0.001, ŋp 2 = 0.78. Furthermore, contrasts of pairs using Bonferroni post-hoc tests indicated that the high-negative (n=11) and high- positive (n=13) sub-groups differed from all other sub- groups, ps≤0.003, whereas the low-negative (n=9) and low- positive (n=11) sub-groups did not differ from each other (p=1). Finally, the high-negative sub-group’s mood change was significantly greater than zero, t(10)=9.38, p<0.001, Fig. 2 Relationship between the change in self-reported mood (pre- to post-mood induction), with the change in tobacco choice (from baseline to extinction test with satiety mood induction in between). Positive mood values reflect increased sadness, whereas negative values reflect in- creased happiness. Positive tobacco choice scores reflect increased tobac- co choice, whereas negative scores reflect decreased tobacco choice. Symbols identify the four sub-groups employed in Fig. Participants Four participants were excluded due to reporting incorrect knowledge of the response-outcome contingencies after the baseline concurrent choice stage, leaving 20 negative and 24 positive participants. Overall, participants were 23.8 years (std=4.2; range=19–34) of age, smoked 9.5 (5.7; 2–23) cig- arettes/day, had smoked for 6.2 years (4.8; 1–21), started at 15.9 years (2.6; 11–26) of age, had a CDS-5 score of 15.0 (4.0; 5–22) out of a maximum score of 25 and a QSU score of 4.2 (1.4; 1–7). There were no significant difference between the negative and positive group with respect to these Breath CO is in parts per million. Pleasure of puffs is per cent of visual analogue scale. Desire scores are on a 7-point scale (positive numbers equal greater desire). Mood state scores are on a 9-point scale (positive numbers equal greater sadness) Numbers are mean and sem in brackets 3241 Psychopharmacology (2015) 232:3235–3247 showed a significant increase in positive mood, F(1.23)=5.41, p=0.03, ŋp 2=0.19. Thus, the two mood induction procedures shifted mood in the expected directions. This association was examined in Fig. 2, which shows that the change self-reported in mood from pre- to post-mood induc- tion correlated significantly with the change in tobacco- seeking from baseline to the extinction test, r=0.43, p= 0.004. This correlation indicates that an increase in self- reported negative mood was associated with an increase in tobacco choice in the extinction test, reversing the effect of satiety, which otherwise decreased tobacco choice. Multiple regression analyses indicated that the absolute level of self- reported mood measured pre-induction (p=0.11) or post- inducted (p=0.073) did not predict the change in tobacco choice, but the change in mood from pre- to post-induction was the effective predictor (t=3.05, p=0.004). Thus, the acute change in negative mood was the salient variable which in- creased tobacco choice, rather than absolute negative mood pre- or post-induction. Relationship between the change in self-reported mood and tobacco-seeking Previous studies have reported correlations between negative subjective reactivity following induction procedures and sub- sequent increases in drug motivation (e.g., Sinha et al. 2008). Fig. 1 Mean per cent tobacco versus chocolate choice (±SEM) at baseline and extinction test for the negative and positive mood induction group (50 %=indifference; >50 %=tobacco preference; <50 %=chocolate preference). The extinction test was conducted after satiety and mood induction, revealing the combined effect of these variables on goal-directed tobacco-seeking. Satiety reduced goal- directed tobacco-seeking in the positive group, but this effect was abolished by negative mood induction Fig. 2 Relationship between the change in self-reported mood (pre- to post-mood induction), with the change in tobacco choice (from baseline to extinction test with satiety mood induction in between). Positive mood values reflect increased sadness, whereas negative values reflect in- creased happiness. Positive tobacco choice scores reflect increased tobac- co choice, whereas negative scores reflect decreased tobacco choice. Symbols identify the four sub-groups employed in Fig. 3: high positive (filled squares), low positive (empty squares), low negative (empty circles) and high negative (filled circles) who did and did not show sig- nificant changes in mood following induction. The scatterplot shows that increased negative mood was associated with increased tobacco choice at test, opposing satiety Fig. 2 Relationship between the change in self-reported mood (pre- to post-mood induction), with the change in tobacco choice (from baseline to extinction test with satiety mood induction in between). Positive mood values reflect increased sadness, whereas negative values reflect in- creased happiness. Positive tobacco choice scores reflect increased tobac- co choice, whereas negative scores reflect decreased tobacco choice. Symbols identify the four sub-groups employed in Fig. 3: high positive (filled squares), low positive (empty squares), low negative (empty circles) and high negative (filled circles) who did and did not show sig- nificant changes in mood following induction. The scatterplot shows that increased negative mood was associated with increased tobacco choice at test, opposing satiety Fig. 2 Relationship between the change in self-reported mood (pre- to post-mood induction), with the change in tobacco choice (from baseline to extinction test with satiety mood induction in between). Positive mood values reflect increased sadness, whereas negative values reflect in- creased happiness. Positive tobacco choice scores reflect increased tobac- co choice, whereas negative scores reflect decreased tobacco choice. Symbols identify the four sub-groups employed in Fig. Relationship between the change in self-reported mood and tobacco-seeking 3: high positive (filled squares), low positive (empty squares), low negative (empty circles) and high negative (filled circles) who did and did not show sig- nificant changes in mood following induction. The scatterplot shows that increased negative mood was associated with increased tobacco choice at test, opposing satiety Fig. 1 Mean per cent tobacco versus chocolate choice (±SEM) at baseline and extinction test for the negative and positive mood induction group (50 %=indifference; >50 %=tobacco preference; <50 %=chocolate preference). The extinction test was conducted after satiety and mood induction, revealing the combined effect of these variables on goal-directed tobacco-seeking. Satiety reduced goal- directed tobacco-seeking in the positive group, but this effect was abolished by negative mood induction 3242 Psychopharmacology (2015) 232:3235–3247 Fig. 3 a Change in self-reported mood (pre- to post-induction) in four sub-groups. Positive mood values reflect increased sadness, whereas neg- ative values reflect increased happiness. The high-negative group showed a significant increase in sadness, the high-positive group showed a sig- nificant increase in happiness, and the two low groups showed no signif- icant change in mood. b Change in tobacco choice (from baseline to extinction test with satiety mood induction in between) in the four sub- groups. Positive tobacco choice scores reflect increased tobacco choice, whereas negative scores reflect decreased tobacco choice. The high- negative group showed a significant increase in tobacco choice despite satiety, whereas the remaining three sub-groups all showed a significant decrease in tobacco choice Fig. 3 a Change in self-reported mood (pre- to post-induction) in four sub-groups. Positive mood values reflect increased sadness, whereas neg- ative values reflect increased happiness. The high-negative group showed a significant increase in sadness, the high-positive group showed a sig- nificant increase in happiness, and the two low groups showed no signif- icant change in mood. b Change in tobacco choice (from baseline to extinction test with satiety mood induction in between) in the four sub- groups. Positive tobacco choice scores reflect increased tobacco choice, whereas negative scores reflect decreased tobacco choice. The high- negative group showed a significant increase in tobacco choice despite satiety, whereas the remaining three sub-groups all showed a significant decrease in tobacco choice Fig. 3 a Change in self-reported mood (pre- to post-induction) in four sub-groups. Positive mood values reflect increased sadness, whereas neg- ative values reflect increased happiness. Relationship between the change in self-reported mood and tobacco-seeking The high-negative group showed a significant increase in sadness, the high-positive group showed a sig- nificant increase in happiness, and the two low groups showed no signif- icant change in mood. b Change in tobacco choice (from baseline to extinction test with satiety mood induction in between) in the four sub- groups. Positive tobacco choice scores reflect increased tobacco choice, whereas negative scores reflect decreased tobacco choice. The high- negative group showed a significant increase in tobacco choice despite satiety, whereas the remaining three sub-groups all showed a significant decrease in tobacco choice 0.23. Thus, increased self-reported negative mood was the crucial factor increasing goal-directed tobacco choice, oppos- ing satiety. and the high-positive sub-group’s mood change was signifi- cantly less than zero, t(12)=−7.67, p<0.001, whereas the low- negative and low-positive sub-groups combined showed no significant change in self-reported mood, t(19)=1.76, p=0.09. g g p , ( ) , p The important analysis concerns change in tobacco choice by the four sub-groups, the mean and SEM for which is shown in Fig. 3b. ANOVA on tobacco choice data with the variable sub-group (4) yielded a significant main effect, F(3.40)=5.54, p=0.003, ŋp 2=0.29. Furthermore, contrasts of pairs using Bonferroni post-hoc tests indicated that the high-negative sub-group differed from all other sub-groups, ps≤0.02, where- as the remaining three sub-groups did not differ from each other (ps=1.0). Within-groups t tests indicated that whereas the high-negative sub-group showed a significant increase in tobacco choice at test, t(10)=2.32, p=0.04, the remaining three sub-groups combined showed a significant decrease in tobacco choice at test, t(32)=−4.47, p<0.001. These within- group t tests were significant even when the Holm–Bonferroni correction was applied. Finally, there was no reliable differ- ence between the four sub-groups with respect to baseline choice, F<1, nor when baseline choice was contrasted be- tween the high- and low-negative group, F<1, or high- and low-positive group, F<1, indicating comparability of tobacco choice at baseline. Thus, a significant increase in self-reported negative mood augmented tobacco choice opposing satiety, whereas variations in self-reported positive mood had no im- pact on tobacco choice. Discussion The post-hoc nature of the assignment of participants to the sub-groups means that a range of confounding factors could explain the increase in tobacco choice in the high-negative group other than the increase in negative mood per se. Several potential confounds can be excluded however. Multiple regression analysis indicated that the increase in goal-directed tobacco-seeking was uniquely predicted by the increase in negative mood, rather than absolute negative mood either before or after induction. In addition, the sub-groups were comparable in their baseline tobacco choice, demograph- ic variables and experience of the devaluation treatment, sug- gesting these variables were not responsible for the increase goal-directed tobacco choice. Finally, the two positive sub- groups did not differ in their change in tobacco choice, so one might conclude that it was negative mood specifically, rather than general arousal produced by the induction proto- col, which increased goal-directed tobacco choice at test. However, we did not measure the general arousal induced in each sub-group, so it remains possible that greater arousal, rather than negative mood per se, drove the increase in tobac- co choice in the high-negative group (perhaps akin to a stress induction procedure). This issue remains to be explored. Finally, a range of unmeasured individual differences such as depression symptoms (Audrain-McGovern et al. 2014) might have been confounded with the sub-group assignment and driven the increase in tobacco choice at test. Future studies should seek to fully characterise in- dividual differences in sensitivity to negative mood in- duced priming of goal-directed drug-seeking. One final defence of S-R theories would be to suggest that the current evidence that negative mood functions as a moti- vational state raising the expected value of drug reinforcement does not exclude the possibility that negative mood might also undergo S-R learning with respect to responses which have been directly reinforced in the mood state. Moreover, such S- R based control by the negative mood state over drug-seeking might be more pronounced in more dependent or psychiatry ill drug users, accounting for their transition to clinical drug de- pendence. Consistent with this view, there is some evidence that smoking behaviour is more prone to habitual control, indexed by its null correlation with subjective craving, com- pared with a novel tobacco-seeking response, in more impul- sive smokers (Hogarth 2011; see also, Gass et al. 2014). Discussion First, on the S-R account, changes in responding can only be brought about by direct experience of the outcome modifying the S-R asso- ciation controlling the response (Dickinson 1985; Dickinson and Balleine 2010). As outcomes were omitted in the extinc- tion test, the change in responding produced by satiety and negative mood could not be driven by changes in S-R strength driven by a change in the direct experience of the tobacco outcome, but must have been mediated by an expectation of the tobacco outcome. Second, negative mood was induced between baseline concurrent training and the extinction test, such that the tobacco-seeking response was never reinforced in the presence of the negative mood state. Consequently, negative mood could not have formed an S-R association with the tobacco-seeking response to control performance at test. Third, if S-R associations have formed between negative mood and smoking behaviour in the natural environment, such learning could only generalise to promote tobacco over chocolate-seeking at test via a representation of the reinforcer shared between these specific responses, and this proposal is beyond the scope of S-R theory. satiety (Hutcheson et al. 2001). The incentive learning inter- pretation also accords with other animal studies which have shown that if internal states which predict opposite values of the reinforcer are manipulated simultaneously, they produce opposing effects on goal-directed action selection in the ex- tinction test similar to that found here (Balleine et al. 1994; Balleine and Dickinson 1994; see also, DeGrandpre et al. 1992; Troisi et al. 2012; Weiss 1972; White and Stolerman 1996). ) The sub-group analysis distinguishes between the impaired goal-directed control and the incentive learning accounts of Fig. 1. The sub-group analysis of Fig. 3 showed that partici- pants who reported a significant increase in negative mood significantly increase their tobacco choice in the extinction test, whereas all other participants significantly decreased their tobacco choice in the extinction test. This finding cannot be explained by negative mood impairing goal-directed control because this process would produce no change in choice at test. By contrast, the increase in tobacco choice in the high- negative group can be explained by negative mood function- ing as a motivational state raising goal-directed tobacco-seek- ing in opposition to satiety. Discussion The study of young adult daily smokers found that smoking satiety decreased tobacco-seeking in the extinction test of the outcome-devaluation procedure, compared with the abstinent baseline, replicating previous studies (Hogarth 2012; Hogarth and Chase 2011; Hogarth et al. 2013). This devaluation effect indicates that tobacco choice is goal directed in being deter- mined by knowledge of the response-outcome contingencies acquired in concurrent training combined with knowledge of the low value of the tobacco outcome in the sated state (incen- tive learning). The novel finding of the current study was that this devaluation effect on goal-directed tobacco-seeking pro- duced by satiety was abolished by negative mood induction (compared with positive mood induction). This abolition of the devaluation effect shown in Fig. 1 is ambiguous in its interpretation. This effect could be produced by negative mood impairing retrieval of goal-directed knowledge at test, resulting in no change in responding, as has been found with stress induction (Schwabe and Wolf 2009, 2010), acute alco- hol (Hogarth et al. 2012) and alcohol expectancy (Hogarth et al. 2013). Alternatively, the abolition of the devaluation effect in Fig. 1 could be produced by negative mood (relative to positive mood) functioning as a motivational state, raising the expected value of tobacco opposing satiety in the control of goal-directed tobacco-seeking (c.f. Willner and Jones 1996). According to the incentive learning interpretation, neg- ative mood has acquired similar incentive properties to drug withdrawal, which has been shown to raise the expected value of the drug controlling drug-seeking in a direction opposite to Importantly, the two negative sub-groups were not con- founded with respect to their experience of the devaluation treatment—they showed no reliable difference in breath CO, number of puffs consumed, pleasure of puffs consumed or desire for tobacco, Fs<1—nor were they confounded with respect to participant characteristics—they showed no reliable difference in age, cigarettes smoked per day, years smoking, age of smoking onset, CDS-5 or QSU scores, ts(18)<1.2, ps> Psychopharmacology (2015) 232:3235–3247 3243 compulsive associations, including negative reinforcement, allostasis and incentive habit theory (Baker et al. 2004; Belin et al. 2013; Koob 2009; Koob and Volkow 2010), might seek to defend their position by attributing the current effects to the formation of S-R associations within the task. There are three principled objections to this claim. Discussion 1996), there is also little evidence that smoking alleviates either exper- imentally induced negative mood (Conklin and Perkins 2005; Kassel and Shiffman 1997; Kassel and Unrod 2000; Perkins et al. 2008, 2010a; Willner and Jones 1996), or depression (Colman et al. 2011). On the con- trary, long-term abstinence appears to improve depres- sion (Cavazos-Rehg et al. 2014; Mathew et al. 2013; Piper et al. 2013; Prochaska et al. 2008; Shahab et al. 2013). Therefore, in subjective report at least, there is little indication that negative mood and depression are associated with the greater reward value of smoking. This problem for the self-medication hypothesis has been answered in at least two ways (Kassel et al. 2003; Khantzian 1997). One answer is that induced negative mood and depres- sion mimic withdrawal-related negative mood (Baker et al. 2004; Heilig et al. 2010; Hughes 2007; Parrott 1999), which is reliably alleviated by smoking (Hatsukami et al. 1984; Hughes et al. 1984; Parrott 1995; Perkins et al. 2010a). On this view, negative mood is a partially reliable predictor of the greater reward value of smoking. One objection to this claim is that if negative mood was a partial predictor, it should not have competed so effectively against the more reliable predic- tor, satiety, for the control over tobacco-seeking (White and Stolerman 1996). One remaining question concerns the relevance of the cur- rent findings to depression. Depression is a heterogenous dis- order made up of a cluster of symptoms including negative mood, anhedonia, cognitive impairment, concentration diffi- culty, neurovegetion, psychomotor retardation, appetite loss and insomnia. It is unlikely that the mood induction procedure used in the current study impacted on depressive symptoms outside negative mood, particularly the non-affective symp- toms. Therefore, the study cannot speak about the role these other symptoms play in the maintenance of drug use. The other issue is that major depression is more severe and chronic than the transitory induced negative mood state. A diagnosis of depression requires that depression lasts for at least 2 weeks and is associated with a clinically significant change in func- tioning. Thus, there are both qualitative and quantitative dif- ferences between mood induction and depression. This anal- ysis does not discount the importance of the present findings. Discussion Despite this, the burden of proof currently rests on habit the- orists to provide positive evidence that negative mood can acquire S-R control over drug-seeking in ‘ecologically valid’ schedules that mimic complex human learning environments (Kosaki and Dickinson 2010; Sjoerds et al. 2014). A difficult question remains as to how negative mood be- comes established as a signal for the greater reinforcement value of smoking, because there is little evidence that negative mood induction actually increases the liking of smoking (Perkins et al. 2010a); although stress induction does increase smoking liking (McKee et al. 2011; Zinser et al. 1992) and studies are somewhat equivocal as to whether depression prone smokers report increases in smoking liking (Audrain- Theories which claim that negative mood primes tobacco- seeking via direct, automatic, unconscious, habitual or 3244 Psychopharmacology (2015) 232:3235–3247 As depression and anhedonia are associated with reduced learning about natural reward contingencies in smokers (Liverant et al. 2014; Pergadia et al. 2014), it is possible that negative mood simultaneously reduced the value of chocolate and increased the value of smoking, expanding the differential value between these two outcomes, driving up the proportion of tobacco choices (this could explain why the high-negative group chose tobacco above their abstinent baseline). The con- current choice procedure means that we cannot rule out the possibility that negative mood reduced chocolate value. However, there is substantial evidence that negative mood augments the value of smoking when tested independently of a natural reward alternative (Brandon et al. 1996; Conklin and Perkins 2005; Fucito et al. 2010; Payne et al. 1991; Perkins et al. 2013; Vinci et al. 2012; Willner and Jones 1996), so we can be confident in this aspect of the interpreta- tion. By contrast, Conklin and Perkins (2005) found no change in motivation for water following negative mood in- duction, suggesting that negative mood does not reduce natu- ral reward value. Furthermore, depression is prospectively linked to developing obesity, suggesting that negative mood may not undermine food motivation (Luppino et al. 2010). Nevertheless, the possibility that negative mood changes the value of both two outcomes, to bias action selection, remains an important consideration that requires direct testing. McGovern et al. 2014; Perkins et al. 2010b). Furthermore, even though smokers who report depressed mood claim to smoke to alleviate depression (Lerman et al. Discussion Indeed, the present findings may have more broad spanning implications beyond depression, addressing how negative af- fect modulates expected outcome values in healthy people and across those with different clinical diagnoses, including de- pression, anxiety and psychoses, all of which have an affective component. Finally, our major claim that emergent, abnormal incentive learning underpins the transition to drug dependence is undermined by young health smoker cohort studied here. Extensive testing in clinical samples is requires before this claim becomes anything more than a plausible possibility. Another possibility is that the correction of anhedonia (en- gagement with reinforcers) by smoking, rather than correction of negative mood per se, provides the additional reinforce- ment signal. Support for this claim comes from the finding that anhedonia increases during abstinence and predicts re- lapse (Goelz et al. 2014; Leventhal et al. 2008, 2014a) and can be corrected by acute smoking or nicotine (Dawkins et al. 2006; Donny et al. 2003; Liverant et al. 2014; Pergadia et al. 2014; Perkins and Karelitz 2013; Powell et al. 2002). On this view, smoking is more reinforcing in the negative mood state because it alleviates co-occurring anhedonia, increasing en- gagement with natural rewards (Ahmed and Koob 2005) rath- er than ameliorating negative mood per se. As anhedonia in- creases with the transition clinical drug dependence (Koob 2013), and withdrawal-related anhedonia is more pronounced in depressed individuals (Pergadia et al. 2014), negative mood should become a more powerful motivational state, raising the expected value of the drug, as dependence grows. Thus, in- centive learning driving supernormal goal-directed drug choice to acutely correct burgeoning anhedonia could be the learning mechanism that underpins the transition to clinical drug dependence, consistent with choice based theories (Bickel et al. 2014; Henden et al. 2013; Heyman 2013) over automaticity theories of dependence (Belin et al. 2013; Everitt and Robbins 2013; Koob 2013; Koob and Volkow 2010; Tiffany 1990). 3245 Psychopharmacology (2015) 232:3235–3247 reinforcement pathologies and their repair. Annu Rev Clin Psychol 10:641–677 To conclude, numerous studies have shown that negative mood induction increases drug motivation, but it was previ- ously unclear if this effect could compete with primary moti- vational states. Our unique finding that negative mood fully countered satiety demonstrates that negative mood is a strong motivational state, capable of competing with primary moti- vational states. Discussion Furthermore, negative mood increased drug- seeking in an extinction test suggesting that negative mood, relative to positive mood, functions as a motivational state as envisaged by incentive learning theory, raising the expected value of the drug to drive goal-directed drug-seeking. This finding contradicts S-R theories which claim that negative mood controls drug-seeking directly. More speculatively, the data might suggest that psychiatric illness writ large confers vulnerability to drug dependence not by promoting automatic control of behaviour but by enabling incentive learning, wherein psychiatric states drive supernormal goal-directed drug choice in order to acutely correct (but ultimately exacer- bate) those psychiatric states. Bradfield LA, Balleine BW (2013) Hierarchical and binary associations compete for behavioral control during instrumental biconditional discrimination. J Exp Psychol Anim Behav Process 39:2–13 Brandon TH, Wetter DW, Baker TB (1996) Affect, expectancies, urges and smoking: do they conform to models of drug motivation and relapse? Exp Clin Psychopharmacol 4:29–36 Breslau N, Peterson EL, Schultz LR, Chilcoat HD, Andreski P (1998) Major depression and stages of smoking. A longitudinal investiga- tion. Arch Gen Psychiatry 55:161–6 Brody AL, Olmstead RE, Abrams AL, Costello MR, Khan A, Kozman D, Saxena S, Farahi J, London ED, Mandelkern MA (2009) Effect of a history of major depressive disorder on smoking-induced dopa- mine release. Biol Psychiatry 66:898–901 Carmody TP (1989) Affect regulation, tobacco addiction, and smoking cessation. J Psychoactive Drugs 21:331–342 Cavazos-Rehg PA, Breslau N, Hatsukami D, Krauss MJ, Spitznagel EL, Grucza RA, Salyer P, Hartz SM, Bierut LJ (2014) Smoking cessa- tion is associated with lower rates of mood/anxiety and alcohol use disorders. Psychol Med 44:2523–2535 Colman I, Naicker K, Zeng Y, Ataullahjan A, Senthilselvan A, Patten SB (2011) Predictors of long-term prognosis of depression. CMAJ 183: 1969–76 Conklin CA, Perkins KA (2005) Subjective and reinforcing effects of smoking during negative mood induction. J Abnorm Psychol 114: 153–164 Declaration of interest The work was supported by the Medical Research Council (G0701456 to LH). Corbit LH, Nie H, Janak PH (2012) Habitual alcohol seeking: time course and the contribution of subregions of the dorsal striatum. Biol Psychiatry 72:389–395 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appro- priate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Discussion Dawkins L, Powell JH, West R, Powell J, Pickering A (2006) A double- blind placebo controlled experimental study of nicotine: I—effects on incentive motivation. Psychopharmacology 189:355–367 DeGrandpre RJ, Bickel WK, Higgins ST (1992) Emergent equivalence relations between interoceptive (drug) and exteroceptive (visual) stimuli. J Exp Anal Behav 58:9–18 Dickinson A (1985) Actions and habits—the development of behavioral autonomy. Philos Trans R Soc B Biol Sci 308:67–78 Dickinson A, Balleine BW (2010) The cognitive/motivational interface. In: Kringelbach ML, Berridge KC (eds) Pleasures of the brain the neural basis of taste, smell and other rewards. Oxford University Press, Oxford, pp 74–84 References Drug Alcohol Depend 138:67– 74 Koob GF, Volkow ND (2010) Neurocircuitry of addiction. Neuropsychopharmacology 35:1051–1051 and short-term smoking cessation. Drug Alcohol Depend 138:67– 74 Hatsukami D, Hughes J, Pickens R, Svikis D (1984) Tobacco withdrawal symptoms: an experimental analysis. Psychopharmacology 84:231– 236 Kosaki Y, Dickinson A (2010) Choice and contingency in the develop- ment of behavioral autonomy during instrumental conditioning. J Exp Psychol Anim Behav Process 36:334–342 Heilig M, Egli M, Crabbe JC, Becker HC (2010) Acute withdrawal, protracted abstinence and negative affect in alcoholism: are they linked? Addict Biol 15:169–184 Laraway S, Snycerski S, Olson R, Becker B, Poling A (2014) The moti- vating operations concept: current status and critical response. Psychol Rec 64:601–623 Lenton SR, Martin PR (1991) The contribution of music vs instructions in the musical mood induction procedure. Behav Res Ther 29:623–625 Henden E, Melberg H-O, Rogeberg O (2013) Addiction: choice or com- pulsion? Front Psychiatry 4:77 Heyes C, Dickinson A (1990) The intentionality of animal action. Mind Lang 5:87–104 Lerman C, Audrain J, Orleans CT, Boyd R, Gold K, Main D, Caporaso N (1996) Investigation of mechanisms linking depressed mood to nic- otine dependence. Addict Behav 21:9–19 Heyman GM (2013) Addiction and choice: theory and new data. Front Psychiatry 4:31 Leventhal AM, Ramsey SE, Brown RA, LaChance HR, Kahler CW (2008) Dimensions of depressive symptoms and smoking cessation. Nicotine Tob Res 10:507–517 Hitsman B, Papandonatos GD, McChargue DE, DeMott A, Herrera MJ, Spring B, Borrelli B, Niaura R (2013) Past major depression and smoking cessation outcome: a systematic review and meta-analysis update. Addiction 108:294–306 Leventhal AM, Piper ME, Japuntich SJ, Baker TB, Cook JW (2014a) Anhedonia, depressed mood, and smoking cessation outcome. J Consult Clin Psychol 82:122–129 Hogarth L (2011) The role of impulsivity in the aetiology of drug depen- dence: reward sensitivity versus automaticity. Psychopharmacology 215:567–580 Leventhal AM, Trujillo M, Ameringer KJ, Tidey JW, Sussman S, Kahler CW (2014b) Anhedonia and the relative reward value of drug and nondrug reinforcers in cigarette smokers. J Abnorm Psychol 123: 375–386 Hogarth L (2012) Goal-directed and transfer-cue-elicited drug-seeking are dissociated by pharmacotherapy: evidence for independent ad- ditive controllers. J Exp Psychol Anim Behav Process 38:266–278 Liverant GI, Sloan DM, Pizzagalli DA, Harte CB, Kamholz BW, Rosebrock LE, Cohen AL, Fava M, Kaplan GB (2014) Associations among smoking, anhedonia, and reward learning in depression. References Behav Ther 45:651–663 Hogarth L, Chase HW (2011) Parallel goal-directed and habitual control of human drug-seeking: implications for dependence vulnerability. J Exp Psychol Anim Behav Process 37:261–276 Luppino FS, de Wit LM, Bouvy PF, Stijnen T, Cuijpers P, Penninx BWJH, Zitman FG (2010) Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies. Arch Gen Psychiatry 67:220–229 Hogarth L, Attwood AS, Bate HA, Munafò MR (2012) Acute alcohol impairs human goal-directed action. Biol Psychol 90:154–160 Hogarth L, Field M, Rose AK (2013) Phasic transition from goal-directed to habitual control over drug-seeking produced by conflicting rein- forcer expectancy. Addict Biol 18:88–97 Mathew AR, Robinson JD, Norton PJ, Cinciripini PM, Brown RA, Blalock JA (2013) Affective trajectories before and after a quit at- tempt among smokers with current depressive disorders. Nicotine Tob Res 15:1807–1815 Hughes JR (1999) Comorbidity and smoking. Nicotine Tob Res 1(Suppl 2):S149–52 Hughes JR (2007) Effects of abstinence from tobacco: etiology, animal models, epidemiology, and significance: a subjective review. Nicotine Tob Res 9:329–339 McKee SA, Sinha R, Weinberger AH, Sofuoglu M, Harrison EL, Lavery M, Wanzer J (2011) Stress decreases the ability to resist smoking and potentiates smoking intensity and reward. J Psychopharmacol 25:490–502 Hughes JR, Hatsukami D, Pickens R, Krahn D, Malin S, Luknic A (1984) Effect of nicotine on the tobacco withdrawal syndrome. Psychopharmacology 83:82–87 Michael J (1993) Establishing operations. Behav Anal 16:191–206 Hull CL (1943) Principles of behavior. Appleton, New York Miles FJ, Everitt BJ, Dickinson A (2003) Oral cocaine seeking by rats: action or habit? Behav Neurosci 117:927–938 Hursh SR, Silberberg A (2008) Economic demand and essential value. Psychol Rev 115:186–198 Morrison R, O’Connor RC (2008) The role of rumination, attentional biases and stress in psychological distress. Br J Psychol 99:191–209 Hutcheson DM, Everitt BJ, Robbins TW, Dickinson A (2001) The role of withdrawal in heroin addiction: enhances reward or promotes avoid- ance? Nat Neurosci 4:943–947 Parrott AC (1995) Stress modulation over the day in cigarette smokers. Addiction 90:233–244 Kassel JD, Shiffman S (1997) Attentional mediation of cigarette smok- ing’s effect on anxiety. Health Psychol 16:359–368 Parrott AC (1999) Does cigarette smoking cause stress? Am Psychol 54: 817–820 Kassel JD, Unrod M (2000) Smoking, anxiety, and attention: support for the role of nicotine in attentionally mediated anxiolysis. References Ahmed SH, Koob GF (2005) Transition to drug addiction: a negative reinforcement model based on an allostatic decrease in reward func- tion. Psychopharmacology 180:473–490 Dickinson A, Wood N, Smith JW (2002) Alcohol seeking by rats: action or habit? Q J Exp Psychol B 55:331–348 Donny E, Chaudhri N, Caggiula A, Evans-Martin FF, Booth S, Gharib M, Clements L, Sved A (2003) Operant responding for a visual rein- forcer in rats is enhanced by noncontingent nicotine: implications for nicotine self-administration and reinforcement. Psychopharmacology 169:68–76 Audrain-McGovern J, Wileyto EP, Ashare R, Cuevas J, Strasser AA (2014) Reward and affective regulation in depression-prone smokers. Biol Psychiatry 76:689–697 Baker TB, Piper ME, McCarthy DE, Majeskie MR, Fiore MC (2004) Addiction motivation reformulated: an affective processing model of negative reinforcement. Psychol Rev 111:33–51 Dougher MJ, Hackbert L (2000) Establishing operations, cognition, and emotion. Behav Anal 23:11–24 Balleine BW, Dickinson A (1994) Role of cholecystokinin in the moti- vational control of instrumental action in rats. Behav Neurosci 108: 590–605 Etter JF, Le Houezec J, Perneger TV (2003) A self-administered ques- tionnaire to measure dependence on cigarettes: the cigarette depen- dence scale. Neuropsychopharmacology 28:359–370 Balleine BW, Ball J, Dickinson A (1994) Benzodiazepine-induced out- come revaluation and the motivational control of instrumental action in rats. Behav Neurosci 108:573–589 Everitt BJ, Robbins TW (2013) From the ventral to the dorsal striatum: devolving views of their roles in drug addiction. Neurosci Biobehav Rev 37:1946–1954 Belin D, Belin-Rauscent A, Murray JE, Everitt BJ (2013) Addiction: failure of control over maladaptive incentive habits. Curr Opin Neurobiol 23:564–572 Fucito LM, Juliano LM, Toll BA (2010) Cognitive reappraisal and ex- pressive suppression emotion regulation strategies in cigarette smokers. Nicotine Tob Res 12:1156–1161 Berna C, Leknes S, Holmes EA, Edwards RR, Goodwin GM, Tracey I (2010) Induction of depressed mood disrupts emotion regulation neurocircuitry and enhances pain unpleasantness. Biol Psychiatry 67:1083–1090 Gass JC, Motschman CA, Tiffany ST (2014) The relationship between craving and tobacco use behavior in laboratory studies: a meta-anal- ysis. Psychol Addict Behav 28:1162–1176 Goelz PM, Audrain-McGovern JE, Hitsman B, Leone FT, Veluz-Wilkins A, Jepson C, Wileyto EP, D’Avanzo PA, Rivera JG, Schnoll RA (2014) The association between changes in alternative reinforcers Bickel WK, Johnson MW, Koffarnus MN, MacKillop J, Murphy JG (2014) The behavioral economics of substance use disorders: 3246 Psychopharmacology (2015) 232:3235–3247 Koob GF, Volkow ND (2010) Neurocircuitry of addiction. Neuropsychopharmacology 35:1051–1051 and short-term smoking cessation. References Psychopharmacology 210:25–34 Skinner BF (1938) The behavior of organisms: an experimental analysis. Appleton-Century, Oxford, England Perkins KA, Karelitz JL, Giedgowd GE, Conklin CA (2013) Negative mood effects on craving to smoke in women versus men. Addict Behav 38:1527–1531 Spring B, Pingitore R, McChargue DE (2003) Reward value of cig- arette smoking for comparably heavy smoking schizophrenic, depressed, and nonpatient smokers. Am J Psychiatry 160:316– 322 Piper ME, Rodock M, Cook JW, Schlam TR, Fiore MC, Baker TB (2013) Psychiatric diagnoses among quitters versus continuing smokers 3 years after their quit day. Drug Alcohol Depend 128:148–154 Swendsen J, Conway KP, Degenhardt L, Glantz M, Jin R, Merikangas KR, Sampson N, Kessler RC (2010) Mental disorders as risk factors for substance use, abuse and dependence: results from the 10 years follow-up of the National Comorbidity Survey. Addiction 105: 1117–1128 Pittenger ST, Bevins RA (2013) Interoceptive conditioning with a nico- tine stimulus is susceptible to reinforcer devaluation. Behav Neurosci 127:465–473 Pomerleau OF, Pomerleau CS (1984) Neuroregulators and the reinforce- ment of smoking: towards a biobehavioral explanation. Neurosci Biobehav Rev 8:503–513 Tiffany ST (1990) A cognitive model of drug urges and drug-use behav- iour: role of automatic and nonautomatic processes. Psychol Rev 97: 147–168 Powell J, Dawkins L, Davis RE (2002) Smoking, reward responsiveness, and response inhibition: tests of an incentive motivational model. Biol Psychiatry 51:151–163 Tiffany ST, Drobes DJ (1991) The development and initial validation of a questionnaire on smoking urges. Br J Addict 86:1467–1476 Trask S, Bouton M (2014) Contextual control of operant behavior: evi- dence for hierarchical associations in instrumental learning. Learn Behav 42:281–288 Prochaska JJ, Hall SM, Tsoh JY, Eisendrath S, Rossi JS, Redding CA, Rosen AB, Meisner M, Humfleet GL, Gorecki JA (2008) Treating tobacco dependence in clinically depressed smokers: effect of smoking cessation on mental health functioning. Am J Public Health 98:446–8 Troisi JRI (2014) Sensation within the skin. ACS Chem Neurosci Troisi JRI, Bryant E, Kane J (2012) Extinction of the discriminative stimulus effects of nicotine with a devalued reinforcer: recovery following revaluation. Psychol Rec 62:707–718 Ramsay DS, Woods SC (2014) Clarifying the roles of homeostasis and allostasis in physiological regulation. Psychol Rev 121:225–247 Velten E (1968) A laboratory task for induction of mood states. Behav Res Ther 6:473–482 Richell RA, Anderson M (2004) Reproducibility of negative mood in- duction: a self-referent plus musical mood induction procedure and a controllable/uncontrollable stress paradigm. References J Abnorm Psychol 109:161–166 Payne TJ, Schare ML, Levis DJ, Colletti G (1991) Exposure to smoking- relevant cues: effects on desire to smoke and topographical compo- nents of smoking behaviour. Addict Behav 16:467–479 Kassel JD, Stroud LR, Paronis CA (2003) Smoking, stress, and negative affect: correlation, causation, and context across stages of smoking. Psychol Bull 129:270–304 Pergadia ML, Der-Avakian A, D’Souza MS, Madden PAF, Heath AC, Shiffman S, Markou A, Pizzagalli DA (2014) Association between nicotine withdrawal and reward responsiveness in humans and rats. JAMA Psychiatry 71:1238–1245 Kassel JD, Veilleux JC, Wardle MC, Yates MC, Greenstein JE, Evatt DP, Roesch LL (2007) Negative affect and addiction stress and addic- tion: biological and psychological mechanisms. Elsevier Academic Press, San Diego, CA, USA, pp 171–189 Perkins KA, Karelitz JL (2013) Reinforcement enhancing effects of nic- otine via smoking. Psychopharmacology 228:479–486 Perkins KA, Epstein LH, Grobe J, Fonte C (1994) Tobacco abstinence, smoking cues, and the reinforcing value of smoking. Pharmacol Biochem Behav 47:107–112 Khantzian EJ (1997) The self-medication hypothesis of substance use disorders: a reconsideration and recent applications. Harv Rev Psychiatry 4:231–244 Perkins KA, Grobe J, Fonte C (1997) Influence of acute smoking expo- sure on the subsequent reinforcing value of smoking. Exp Clin Psychopharmacol 5:277–285 Koob GF (2009) Neurobiological substrates for the dark side of compul- sivity in addiction. Neuropharmacology 56(Supplement 1):18–31 Perkins KA, Ciccocioppo M, Conklin CA, Milanak ME, Grottenthaler A, Sayette MA (2008) Mood influences on acute smoking responses Koob GF (2013) Addiction is a reward deficit and stress surfeit disorder. Front Psychiatry 4:72 3247 Psychopharmacology (2015) 232:3235–3247 are independent of nicotine intake and dose expectancy. J Abnorm Psychol 117:79–93 Sinha R, Fox HC, Hong KA, Bergquist K, Bhagwagar Z, Siedlarz KM (2008) Enhanced negative emotion and alcohol craving, and altered physiological responses following stress and cue exposure in alco- hol dependent individuals. Neuropsychopharmacology 34:1198– 1208 Perkins KA, Karelitz JL, Conklin CA, Sayette MA, Giedgowd GE (2010a) Acute negative affect relief from smoking depends on the affect situation and measure but not on nicotine. Biol Psychiatry 67: 707–14 Sjoerds Z, Luigjes J, Van Den Brink W, Denys D, Yücel M (2014) The role of habits and motivation in human drug addiction: a reflection. Front Psychiatry 5:8 Perkins KA, Karelitz JL, Giedgowd GE, Conklin CA, Sayette MA (2010b) Differences in negative mood-induced smoking reinforce- ment due to distress tolerance, anxiety sensitivity, and depression history. References J Psychopharmacol 18: 94–101 Vinci C, Copeland AL, Carrigan MH (2012) Exposure to negative affect cues and urge to smoke. Exp Clin Psychopharmacol 20:47–55 Weiss SJ (1972) Stimulus compounding in free-operant and clas- sical conditioning: a review and analysis. Psychol Bull 78: 189–208 SAMHSA (2012) In: NSDUH (ed) Results from the 2011 national survey on drug use and health: mental health findings. Substance Abuse and Mental Health Services Administration, Rockville, MD White JAW, Stolerman IP (1996) Reversal of overshadowing in a drug mixture discrimination in rats. Psychopharmacology 123:46–54 Schwabe L, Wolf OT (2009) Stress prompts habit behavior in humans. J Neurosci 29:7191–7198 Willner P, Jones C (1996) Effects of mood manipulation on subjective and behavioural measures of cigarette craving. Behav Pharmacol 7: 355–363 Schwabe L, Wolf OT (2010) Socially evaluated cold pressor stress after instrumental learning favors habits over goal-directed action. Psychoneuroendocrinology 35:977–986 Zinser MC, Baker TB, Sherman JE, Cannon DS (1992) Relation between self-reported affect and drug urges and cravings in continuing and withdrawing smokers. J Abnorm Psychol 101:617–629 Shahab L, Andrew S, West R (2013) Changes in prevalence of depression and anxiety following smoking cessation: results from an interna- tional cohort study (ATTEMPT). Psychol Med 44:127–141
https://openalex.org/W3200607721	https://www.researchsquare.com/article/rs-919197/latest.pdf	English	null	Characteristic Analysis and Electrical Conductivity of Graphene and Graphene-Nanosilver Prepared by Electrical Spark Discharge Method	Research Square (Research Square)	2,021	cc-by	4,367	Characteristic Analysis and Electrical Conductivity of Graphene and Graphene-Nanosilver Prepared by Electrical Spark Discharge Method Kuo-Hsiung Tseng (  khtseng@ee.ntut.edu.tw ) National Taipei University of Technology Chu-Ti Yeh National Taipei University of Technology Hsueh-Chien Ku National Taipei University of Technology Der-Chi Tien National Taipei University of Technology Leszek Stobinski Warsaw University of Technology ABSTRACT This study used an electric discharge machine (EDM) to perform the electrical spark discharge method (ESDM) to prepare a graphene colloid and a graphene-Ag colloid. The characteristic wavelengths of graphene, and graphene-Ag are both 262 nm. They had the properties of high dispersion and are unlikely to aggregate. The XRD patterns of graphene and graphene-Ag are typical carbon diffraction peak angles and crystal orientations. Graphene-Ag in DW can increase the Raman signal intensity of graphene. Regarding the graphene colloid and graphene-Ag colloid, their average sheet resistance values are 0.0329 MΩ/sq and 0.00136 MΩ/sq, respectively. Moreover, when AgNPs composited with graphene layers, the average sheet resistance is only 1/24 that of graphene layers, indicating that graphene-Ag has superior conductivity. Keywords: Electrical spark discharge method, Graphene, Graphene-Ag, Conductivity Characteristic Analysis and Electrical Conductivity of Graphene and Graphene-Nanosilver Prepared by Electrical Spark Discharge Method Kuo-Hsiung Tsenga,, Chu-Ti Yeha, Hsueh-Chien Kua, Der-Chi Tiena, and Leszek Stobinskib aDepartment of Electrical Engineering, National Taipei University of Technology, Taipei 10608, Taiwan, R.O.C. bMaterials Chemistry, Warsaw University of Technology, Warynskiego 1, 00-645 Warsaw, Poland Correspondence should be addressed to Kuo-Hsiung Tseng: khtseng@ee.ntut.edu.tw Research Article Keywords: Electrical spark discharge method, Graphene, Graphene-Ag, Conductivity Posted Date: September 23rd, 2021 DOI: https://doi.org/10.21203/rs.3.rs-919197/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License License:   This work is licensed under a Creative Commons Attribution 4.0 International License. 1. Introduction EDM is an unconventional machining method. It utilizes the ESDM to process conductive materials. The principle is to convert electricity into heat and use high energy to cut conductive materials[16-17]. First, a conductive material is placed between the upper electrode (the item that conducts processing) and the lower electrode (the item to be processed) and is soaked in a dielectric fluid with high insulation performance [18]. The dielectric fluid can be DW or an insulating liquid. DC pulse voltage is applied between the two electrodes to form an electric field. Subsequently, the servo control system pushes the upper electrode slowly toward the lower electrode. The discharge state and distance between the two electrodes are critical factors affecting the machining quality. The heat generated by the electric arc increases the temperature of the electrode surface up to 6000– 10,000°C. The high temperature causes the electrodes to melt and vaporize. Simultaneously, minute particles peel from their surface. These particles rapidly cool, condense, and spread in the dielectric fluid. At this moment, the discharge pulse stops. The surfaces of the two electrodes return to being insulated, awaiting the discharge pulse of the next cycle. Through the aforementioned method, using periodic discharge, 1. Graphene consists of a single layer of graphite. Its structure is a two-dimensional thin film honeycomb lattice formed with carbon atoms in an sp2 hybrid orbital [1]. The thickness of graphene is only one carbon atom. The thickness of perfect one-layer graphene is 0.34 nm [2]. The distance of the C–C bond is 0.14 nm, which is the thinnest among all existing materials. It is a structurally stable and tough two-dimensional material. The potential applications of graphene, including integrated circuits (ICs)[3-4], super capacitors[5-6], transparent conductive electrodes[7-8], dye-sensitized solar cells[9-10], thermal conductive materials[11-12], and antibacterial applications[13-14].Graphene is a type of semimetal or zero-band-gap semiconductor. At room temperature, when electron transport is measured, the results indicate that graphene exhibits high electron mobility, with a mobility rate of approximately 200,000 cm2·V-1·s-1. The resistivity that corresponds to this value is 10−8 Ω·m, whereas the resistivity of silver is 10−8 Ω·m. Because the resistivity of graphene is slightly lower than silver, this means that it has superior conductivity to silver, making it the material with the highest conductivity in existence [15]. 1 1 Figure 1. Schematic of EDM conductive materials can be processed [19]. 2. Experimental system and method 2.1. Experimental system Figure 1 is a composition schematic of an EDM. When creating nanocolloids, the conductive materials prepared in advance are loaded onto the electrodes, and the lower electrode is connected to a DC power negative. Corresponding parameters values are set on the control panel to output different levels of discharge pulse, current, power, and stability. When the two electrodes are emerged in the dielectric fluid to initiate discharge processing, electrons in the dielectric fluid move from the negative electrode to positive electrode. The amount of metal consumed at the lower electrode is greater than that at the upper electrode; thus, the lower electrode is called the processed end, whereas the upper electrode is called the processing end. To ensure that tip of the electrode surface discharges stably, the lower and upper electrodes are usually made of conductive materials with a wire diameter of 2 mm and 1 mm, respectively. The two electrodes are placed in a dielectric fluid for the ESDM. The prepared nanocolloids are stored in scintillation counting vials. To prevent the carbon depositing of the two electrodes during the preparation process, which would cause a short circuit, a magnetic stirrer is placed underneath the storage container. The magnetic stirrer rotates the stir bar in the storage container so that the generated nanoparticles distribute in the dielectric fluid evenly. Before the discharge procedure begins, the upper and lower electrodes must be aligned. This can be done through adjusting the cross slide by rotating the bidirectional wheel axle. Finally, the control panel is used to set parameters, and the servo control system is used to control the distance between the two electrodes. An extremely small distance can effectively form spark discharge, enabling a smooth discharge process. Figure 1. Schematic of EDM Figure 1. Schematic of EDM EDM is a discharge process that involves the generation of sparks [20]. Two electrodes are kept in the dielectric fluid in an insulating state before being provided with a DC pulse power source. This forms a current channel between the two electrodes, destroying the insulating state of the dielectric fluid and generating sparks. Next, the electric energy is converted to heat energy, and the high temperature generated by the heat energy melts the surface of the material of the electrode. The material on the surface peels off, cools, and forms nanoparticles. Finally, the particles disperse in the dielectric fluid, which returns to the insulating state. The EDM discharge process is illustrated in Figure 2 and explained below. EDM is a discharge process that involves the generation of sparks [20]. Two electrodes are kept in the dielectric fluid in an insulating state before being provided with a DC pulse power source. This forms a current channel between the two electrodes, destroying the insulating state of the dielectric fluid and generating sparks. Next, the electric energy is converted to heat energy, and the high temperature generated by the heat energy melts the surface of the material of the electrode. The material on the surface peels off, cools, and forms nanoparticles. Finally, the particles disperse in the dielectric fluid, which returns to the insulating state. The EDM discharge process is illustrated in Figure 2 and explained below. (a) Preparation for discharge: The upper and lower electrodes of the electrode materials are soaked in the dielectric fluid. The two ends of the electrodes are connected to a DC power source and are confirmed to be aiming directly at each other. In this state, there is no voltage or current, and the dielectric fluid is in an insulating state. (b) Discharge initiation: When the discharge cycle begins, Ton begins and Vgap increases. When the upper electrode slowly moves toward the lower electrode until the distance between the two are extremely small, the electric field strength at the gap between the two electrodes exceeds the dielectric strength of the dielectric fluid. Electrons are injected from the surface of the lower electrode into the upper electrode. Molecules in the dielectric fluid become ionized, and the gap between the electrodes gradually establishes a discharge channel. Figure 1. Schematic of EDM 2 (c) Ionization: When the insulating state in the gap between the two electrodes is destroyed, electrons depart from the lower electrode, impact the neutral atoms in the dielectric fluid between the electrodes, and excite valence electrons at the most outer layer of the atoms to form positive ions and free electrons. At this time, electrons rapidly flow to the upper electrode, conduct ionization again, and form an ionization channel. Igap increases and Vgap decreases. with features and size suitable for discharge processing. Graphite-silver electrode materials are illustrated in Figure 4. Entire graphite of 99.9% purity was used. During processing, silver was added to burn into a chunk to generate a graphite-silver rod with features and size suitable for discharge processing. for discharge processing. Figure 3. Graphite rod electrodes Figure 4. Graphite-silver rod electrodes g p g Figure 3. Graphite rod electrodes g p (d) Melting: When discharge succeeds, Igap is maintained at its maximum value. The positive ions and free electrons keep impacting the upper and lower electrodes. When they impact the surface of the electrodes, the kinetic energy instantly turns into heat, forming a discharge spark. This high heat melts and evaporates the surface of the electrodes, and the metal electrode surface peels off nanoparticles. Figure 3. Graphite rod electrodes Figure 4. Graphite-silver rod electrodes (e) End of discharge: Toff begins, and Vgap and Igap decrease. When the pulse voltage and pulse current enter the Toff state, electrodes stop releasing electrons and the current starts to reduce. The ionized channel rapidly disappears. The nanoparticles that melted and peeled off from the electrode surface are suspended in the dielectric fluid. The gap voltage and gap current reduce. Figure 4. Graphite-silver rod electrodes Figure 4. Graphite-silver rod electrodes 2.3 This study used EDM to prepare a graphene colloid and a graphene-Ag colloid. The environmental parameter settings for the preparation are listed in Table 1. The dielectric fluid used was DW and the amount was 150 mL. Ton:Toff were set at 30:30 μs and the voltage was set at 140 V. The current was set as the second level. The discharge time for the graphene colloid was 27 min. To fix the lower electrode in a container, entire graphite of 99.9% purity was processed into a shape that could be fixed in the base of the lower electrode and then placed in the container. The environmental parameter settings for preparing the graphene-Ag colloid are listed in Table 2. The dielectric fluid used was DW and the amount was 150 mL. Ton:Toff were set at 30:30 μs and the voltage was set at 140 V. The current was set as the second level. The discharge time was 3.5 min. The proportion of graphite and silver in the graphite-silver rod was 70% and 30%. The upper and lower electrodes of g g p g g p Figure 2. The EDM discharge process: (a) preparation for discharge, (b) discharge initiation, (c) ionization, (d) melting, (e) end of discharge Figure 2. The EDM discharge process: (a) preparation for discharge, (b) discharge initiation, (c) ionization, (d) melting, (e) end of discharge g 2.2 Dielectric fluid and material The dielectric fluid used are DW. The graphite electrode materials are presented in Figure 3. The upper and lower electrodes both used graphite rods, which were made from entire graphite of 99.9% purity; through lathe processing, the graphite was burned into a chunk 3 the EDM were both graphite-silver rods. Table 1. Environmental Parameter Settings for Preparing Graphene Colloid Title Parameter Electrode Graphite rod Ton:Toff 30:30 μs Current segment 2 Voltage 140 V Dielectric fluid DW Volume of the dielectric fluid 150 mL Discharge time 27 min Temperature 25 ℃ Atmospheric pressure 1 atm stability of graphene-Ag improved. In summmary, the zeta potential of graphene-Ag is greater than that of graphene, demonstrating that graphene-Ag had superior suspension stability to graphene, enabling its wider applicability. (a) (b) Figure 5. Zeta potential (a)graphene, (b)graphene-Ag (a) Table 2. Environmental Parameter Settings for Table 2. Environmental Parameter Settings for Preparing Graphene-Ag Colloid Title Parameter Electrode Graphite-silver rod Ton:Toff 30:30 μs Current segment 2 Voltage 140 V Dielectric fluid DW Volume of the dielectric fluid 150 mL Discharge time 3.5 min Temperature 25 ℃ Atmospheric pressure 1 atm (a) (b) (b) Figure 5. Zeta potential (a)graphene, (b)graphene-Ag 3.2 Characteristic analysis 3. Results and Discussion 3.1 Suspension stability UV–Vis is used to analyze the optical properties of graphene and graphene-Ag. The analysis results are presented in Figure 6. The figure shows that graphene had a clear absorption peak that corresponded to the wavelength at 262 nm, and its absorbance is 0.822. This is formed due to sp2 carbon atoms. This absorbance peak is typical for the UV–Vis spectra of graphene, verifying that graphene can indeed be obtained from EDM. The absorption peaks of graphene-Ag and graphene are very close. The wavelength of the absorption peak of graphene-Ag is also located at 262 nm, which is identical to the characteristic wavelength of graphene, and its absorbance is 0.803. As a result, when the light source of a wolfram lamp passes the sample, it cannot analyze the characteristic wavelength of AgNPs. In summary, we verified that using graphite rods and graphite-silver rods 3.1 Suspension stability The Zetasizer is used to analyze the suspension stability of graphene and graphene- Ag. The analysis results are presented as zeta potential. The suspension stability analysis results of graphene is presented in Figure 5(a). The analysis results in Figure 5(a) reveal that the zeta potential of graphene is −34.6 mV. The absolute zeta potential value is greater than 30 mV, indicating that this colloid had favorable suspension stability. The suspension stability analysis results of graphene-Ag is presented in Figure 5(b). Figure 5(b) reveals that the zeta potential of graphene-Ag is −40.6 mV, meaning that this colloid also had favorable suspension stability. Because AgNP had an excellent suspension ability, when AgNPs formed a composite with graphene layers, the suspension 4 can prepare graphene and graphene-Ag, respectively, through EDM. The preparation time is short, and no additional chemical agents are required. Figure 7. XRD patterns of the graphene and graphene-Ag Figure 6. UV–Vis spectra of the graphene and graphene-Ag Figure 7. XRD patterns of the graphene and graphene-Ag Raman is used to analyze the defects and SERS of graphene and graphene-Ag, and the analysis results are presented in Figure 8. The figure shows that the Raman spectra of graphene has two characteristic peaks (D-band and G- band). Their characteristic peaks correspond to the wavenumbers of 1353 cm-1 and 1577 cm-1. Using the D-band and G-band intensity ratio, we calculated the intensity ratio of graphene as 0.994 (5910/5944). The Raman spectrum of the graphene-Ag is special in that it has a very small characteristic peak D-band and a very large G- band, meaning that the graphene-Ag is highly graphitized and had few defects. The D-band and G-band corresponded to the wavenumbers of 1353 cm-1 and 1575 cm-1, respectively. The intensity ratio of graphene-Ag is 0.854 (10297/12050). According to the intensity ratio, we learned that graphene had more defects compared with graphene-Ag. This is because the electrodes used to prepare graphene-Ag are graphite-silver rods. Silver is a conductor and has favorable conductivity, so its discharge process is smooth and may peel more complete graphene-Ag to be dispersed in DW. From the figure, we could also determine the impact of SERS on graphene and graphene-Ag. The G- band Raman signal intensity of graphene-Ag substantially increased, verifying that using the physical method of the ESDM to prepare graphene-Ag in DW can increase the Raman signal intensity of graphene. Figure 6. 3.1 Suspension stability UV–Vis spectra of the graphene and graphene-Ag XRD is used to analyze the crystal structures of graphene and graphene-Ag. Figure 7 presents the analysis results of the diffraction peak and crystal structure. The figure reveals a diffraction peak of graphene at the position where 2θ is 26.3°. The crystal orientation is (002). From the figure 7, we observed another XRD pattern, which is the graphene-Ag analysis result. The figure 7 shows a clear diffraction peak, whose 2θ (degree) angle is located at 26.4°, and the corresponding crystal orientation is (002). Notably, in the graphene-Ag crystal, a diffraction peak of silver is not observed, meaning that the silver crystal is not detected. In summary, Figure 7 reveals that when using two different types of graphite rods and electrodes, through EDM we are able to prepare carbon- structure colloids with comprehensive crystals with an orientation of (002). Regarding the diffraction peak intensity, because they are two completely different materials and products, the intensities of the two diffraction peaks can be overlooked. 5 5 (a) (b) Figure 9. (a)TEM image of graphene, (b) magnified image of graphene Figure 8. Raman of the graphene and graphene- Ag (a) (b) (a) Figure 8. Raman of the graphene and graphene- Ag TEM is used to analyze the surface properties of graphene and graphene-Ag. Figure 9 presents the TEM image of graphene. Figure 9(a) illustrates clearly the stacking of graphene layers, and we can identify the differences represented by color. The peripheral had less graphene stacked together, so the color is a light gray. Thicker stacks of graphene layers are observed, which exhibited a black color. From the Figure 9 (b), one can see the lattice line of part of the graphene. This lattice line indicates the distance between the graphene layers. In this sample, after measurement, the inter-layer distance is 0.33 nm. Figure 10 presents the TEM image of graphene-Ag. In the Figure 10 (a), numerous round and dark small AgNPs are observed on the graphene layers. The particles are embedded in the graphene layers. The size distribution of AgNPs is presented in Figure 10 (b). This distribution graph is mainly from the AgNP size in the TEM image of Figure 10 (a). The size of small AgNPs is approximately 5 nm; they account for almost half of the AgNPs. 3.1 Suspension stability This means that when using graphite-silver rods to conduct ESDM processing, we could prepare AgNPs of a very small size embedded in graphene layers. (b) ( ) Figure 9. (a)TEM image of graphene, (b) magnified image of graphene Figure 9. (a)TEM image of graphene, (b) magnified image of graphene (a) (a) 6 increased the possibility of the future application of graphene-Ag to conductive matters, and also met our original goal of not using chemical methods but rather purely using EDM to prepare this composite material. (b) Figure 10. (a) TEM image of graphene-Ag, (b) particle size distribution of the AgNPs. (b) Table 3. Sheet Resistance Values (MΩ/sq) of Graphene and Graphene-Ag Table 3. Sheet Resistance Values (MΩ/sq) of Graphene and Graphene-Ag Area No. Graphene Graphene-Ag 1 0.04 0.002 2 0.076 0.002 3 0.029 0.001 4 0.013 0.0006 5 0.03 0.001 6 0.03 0.001 7 0.036 0.001 8 0.02 0.001 9 0.027 0.001 10 0.028 0.003 Average 0.0329 0.00136 (b) Figure 10. (a) TEM image of graphene-Ag, (b) particle size distribution of the AgNPs. 3.3 Conductivity Table 3 lists the conductivity test results of graphene and graphene-Ag. The table lists the sheet resistance of graphene and graphene-Ag measured at 10 different areas on the films. For graphene, the minimum, maximum, and average sheet resistance values are 0.013 MΩ/sq, 0.076 MΩ/sq, and 0.0329 MΩ/sq, respectively. For graphene-Ag, the minimum, maximum, and average sheet resistance values are 0.0006 MΩ/sq, 0.003 MΩ/sq, and 0.00136 MΩ/sq, respectively. The graphene colloid prepared using the ESDM with graphite rods as electrodes is prepared into coating samples. Its average sheet resistance value is 0.0329 MΩ/sq, verifying that under this preparation environment, graphene flakes that peeled in DW are conductive and had favorable dispersion. Regarding the graphene-Ag colloid prepared using graphite-silver as the electrode, its average sheet resistance is 0.00136 MΩ/sq, which is 1/24 that of graphene, indicating that graphene- Ag has superior conductivity and dispersion compared with graphene. In short, we confirmed that using EDM to conduct the ESDM and prepare graphene colloid samples as well as using the coating method to prepare coating samples obtained graphene products with conductivity. In addition, after mixing silver into graphite to prepare the graphene-Ag colloid, when AgNPs and graphene layers became a new composite, the conductivity of graphene-Ag is superior to graphene and rGO-Ag because Ag has excellent conductivity. This substantially Availability of data and materials The data used to support the findings of this study are included within the article. [4] Saeed, M., Hamed, A., Wang, Z., Shaygan, M., Neumaier, D., & Negra, R. (2018). Graphene integrated circuits: new prospects towards receiver realisation. Nanoscale, 10(1), 93-99. Funding The Ministry of Science and Technology (MOST 110-2221-E-027-042-). The Ministry of Science and Technology (MOST 110-2221-E-027-042-). [6] Huang, Y., Liang, J., & Chen, Y. (2012). An overview of the applications of graphene‐ based materials in supercapacitors. small, 8(12), 1805-1834. Acknowledgements The authors would like to thank the Precision Research and Analysis Center, National Taipei University of Technology for technical supporting this research. [7] Wang, Y., Chen, X., Zhong, Y., Zhu, F., and Loh, K. P., "Large area, continuous, few- layered graphene as anodes in organic photovoltaic devices," Applied Physics Letters, vol. 95, no. 6, 2009, 209. 4. Conclusions 4. This study proposed a method for preparing graphene and graphene-Ag by using EDM. The ESDM is a physical preparation method. Compared with chemical synthesis methods, this method is simple, fast, and environmentally friendly. The research results of this dissertation are as follows: 1. The suspension stability test of graphene, and graphene-Ag revealed that the absolute values of their zeta potential are all greater than 30 mV, indicating that the graphene products prepared using the ESDM all had favorable suspension stability. Favorable suspension stability means there is a high mutual repulsion force between graphene layers and nanoparticles. The layers and particles do not aggregate easily, so the colloid can temporarily retain the properties of the nanomaterial, which increases the feasibility of using it in the industry in the future. 2. Through the analysis of surface properties of graphene products, we clearly observed that the shape of graphene is in flakes, and the AgNPs in graphene-Ag are composited on the graphene layers, forming a new type of graphene composite with different properties. 3. Regarding the conductivity analysis, the 7 [2] Sidorov, A. N., Yazdanpanah, M. M., Jalilian, R., Ouseph, P. J., Cohn, R. W., and Sumanasekera, G. U., "Electrostatic deposition of graphene," Nanotechnology, vol. 18, no. 13, 2007, 135301. average sheet resistance values of graphene, and graphene-Ag are 0.0329 MΩ/sq and 0.00136 MΩ/sq, respectively. The average sheet resistance value of graphene-Ag is only 1/24 that of graphene, indicating that graphene-Ag had a conductivity far superior to that of graphene. [3] Han, S. J., Garcia, A. V., Oida, S., Jenkins, K. A., and Haensch, W., "Graphene radio frequency receiver integrated circuit," Nature communications, vol. 5, no. 1, 2014, pp. 1-6. Author Contributions Project administration, Kuo-Hsiung Tseng; resources, Kuo-Hsiung Tseng; supervision, Kuo-Hsiung Tseng; funding acquisition, Kuo- Hsiung Tseng; data curation, Chu-Ti Yeh and Hsueh-Chien Ku; formal analysis, Chu-Ti Yeh and Hsueh-Chien Ku; methodology, Der-Chi Tien and Leszek Stobinski; validation, Hsueh- Chien Ku and Der-Chi Tien; writing—original draft, Hsueh-Chien Ku; writing—review and editing, Chu-Ti Yeh. All authors have read and agreed to the published version of the manuscript. [8] Woo, Y. S. (2019). Transparent conductive electrodes based on graphene-related materials. Micromachines, 10(1), 13. [9] Longo, C., and De Paoli, M. A., "Dye- sensitized solar cells: a successful combination of materials," Journal of the Brazilian Chemical Society, vol. 14, no. 6, 2003, pp. 898-901. [10] Roy-Mayhew, J. D., & Aksay, I. A. (2014). Graphene materials and their use in dye- sensitized solar cells. Chemical reviews, 114(12), 6323-6348. Competing interests The authors declare that there is no conflict of interest regarding the publication of this paper. [5] Jost, K., Dion, G., and Gogotsi, Y., "Textile energy storage in perspective," Journal of Materials Chemistry A, vol. 2, no. 28, 2014, pp. 10776-10787. References [11] Matsumoto, T., Koizumi, T., Kawakami, Y., Okamoto, K., and Tomita, M., "Perfect blackbody radiation from a graphene nanostructure with application to high- temperature spectral emissivity measurements," Optics express, vol. 21, no. 25, 2013, pp. 30964-30974. [1] Wang, J. T., Qian, Y., Weng, H., Wang, E., and Chen, C., "Three-dimensional crystalline modification of graphene in all- sp2 hexagonal lattices with or without topological nodal lines," The Journal of Physical Chemistry Letters, vol. 10, no. 10, 2019, pp. 2515-2521. [12] Shahil, K. M., & Balandin, A. A. (2012). Thermal properties of graphene and 8 multilayer graphene: Applications in thermal interface materials. Solid state communications, 152(15), 1331-1340. [13] Ji, H., Sun, H., and Qu, X., "Antibacterial applications of graphene-based nanomaterials: recent achievements and challenges," Advanced drug delivery reviews, vol. 105, 2016, pp. 176-189. [14] Ma, J., Zhang, J., Xiong, Z., Yong, Y., and Zhao, X. S., "Preparation, characterization and antibacterial properties of silver- modified graphene oxide," Journal of Materials Chemistry, vol. 21, no. 10, 2011, pp. 3350-3352. pp [15] Balandin, A. A., Ghosh, S., Bao, W., Calizo, I., Teweldebrhan, D., Miao, F., and Lau, C. N., "Superior thermal conductivity of single-layer graphene," Nano letters, vol. 8, no. 3, 2008, pp. 902-907. [16] Tseng, K. H., Ku, H. C., Tien, D. C., and Stobinski, L., "Novel preparation of reduced graphene oxide–silver complex using an electrical spark discharge method," Nanomaterials, vol. 9, no. 7, 2019, 979. [17] Tseng, K. H., Chang, C. Y., Chung, M. Y., & Cheng, T. S., "Fabricating TiO2 nanocolloids by electric spark discharge method at normal temperature and pressure." Nanotechnology, vol. 28, no. 46, 2017, 465701. [18] Hockenberry., T. O., "The Role of the Dielectric Fluid in Electrical Discharge Machining," SAE Technical Paper, 1968. [19] Tseng, K. H., Ku, H. C., Tien, D. C., and Stobinski, L., "Parameter control and concentration analysis of graphene colloids prepared by electric spark discharge method," Nanotechnology Reviews, vol. 8, no. 1, 2019, pp. 201-209. [20] Tseng, K. H., Lin, Y. H., Tien, D. C., Ku, H. C., and Stobinski, L., "Stability analysis of platinum nanoparticles prepared by ESDM in deionised water," Micro & Nano Letters, vol. 13, no. 11, 2018, pp. 1545- 1549. 9 9
https://openalex.org/W4387622831	https://www.scielo.br/j/bjb/a/BtcXNr7bLnbDjhWGT3NKhTG/?lang=en&format=pdf	English	null	Exposure of insects and host plants to different concentrations of CO2 affects the performance of Mahanarva spectabilis (Hemiptera: Cercopidae) in successive insect generations	Brazilian Journal of Biology	2,023	cc-by	8,663	Resumo Avaliou-se o desempenho de três gerações sucessivas de Mahanarva spectabilis (Distant) (Hemiptera: Cercopidae) alimentadas com quatro forrageiras expostas à ambientes com diferentes concentrações de CO2. No primeiro bioensaio, utilizamos os seguintes cenários: A) plantas e insetos foram mantidos em fitotron com CO2 alto e constante (média de 700 ppm) e, B) os insetos foram mantidos em nível de 700 ppm de CO2 e alimentados com plantas advindas da casa de vegetação (média de 390 ppm). No segundo bioensaio, utilizamos os seguintes cenários: C) plantas e insetos foram mantidos em casa de vegetação e, D) os insetos foram mantidos em casa de vegetação e alimentados com plantas mantidas em fitotron. A sobrevivência e duração das fases ninfal e adulta e o número de ovos/fêmea de M. spectabilis foram avaliados. Só foi possível avaliar os efeitos cumulativos do aumento do CO2 em três gerações sucessivas de M. spectabilis mantidas em casa de vegetação, devido à reduzida sobrevivência dos insetos na primeira geração quando mantidos em fitotron. Observou-se maior efeito direto do que indireto do nível de CO2 sobre o desempenho de M. spectabilis. Além disso, deve-se considerar que o efeito da elevação de CO2 sobre a sobrevivência, períodos de desenvolvimento e fecundidade, quando tomados em conjunto, podem impactar significativamente a dinâmica populacional de M. spectabilis em cenários climáticos futuros. Palavras-chave: cigarrinhas, forragens, alterações climáticas. THE INTERNATIONAL JOURNAL ON NEOTROPICAL BIOLOGY THE INTERNATIONAL JOURNAL ON NEOTROPICAL BIOLOGY ISSN 1519-6984 (Print) ISSN 1678-4375 (Online) Original Article Exposure of insects and host plants to different concentrations of CO2 affects the performance of Mahanarva spectabilis (Hemiptera: Cercopidae) in successive insect generations A exposição de insetos e plantas hospedeiras às diferentes concentrações de CO2 afeta o desempenho de Mahanarva spectabilis (Hemiptera: Cercopidae) em sucessivas gerações do inseto M. G. Fonsecaa , A. M. Auada* , T. T. Resendea , B. A. Veríssimob  and C. M. Oliveirac  aEmbrapa Gado de Leite, Laboratório de Entomologia, Juiz de Fora, MG, Brasil bUniversidade Federal de Juiz de Fora, Departamento de Biodiversidade e Conservação da Natureza, Juiz de Fora, MG, Brasil cEmbrapa Cerrados, Brasília, DF, Brasil M. G. Fonsecaa , A. M. Auada* , T. T. Resendea , B. A. Veríssimob  and C. M. Oliveirac  aEmbrapa Gado de Leite, Laboratório de Entomologia, Juiz de Fora, MG, Brasil bUniversidade Federal de Juiz de Fora, Departamento de Biodiversidade e Conservação da Natureza, Juiz de Fora, MG, Brasil cEmbrapa Cerrados, Brasília, DF, Brasil Abstract The performance of three successive generations of Mahanarva spectabilis (Distant) (Hemiptera: Cercopidae) fed on four forages exposed to environments with different CO2 concentrations was evaluated. In the first bioassay, we utilized the following scenarios: A) plants and insects were kept at high and constant CO2 (700 ppm) and B) the insects were kept at CO2 700 ppm and fed on plants from the greenhouse (average of 390 ppm). In the second bioassay, we utilized the following scenarios: C) plants and insects were kept in a greenhouse and D) the insects were kept in the greenhouse and fed on plants kept at CO2 700 ppm. The survival and duration of the nymphal and adult stages and the number of eggs/female of M. spectabilis were evaluated. It was only possible to evaluate the cumulative effects of the increase of CO2 on three successive generations of M. spectabilis kept in a greenhouse, due to the reduced survival of the insects in the first generation in the laboratory. A greater direct than indirect effect of the CO2 level on the performance of M. spectabilis was observed. Furthermore, it should be considered that the effect of CO2 elevation on the survival, periods of development, and fecundity, when taken together, can significantly impact the population dynamics of M. spectabilis in future climate scenarios. Keywords: spittlebugs, forages, climate change. e-mail: alexander.auad@embrapa.br Received: March 30, 2023 – Accepted: August 9, 2023 This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. p g 1/10 B ili J l f Bi l 2023 l 83 273470 \| htt //d i /10 1590/1519 6984 273470 This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. e-mail: alexander.auad@embrapa.br Received: March 30, 2023 – Accepted: August 9, 2023 2.2. Plants and growth conditions Two species of brachiaria were tested; where one was susceptible to spittlebugs, Brachiaria ruziziensis, and the other was resistant to B. brizantha (Cardona et al., 2004). In addition to elephant grass (P. purpureum), Pioneiro and Roxo de Botucatu cultivars were characterized as resistant and susceptible to M. spectabilis, respectively (Auad et al., 2007). ) It is worth mentioning that insect-plant interaction can be influenced by the increase in CO2 concentration due to changes in the nutritional and physiological quality of the host plants (Dáder et al., 2016; Satishchandra et al., 2018) and by the direct effects of elevated CO2 levels on the fitness performance of the insects (Xie et al., 2015). The increase in CO2 increases the plants’ carbon/ nitrogen balance, which affects insect feeding (Coviella and Trumble 1999). The abundance of carbon and the scarcity of nitrogen lead to other changes in a plant that can further affect the insects. A group of carbon- based compounds (such as phenolic compounds and tannins) tend to slow insect growth. Thus, high levels of CO2 allow many plant species to significantly increase their defenses based on the carbon availability (Trumble and Butler, 2009). Such alterations can modify the pattern of resistant plants against insect pests (Karthik et al., 2021). In this regard, little is known about the intrinsic mechanisms between plant resistance and insect survival strategies under future climate scenarios (Xie et al., 2022). Therefore, considering the use of resistant plants as the main control strategy for spittlebugs, it is essential to know the interference of climate change on the control strategies of this insect pest, to avoid any compromise on the sustainability of the production system. Therefore, based on the future climate scenarios in which CO2 elevation is expected, the objective was to evaluate the direct and indirect effects of the CO2 elevation on the fitness performance of multiple generations of M. spectabilis fed on four different forages. Brachiaria was cultivated from seeds in Styrofoam trays with 140 cells, and on the 15th day after cultivation, it was transplanted in tubes. On the 50th day after sowing, the plants were transplanted in plastic pots with a capacity of 1 kg. Elephant grass was grown by using stakes of 10 cm (with one node) and planted directly in plastic pots. 2.1. Insects 2.1. Insects 2.1. Insects Brazil has privileged conditions for the cultivation of forage grasses. It’s wide territorial extension occupied by the pastures reflects its high productive potential (Rabinovitch et al., 1998). However, the spittlebug, Mahanarva spectabilis (Distant) (Hemiptera: Cercopidae), reduces the growth of these grasses and thereby causes a drop in the production of dry mass as well as the quality of the forage, which amounts to a major problem within the cattle industry throughout tropical America (Auad et al., 2007; Souza et al., 2008; Ferreira et al., 2013; Resende et al., 2014; Fonseca et al., 2016; Alvarenga et al., 2017; Silva et al., 2017). Given the impact of this species on grasses, it is vital to understand how global changes may influence its performance and thus assess whether the increase in CO2 levels may lead to a better adaptation of this pest, triggering the need to search for new control strategies. Adults of the M. spectabilis were collected in the experimental field of Embrapa Gado de Leite in Coronel Pacheco-MG. They were then taken to the entomology laboratory in Juiz de Fora-MG, sexed, and kept in acrylic cages (30×30×60 cm). One Pennisetum purpureum plant was cultivated in a pot (1L) with a mixture of soil, sand and organic fertilizer (3:1:1), and was placed in the cage. The bottom of the pot was wrapped in gauze and moistened with distilled water, which served as a substrate for oviposition. To remove the eggs retained in the substrate, the gauze was placed on a set of sieves and subjected to running water, where the eggs were retained in the thinnest sieve (400 mesh opening). The obtained eggs were placed in Petri dishes 5 cm in diameter and lined with filter paper. They were kept in a climatic chamber (25 ± 2 °C, 14 hours of photophase, and relative humidity of 70 ± 10%) until they became close to hatching, i.e., when they showed the stage of advanced embryo (Stage 4- S4), with expected hatching in two days. At that stage, the eggs were used in the experiments. Most studies published on the response of insects that feed on plants grown under high concentrations of CO2 are based on short-term experiments, recording the rates of development in a single generation of an insect (Bezemer and Jones, 1998). 2.1. Insects Experiments carried out over several generations can reveal differences in responses between generations (Brooks and Whittaker, 1998). Few studies have addressed the impact of high CO2 on multiple generations of herbivorous insects (Brooks and Whittaker, 1999). 2.2. Plants and growth conditions The pots of Brachiaria and elephant grass were kept in environments with different levels of CO2 as follows: 1) greenhouse (average 390 ppm; minimum of 50 ppm and a maximum of 2499 ppm, i.e., the current level CO2) and 2) climatized chamber with constant CO2 level (700 ppm, which is the level expected for the year 2100 by International Panel on Climate Change, IPCC). Each climatized chamber (2.5 x 2.20 x 2.80 m) was maintained at 25 ± 2ºC, 70 ± 10 relative Humidity (RH), and 14:10 (light: dark). In the greenhouse, averages of 27 ± 2ºC and 73 ± 10 RH were recorded. Conditions in the chamber with constant CO2 were maintained with an automatic control system, COEL HW 4200 (Manaus-Amazonas, Brazil) and an injection system, using a CO2 cylinder to maintain the desired CO2 concentration. The CO2 levels in the floating chamber and greenhouse were recorded with a data logger every 2 minutes, and the data were transferred to the computer using HOBO ware software (Onset Co., Pocasset, Massachusetts, USA). 1. Introduction 800 ppm by 2100 (IPCC, 2014). This change in climate can significantly alter the relationships between plants and insects in important agricultural production areas (Karthik et al., 2021; Skendžić et al., 2021; Theurillat and The atmospheric concentration of carbon dioxide (CO2) has steadily increased from approximately 280 ppm before the industrial revolution (Qian et al., 2021) to 418 ppm in April 2022 (NOAA, 2022). It is predicted to rise to around 1/10 Fonseca, M.G. et al. 2. Material and Methods Guisan, 2001) and influence the geographic distribution of insect as example, the eucalyptus yellow beetle (Vasconcelos et al., 2022). Brazilian Journal of Biology, 2023, vol. 83, e273470 2.3. Scenarios used in the experiments The experiments were divided into two bioassays. The first bioassay was carried out in climatized chambers, 2/10 Brazilian Journal of Biology, 2023, vol. 83, e273470 Concentrations of CO2 on the performance of spitllebug fed on cultivars Pioneiro and Roxo de Botucatu in Scenario C and Roxo de Botucatu in Scenario D (Table 1). In these treatments, the survival and duration of the nymphal phase, number of eggs/female, and longevity of the adults of M. spectabilis in the three generations were evaluated daily. phytotron type, in the laboratory under two scenarios: A) plants and insects were maintained at high and constant CO2 (700 ppm) and B) the insects were kept at high and constant CO2 (700 ppm) and fed on plants from the greenhouse (average of 390 ppm CO2). The second bioassay was carried out in a greenhouse under two scenarios: C) plants and insects were kept in a greenhouse (average of 390 ppm CO2), and D) the insects were kept in a greenhouse (average of 390 ppm CO2) and fed on plants that came from the climatized chamber with high and constant CO2 (700 ppm). To evaluate the number of eggs and the longevity of M. spectabilis from each treatment, the adults were sexed and separated by pairs. The pairs were individualized, kept in the shoot of the plant of each treatment (scenario), and covered with a cage (45 cm diameter x 40 cm long) made of “voile”- type fabric (100% polyester). The base of the plant was wrapped with gauze moistened with distilled water, which served as a substrate for oviposition. The plants were exchanged when they showed yellowing of the leaves. 2.4. Bioassays In the first generation, 160 plants, at 40 days, were submitted to water jets, aiming to expose the roots for the fixation of spittlebug nymphs. Thirty eggs were placed per pot, which were individually covered by a cage (45 cm diameter × 40 cm long) made of “voile”- type fabric (100% polyester) to prevent the insects from escaping. The assays were carried out using four types of forage (two species of Brachiaria and two cultivars of elephant grass), two scenarios (bioassay 1 and bioassay 2), and three successive generations, with ten replications, totaling 4,800 eggs in state S4 in the first generation. In subsequent generations, the number of repetitions varied as a function of the number of eggs produced by the females. All data analysis were performed with the R program (R Core Team, 2022). The obtained data did not meet the normality assumptions related to the residues and the homogeneity of the variance (Shapiro-Wilk p <0.0001). Therefore, nonparametric analysis was used to verify the differences between the treatments. A Kruskal-Wallis (Hollander and Wolfe, 1973) analysis was used, and a comparison among forages was performed by Nemenyi post hoc comparison with a Tukey correction using the ‘kwAllPairsNemenyiTest’ function in the R package ‘PMCMRplus’ (Pohlert 2014). The Wilcoxon-Mann-Whitney U-test (Conover 1980) was used for comparison between scenarios, generations, and forages, using the “Wilcox.text” function in the R package ‘PMCMRplus’ (Pohlert 2014). Due to the low insect survival in the bioassay carried out in the climatic chamber (bioassay 1), in which forages and insects were submitted to Scenarios A and B, it was only possible to analyze the nymphal survival in the first generation in these scenarios (Table 1). In Scenarios C and D, bioassay 2, when insects were always kept in a greenhouse, it was possible to analyze the biological parameters, survival and nymphal duration, longevity of adults, and the number of eggs/female in the first generation of M. spectabilis, fed on the four forages. The comparison of the pests’ performance between the three generations was possible when they were 3. Results nal of Biology, 2023, vol. 83, e273470 3/1 x = adequate number of individuals for analysis; 0 = Insufficient number of individuals for analysis. 3.2. Performance of M. spectabilis kept in a greenhouse, subjected to two food sources and two scenarios in three generations No significant difference was observed in the nymphal survival of M. spectabilis when the insects were kept in the greenhouse and fed on B. ruziziensis (W=69.5, P=0.14), B. brizantha (W=52, P=0.91), and Pioneiro (W=58.5, P=0.54), regardless of the scenarios in which the plants were kept. On the other hand, Roxo de Botucatu showed greater survival (W=18, P=0.017) of the nymphs in the Scenario D (where the insects were kept in the greenhouse and fed on plants from the climatic chamber with constant CO2 at 700 ppm) (Figure 2). When the four forages are compared within each scenario, the nymphal survival of those kept and fed in the greenhouse showed no difference (x2=5.51; df=3; P=0.137) among the forages. However, greater nymphal survival (x2=18.017; df=3; P=0.0004) was observed for the insects which were fed on Roxo de Botucatu, maintained at constant and high CO2 (700ppm), compared with B. ruziziensis (Scenario D) (Figure 2). No difference was observed in the number of eggs when the females were fed on B. ruziziensis, B. brizantha, regardless of the scenario; however, females laid fewer eggs when they were fed on Roxo de Botucatu (W=2357, P<0.039) and Pioneiro (W=612, P<0.0003) that was grown at constant CO2 and kept with the insects in the greenhouse (Scenario D) (Figure 2). Furthermore, females laid a significantly higher number of eggs (P<0.01, F=3.858) when kept in Roxo de Botucatu compared with B. ruziziensis (Scenario C) (x2=11.65; df=3; P=0.042) and Pioneiro (Scenario D) (x2=15.45; df=3; P=0.0014). (Figure 2). In Scenarios C and D, evaluating the elephant grass cultivars was possible, which provided a sufficient number of individuals for analysis in subsequent generations (Figure 3). No significant difference was found in the nymphal survival of M. spectabilis between the three generations when the insects and forages Pioneiro (x2=1.696; df=2; P=0.43) and Roxo de Botucatu (x2=3.47;df=2; P=0.176) were kept indoors in the greenhouse under current CO2 conditions (Scenario C). The same occurred between the forages in generations 1 and 2. However, lower nymphal survival was found (W=131, P<0.0019) in Roxo de Botucatu compared to Pioneiro in generation 3 (Figure 3). Figure 1. Biological aspects of M. 3. Results 3.1. Performance of M. spectabilis maintained in a phytotron and subjected to four food sources in two scenarios 3.1. Performance of M. spectabilis maintained in a phytotron and subjected to four food sources in two scenarios Nymphal survival of M. spectabilis that were always kept in the future climate scenario (700ppm CO2) showed Table 1. Evaluation of the number of individuals sufficient for statistical analysis when Mahanarva spectabilis was fed on four forages and its three generations were subjected to four scenarios. Generation Scenarios Forages B. ruziziensis B. Brizantha Cv. Pioneiro Cv. Roxo de Botucatu G1 A x x x x G2 0 0 0 0 G3 0 0 0 0 G1 B x x x x G2 0 0 0 0 G3 0 0 0 0 G1 C x x x x G2 0 0 x x G3 0 0 x x G1 D x x x x G2 0 0 0 x G3 0 0 0 x x = adequate number of individuals for analysis; 0 = Insufficient number of individuals for analysis. Table 1. Evaluation of the number of individuals sufficient for statistical analysis when Mahanarva spectabilis was fed on four forages and its three generations were subjected to four scenarios. Brazilian Journal of Biology, 2023, vol. 83, e273470 3/10 Fonseca, M.G. et al. The duration of the nymphal phase of M. spectabilis showed no difference when fed on cultivars Pioneiro (W=3069, P=0.85) and Roxo de Botucatu (W=8765, P=0.49), regardless of the scenario. However, shorter durations of the pest insect were found when they were fed on B. ruziziensis (W=654, P<0.001) and B. brizantha (W=3764, P=0.003) from climatic chambers with constant CO2 and kept in the greenhouse (Scenario D) (Figure 2). When the four forages within each scenario are compared, the shortest durations (x2=25.04; df=3; P<0.001) are found for those that were fed on Pioneiro and Roxo de Botucatu kept in the greenhouse (Scenario C), and when fed on B. ruziziensis, which was maintained at high CO2 condition (Scenario D) (x2=9.26; df=3; P=0.026) (Figure 2). no significant difference when they were fed on four forages regardless of the scenarios in which the plants were kept. (Figure 1). 3. Results Performance of M. spectabilis kept in a greenhouse, subjected to two food sources and two scenarios in three generations Brazilian Journal of Biology, 2023, vol. 83, e273470 3. Results When the nymphal survival is compared among the four foragers within each scenario, it is found that in Scenarios A (x2=12.95; df=3; P=0.004) and B (x2=16.66; df=3; P=0.0008), greater nymphal survival is observed in the Roxo de Botucatu cultivar compared with B. brizantha and Pioneiro forages (Scenario B) and B. ruziziensis (Scenario A) (Figure 1). In the laboratory (bioassay 1), where insects were always kept at higher CO2 level, the evaluation of the cumulative effect of the increase in CO2 concentration on three successive generations of M. spectabilis in the scenarios in which the insects were fed on B. ruziziensis and B. brizantha was not possible. The low nymphal survival and low fertility of the females of the first generation made it impossible to assess the next generations. The longevity of M. spectabilis adults was observed to be significantly low for the insects which were fed on B. ruziziensis (W=527, P=0.0010), B. brizantha (W=3953, P<0.001), Pioneiro (W=3898, P=0.001), and Roxo de Botucatu (W=9712, P<0.001) from the climatized chambers with constant CO2 (Scenario D). In Scenario C (x2=9.90; df=3; P<0.019), the longevity of the adults came greater when they were fed on Roxo de Botucatu compared with B. ruziziensis, and in Scenario D (x2=27.74; df=3; P<0.001), when fed on B. brizantha and Roxo de Botucatu (Figure 2). The longevity of M. spectabilis adults was observed to be significantly low for the insects which were fed on B. ruziziensis (W=527, P=0.0010), B. brizantha (W=3953, P<0.001), Pioneiro (W=3898, P=0.001), and Roxo de Botucatu (W=9712, P<0.001) from the climatized chambers with constant CO2 (Scenario D). In Scenario C (x2=9.90; df=3; P<0.019), the longevity of the adults came greater when they were fed on Roxo de Botucatu compared with B. ruziziensis, and in Scenario D (x2=27.74; df=3; P<0.001), when fed on B. brizantha and Roxo de Botucatu (Figure 2). No difference was observed in the number of eggs when the females were fed on B. ruziziensis, B. brizantha, regardless of the scenario; however, females laid fewer eggs when they were fed on Roxo de Botucatu (W=2357, P<0.039) and Pioneiro (W=612, P<0.0003) that was grown at constant CO2 and kept with the insects in the greenhouse (Scenario D) (Figure 2). Furthermore, females laid a significantly higher number of eggs (P<0.01, F=3.858) when kept in Roxo de Botucatu compared with B. ruziziensis (Scenario C) (x2=11.65; df=3; P=0.042) and Pioneiro (Scenario D) (x2=15.45; df=3; P=0.0014). (Figure 2). 3.2. 3.2. Performance of M. spectabilis kept in a greenhouse, subjected to two food sources and two scenarios in three generations spectabilis, first generation, fed on different forages in the following scenarios: (A) plants and insects were maintained at a high and constant CO2 (700 ppm) and B) the insects were kept at a high and constant CO2 (700 ppm) and fed on plants from the greenhouse (average of 390 ppm CO2). Means followed by the same lowercase letter between the forages within each scenario do not differ from each other, by Kruskal-Wallis or Nemenyi post hoc test (P<0.01). Means compared between scenarios within each forage did not differ from each other. No difference was found in the nymphal duration of the insects fed on forages Pioneiro or Roxo de Botucatu within the first and second generations, whereas it was found longer in the third generation that was fed on Roxo de Botucatu (W=2942, P=0.014). A significantly lower nymphal duration was observed when the insects were fed on both Pioneiro (x2=24.62; df=2; P<0.0001) and Roxo de Botucatu (x2=44.69; df=2; P<0.0001) in generation 1, compared to the other generations (Figure 3). Significantly lower adult longevity was observed for the insects in the first generation that were fed on both Pioneiro (x2=44.69; df=2; P<0.0001) and Roxo de Botucatu (x2=37.07; df=2; P<0.0001) when compared to the subsequent generations. Figure 1. Biological aspects of M. spectabilis, first generation, fed on different forages in the following scenarios: (A) plants and insects were maintained at a high and constant CO2 (700 ppm) and B) the insects were kept at a high and constant CO2 (700 ppm) and fed on plants from the greenhouse (average of 390 ppm CO2). Means followed by the same lowercase letter between the forages within each scenario do not differ from each other, by Kruskal-Wallis or Nemenyi post hoc test (P<0.01). Means compared between scenarios within each forage did not differ from each other. Brazilian Journal of Biology, 2023, vol. 83, e273470 4/10 Concentrations of CO2 on the performance of spitllebug Concentrations of CO2 on the performance of spitllebug Figure 2. Biological aspects of the first generation of M. spectabilis fed on different forages in the following scenarios: (C) plants and insects were kept in a greenhouse (average of 390 ppm CO2) and (D) the insects were kept in a greenhouse (average of 390 ppm CO2) and fed on plants that came from the climatized chamber with high and constant CO2 (700 ppm). 3.2. Performance of M. spectabilis kept in a greenhouse, subjected to two food sources and two scenarios in three generations Means followed by the same capital letter are compared between scenarios within each forage, and those followed by the same lowercase letter between the forages within each scenario do not differ from each other by Wilcoxon-Mann-Whitney U test (P<0.01) or Kruskal-Wallis, Nemenyi post hoc test (P<0.01), respectively. Figure 2. Biological aspects of the first generation of M. spectabilis fed on different forages in the following scenarios: (C) plants and insects were kept in a greenhouse (average of 390 ppm CO2) and (D) the insects were kept in a greenhouse (average of 390 ppm CO2) and fed on plants that came from the climatized chamber with high and constant CO2 (700 ppm). Means followed by the same capital letter are compared between scenarios within each forage, and those followed by the same lowercase letter between the forages within each scenario do not differ from each other by Wilcoxon-Mann-Whitney U test (P<0.01) or Kruskal-Wallis, Nemenyi post hoc test (P<0.01), respectively. (x2=12.79; df=2; P=0.001662) in the first generation. In this same generation, greater nymphal survival was observed (W=18, P=0.017) in the respective scenario (Figure 4). Among the forages, Roxo de Botucatu increased the longevity of the adults of the insect pest in the three generations (Figure 3). The duration of the nymphal phase was observed to be shorter in generation 1 when the insects were fed on Roxo de Botucatu and maintained in the C (x2=44.69; df=2; P<0.0001) and D (x2=118.83; df=2; P<0.0001) scenarios compared to the other generations. This forage maintained in Scenario D reduced (W=7213, P=0.049) the duration of the nymphs of the second generation; however, there was no difference in this parameter in the other generations (Figure 4). No difference was observed in the nymphal duration (x2=5.89; df=2; P<0.052) in the three generations when the insects were fed on Pioneiro; however, it became considerably higher in generation 3, when insects were fed on Roxo de Botucatu (x2=15.05; df=2; P<0.0001) compared with the first generation. Within each generation, the number of eggs was observed to be significantly higher when kept in the Roxo de Botucatu cultivar compared to the Pioneiro in generations 1 (W=565, P=0.046) and 3 (W=252, P<0.0001) (Figure 3). In both scenarios, adult longevity was lower in generation 1 when insects were fed on Roxo de Botucatu. Brazilian Journal of Biology, 2023, vol. 83, e273470 3.2. Performance of M. spectabilis kept in a greenhouse, subjected to two food sources and two scenarios in three generations In generations 1 (W=9712, P<0.0001) and 2 (W=7504, P=0.0003), the longevity of the adults was found to be significantly higher when they were fed on Roxo de Botucatu and kept in Scenario C. This difference was not observed (W=2779, P=0.059) in generation 3 (Figure 4). Significant differences were observed in the number of eggs in generation 2 (W=2091, P=0.0019) and 3 (W=809, P<0.0001) since the females laid more eggs when the insects were fed on Roxo de Botucatu and kept in the greenhouse (Scenario C). Between the generations, only In both scenarios, adult longevity was lower in generation 1 when insects were fed on Roxo de Botucatu. In generations 1 (W=9712, P<0.0001) and 2 (W=7504, P=0.0003), the longevity of the adults was found to be significantly higher when they were fed on Roxo de Botucatu and kept in Scenario C. This difference was not observed (W=2779, P=0.059) in generation 3 (Figure 4). Comparing these two factors, generations and scenarios (Figure 4), was only possible when the cultivar Roxo de Botucatu was offered due to the reduced survival of the spittlebugs in the Pioneiro forage in Scenario D. As can be seen in Figure 4, no difference was observed in the nymphal survival in three generations (x2=3.52; df=2; P<0.172) when insects were fed on Roxo de Botucatu, that was grown and kept in the greenhouse (Scenario C); however, when fed the same forage but grown at high and constant CO2 and kept in the greenhouse (Scenario D), significantly higher nymphal survival was obtained Significant differences were observed in the number of eggs in generation 2 (W=2091, P=0.0019) and 3 (W=809, P<0.0001) since the females laid more eggs when the insects were fed on Roxo de Botucatu and kept in the greenhouse (Scenario C). Between the generations, only 5/10 Fonseca, M.G. et al. Figure 3. Biological aspects of M. spectabilis in three consecutive generations fed on Pioneiro and Roxo de Botucatu cultivars in the following scenario: (C) plants and insects were kept in a greenhouse (average of 390 ppm CO2). Means followed by the same capital letter are compared between scenarios within each forage, and those by the same lowercase letter between the forages within each scenario do not differ from each other by Wilcoxon-Man-Whitney U test (P<0.01) and Kruskal-Wallis, Nemenyi post hoc test (P<0.01), respectively. Figure 3. Biological aspects of M. 3.2. Performance of M. spectabilis kept in a greenhouse, subjected to two food sources and two scenarios in three generations spectabilis in three consecutive generations fed on Pioneiro and Roxo de Botucatu cultivars in the following scenario: (C) plants and insects were kept in a greenhouse (average of 390 ppm CO2). Means followed by the same capital letter are compared between scenarios within each forage, and those by the same lowercase letter between the forages within each scenario do not differ from each other by Wilcoxon-Man-Whitney U test (P<0.01) and Kruskal-Wallis, Nemenyi post hoc test (P<0.01), respectively. changes in the number of eggs were observed in Scenario C (x2=15.069; df=2; P<0.0005), which was higher in generation 3 (Figure 4). of eggs/female) by making the number of repetitions insufficient for statistical analysis in the first, second, and third generations (Table 1). Ferreira et al. (2013) also recorded low survival of this insect pest in the first generation, which also varied depending on the scenarios to which the insects and plants were subjected and on the forage species offered, ranging from 2% (B. brizantha) to 65% (Roxo de Botucatu). 4. Discussion Climate change, with rising CO2 and temperature, can affect the pasture production system through chemical and biological changes in terms of the interaction between the forage and pasture spittlebugs. The defense system of forage plants can influence the survival, fecundity, and development of phytophagous insects and, consequently, their distribution and abundance in the environment (Gould, 1998). The investigation of this process would support control strategies, in which the primary purpose is to find out the best interaction between the organisms and the search for proximity to environmental balance, ensuring low production costs, considering that, according to Dias-Filho (2014), forages are the most practical and economical means of feeding cattle. In the two Scenarios A and B, in which the insects were always kept under controlled conditions, it was possible to verify that the forages provided the same nymphal survival of the insect pest. This result suggests that the quality of forage under high levels of CO2 did not influence the development of M. spectabilis. However, in general, an alteration in the performance of M. spectabilis was observed when they were kept in the greenhouse in Scenario D compared to Scenario C (Figure 2). It should be taken into account that the only difference between these scenarios is the place where the four forages were submitted. When they were kept in a scenario with high and constant concentration of CO2, the nymphal duration, adult longevity, and the number of eggs/female reduced in one or more forages. This proves the indirect effect of plants kept at different CO2 levels on the performance of the insect pest kept in the current climate scenario (greenhouse with 390 ppm CO2). Insect survival is an important factor in assessing insect-plant interaction. The low nymphal survival of M. spectabilis in Scenarios A and B used in the present study prevented the analysis of some biological parameters of this insect pest (duration, longevity and the average number Brazilian Journal of Biology, 2023, vol. 83, e273470 6/10 Concentrations of CO2 on the performance of spitllebug Concentrations of CO2 on the performance of spitllebug Figure 4. Biological aspects of M. 4. Discussion However, elevated levels of CO2 also affect insects directly (Xie et al., 2015), which can be evidenced by the low survival of M. spectabilis nymphs when they were kept always under high CO2 level (Figure 1). It is noteworthy that of the four forages tested, Roxo de Botucatu proved to be the best food for developing M. spectabilis in the future (with an expected increase in CO2). Similar results were reported by Ferreira et al. (2013). They found no significant difference in the nymphal survival of M. spectabilis when the insects were fed on B. decumbens (susceptible) and B. brizantha (resistant) and kept in environments with different concentrations of CO2. On the other hand, elephant grass cultivars that are susceptible (Roxo de Botucatu) and resistant (Pioneiro) to M. spectabilis resulted in a significantly higher percentage of nymphal survival when maintained at 500 ppm than in environments with CO2 at 250 ppm, stating that the susceptibility of the Roxo de Botucatu increased under high CO2 levels. This information is vital, since the plant-resistance strategy has been the most indicated strategy for reducing the commitment of the spittlebugs in the productive systems of pastures (Auad et al., 2007; Souza Sobrinho et al., 2010; Resende et al., 2014; Alvarenga et al., 2017). g ) Still, in pastures, Silva et al. (2019) recorded a direct and indirect effect of the increase in CO2, which promoted longer duration and lower survival of the Collaria oleosa (Distant) (Hemiptera: Miridae) bug, which led to recording the maintenance of the resistance of the forage species B. brizantha in the future climate scenario. In addition, Auad et al. (2012) concluded that the population of the aphid Sipha flava (Forbes) (Hemiptera: Aphidae), which is also found in forage, would decrease significantly in future climatic conditions, in which the concentration of the CO2 level and temperature is expected to increase. Thus, despite research suggesting that global climate change may disrupt insect resistance in specific crops, with profound food security implications (Guo et al., 2014, 2017), resistance patterns will remain the same for the primary insect pest species in pastures (pasture spittlebugs, stink bug, and aphid). However, it should be taken into account that the effects of CO2 elevation on the survival, developmental periods, and fecundity, when It is noteworthy that of the four forages tested, Roxo de Botucatu proved to be the best food for developing M. Brazilian Journal of Biology, 2023, vol. 83, e273470 4. Discussion spectabilis in three consecutive generations fed on different forages in the following scenarios: (C) the Roxo de Botucatu cultivar and insects were kept in a greenhouse (average of 390 ppm CO2) and (D) the insects were kept in a greenhouse (average of 390 ppm CO2) and fed on the Roxo de Botucatu cultivar that came from a climatized chamber with high and constant CO2 (700 ppm). Means followed by the same capital letter are compared between scenarios within each forage, and those followed by the same lowercase letter between the forages within each scenario do not differ from each other by Wilcoxon-Man Whitney U test (P<0.01) and Kruskal-Wallis, Nemenyi post hoc test (P<0.01), respectively. Figure 4. Biological aspects of M. spectabilis in three consecutive generations fed on different forages in the following scenarios: (C) the Roxo de Botucatu cultivar and insects were kept in a greenhouse (average of 390 ppm CO2) and (D) the insects were kept in a greenhouse (average of 390 ppm CO2) and fed on the Roxo de Botucatu cultivar that came from a climatized chamber with high and constant CO2 (700 ppm). Means followed by the same capital letter are compared between scenarios within each forage, and those followed by the same lowercase letter between the forages within each scenario do not differ from each other by Wilcoxon-Man Whitney U test (P<0.01) and Kruskal-Wallis, Nemenyi post hoc test (P<0.01), respectively. According to Zavala et al. (2013) and Hamann et al. (2021), the effects on herbivorous insects caused by the climate change occur due to changes in the host plants. Yin et al. (2010) emphasized that under high CO2 concentrations, the lower nutritional quality of plants results in higher insect mortality, reducing the extent of their damage to plants. However, elevated levels of CO2 also affect insects directly (Xie et al., 2015), which can be evidenced by the low survival of M. spectabilis nymphs when they were kept always under high CO2 level (Figure 1). According to Zavala et al. (2013) and Hamann et al. (2021), the effects on herbivorous insects caused by the climate change occur due to changes in the host plants. Yin et al. (2010) emphasized that under high CO2 concentrations, the lower nutritional quality of plants results in higher insect mortality, reducing the extent of their damage to plants. 4. Discussion spectabilis in the future (with an expected increase in CO2). Similar results were reported by Ferreira et al. (2013). They found no significant difference in the nymphal survival of M. spectabilis when the insects were fed on B. decumbens (susceptible) and B. brizantha (resistant) and kept in environments with different concentrations of CO2. On the other hand, elephant grass cultivars that are susceptible (Roxo de Botucatu) and resistant (Pioneiro) to M. spectabilis resulted in a significantly higher percentage of nymphal survival when maintained at 500 ppm than in environments with CO2 at 250 ppm, stating that the susceptibility of the Roxo de Botucatu increased under high CO2 levels. This 7/10 Fonseca, M.G. et al. taken together, can significantly impact the population dynamics of M. spectabilis. high CO2, and it was 10% higher in the fourth generation compared to the first generation. Lower survival and fecundity were also reported by Li et al. (2022) for L. erysimi kept under 800 µL/L CO2 compared to 400 µL/L CO2 in the third generation of this insect pest. Li et al. (2022) reported that the population of the aphid Lipaphis erysimi (Kaltnbach) (Hemiptera: aphidae) decreases under high CO2 emissions, which indicates that the damage caused by the aphids may be reduced in the future with the increase in CO2 levels. In contrast, an increase in CO2 concentration reduced the resistance in wild-type tomato plants infested with Helicoverpa armigera Hübner (Lepidoptera: Noctuidae) (Guo et al., 2012) by lowering Jasmonic acid, an important hormone and polyphenol oxidase in these plants, which could likely exacerbate the problem of this pest in future climate scenarios. A longer duration and no effect on the adult survival and fecundity of L. sativa in environments of high CO2 concentration were found by Santos et al. (2021). It was emphasized that the lengthening of the long developmental duration of phytophagous insects is an important factor that can assist in evaluating the host quality, which suggests poor nutritional quality. Adverse effects on biological parameters also indicate that the damage caused by L. erysimi (Li et al., 2022) and Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) (De Paulo et al., 2020) will decrease in the future with the increase in the CO2 level. Hence, there is variation in the responses of different insect species to changes in CO2 concentration, which corroborates the results of Lindroth et al. 5. Conclusion Mahanarva spectabilis kept at the current CO2 level (Scenario C and D) had altered survival and fecundity as a function of the scenarios in which the plants were maintained. On the other hand, when the insects were always kept in the conditions of the future scenario (700 ppm), regardless of the scenarios where the plants were kept, survival was low, and the females of the first generation were not fertile, which made it impossible to assess the next generations, characterized by a greater direct than indirect effect of CO2 level. Furthermore, it should be considered that the effects of CO2 elevation on the survival, periods of development, and fecundity, when taken together, can significantly impact the population dynamics of M. spectabilis in future climate scenarios. Acknowledgements 2 Much of the research concerning insect responses to future climate scenarios refers to a single generation with plant fragments. However, experiments carried out over more than one generation and on intact plants (Brooks and Whittaker, 1998) reveal differences in responses between the generations and the growth conditions of the host plants, in addition to the insects’ freedom of movement in search of feeding places. These characteristics are taken into account in the present research, but it was yet only possible to analyze the performance of the spittlebugs in three successive generations, fed on forages Roxo de Botucatu (in Scenarios C and D) and Pioneiro (Scenario C). Effects on the interaction of generations of M. spectabilis in different scenarios are observed in the present research, in which those fed on Roxo de Botucatu plants derived from the phytotron with a high concentration of CO2 provided a reduction in the nymphal survival (63%) and the number of eggs (80%), in addition to an increase in the nymphal duration (14%) and adult longevity (75%) in the third generation compared to the first generation. Herein, Neophilaenus lineatus L (Hemiptera: Aphrophoridae) showed a constant response in an environment of high CO2 concentration over three generations (Brooks and Whittaker, 1998). In contrast, Gastrophysa viridula (De Geer) (Coleoptera: Chrysomelidae) showed reduced performance at high CO2 emission levels in the second generation of exposure but not in the first generation (Brooks and Whittaker, 1998). Srinivasa Rao et al. (2015) found significant variation in the growth and development of Spodoptera litura (Fabricius) (Lepidoptera: Noctuidae) as a function of CO2 concentration and generation. Larval duration increased by 6-8% over four generations under The authors express their gratitude to the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Brazil (Finance Code 304281/2019-0). 4. Discussion (1995), who reported that different species of saturniidae are differently affected by rising CO2 concentrations. Thus, it is essential to know the effect of climate changes on the control strategies of the spittlebugs in pastures, in a way that does not compromise the sustainability of the production system. References ALVARENGA, R., AUAD, A.M., MORAES, J.C., SILVA, S.E.B., RODRIGUES, B.S. and SILVA, G.B., 2017. Spittlebugs (Hemiptera: Cercopidae) and their host plants: a strategy for pasture diversification. Applied Entomology and Zoology, vol. 52, no. 4, pp. 653-660. http://dx.doi.org/10.1007/s13355-017-0521-0. AUAD, A.M., FONSECA, M.G., RESENDE, T.T. and MADDALENA, Í.S.C.P., 2012. Effect of climate change on longevity and reproduction of Sipha flava (Hemiptera: aphididae). The Florida Entomologist, vol. 95, no. 2, pp. 433-444. http://dx.doi.org/10.1653/024.095.0227. AUAD, A.M., SIMÕES, A.D., PEREIRA, A., BRAGA, A.L.F., SOUZA SOBRINHO, F., LÉDO, F.J.S., PAULA-MORAES, S.V., OLIVEIRA, S.A. and FERREIRA, R.B., 2007. Seleção de genótipos de capim-elefante quanto à resistência à cigarrinha-das-pastagens. Pesquisa Agropecuária Brasileira, vol. 42, no. 8, pp. 1077-1081. http:// dx.doi.org/10.1590/S0100-204X2007000800003. BEZEMER, T.M. and JONES, T.H., 1998. Plant-insect herbivore interactions in elevated atmospheric CO2: quantitative analysis and guild effect. Oikos, vol. 82, no. 2, pp. 212-222. http://dx.doi. org/10.2307/3546961. BROOKS, G. and WHITTAKER, J.B., 1998. Responses of multiple generations of Gastrophysa viridula, feedng on Rumex obtusifolis, to elevated CO2. Global Change Biology, vol. 4, no. 1, pp. 63-75. http://dx.doi.org/10.1046/j.1365-2486.1998.00111.x. BROOKS, G.L. and WHITTAKER, J.B., 1999. Responses of three generations of a xylem-feeding insect, Neophilaenus lineatus 8/10 Brazilian Journal of Biology, 2023, vol. 83, e273470 Concentrations of CO2 on the performance of spitllebug (Homoptera), to elevated CO2. Global Change Biology, vol. 5, no. 4, pp. 395-401. http://dx.doi.org/10.1046/j.1365- 2486.1999.00239.x. INTERGOVERNMENTAL PANEL ON CLIMATE CHANGE – IPCC, 2014 [viewed 30 March 2023]. Climate change 2014: synthesis report. Contribution of working groups I, II and III to the ffth assessment report of the Intergovernmental Panel on Climate Change [online]. Geneva: IPCC. Available from: https://archive. ipcc.ch/report/ar5/syr/ CARDONA, C., FORY, P., SOTELO, G., PABON, A., DIAZ, G. and MILES, J.W., 2004. Antibiosis and tolerance to five species of spittlebug (Homoptera: Cercopidae) in Brachiaria spp.: implications for breeding for resitance. Journal of Economic Entomology, vol. 97, no. 2, pp. 635-645. http://dx.doi.org/10.1603/0022-0493- 97.2.635. PMid:15154493. KARTHIK, S., REDDY, M.S. and YASHASWINI, G., 2021. Climate change and its potential impacts on insect-plant interactions. In: S.A. HARRIS, ed. The nature, causes, effects and mitigation of climate change on the environment. 10th ed. London: IntechOpen, pp. 393-415. http://dx.doi.org/10.5772/intechopen.98203. CONOVER, W.J., 1980. Practical nonparametric statistics. New York: John Wiley & Sons. Wiley Series in Probability and Statistics, no. 350. LI, X.M., ZHAO, M.H., HUANG, F., SHANG, F.G., ZHANG, Y.H., LIU, C.M., HE, S.J. and WU, G., and the LI, 2022. References Effects of elevated CO2 on the fitness of three successive generations of Lipaphis erysimi. Insects, vol. 13, no. 4, pp. 333. http://dx.doi.org/10.3390/ insects13040333. PMid:35447775. COVIELLA, C.E. and TRUMBLE, J.T., 1999. Effects of elevated atmospheric carbon dioxide on insect‐plant interactions. Conservation Biology, vol. 13, no. 4, pp. 700-712. http://dx.doi. org/10.1046/j.1523-1739.1999.98267.x. LINDROTH, R.L., ARTEEL, G.E. and KINNEY, K.K., 1995. Responses of three saturniid species to paper birch grown under enriched CO2 atmospheres. Functional Ecology, vol. 9, no. 2, pp. 306-311. http://dx.doi.org/10.2307/2390578. DÁDER, B., FERERES, A., MORENO, A. and TRĘBICKI, P., 2016. Elevated CO2 impacts bell pepper growth with consequences to Myzus persicae life history, feeding behaviour and virus transmission ability. Scientific Reports, vol. 6, no. 1, pp. 19120. http://dx.doi. org/10.1038/srep19120. PMid:26743585. NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION – NOAA, 2022 [viewed 30 March 2023]. Trends in atmospheric carbon dioxide [online]. Boulder, CO: NOAA. Available from: https:// gml.noaa.gov/ccgg/trends/global.html DE PAULO, P.D., PEREIRA, E.J.G., OLIVEIRA, E.E., FERERES, A. and GARZO, E., 2020. Indirect effect of elevated CO2 concentration on Bemisia tabaci MEAM1 feeding on Bt soybean plants. Journal of Applied Entomology, vol. 144, no. 10, pp. 941-951. http:// dx.doi.org/10.1111/jen.12822. POHLERT, T., 2014. The Pairwise Multiple Comparison of Mean Ranks Package (PMCMR). Vienna: R Foundation for Statistical Computing, pp. 9. DIAS-FILHO, M.B., 2014. Diagnóstico das pastagens no Brasil. Belém: Embrapa Amazônia Oriental. QIAN, L., LIU, X., HUANG, Z., WANG, L., ZHANG, Y., GAO, Y., GUI, F. and CHEN, F., 2021. Elevated CO2 enhances the host resistance against the western flower thrips, Frankliniella occidentalis, through increased callose deposition. Journal of Pest Science, vol. 94, no. 1, pp. 55-68. http://dx.doi.org/10.1007/s10340- 019-01123-7. FERREIRA, R.B., MORAES, J.C., AUAD, A.M. and FONSECA, M.G., 2013. Interaction of spittlebug and forage grass under different carbon dioxide concentrations. Journal of Pest Science, vol. 86, no. 2, pp. 161-166. http://dx.doi.org/10.1007/s10340-012-0449-7. FONSECA, M.G., AUAD, A.M., RESENDE, T.T., HOTT, M.C. and BORGES, C.A., 2016. How will Mahanarva spectabilis (Hemiptera: Cercopidae) respond to global warming? Journal of Insect Science, vol. 16, no. 1, pp. 32. http://dx.doi.org/10.1093/jisesa/ iew005. PMid:27012869. R CORE TEAM 2022 [viewed 11 December 2022]. R: a language and environment for statistical computing [online]. Vienna: R Foundation for Statistical Computing. Available from: https:// www. R-project.org/ RABINOVITCH, L., CAVADOS, C.F.G. and LIMA, M.M., 1998. O controle biológico de insetos nocivos a agricultura com o emprego de fungos imperfeitos ou hifomicetos. Biotecnologia Ciencia & Desenvolvimento, vol. 2, pp. 10-12. GOULD, F., 1998. References Sustainability of transgenic insecticidal cultivars: integrating pest genetics and ecology. Annual Review of Entomology, vol. 43, no. 1, pp. 701-726. http://dx.doi.org/10.1146/ annurev.ento.43.1.701. PMid:15012402. RESENDE, T.T., AUAD, A.M. and FONSECA, M.G., 2014. How many adults of Mahanarva spectabilis (Hemiptera: Cercopidae) should be used for screening Brachiaria ruziziensis (Poales: Poaceae) resistance? Journal of Economic Entomology, vol. 107, no. 1, pp. 396-402. http://dx.doi.org/10.1603/EC13297. PMid:24665725. GUO, H., PENG, X., GU, L., WU, J., GE, F. and SUN, Y., 2017. Up- regulation of MPK4 increases the feeding efficiency of the green peach aphid under elevated CO2 in Nicotiana attenuata. Journal of Experimental Botany, vol. 68, no. 21-22, pp. 5923- 5935. http://dx.doi.org/10.1093/jxb/erx394. PMid:29140446. GUO, H., SUN, Y., LI, Y., LIU, X., ZHANG, W. and GE, F., 2014. Elevated CO2 decreases the response of the ethylene signaling pathway in Medicago truncatula and increases the abundance of the pea aphid. The New Phytologist, vol. 201, no. 1, pp. 279-291. http:// dx.doi.org/10.1111/nph.12484. PMid:24015892. SANTOS J.O., ANGELOTTI, F., COSTA-LIMA, T.C., 2021. O aumento da concentração de CO2 interfere em aspectos biológicos de Liriomyza sativae em meloeiro? Ciência Agrária, vol. 42, pp. 2151- 2162. http://dx.doi.org/10.5433/1679-0359.2021v42n4p2151. SATISHCHANDRA, N., VADDI, S., NAIK, S.O., CHAKRAVARTHY, A.K. and ATLIHAN, R., 2018. Effect of temperature and CO2 on population growth of South American Tomato Moth, Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae) on tomato. Journal of Economic Entomology, vol. 111, no. 4, pp. 1614-1624.. http:// dx.doi.org/10.1093/jee/toy143. PMid:29868891. GUO, H., SUN, Y., REN, Q., ZHU-SALZMAN, K., KANG, L., WANG, C., LI, C. and GE, F., 2012. Elevated CO2 reduces the resistence and tolerance of tomato plants to Helicoverpa armigera by suppressing the JA signaling pathway. PLoS One, vol. 7, no. 7, e41426. http://dx.doi.org/10.1371/journal.pone.0041426. PMid:22829948. SILVA, D., AUAD, A.M., MORAES, J.C. and SILVA, S.E.B., 2019. How do Collaria oleosa and Brachiaria spp. respond to increase in carbon dioxide levels? Neotropical Entomology, vol. 48, no. 2, pp. 340-348. http://dx.doi.org/10.1007/s13744-018-0640-5. PMid:30374739. HAMANN, E., BLEVINS, C., FRANKS, S.J., JAMEEL, M.I. and ANDERSON, J.T., 2021. Climate change alters plant-herbivore interactions. The New Phytologist, vol. 229, no. 4, pp. 1894-1910. http:// dx.doi.org/10.1111/nph.17036. PMid:33111316. SILVA, S.E.B., AUAD, A.M., MORAES, J.C., ALVARENGA, R., CLAUDINO, S.S. and RESENDE, T.T., 2017. Biological Performance and HOLLANDER, M. and WOLFE, D.A., 1973. Nonparametric statistical methods. Hoboken: Wiley, pp. 115-120. 9/10 Brazilian Journal of Biology, 2023, vol. 83, e273470 Fonseca, M.G. et al. TRUMBLE, J. and BUTLER, C., 2009. Climate change will exacerbate California’s insect pest problems. California Agriculture, vol. 63, no. References 2, pp. 73-78. http://dx.doi.org/10.3733/ca.v063n02p73. preference of Mahanarva spectabilis (Hemiptera: Cercopidae) for feeding on different forage plants. Journal of Economic Entomology, vol. 110, no. 4, pp. 1877-1885. http://dx.doi. org/10.1093/jee/tox180. PMid:28854657. VASCONCELOS, D.C., KÜNAST, T.B.S., PIRES, E.M. and CORASSA, J.N., 2022. How can global climate change influence the geographic distribution of the eucalyptus yellow beetle? Modeling and prediction for Brazil. Brazilian Journal of Biology = Revista Brasileira de Biologia, vol. 82, e265046. http://dx.doi. org/10.1590/1519-6984.265046. SKENDŽIĆ, S., ZOVKO, M., ŽIVKOVIĆ, I.P., LEŠIĆ, V. and LEMIĆ, D., 2021. The impact of climate change on agricultural insect pests. Insects, vol. 12, no. 5, pp. 440. http://dx.doi.org/10.3390/ insects12050440. PMid:34066138. SOUZA SOBRINHO, F., AUAD, A.M. and LÉDO, F.J.S., 2010. Genetic variability in Brachiaria ruziziensis for resistance to spittlebugs. Crop Breeding and Applied Biotechnology, vol. 10, no. 1, pp. 83-88. http://dx.doi.org/10.12702/1984-7033.v10n01a11. XIE, H., ZHAO, H., SUN, Y., WANG, X. and LIN, X., 2022. Effect of elevated CO2 and temperature on maize resistance against the Asian corn borer. Journal of Applied Entomology, vol. 146, no. 8, pp. 990-1002. http://dx.doi.org/10.1111/jen.13038. SOUZA, A.P., GASPAR, M., SILVA, E., ULIAN, E.C., WACLAWOVSKY, A.J., NISHIYAMA JUNIOR, M.Y., SANTOS, R.V., TEIXEIRA, M.M., SOUZA, G.M. and BUCKERIDGE, M.S., 2008. Elevated CO2 increases photosynthesis, biomass and productivity, and modifies gene expression in sugarcane. Plant, Cell & Environment, vol. 31, no. 8, pp. 1116-1127. http://dx.doi.org/10.1111/j.1365- 3040.2008.01822.x. PMid:18433443. XIE, H., ZHAO, L., YANG, Q., WANG, Z. and HE, K., 2015. Direct effects of elevated CO2 levels on the fitness performance of Asian corn borer (Lepidoptera: Crambidae) for multigenerations. Environmental Entomology, vol. 44, no. 4, pp. 1250-1257. http:// dx.doi.org/10.1093/ee/nvv102. PMid:26314071. YIN, J., SUN, Y., WU, G. and GE, F., 2010. Effects of elevated CO2 associated with maize on multiple generations of the cotton bollworm, Helicoverpa armigera. Entomologia Experimentalis et Applicata, vol. 136, no. 1, pp. 12-20. http://dx.doi.org/10.1111/ j.1570-7458.2010.00998.x. SRINIVASA RAO, M., MANIMANJARI, D., VANAJA, M., RAMA, R.C.A., SRINIVAS, K., RAJU, B.M.K., MAHESWARI, M. and VENKATESWARLU, B., 2015. Response of multiple generations of tobacco caterpillar Spodoptera litura fab, feeding on peanut, to elevated CO2 peanut, to elevated CO2. Applied Ecology and Environmental Research, vol. 13, no. 2, pp. 373-386. http:// dx.doi.org/10.15666/aeer/1302_373386. ZAVALA, J.A., NABITY, P.D. and DELUCIA, E.H., 2013. An emerging understanding of mechanisms governing insect herbivory under elevated CO2. Annual Review of Entomology, vol. 58, no. 1, pp. 79-97. http://dx.doi.org/10.1146/annurev-ento-120811-153544. PMid:22974069. THEURILLAT, J.P. and GUISAN, A., 2001. Potential impact of climate change on vegetation in the European Alps. Brazilian Journal of Biology, 2023, vol. 83, e273470 References Climatic Change, vol. 50, no. 1-2, pp. 77-109. http://dx.doi.org/10.1023/A:1010632015572. 10/10
W1931679804.txt	https://journals.library.brocku.ca/index.php/voixplurielles/article/download/484/458	fr		La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine	Voix plurielles	2,008	cc-by	5,579	Voix plurielles Volume 5, Numéro 1 : mai 2008 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine Citation MLA : Vitali, Ilaria. «La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine.» Voix plurielles 5.1 (mai 2008). © Voix plurielles, revue électronique de l'APFUCC 2008. La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine Ilaria Vitali Université de Bologne - Université Sorbonne-Paris IV Mai 2008 Résumé Au carrefour de l’Orient et de l’Occident, la littérature algérienne de langue française a su s’approprier la bibliothèque du recueil des Mille et une Nuits pour en faire un espace scripturaire complexe. Véritable livre-édifice riche en symboles, le célèbre recueil a construit par sa structure un ensemble foisonnant de textes qui ont enchanté le lecteur et ont fait l’objet d’innombrables remaniements et emprunts. Il suffit de penser aux romans d’Assia Djebar, de Leïla Sebbar et plus récemment à ceux de Salim Bachi et de Mourad Djebel. Nous proposons d’explorer l’étoilement de cette bibliothèque dans la littérature algérienne contemporaine, qui a su exploiter les étagères des Mille et une Nuits comme un véritable espace à habiter et à réinventer. Au commencement, une bibliothèque Véritable édifice riche en symboles, le recueil des Mille et une Nuits construit par sa structure un ensemble foisonnant de textes qui enchantent le lecteur. Mais d’où viennent ces histoires à l’imagination flamboyante ? Contrairement à ce que l’on pourrait penser, la bibliothèque « virtuelle » de Shéhérazade est à l’origine une bibliothèque bien réelle. Car la « Sultane des aubes » n’invente pas, elle lègue les histoires qu’elle connaît : « Elle avait dévoré bien des livres : annales, vies de rois anciens, histoire des peuples passés, ouvrages de médecine. On dit qu’elle avait réuni mille livres touchant à ces peuples, aux rois de l’Antiquité et à leurs poètes. » (Les Mille et Une Nuits, t. I, 41.) Shéhérazade est donc porteuse des paroles de sa bibliothèque, car « le statut qu’elle revendique est celui de transmetteuse, et justement chacune de ses histoires commence par un « protocole d’ouverture », un syntagme rituel qui la rattache à une voix anonyme et mystérieuse : « On raconte » (yuhkâ), ou bien : « Il m’est parvenu » (balaghanî) ». (Kilito 14) Voix plurielles 5.1, mai 2008 1 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine L’origine de ces contes, précise Kilito, n’est pourtant jamais prise en compte : ils sont tout simplement là, connus et racontés de tout temps. Ce qui est cependant souligné avec force est le fait qu’ils n’ont pas été transmis dans l’oralité, mais qu’ils sont le fruit d’une œuvre de lecture. Shahrazâd n’a cependant pas recueilli ses contes à la faveur d’une transmission orale, mais dans les livres. Ses seuls maîtres sont des livres, au nombre de mille, où elle a appris la médecine, la poésie, l’histoire et les dits des sages et des rois. C’est dans cette immense mémoire écrite qu’elle a puisé la matière de ses récits, de l’enseignement qu’elle a dispensé à Shâhriyâr, nuit après nuit. Au commencement, une bibliothèque. (Kilito, 14-5). Il y a deux termes principaux à retenir au préalable dans ce passage : « livres » et « mémoire ». L’importance des livres et de leur transmission dans les contes des Nuits, ainsi que dans la constellation des textes qui s’y inspirent, est manifeste. La « bibliothèque » de Shéhérazade est donc un récipient, un réservoir, mais aussi un réceptacle, un miroir convexe qui reflète en abyme sa propre histoire, jusqu’au point de pousser Borges à imaginer la nuit centrale du recueil, dans laquelle Shéhérazade, par une distraction magique du copiste, raconte sa propre histoire « qui embrasse toutes les autres, qui – monstrueusement – s’embrasse elle-même ». (Borges 68).1 L’idée de l’infinité, évoquée par le titre même de l’ouvrage, qui rappelle par la finitude d’un numéro bien précis – 1001 – l’illimité, à mené la bibliothèque de Shéhérazade à voyager dans le temps et l’espace. Par ailleurs, cet infini dédaléen évoque également l’image du labyrinthe. Dans son essai Il labirinto tra Medioevo e Rinascimento, Umberto Eco concentrait sa réflexion sur trois des métaphores les plus anciennes, puissantes et omniprésentes dans la littérature, qui remettent sans cesse en cause l’identité de l’homme et sa représentation de la réalité. La première envisageait le monde comme un livre, la deuxième le livre comme un monde, et la troisième le monde – et le livre – comme un labyrinthe. L’infini labyrinthique des Mille et une Nuits, prévu en quelque sorte par la structure du recueil lui-même, a porté cette somme à sa dissémination dans les siècles. Ne pouvant pas analyser toute la constellation de textes qui se sont inspirés du recueil, nous proposons d’étudier ses reprises plus récentes dans quelques romans-clefs de la littérature algérienne de langue française. Au carrefour de l’Orient et de l’Occident, cette littérature a su s’approprier la bibliothèque du célèbre recueil pour en faire un édifice scripturaire complexe, qui dit, entre autres, l’opposition entre l’espace traditionnel et l’espace colonisé, guidée par les dynamiques antagonistes d’ouverture et de clôture, d’intérieur et d’extérieur. À la lumière de ces dynamiques sous-jacentes il est intéressant d’explorer l’étoilement de la bibliothèque de Shéhérazade dans la littérature algérienne contemporaine, qui a exploité les étagères et les couloirs labyrinthiques des Nuits comme un véritable espace à habiter et à réinventer. Nous pensons notamment aux romans d’Assia Djebar, de Leïla Sebbar et plus récemment à ceux de Salim Bachi et de Mourad Djebel, qui constituent le corpus que nous analyserons au cours de cet article. Voix plurielles 5.1, mai 2008 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine Naissance et étoilement de la bibliothèque des Milles et une Nuits De nombreuses études ont été consacrées à déterminer l’origine des Nuits, qui demeure néanmoins chargée de mystère : un écrit arabe ancien, le Kitab al-Fihrist, rédigé en 987, relate l’existence d’un volume persan contenant l’histoire de Shahrâzâd, intitulé Hezar Efsane (Les Milles Contes) dont pourtant il n’existe nulle trace. De fait, les manuscrits arabes n’ont été retrouvés que partiellement et ont eux-mêmes des sources différentes. Par l’emploi du préfixe «Shah», « Roi » en persan, les noms de Shahrâzâd (Shéhérazade en français) et de Shâhriyâr révéleraient l’origine persane des contes. Cependant, d’autres éléments, tels que le fait de retarder la mort en contant des fables, les métamorphoses en animaux, et les génies demi-dieux font référence au polythéisme hindou et témoigneraient d’une origine indienne, remontant au IIIe siècle. L’hypothèse veut donc que les contes soient nés en Inde et que, par voie orale, ils auraient atteint la Perse où un premier recueil, le Hezar Efsane aurait été écrit. Ce recueil se serait ensuite propagé dans le monde arabe grâce, entre autres, aux marchands avides de récits pour briser la monotonie de leurs voyages. Autour du VIIIe siècle, les conteurs arabes auraient traduit le livre et répandu ces histoires en les « arabisant », et en y ajoutant un bon nombre de contes, tout en conservant plusieurs éléments du texte primitif. Parmi les éléments arabes présents dans le recueil, mentionnons les éloges au Prophète, la forte présence des références toponymiques arabes (Bagdad, Le Caire, Bassora, Damas), et les renvois à des poètes célèbres, à des califes et à des savants du monde arabe. Ainsi s’est constituée la bibliothèque des Nuits. Par un jeu continu de reprises, d’emprunts, de résurgences et de réinterprétations, ce livreédifice figure comme le lieu privilégié d’une rencontre entre l’Orient arabo-islamique et l’Occident, allant plus loin qu’un jeu stérile d’auto-contemplation, en mettant à jour un processus de réécriture ininterrompu. Au sein de ce mouvement, le complexe entrelacs de relations que l’imaginaire littéraire du XXe siècle entretient avec le recueil des Nuits, peut aller « d’une fidélité-calque à une originalité impulsant une œuvre nouvelle. » (Chaulet-Achour 8). Cependant, la figure de la conteuse, qui retarde sa mort par la narration des contes, demeure sans doute l’élément le plus retravaillé. L’exploitation de la bibliothèque de Shéhérazade pour raconter les femmes du Maghreb C’est précisément par la figure de Shéhérazade que des romancières comme Assia Djebar et Leïla Sebbar introduisent en Occident toute une tradition orientale en l’apprivoisant, en l’asservissant à leurs voix nouvelles. Figure multiple, Shéhérazade se prête sans cesse au remaniement et offre aux romancières la matière pour rebrousser chemin, revenir en arrière, autant sur les enjeux que sur les pièges d’un Orient trop souvent mal connu et mal compris. Shéhérazade devient ainsi la voix dans un univers muet, celle qui sauva ses congénères, héroïne triomphante, dans un monde qui s’organise en principe autour de la figure de « l’homme », de « l’ennemi »2. Elle incarne la capacité de raconter, de nommer, de dire et de se dire. « Écrivain-femme porte parole des femmes séquestrées (Chikhi, 2007 14), Assia Djebar a exploité ce mythe protéiforme pour libérer les femmes algériennes, enfermées « dans leur Voix plurielles 5.1, mai 2008 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine appartement ». Dialoguant avec la tradition culturelle occidentale et orientale, ses romans se distinguent par la primauté exceptionnelle qu’ils accordent au féminin. Second volet du quatuor romanesque à vocation autobiographique commencé avec L’amour, la fantasia, le roman Ombre sultane semble être un appel à l’intention de ces femmes en attente, « prisonnières résignées d’un lieu clos », invitation à ôter leur voile physique et symbolique.3 Évoquant le fait que la « sultane des aubes » n’aurait pas survécu sans l’aide de sa sœur Dunyazad, la narratrice de Djebar, Isma, Shéhérazade contemporaine, porte secours à Hajila, sa coépouse, en l’aidant ainsi à sortir de l’isolement. La voix comme libération est ensuite reprise par Djebar dans un autre ouvrage, Loin de Médine, roman écrit dans l’urgence et qui vient interrompre la rédaction du quatuor, dans lequel l’auteure examine la situation des femmes depuis les origines de l’Islam, dans le but de rétablir leur rôle de premier plan aussi bien dans la religion que dans l’Histoire, en en faisant l’aboutissement de tout principe de vie. Pour rétablir la mémoire ancienne, Djebar tisse une nouvelle histoire, en exploitant la figure des femmes rawiyates4, « les transmetteuses », et dévoile en creux que la parole est un pouvoir et que la « voix » peut devenir « voie » : Du coup, la fiction contamine l’Histoire ou plutôt la corrige car ces femmes ressuscitées, ces femmes qui ont connu directement ou indirectement le Prophète s’emparent de la « parole vive » et devenues paradoxalement personnages romanesques puis narratrices, profèrent la vérité historique, vérité que le discours officiel ou qu’Assia Djebar identifie comme la Tradition a ensevelie. (Regaieg 2000) Reconsidérant la place de la femme dans la société musulmane, Djebar, narratrice-historienne, opère en même temps une renégociation des places dans le processus de communication littéraire, en invitant le lecteur à y jouer un rôle actif, à se frayer un chemin dans un récit incapable de séparer le fictionnel de l’historique, dans une écriture entre deux rivages qui ne peuvent être opposés. Dessiné en filigrane tout le long de son itinéraire littéraire, le mythe de Shéhérazade permet à Djebar de retracer le parcours des femmes qui va de la claustration à la liberté, un chemin qui se termine dans le recueil de nouvelles Femmes d’Alger dans leur appartement, qui semble finalement « libérer » les femmes de leur isolement.5 D’après Winifred Woodhull, malgré cette libération, l’œuvre de Djebar demeure pourtant « pessimiste », contrairement, par exemple, à celle de Fatima Mernissi : Yet even if it offers itself as an instrument of Algerian women’s liberation, Djebar’s text reinscribes a pessimistic view of her maghrebians sisters that many feminists, such as Fatima Mernissi, contest, emphasizing not their mutilation and dispossession but rather their capacity to speak and act on their on behalf today. (Woodhull 84.) Voix plurielles 5.1, mai 2008 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine Djebar n’est pourtant pas la seule à vouloir faire sortir les femmes d’Alger de leur appartement en exploitant la bibliothèque de Shéhérazade. Dans Fatima ou les Algériennes au square (1981), ou encore dans Parle mon fils, parle à ta mère (1985), Leïla Sebbar attribue une voix aux femmes immigrées de la première génération, en reconstruisant leur mémoire – ainsi que la mémoire collective de tout un peuple – par une sorte de bricolage identitaire. C’est pourtant dans sa trilogie de « Shéhérazade »6 que Leïla Sebbar explore les richesses et les rouages d’un Orient qui fascine et repousse à la fois, dont la représentante est la jeune beurette Shérazade, qui insiste sur la prononciation non-orientale de son prénom, et arrive jusqu’à adopter des pseudonymes.7 Nomade et picaresque, la fugitive Shérazade questionne les mythes du passé pour construire son identité : en traçant une topographie de la tradition elle finit par découvrir une cartographie nouvelle. Si le premier roman de la trilogie pose les cadres de la série, le deuxième, Les carnets de Shéhérazade, continue cet itinéraire fictionnel qui veut reconstruire les liens et les contrastes entre deux cultures en présence. Cette reconstruction au carrefour de deux mondes, comporte une réflexion qui se fonde d’une part sur la reprise d’un orient légendaire, et de l’autre sur celle d’une France également mythique. Dans ce processus créatif, l’imagination occupe une place de premier plan. A ce sujet, Inge Boer écrit : Sherazade and her friends suffer not from national trauma per se, but from their inability to memorialize and mourn Algeria’s traumatic past. Denied acces to the past, Sherazade is both limited to, and freed by, her imagination. She must invent the community of Algeria and its implications for her identity. (Boer 33). Ce besoin de renouer les fils de l’histoire, se lie, chez Sebbar, à la volonté de fonder une nouvelle identité féminine, entre tradition et modernité, entre Orient et Occident, dans l’isthme de l’entre-deux. Dans cette quête des origines le passé demeure pourtant lointain et les qualités de la conteuse mythique sont impossibles à atteindre. A ce sujet, il est intéressant de remarquer que dans le dernier volet de la trilogie, Le fou de Shérazade, la seule fois où la protagoniste prononce son nom comme celui de l’authentique conteuse des Nuits – « Shé-hér-azade », à la façon orientale, avec l’insertion de la syllabe médiane « her » – elle est soudainement incapable de raconter, et son auditoire de l’écouter : - C’est quoi ton nom ? - Shéhérazade. - Quoi ? disent ensemble les hommes. - Shéhérazade. Elle dit son nom comme on prononce celui de la sultane des Mille et une Nuits. - Vous entendez ? les hommes parlent tous à la fois. - On ne se trompait pas, une maison pour une sultane, qui va vous raconter des histoires la nuit, dans la langue la plus belle, la plus pure…Ah ! Quelles nuits… Quelles nuits ! (Sebbar 1991, 19) Voix plurielles 5.1, mai 2008 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine Et, pourtant, ils seront déçus, car ils ne seront pas à même de comprendre son discours « hybride », sa langue arabe parsemée de mots français : « Qu’est-ce qu’elle raconte encore ? On comprend à peine ce qu’elle dit. Quelle langue bâtarde elle parle ? Faites-la taire. » (p. 37)8 Noyau fécond, la scénographie de la bibliothèque des Nuits ne s’épuise pas, « ouverture d’une scène double, elle inspire l’infinité sémantique, romanesque et poétique. . . » (Chikhi 2007 79) Et pourtant, dans cette constellation littéraire de reprises, dans cet agrégat de textes multiples, « les nuits se suivent mais ne se ressemblent pas ; nostalgiques, quêteuses, ou critiques chez Boudjedra, elles se font soupçonneuses chez Khatibi, et deviennent motif d’écriture. » (p. 116) Car les hommes-écrivains aussi se sont intéressés à la conteuse et à son livre-édifice. Après tant de femmes qui ont consacré leur œuvre à se mettre dans la peau de Shéhérazade – ou à en sortir – , quelques hommes ont accepté de prendre la relève. Leur regard sur la sultane des Nuits est différent, leurs reprises de la bibliothèque le sont aussi. La bibliothèque de Shéhérazade vue par les hommes De toute évidence, Shéhérazade demeure assimilée à l’extraordinaire capacité de raconter. Voici ce qu’écrit Tahar Ben Jelloun : J’adore raconter des histoires. C’est un métier et une passion. . . Le principe littéraire le plus fondamental de tous les temps, c’est celui des Mille et une nuits. Raconte-moi une histoire ou je te tue... Nous sommes condamnés à raconter des histoires sous peine de disparition. Et une société sans romanciers, sans créateurs, sans raconteurs d’histoires, est une société déjà morte. (Ben Jelloun 107) Corps-récit immémorial, Shéhérazade est conçue par Khatibi au niveau purement théorique : « nul besoin de l’imaginer comme corps réel, nul besoin de reconstruire cet Orient de pacotille » (Khatibi 11). De prime abord, Khatibi opère une séparation inhérente au corps-conducteur de Shéhérazade : D’une part, le corps « réel, tatoué de perles et de parfums enivrant dans un harem de paradis où dansent et partouzent les princes (les principes), les concubines et les esclaves en un coït sériel… ». Ce corps, appartenant à un « Orient de pacotille », est à écarter, à oublier, à refouler. D’autre part, le corps « théorique », celui qu’elle-même a mis en jeu sur une scène « hautement théorique », parce que « pensée somptueuse d’un principe absolu, et somptueuse dans son luxe narratif – récits de récits – pour envahir, enchanter et ravir, comme si le narrataire était un mort ensorcelé ». (Khatibi 11 ) Voix plurielles 5.1, mai 2008 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine Dans les mots de Khatibi, Shéhérazade devient ainsi la théorie fastueuse du récit, le principe du conte comme séduction absolue, principe de la littérature tout court. Le titre même de l’ouvrage qui la voit protagoniste s’étoffe dans une poche d’ombre magique, dans l’interstice entre le sommeil et la veille, entre le jour et la nuit, la fiction et la réalité, se voulant ouverture vers un infini romanesque qui devient ainsi motif d’écriture, objet d’écriture. Par ailleurs, l’essence double de Shéhérazade permet une double – voire multiple – interprétation de sa figure. Si d’une part elle consolide la prise de conscience des « femmes séquestrées » et se fait porteuse de la fascination imaginative du récit, de l’autre, elle est aussi « corps à oublier », et quelques romancières des générations plus récentes semblent s’être lassées de la figure de la narratrice orientale, désormais devenue, grâce – ou à cause de – son succès universel, la messagère de cet « orient de pacotille », qui ne fait que recycler et renforcer des clichés abusés auprès du lecteur occidental9. Zineb Ali-Benali écrit : Progressivement le gel a gagné le verbe de Shéhérazade. Son personnage, de ne jamais bouger, est devenu une momie dans une galerie de personnages de cire, un cliché qui a fonctionné – qui fonctionne encore – dans une lecture toute faite. Shéhérazade est devenue le stéréotype de l’orientale, belle et apparemment soumise, personnage-clef des fantasmes masculins.(Ali-Benali : 2004, 217) En considérant exclusivement le côté « divertissement » de la sultane des aubes, ces auteurs semblent pourtant passer à côté de la complexité des messages qu’elle véhicule, et qui ne peuvent se réduire au fantasme masculin. Car Shéhérazade n’est pas seulement la maîtresse parfaite qui restaure le pouvoir royal grâce à ses contes, elle est aussi l’incarnation de la Qayna, poétesse et musicienne qui entre dans l’espace viril, rendant la parole « incontournable, indépassable, puisque le désir de fiction inoculé à l’homme reste à jamais inassouvi. » (Chikhi 128) Livre-édifice, la bibliothèque garantit à Shéhérazade non seulement sa survie, mais son immortalité. Le livre-édifice des Nuits dans l’extrême contemporain Grâce à son visage double et aux aspects hétéroclites que sa figure comporte, la sultane des aubes n’a pas été condamnée à l’oubli et revient sans cesse transformée dans les nouvelles voix de la littérature maghrébine. À l’instar de Khatibi et d’Assia Djebar, les auteurs qui suivent, notamment Salim Bachi et Mourad Djebel, considèrent Shéhérazade dans sa capacité imaginative suprême, comme principe prodigieux de l’infinité sémantique et romanesque : « Les contes appellent d’autres contes. Shéhérazade ne s’arrête jamais. Le monde n’est jamais figé. La vérité est multiple. Tout est dans la relation. » (Vitali 372) Si d’une part, la bibliothèque de Shéhérazade est évoquée sur le plan symbolique, de l’autre on retrouve aussi des reprises de la bibliothèque réelle, physique. Chez Salim Bachi, par exemple, plusieurs genres de bibliothèque émaillent le texte narratif. Voix plurielles 5.1, mai 2008 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine Rappelons ici la première apparition d’une bibliothèque dans Le Chien d’Ulysse, le premier roman de l’auteur. Il s’agit de la bibliothèque d’Ali Khan, professeur de lettres des protagonistes Hocine et Mourad, bibliothèque lourde de sens qui semble in fine crouler « sous son poids de mots imprimés » (Salim Bachi : 2001, 75-6.) A la valeur symbolique de la bibliothèque de ce professeur s’allient les connotations de la toile du Christ mort de Hans Holbein le jeune, situé dans la chambre, à proximité des étagères chargées de livres. Mesurant deux mètres de longueur, le tableau représente le Christ de façon inédite, cadavre humain écrasé par la tombe. D’après la figure de rhétorique de l’ekphrasis, qui fait confluer la description picturale dans l’espace littéraire, le tableau est ainsi retracé : . . . la bibliothèque croulant sous son poids de mots imprimés, la tranche cassée de chaque livre malmené par un usage frénétique, une reproduction du Christ mort de Holbein le Jeune prenaient un relief surnaturel selon qu’ils étaient ou non exposés au soleil. Le Christ décharné, dans sa tombe, ne jouissait plus d’aucune lumière, une éternité de claustration lui semblait promise. Le miracle ne s’accomplirait jamais, la lumière se refusait au sépulcre. (Salim Bachi : 2001, 75-6.) Décrit par Hocine, narrateur principal dans Le Chien d’Ulysse, le tableau à côté de la bibliothèque acquiert une valeur grandissante et convie néanmoins un renvoi intertextuel majeur à Dostoïevski, très impressionné par cette peinture qu’il avait pu voir à Bâle, et qui marque les pages de L’Idiot, dans lequel les personnages, et notamment le prince Michkine, en font un commentaire à plusieurs reprises, ce qui ne fait qu’amplifier la valeur de la référence intertextuelle. Par ailleurs, d’après Bernard Aresu : Combat allégorique entre la lumière vitale et la claustrophobie sépulcrale, de facture brutalement naturaliste, le tableau du Christ mort de Hans Holbein le jeune, dont le récit remarque la violence réformatrice et hétérodoxe, s’impose ainsi comme une référence fictionnellement et métaphysiquement spéculaire, comme ajout pictural densifiant le leitmotiv thanatographique du récit (CU, p. 76). L’objectivité que l’on pourrait qualifier de médico-légale du tableau, sa naturalité cadavérique, son message foncièrement antiidéaliste ne sauraient mieux souligner les angoisses politique et métaphysique que différentes voix ne cessent d’évoquer. (Aresu 180.) A cette bibliothèque physique s’en ajoutent ensuite plusieurs autres au cours des pages, bibliothèques formées par les noms d’auteurs ou les titres d’ouvrages qui affleurent à la surface du texte, à maintes reprises délivrées au lecteur. C’est le cas de la bibliothèque, cette fois virtuelle, de Hamid Kaïm, autre personnage majeur du Chien d’Ulysse, qui « grandit seul, entouré de livres. » (Bachi 2001, 102) : Voix plurielles 5.1, mai 2008 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine L’Iliade, où Achille combattait et mourait après avoir traîné le corps vaillant et pourtant mort d’Hector. Le chevalier à la triste figure s’élançait contre son dernier moulin. Le père Goriot, sur son lit de misère, attendait vainement la venue de ses chères filles. Il devenait une ville et, sur la colline, lançait comme un refrain un « A nous deux » bien creux. (102) Dans la bibliothèque de Kaïm, c’est donc d’abord L’Iliade qui a façonné la formation du protagoniste, suivie par Cervantès et Balzac, textes majeurs occidentaux, voire universels. D’ailleurs, le rôle épiphanique des livres est souligné par l’importance attribuée à Ulysses de Joyce, dont Kaïm entame la lecture lors de la mort de la sœur d’Ali Khan. Cyrtha, transformée en Ithaque par les multiples échos de l’Odyssée, devient ainsi Dublin, en se chargeant de toutes les valeurs symboliques que ces rapprochements entraînent. La bibliothèque de Shéhérazade revient aussi sous d’autres plumes algériennes de la nouvelle génération, comme celle de Mourad Djebel, qui, dans Les cinq et une nuits de Sharazède, raconte l’histoire de l’Algérie depuis l’Indépendance jusqu’aux événements des années 1990 par l’intermédiaire d’une femme : Ma génération est habitée par une révolte incommensurable, affirme Mourad Djebel, né en 1967, Nous sommes nés après l’indépendance et nous avons connu une sorte de dictature. Notre passage de l’adolescence à l’âge adulte a été ponctué par des explosions – celle des étudiants en 1986, celle de 1988, puis celle qu’on nomme « les événements ». Nous vivions alors la peur au ventre. (Le Monde, « Spécial Salon du Livre » 17 Mars 2006.) De toute évidence, ce roman présente nombre de ressemblances avec La Kahéna, deuxième roman de Salim Bachi. De même que chez Bachi, la reprise de la figure de Shéhérazade chez Djebel comporte le recyclage d’un autre élément topique des Nuits, à savoir les qualités presque « thaumaturgiques » du livre, son potentiel salvateur, sa capacité de dépasser la réalité, en en faisant oublier ses aspects les plus redoutables. Comme Shéhérazade, les personnages de ces romans luttent pour préserver des histoires de l’oubli, pour retarder leur mort, sabre suspendu au dessus de leur tête. Comme La Kahéna de Bachi, le roman de Djebel met en scène une structure divisée en plusieurs « nuits ». Du point de vue thématique, les romans des deux auteurs traitent de l’Algérie depuis l’Indépendance jusqu’aux années 1990. Pivot du récit, la protagoniste féminine cherchera à sauver un homme plongé dans une profonde dépression au cours des « années noires », à l’aide du verbe et de la poésie. Comme dans les Nuits, plusieurs récits se mêlent dans la construction romanesque orchestrée par Djebel : au premier degré, l’histoire d’amour entre les deux protagonistes, qui permet de retracer l’histoire de l’Algérie, est ensuite compliquée par les récits des nuits qui renouent avec la tradition des conteuses ancestrales. Sous l’angle stylistique, Djebel ne se contente pas de phagocyter les règles romanesques traditionnelles et préfère tresser les genres, en insérant ainsi des carnets (les carnets de Voix plurielles 5.1, mai 2008 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine Shâhriyâr), des poèmes et des chansons tirés des Nuits (par exemple, la Chanson de Oum Kalthoum, p. 294). L’intérêt et l’attention pour le texte et tout ce qui l’entoure sont soulignés également par la forte présence des notes, des citations, des traductions qui émaillent le roman, ainsi que par les remarques métacritiques (comme les commentaires sur la traduction des Nuits d’Antoine Galland.) Par l’analyse de quelques romans-clefs, ce court voyage dans la littérature algérienne francophone nous a montré de quelle manière le livre-édifice de Shéhérazade continue d’influencer les auteurs contemporains. Nous avons pu apprécier de quelle façon ces livres ne figurent pas seulement comme récipients de « toutes les histoires, de l’intégralité du raconté et du racontable » (Kilito 15), mais deviennent aussi des éléments romanesques capables de propulser l’histoire. Ainsi, la bibliothèque de la célèbre conteuse a joué un rôle de premier plan dans la constellation de textes qui s’inspirent des Nuits. En voyageant à travers l’espace et le temps, le livre-édifice de la Sultane des Aubes est devenu le lieu idéal de la rencontre et de l’étoilement des histoires et des savoirs. Bibliographie Œuvres Littéraires Bachi Salim, (2001). Le Chien d’Ulysse, Paris : Gallimard. ---. La Kahéna, (2003). Paris : Gallimard. Chouaki Aziz, (1998). Les Oranges, Paris : Les Mille et Une Nuits. Djebar Assia, (1980). Femmes d’Alger dans leur appartement, Paris : Albin Michel. ---. (1987). Ombre sultane, Paris : Jean-Claude Lattès. ---. (1995). Vaste est la prison, Paris : Albin Michel. Djebel Mourad, (2005). Les cinq et une nuits de Sharazède, Paris : La Différence. Sebbar Leïla, (1982). Shérazade, 17 ans, brune, frisée, les yeux verts, Paris : Stock. ---. (1991). Le Fou de Shérazade, Paris : Stock. Ouvrages critiques Bernard Aresu, (2004). « Arcanes algériens entés d’ajours helléniques : Le chien d’Ulysse de Salim Bachi », in Charles Bonn (dir.), Echanges et mutations des modèles littéraires entre Europe et Algérie, Paris : L’Harmattan, 177-187. Bencheikh Jamel Eddine; André Miquel, (1996). Les Mille et Une Nuits I, II et III, (trad.), Paris : Gallimard. Galland Antoine (trad.), (1965). Les Mille et Une Nuits, Paris : Flammarion. Bencheikh Jamel Eddine, Les Mille et Une Nuits ou la parole prisonnière, Paris : Gallimard, 1988. Voix plurielles 5.1, mai 2008 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine Ben Jelloun Tahar, (1995). Tahar Ben Jelloun. Deux cultures, une littérature, entretien avec Pierre Maury, Magazine Littéraire, Paris, nº 329. Bishop Michael, (1996). Thirty Voices in the Feminin, Amsterdam/Atlanta : Rodopi. Boer Inge E. (2003). After Orientalism: Critical Entanglements, Productive looks, Amsterdam/New York : Rodopi. Borges Jorge Luis, (1986). Enquêtes, Paris : Gallimard. « Ces nouvelles écritures nées de la douleur de l’Algérie », Le Monde, « Spécial Salon du Livre », 17 Mars 2006. Chaulet-Achour Christane (dir.), (2004). Les 1001 Nuits et l’imaginaire du XXe siècle, Paris : L’Harmattan.. Chikhi Beïda, (1996). Assia Djebar, extrait de La littérature maghrébine de langue française, Charles Bonn (dir.), Naget Khadda & Abdallah Mdarhri-Alaoui, Paris, EDICEF-AUPELF, <http://www.ziane-online.com/assia_djebar/textes/assia_djebar2.htm>. ---. (2007). Assia Djebar. Histoires et fantaisies, Paris : PUPS. Khatibi Abdelkébir, (1988). Ombres japonaises précédé de Nuits blanches, Montpellier : Fata Morgana. Kilito Abdelfattah, (1994). L’œil et l’aiguille. Essai sur les Mille et Une Nuits, Paris : La découverte. Laronde Michel, (1993). Autour du roman beur, Paris : L’Harmattan. Najiba Regaieg, (2000). « L’histoire sans les femmes, l’histoire des femmes, l’histoire par les femmes dans Loin de Médine d’Assia Djebar », Expressions maghrébines, n° 20-21. Vitali Ilaria, (2007). « Si Shéhérazade ne s’arrete jamais ». Entrevue avec Salim Bachi, en annexe à la thèse Entre les Mille et une nuits et Internet, Université de Bologne/Université Paris IV : 366-72. Woodhull Winifred, (1993). Transfigurations of the Maghreb: Feminism, Decolonization, and Literatures, Minneapolis : University of Minnesota Press. 1 Il est sans doute intéressant de remarquer que cette nuit, désignée par Borges comme la 602 ème, n’existe dans aucune des versions du recueil des Mille et une Nuits (voir à ce sujet A. Kilito 15). Séduit par l’invention borgésienne, même Italo Calvino n’hésite pas à s’attacher à cette fascinante possibilité (voir à ce sujet Livelli di realtà, sous la direction de Piattelli Palmarini, Feltrinelli, 1984, p. 441.) 2 « e’dou », « ennemi ». C’est ainsi que les femmes au hammam définissent leurs maris. Voir A. Djebar, (1995). Vaste est la prison, Paris : Albin Michel, p. 14. 3 « Ombre sultane porte à son paroxysme le sentiment de révolte féminine contre toutes les oppressions, qui se traduit par un encouragement manifeste à changer la société par l’affirmation féminine individuelle d’abord, collective ensuite. » ( Chikhi, Assia Djebar, extrait de La littérature maghrébine de langue française, Charles Bonn (dir.), Naget Khadda & Abdallah Mdarhri-Alaoui, Paris, EDICEF-AUPELF, 1996, <http://www.zianeonline.com/assia_djebar/textes/assia_djebar2.htm>. 4 En effet, Djebar paraît entendre par ce terme non seulement les « transmetteurs » de la vie du Prophète, mais plutôt les conteurs de toute la tradition transmise. 5 Ce que Picasso avait déjà fait en peinture, à travers la revisitation de la toile de Delacroix, comme l’explique bien Djebar dans la postface à son recueil, « Regard interdit, son coupé » : « Picasso renverse la malédiction, fait éclater le malheur, inscrit en lignes hardies un bonheur totalement nouveau. Prescience qui devrait, dans notre Voix plurielles 5.1, mai 2008 Ilaria Vitali La bibliothèque de Shéhérazade en voyage. Reprises et réécritures d’un livre-édifice dans la littérature algérienne contemporaine quotidien, nous guider. », Djebar, (2002). Femmes d'Alger dans leur appartement, Paris : Albin Michel, p. 244. 6 La trilogie qui voit protagoniste Shérazade est constituée par trois romans : Shérazade, 17 ans, brune, frisée, les yeux verts (1982), Les carnets de Shérazade (1985) et Le fou de Shérazade (1991). 7 Au sujet de la substitution nominale fonctionnelle opérée par la Shérazade de Sebbar, voir M. Laronde, (1993). Autour du roman beur, Paris : L’Harmattan, pp. 197-205. 8 Voir aussi à ce sujet M. Bishop, (1996). Thirty Voices in the Feminin, Amsterdam/Atlanta : Rodopi, p. 244. 9 Et pourtant nombreuses sont les reprises qui réduisent Shéhérazade à la femme séductrice. Voir aussi la pièce Les Oranges Aziz Chouaki : Tiens tout s’arrête, qu’est-ce qui se passe ? Le match s’est arrêté, mais qu’est-ce qui se passe ? Tous les mâles du quartier sont immobiles. Ah, ça y est, c’est Schéhérazade, qui vient du bout de la rue, comme elle est belle… jambes luisantes et bronzées, ses seins moulés dans une casaca blanche, et sa gueule d’amour… Tout le monde bave et bande, normal, faut la voir, Schéhérazade, frimousse brune, sourire marine. (Chouaki 38.) Voix plurielles 5.1, mai 2008
W2143540650.txt	https://link.springer.com/content/pdf/10.1007/s00210-015-1145-y.pdf	en		A streptozotocin-induced diabetic neuropathic pain model for static or dynamic mechanical allodynia and vulvodynia: validation using topical and systemic gabapentin	Naunyn-Schmiedeberg's archives of pharmacology	2,015	cc-by	8,821	Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 DOI 10.1007/s00210-015-1145-y ORIGINAL ARTICLE A streptozotocin-induced diabetic neuropathic pain model for static or dynamic mechanical allodynia and vulvodynia: validation using topical and systemic gabapentin Gowhar Ali 1 & Fazal Subhan 1 & Muzaffar Abbas 2 & Jehan Zeb 1 & Muhammad Shahid 1 & Robert D. E. Sewell 3 Received: 26 August 2014 / Accepted: 10 June 2015 / Published online: 3 July 2015 # The Author(s) 2015. This article is published with open access at Springerlink.com Abstract Neuropathic vulvodynia is a state of vulval discomfort characterized by a burning sensation, diffuse pain, pruritus or rawness with an acute or chronic onset. Diabetes mellitus may cause this type of vulvar pain in several ways, so this study was conducted to evaluate streptozotocin-induced diabetes as a neuropathic pain model for vulvodynia in female rats. The presence of streptozotocin (50 mg/kg i.p.)-induced diabetes was initially verified by disclosure of pancreatic tissue degeneration, blood glucose elevation and body weight loss 5–29 days after a single treatment. Dynamic (shortened paw withdrawal latency to light brushing) and static (diminished von Frey filament threshold pressure) mechanical allodynia was then confirmed on the plantar foot surface. Subsequently, both static and dynamic vulvodynia was detected by application of the paradigm to the vulval region. Systemic gabapentin (75 mg/kg, i.p.) and topical gabapentin (10 % gel) were finally tested against allodynia and vulvodynia. Topical gabapentin and the control gel vehicle significantly increased Electronic supplementary material The online version of this article (doi:10.1007/s00210-015-1145-y) contains supplementary material, which is available to authorized users. * Robert D. E. Sewell sewell@cardiff.ac.uk Fazal Subhan fazal_subhan@upesh.edu.pk 1 Department of Pharmacy, University of Peshawar, Peshawar 25120, Pakistan 2 Fulbright Graduate Student, Department of Pharmaceutical Sciences, College of Pharmacy, South, Dakota State University, Brookings, SD 57007, USA 3 Cardiff School of Pharmacy and Pharmaceutical Sciences, Cardiff University, Redwood Building, King Edward VII Ave., Cardiff CF10 3NB, UK paw withdrawal threshold in the case of the static allodynia model and also paw withdrawal latency in the model for dynamic allodynia when compared with the streptozotocinpretreated group. Likewise, in the case of static and dynamic vulvodynia, there was a significant antivulvodynia effect of systemic and topical gabapentin treatment. These outcomes substantiate the value of this model not only for allodynia but also for vulvodynia, and this was corroborated by the findings not only with systemic but also with topical gabapentin. Keywords Vulvodynia model . Allodynia . Streptozotocin . Diabetic neuropathic pain . Gabapentin Introduction The prevalence of diabetes mellitus has been increasing globally, and it was estimated by the WHO that the occurrence of diabetes in adults was 173 million in 2002 (Shaw et al. 2010). Painful diabetic neuropathy is a complication of diabetes mellitus, and it is encountered in 60 % of patients adversely affecting their quality of life (Said 2007). In neuropathic pain, spontaneous and pathologically exaggerated responses occur to noxious and non-noxious stimuli (Costigan et al. 2009). Underlying mechanisms implicated in neuropathic pain include synaptic facilitation which contributes to secondary hypersensitivity and tactile allodynia (Campbell and Meyer 2006). Vulvodynia is a state of vulval discomfort characterized by burning, diffuse moderate to severe pain, pruritus or rawness with acute or chronic onset. Usually, there are few visible symptoms, although varying degrees of erythema have been reported (Bachmann et al. 2006; Lynch 1986; Paavonen 1995; Young et al. 1984), and the most frequently identified types are generalized unprovoked vulvodynia or vestibulodynia (Backonja et al. 1998). Diabetes mellitus may cause vulvar 1130 pain of this type in several ways (Kalra et al. 2013), and currently, the management of neuropathic pain involves adjunctive therapy. It is a challenge to define an exact course of treatment for instance in diabetic neuropathy since active pain mechanisms in patients are still not clearly defined and often expressed as complex pain phenotypes. Consequently, only partial relief is achievable in most individuals (Goodnick et al. 2000; Field et al. 1999). It has been shown that diabetic and para-diabetic neuropathic pain is invariably resistant to classical non-steroidal anti-inflammatory drugs (NSAIDs) and opioids or that it has a poor response to these agents (James and Page 1994). In this connection, numerous drugs have been extensively investigated against painful diabetic neuropathy including gabapentin (Gorson et al. 1999; Field et al. 1999; Zhang et al. 2013). Moreover, analgesic antidepressants have also been shown to promote the responsiveness of neurons in the locus coeruleus to noxious stimulation in a chronic neuropathy model (Alba-Delgado et al. 2012) though gabapentin has not been studied in this respect. An earlier retrospective clinical study concluded that topical gabapentin may be useful in neuropathic pain including allodynia in the area of the vulva (vulvodynia) (Boardman et al. 2008). Gabapentin [1-(aminomethyl) cyclohexane acetic acid] is an antiepileptic drug and structural analogue of the neurotransmitter gammaaminobutyric acid (GABA). It was introduced in 1993 by the US FDA as an adjuvant anticonvulsant (Barrueto et al. 2002) and subsequently for various chronic pain conditions including diabetic neuropathy (Apkarian et al. 2009; Bennett and Simpson 2004; Gilron 2007; Gilron et al. 2005; Hunter et al. 1997; Mao and Chen 2000; Wiffen et al. 2005; Leo 2013). In this context, analgesic effects on peripheral nociception in rats have been reported to injectable gabapentin (Carlton and Zhou 1998; Cesena and Calcutt 1999). The primary aim of the study was to employ a streptozotocin-induced neuropathic pain model to establish the existence of mechanical allodynia and then to investigate the presence of any potential static or dynamic mechanical vulvodynia. Subsequently, the objective was to evaluate the activity of gabapentin either administered systemically or applied locally as a topical gel preparation against vulvodynia as a mode of further validating the model for predicting relevance to the clinic (Bates and Timmins 2002). Methods and materials Chemicals and instruments A von Frey filament (hair) kit was obtained from Stoelting (USA). The following were purchased from suppliers: streptozotocin (Sigma-Aldrich), glucometer (Roche Diagnostic Corporation, Germany), blood glucose strips (Optium Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 Xceed, Abbott), gabapentin gel 10 % w/w and the control gel base (a liposome-containing oil in water gel comprising xanthan gum hydrocolloid with polyacrylamide) minus the active pharmaceutical ingredient [API]) were supplied by St Mary’s Pharmaceutical Unit (SMPU, Cardiff, UK) under their Manufacturer’s Special License (MS). Gabapentin was obtained from MKB Pharmaceuticals (Pvt.) Ltd., Peshawar, Pakistan. All other chemicals and reagents used were of analytical grade. Animals Sprague-Dawley female rats, bred in the animal house and bioassay laboratories of the Department of Pharmacy, University of Peshawar, were used throughout experimental studies. Animals were housed in transparent cages with free access to standard laboratory food and water available ad libitum. All experimental procedures were carried out between 0800 and 1700 h. A 12–12-h light and dark cycle was provided with an ambient temperature maintained at 22.0±2.0 °C through a reversible air conditioning system. Ethical approval The study of the topical gabapentin application was approved under a project entitled ‘Studies on the effects of topical formulations of different analgesics on allodynia and vulvodynia components of diabetic neuropathic pain’ by the Ethical Committee of the Department of Pharmacy, University of Peshawar, Peshawar, Khyber Pakhtunkhwa (KP), Pakistan, who issued approval certificate no 13/EC-12/Pharm. The study of systemic gabapentin administration was approved under a project entitled ‘Evaluation of the anti-inflammatory, antinociceptive and antipyretic activities of selected synthetic and natural compounds in animal models’ (approval certificate no 15/EC/Pharm). Moreover, all animal procedures were conducted strictly according to the Animals Scientific Procedure Act (1986) of UK. Induction of diabetes Female Sprague-Dawley rats, weighing 175–200 g (age range=7.0±1.0 weeks), were randomly divided into saline vehicle and treatment groups. Sixteen hours after fasting (Babu et al. 2003), animals were weighed and a single intraperitoneal (i.p.) injection of streptozotocin (STZ) (50 mg/kg) was administered for the induction of diabetes (Field et al. 1999). Since streptozotocin has a stability problem (Rakieten et al. 1963), fresh solutions were prepared for each period of administration. Sixty minutes after streptozotocin administration, animals were allowed free access to food and water and then kept under close observation for the next 5 days during the development of diabetic symptoms. To maintain Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 cleanliness and avoid development of any infection due to excessive urination, animal bedding was changed frequently. The singly administered dose of streptozotocin in our study was lower than that employed by other research groups in rats (Indolfi et al. 2001; Wei et al. 2003) and was in point of fact only half of that given by Indolfi et al. (2001). Additionally, at the 50 mg/kg streptozotocin dose level, no undue adverse behavioural effects have been previously reported (Field et al. 1999; Jolivalt et al. 2006). Consequently, the degree of insulinopoenia in this study was not supplemented by exogenous insulin administration, and this has been corroborated by chronological survival rates even up to 6-fold longer than our study duration (Wei et al. 2003). Measurement of blood glucose and body weight Blood glucose levels of animals were measured at predetermined time intervals, i.e. experimental days 0, 5, 15 and 29 using a glucometer (Roche Diagnostic Corporation, Germany). Blood samples were collected from the rats by the tail tip method. Body weights of the diabetic and salinetreated animals were also recorded throughout the experimental protocol. Assessment of static allodynia in diabetic female rats Pretreatment schedule On the 29th day post-streptozotocin administration, animals were acclimatized to cages with wire mesh bottoms for 15– 45 min. A series of von Frey filaments (0.4, 0.6, 1, 1.4, 2, 4, 6, 8, 10 and 15 g) were applied perpendicularly to the midplantar surface of the right hind paw to an extent that caused the hairs to bend (Calcutt and Chaplan 2009). Each von Frey filament was applied for a period of up to 6 s as a cut-off time or until an escape response (paw withdrawal threshold (PWT) or licking) occurred. The withdrawal threshold was defined as the minimum applied force (g±SD) required to induce a paw withdrawal reflex (Field et al. 1999). Lifting of the paw or licking was recorded as a positive response, and a succeeding von Frey filament of lower force was applied for the following recording. However, in the case of an absence of response, the subsequent von Frey filament of higher force was applied. This procedure was continued until four measurements were taken after an initial change including the first one, in response (positive response) or five consecutive negative responses (2, 4, 6, 8 and 10-g force) or four consecutive positive responses (1.4, 1, 0.6 and 0.4-g force). A force of 15g was selected as the cut-off force at which point further application ceased. von Frey filaments were applied at intervals of several seconds in order to avoid any influence of previous stimuli on behaviour. Any ambulation was noted as an indefinite response, and 1131 the stimulus was repeated. Static allodynia was indicated by a reduced force required to induce a paw withdrawal reflex. Post-treatment schedule Animals included for the post-drug-treatment study were those which exhibited a mean response of ≤3.63g (see Table 1) (Field et al. 1999). In the topical gabapentin investigation, a uniform amount of 10 % gabapentin gel (1.0 mL/cm2) was applied topically for three times daily (0800, 1200 and 1600 h) on the plantar surface of the right hind paw. Static allodynia was measured 1 h post-drug application. The concentration of gabapentin applied in the form of gel (i.e. 100 mg/cm 2 ) was 3.3-fold higher than that previously employed for topical lidocaine in hypoalgesic studies on mucous membranes (Schønemann et al. 1992). Control animals received an equivalent application of control gel (1.0 mL/ cm2). In the case of the systemic study, gabapentin (75 mg/kg, i.p.) was administered to animals meeting the selection criterion, and static allodynia was measured at 1 and 2 h post-drug administration. Control animals received an equivalent amount of normal saline vehicle. The von Frey PWT inclusion criterion for static vulvodynia is shown in Table 1. The overall mean PWT values were calculated from the pooled post-treatment triplicate daytime readings (0800, 1200 and 1600 h) of static allodynia in the topical study or the duplicate readings (1 and 2 h) post-administration in the systemic study. Assessment of dynamic allodynia in diabetic female rats On the 29th day post-streptozotocin injection, dynamic allodynia was assessed by lightly brushing the plantar surface of the hind paw of each animal with a cotton bud. The time to flinching or licking of the paw was regarded as the paw withdrawal latency (PWL) in seconds. Animals responding to the cotton bud within 8 s were selected for the study (Field et al. 1999). Fifteen seconds was selected as cut-off time beyond which the assessment was terminated. In the topical study, gabapentin gel (10 %; 100 mg/cm2) or control gel (1.0 mL/cm2) was then applied to the hind paw, three times daily (0800, 1200 and 16.00 h), and animals Table 1 Inclusion criteria for static and dynamic allodynia or vulvodynia Static allodynia Dynamic allodynia Static vulvodynia Dynamic vulvodynia ≤3.63-g von ≤8.0-s brushing ≤0.16-g von ≤4.0-s brushing Frey paw paw withdrawal Frey flinching flinching withdrawal latency (PWL) response response threshold threshold latency (FRL) (PWT) (FRT) 1132 were tested for dynamic allodynia, by the method described above, 1 h post-gabapentin gel or control gel application. Systemic gabapentin (75 mg/kg, i.p) was administered and dynamic allodynia was measured by the method described above at 1 and 2 h post-drug administration. The brushing PWL inclusion criterion for dynamic allodynia is shown in Table 1. The overall mean PWL values were calculated from the pooled post-treatment triplicate daytime readings (08.00, 12.00 and 16.00 h) of dynamic allodynia in the topical study or the duplicate readings (1 and 2 h) post-administration in the systemic study. Assessment of static vulvodynia in diabetic female rats Pretreatment schedule The anogenital area including the mons pubis area of female rats was carefully shaved. The animals were then acclimatized to cages with wire mesh bottoms for 15–45 min (to prevent excessive exploration and major grooming activities). A series of von Frey filaments (0.008, 0.02, 0.04, 0.07, 0.16, 0.4, 0.6 and 1.0 g) were applied perpendicularly for a period of 4 s and the flinching response threshold (FRT, g) recorded. This procedure was continued for up to four measurements after an initial change in response (positive response) or five consecutive negative responses (0.07, 0.16, 0.4, 0.6 and 1-g force) or two consecutive positive responses (0.008 and 0.02-g force). Ambulation was noted as an ambiguous response, and if it occurred, the stimulus was repeated. A von Frey force of 1g was selected as the cut-off beyond which the assessment was terminated. Post-treatment schedule On the 29th day post-streptozotocin administration, a standardized testing paradigm (Chaplan et al. 1994) was used with modifications for the vulvodynia study. The von Frey FRT inclusion criterion for static vulvodynia is shown in Table 1. Gabapentin gel (10 %; 100 mg/cm2) was then applied topically for three times daily (0800, 1200 and 1600 h) to the anogenital region. Static vulvodynia was assessed 1 h after gabapentin gel application. Control gel (1.0 mL/cm2) was applied in the control animal group. Systemic gabapentin (75 mg/kg, i.p) was administered, and static vulvodynia was measured by the method described above at 1 and 2 h postdrug administration, the controls receiving an equivalent amount of i.p. normal saline. The overall mean FRT values were calculated from the pooled post-treatment triplicate daytime readings (800, 1200 and 1600 h) of static vulvodynia in the topical study or the duplicate readings (1 and 2 h) postadministration in the systemic study. Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 Assessment of dynamic vulvodynia in diabetic rats The anogenital area of animals was shaved as before. The animals were then acclimatized on wire mesh cages for 15– 45 min. On the 29th day post-streptozotocin administration, dynamic vulvodynia was assessed by lightly brushing the anogenital region for a period of 10 s or until a flinching response occurred (flinching response latency (FRL)). If no response occurred within 10 s (cut-off time), then the procedure was terminated and the animal was excluded from the study. The brushing FRL inclusion criterion for dynamic vulvodynia is shown in Table 1. Animals meeting the inclusion criteria in the topical study had gabapentin gel (10 %; 100 mg/cm2) or control gel (1.0 mL/cm2) applied for three times daily (0800, 1200 and 1600 h) and were tested for dynamic vulvodynia 1 h later. Systemic gabapentin (75 mg/kg, i.p) was administered, and animals were tested for dynamic vulvodynia by the method described above at 1 and 2 h post-drug administration, the controls receiving an equivalent amount of i.p. normal saline. The overall mean FRL values were calculated from the pooled post-treatment triplicate daytime readings (0800, 1200 and 1600 h) of dynamic vulvodynia in the topical study or the duplicate readings (1 and 2 h) post-administration in the systemic study. Microscopic examination of the pancreas The pancreas was removed from each animal at the end of the experiment and stored in 10 % formalin solution for microscopic examination of islet of Langerhans cells. The tissues were then processed and slides of 5-μm thickness were prepared employing routine haematoxylin and eosin staining procedures (Falkeholm et al. 2001) (Fig. 1). Statistical analysis Data were expressed as means and were analyzed by one-way ANOVA and Tukey’s multiple comparison post hoc test. Mann-Whitney U test was applied to data where applicable. Values were considered significant at P<0.05. Results Development of diabetes in streptozotocin-pretreated animals As disclosed in Fig. 2, animals administered streptozotocin manifested diabetic symptoms reproducibly by increasing the group mean blood glucose level from 91.0±1.7 mg/dL on day 0 to 342.6 ± 8.0 mg/dL on the fifth day, 350.1 ± 8.8 mg/dL on day 15, and 383.3±9.3 mg/dL on day 29 of Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 1133 the experimental schedule. However, at a streptozotocin dose of 50 mg/kg (Field et al. 1999), animals exhibiting mean blood glucose levels ≥270 mg/dL on day 5 were selected for the study (Jolivalt et al. 2006). Development of diabetes was further confirmed from the degradation of pancreatic islet cells in streptozotocin-pretreated diabetic rats (Fig. 1). In addition to hyperglycaemia, diabetic animals exhibited mild polyphagia, polydipsia and polyuria and some body weight loss (Fig. 3). Activity of topical gabapentin gel (10 %) and systemic gabapentin (75 mg/kg i.p.) on static allodynia in female diabetic rats Fig. 1 Photomicrographs of the pancreas from rats treated with a saline or b streptozotocin (50 mg/kg) i.p. showing intact or degenerated islets of Langerhans, respectively, 29 days after treatment Fig. 2 The effect of streptozotocin (STZ, 50 mg/kg i.p.) or saline vehicle on blood glucose levels of rats (mean±SEM) measured on days 0, 5, 15 and 29 of the experimental protocol. Values for streptozotocin-treated animals were significantly different compared to saline-vehicle-treated controls (*P<0.001, Mann–Whitney U test). The animal selection level for subsequent allodynia studies was ≥270 mg/dL on day 5 The pooled mean von Frey filament pressures (g±SD) calculated as PWT values (static allodynia) are plotted in Fig. 4. In the topical application experiments (Fig. 4a), the treatment groups consisted of saline, gabapentin gel alone (GG), streptozotocin pretreatment alone (STZ), and streptozotocin pretreatment plus either topical gabapentin gel (STZ + GG 10 %) or topical control gel (STZ + CG) (F(4,31) =36.41, P<0.001). Tukey’s post hoc analysis revealed a significant decrease (P<0.001) in the PWT responses of streptozotocinpretreated animals versus saline-treated controls, and this was reversed in the group pretreated with streptozotocin and then treated with topical gabapentin gel (STZ + GG, P<0.001). There were also increased PWT values for both gabapentin gel (P<0.001) and control gel vehicle (P<0.001) topical treatments compared to the streptozotocin-pretreated control (STZ) group (Fig. 4a). Fig. 3 Effect of streptozotocin (STZ, 50 mg/kg i.p.) or saline vehicle on rat body weights (mean±SEM) measured on days 0, 5, 15 and 29 of the experimental protocol. Values for streptozotocin-treated animals were significantly different compared to saline-vehicle-treated controls (P<0.001, Mann–Whitney U test) 1134 Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 pretreatment alone (STZ) and streptozotocin plus gabapentin (STZ + GP (75 mg/kg i.p.) (F (3,16) = 15.92, P < 0.001) (Fig. 4b). Tukey’s post hoc analysis test revealed a decrease (P<0.001) in the PWT response to streptozotocin pretreatment versus saline-treated controls. A reversal of this streptozotocin-decreased PWT occurred when combined with systemic gabapentin (STZ + GP, P<0.01) at the pooled 1 and 2-h post-treatment times, whereas gabapentin alone induced a response that was greater than streptozotocin alone (P<0.05) (Fig. 4b). Activity of topical gabapentin gel (10 %) and systemic gabapentin (75 mg /kg i.p) on dynamic allodynia in female diabetic rats Fig. 4 The static antiallodynia activity of topical gabapentin gel (10 %) and systemic gabapentin (75 mg/kg i.p.) in streptozotocin-induced female diabetic rats. Scatter plots showing mean paw withdrawal thresholds (PWTs, g±SD) in response to von Frey filaments were determined in all groups. a Topical gabapentin gel (10 %) alone without streptozotocin pretreatment (triangle, GG, n=8), streptozotocin pretreatment followed by control gel (inverted triangle, STZ + CG, n=9) or streptozotocin pretreatment followed by gabapentin gel (diamond, STZ + GG, n=5) was applied on the plantar surface of the right hind paws of rats for three times daily and PWT was measured 1 h after application. Significance of differences in PWT is shown between topical saline (white circle, n=15) and streptozotocin pretreatment (white square, STZ, n=6; ###P<0.001) and between streptozotocin pretreatment (white square, STZ, n=6) versus gabapentin gel alone (triangle, GG, n=8; P<0.001), streptozotocin pretreatment with control gel (inverted triangle, STZ + CG, n= 9; P< 0.001) and streptozotocin pretreatment plus gabapentin gel (diamond, STZ + GG, n=5; *P<0.001), ANOVA followed by Tukey’s post hoc test. b In the case of the systemic study, gabapentin (75 mg/kg) was administered intraperitoneally (i.p.) and PWT was measured using the same protocol as gabapentin gel. ANOVA followed by Tukey’s post hoc test revealed statistical differences in PWT between saline (white circle, n=15) and streptozotocin pretreatment (white square, STZ, n=6; ###P<0.001), between streptozotocin pretreatment (white square, STZ, n=6) and gabapentin alone (triangle, GP 75 mg/kg, n=5; P<0.05) and also between streptozotocin plus gabapentin (inverted triangle, STZ + GP 75 mg/kg, n=8; P<0.01) In the systemic study, the mean pooled von Frey filament pressures (g) calculated as PWT values (static allodynia) were graphed. The four groups consisted of saline controls, gabapentin alone (GP 75 mg/kg i.p), streptozotocin Dynamic allodynia was assessed by lightly brushing a cotton bud over the plantar surface of the hind paw of rats, and the pooled PWL (mean s±SD) was determined (F(4,34) =13.99, P<0.001) and plotted as shown in Fig. 5a. There was a significant decrease in PWL caused by streptozotocin pretreatment in comparison with the saline-treated controls (P<0.001). This reduced PWL induced by streptozotocin was totally reversed by combination with topical gabapentin gel (STZ + GG, P<0.001) and also somewhat unpredictably by the control gel (STZ + CG, P<0.001). In the systemic study, dynamic allodynia was assessed using the same protocol. PWL (mean s±SD) with respect to time after treatment at the grouped 1 and 2-h intervals (F(3, 16) =15.92, P<0.001) is depicted in Fig. 5b. Streptozotocin pretreatment reduced the PWL compared to saline treatment (P<0.001), and this was reversed by combination with systemic gabapentin administration (STZ + GP, P<0.01). However, as might be expected, there was also a statistical difference between the group treated with gabapentin alone (GP) compared with those pretreated with streptozotocin (P<0.05, Fig. 5b). Activity of topical gabapentin gel (10 %) and systemic gabapentin (75 mg/kg i.p) on static vulvodynia in female diabetic rats The mean von Frey filament force (g±SD) required to induce an FRT (static vulvodynia) was calculated and then plotted for the five treatment groups at the three merged daily application times (F(4,46) =7.920, P<0.001) shown in Fig. 6a. ANOVA coupled with Tukey’s post hoc analysis revealed a significant reduction in the FRT force expressed after streptozotocin pretreatment compared to the saline-treated group (P<0.05). This decrease was reversed in the group treated with the streptozotocin plus topical gabapentin gel combination (P < 0.05) but not in those animals pretreated with streptozotocin plus subsequent control gel. The gabapentingel-alone group (GG) did display a difference from animals Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 Fig. 5 The dynamic antiallodynia activity of gabapentin gel 10 % and systemic gabapentin (75 mg/kg i.p.) in streptozotocin-induced female diabetic rats. Scatter plots showing mean paw withdrawal latency (PWL, s±SD) in response to light brushing were determined in all groups (n=animal group numbers meeting the inclusion criteria are shown in brackets). a Topical control gabapentin gel alone without streptozotocin treatment (triangle, GG, n=8), streptozotocin pretreatment followed by gabapentin gel (10 %) (diamond, STZ + GG, n=10) or streptozotocin pretreatment followed by control gel (inverted triangle, STZ + CG, n=7) was applied on the plantar surface of the right hind paw of rats for three times daily and PWL was measured 1 h after application. Significance of differences in PWL is shown between topical saline (white circle, n=15) and streptozotocin pretreatment (white square, STZ, n=6, ###P<0.001) and between streptozotocin pretreatment (white square, STZ, n=6) versus gabapentin gel alone (triangle, GG, n=8, P<0.001), streptozotocin treatment followed by control gel, i.e. STZ + CG (inverted triangle, n=7; P<0.001) and streptozotocin treatment followed by gabapentin gel, i.e. STZ + GG (10 %) (diamond, n=10; P<0.001), ANOVA followed by Tukey’s post hoc test. b In the case of the systemic study, the same protocol as topical gabapentin was used but gabapentin (triangle, GP, 75 mg/kg) was administered i.p. and PWL was measured. Statistical significance of differences in PWL is shown between saline (white circle, n=15) and streptozotocin control (white square, STZ, n=6) (###P<0.001) and between streptozotocin control (white square, STZ, n=6) and gabapentin alone (triangle, GP, n= 8; P<0.05) and STZ + GP 75 mg/kg (inverted triangle, n = 5; *P<0.01) receiving streptozotocin pretreatment (P<0.05) which was not unexpected (Fig. 6a). In the systemic gabapentin study, mean von Frey filament pressures (g) calculated as FRT values (static vulvodynia) were plotted for treatment groups at the pooled 1 and 2-h intervals (F(3,20) =51.02, P<0.001) shown in Fig. 6b. There was a significant reduction of FRT in the group pretreated with 1135 Fig. 6 The static antivulvodynia effect of gabapentin gel (10 %) and systemic gabapentin (75 mg/kg i.p.) in streptozotocin-induced diabetic female rats. Scatter plots showing mean von Frey filament force flinching response thresholds (FRTs, g±SD) were determined in all groups (n=animal group numbers meeting the inclusion criteria are shown in brackets). a Topical control gabapentin gel (10 %) alone without streptozotocin pretreatment (triangle, GG, n = 8) or streptozotocin pretreatment with either control gel (inverted triangle, STZ + CG, n=13) or with gabapentin gel (diamond, STZ + GG, n=14) was applied on the anogenital area including mons pubis of rats for three times daily and responses were measured 1 h after application. Significance of differences in FRT is shown between topical saline control (white circle, n=15) and streptozotocin pretreatment (white square, STZ, n=8; #P<0.05) and between streptozotocin pretreatment (white square, STZ, n=8) versus gabapentin gel alone (triangle, GG, n=8;P<0.05), streptozotocin pretreatment followed by control gel (inverted triangle, STZ + CG, n=13; P>0.05) and streptozotocin pretreatment followed by gabapentin gel (diamond, STZ + GG, n=14; P<0.05) application, ANOVA followed by Tukey’s post hoc test. b In the systemic study, mean von Frey filament force flinching response thresholds (FRTs, g ± SD) were determined in all groups, and gabapentin (triangle, GP, 75 mg/kg, n=8) was administered i.p. The responses were measured post-treatment. Statistical significance of differences in FRT is shown between saline control (white circle, n=15) and streptozotocin pretreatment (white square, STZ, n=8; ###P<0.001) and between streptozotocin pretreatment (white square, STZ, n=8) and gabapentin alone (triangle, GP, n = 8; P < 0.05), and also from streptozotocin combined with gabapentin (STZ + GP, inverted triangle, n=6) post-drug treatment (P<0.001), ANOVA with Tukey’s post hoc analysis test streptozotocin (STZ) compared to the saline-administered group (P<0.001). This action was reversed in the group that received streptozotocin pretreatment plus ensuing gabapentin (STZ + GP, P<0.001). In addition, there was a difference in FRT between 1136 Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 the group administered gabapentin alone (GP) versus STZ pretreatment (P<0.05) and this is presented in Fig. 6b. Activity of topical gabapentin gel (10 %) and systemic gabapentin (75 mg/kg i.p) on dynamic vulvodynia in female diabetic rats As depicted in Fig. 7a, the FRL in seconds (dynamic vulvodynia) from the rat anogenital area (vulva) to light brushing was determined in the morning, noon and afternoon pooled groups and then graphed (F(4,35) =31.14, P<0.001). The latency values following saline control treatment were reduced in the streptozotocin-pretreated group (STZ) (P<0.001) and subsequently reversed by combined topical gabapentin gel application (STZ + GG, P<0.001). In the group treated with control gel along with streptozotocin pretreatment (STZ + CG), there was no difference noted in comparison with streptozotocin-alone pretreatment (STZ). Moreover, topical gabapentin gel by itself (GG) produced virtually no change in FRL compared to the saline controls, the latency values additionally being significantly greater than in animals pretreated with streptozotocin (P<0.001, Fig. 7a). In the systemic study, the FRL values (dynamic vulvodynia) were plotted for merged groups at 1 and 2-h treatment intervals (F(3,20) =51.02, P<0.001). The outcome revealed a significant difference between the latency following saline control treatment which was reduced by streptozotocin pretreatment (STZ) (P<0.001) and subsequently reversed by combined systemic gabapentin treatment (STZ + GP, 75 mg/kg, P<0.001) as shown in Fig. 7b. In addition, the gabapentin-alone treatment group (GP, 75 mg/kg) displayed greater FRL values (P<0.05) than the STZ-pretreated controls in the protocol (Fig. 7b). Discussion Diabetic neuropathy is a common and serious complication stemming from metabolic abnormalities (Elliott 1994) causing lesions of the peripheral and/or central nervous system leading to sensory signs and symptoms (Backonja 2003). It is characterized by progressive chronic neuropathic pain that is tingling and burning in nature with hyperesthesia and paresthesia with deep aching and it is increased by touch (Krause and Backonja 2003). It is greater at night (Boulton et al. 2005) and occurs in the feet and lower legs and may involve the hands (Vinik et al. 2000) affecting patient sleep, mood, self-esteem, social life (Said 2007) and ability to work (Vileikyte et al. 2003). Neuropathic pain is often associated with allodynia, which is a distinct feature and represents an abnormal pain to a stimulus that does not normally provoke pain (Merskey 1986). It has also been reported that small unmyelinated and large myelinated primary nerves (Aβ and small-diameter nociceptive Fig. 7 The dynamic antivulvodynia effect of gabapentin gel 10 % and systemic gabapentin (75 mg/kg i.p.) in streptozotocin-induced female diabetic rats. Scatter plots showing mean flinching response latency (FRL, s±SD) from the anogenital area (vulva) in response to light brushing determined in all groups (n=animal group numbers meeting the inclusion criteria are shown in brackets). a Topical control gabapentin gel alone (triangle, GG, n=8), streptozotocin pretreatment followed by gabapentin gel (10 %) (diamond, STZ + GG, n=8) or streptozotocin pretreatment followed by control gel (inverted triangle, STZ + CG, n=8) was applied to the anogenital area (vulva) of rats for three times daily and FRL was measured 1 h after application. Significance of differences in FRL is shown between topical saline control (white circle, n=15) versus streptozotocin pretreatment (white square, STZ, n = 8; ### P < 0.001), gabapentin gel alone versus streptozotocin pretreatment (GG, triangle, n= 8; P< 0.001) and streptozotocin pretreatment followed by gabapentin gel (STZ + GG, diamond, n=8, *P<0.001), ANOVA with post hoc Tukey’s test. b In the systemic study, the flinching response latency (FRL) was determined in each group using the same protocol as for gabapentin gel and this was plotted against treatment groups. The FRL was measured for gabapentin (triangle, GP, 75 mg/kg i.p.). ANOVA with Tukey’s post hoc analysis revealed significant differences in FRL between either the control gabapentin-alone group (triangle, GP, 75 mg/kg, n=8; P<0.05) or the saline controls (white circle, n=15) and streptozotocin controls (white square, STZ, n=8; ###P<0.001). There were subsequent increases in FRL caused by gabapentin following streptozotocin pretreatment (inverted triangle, STZ + GP, n=6; **P<0.001) fibres) are implicated in allodynia (Field et al. 1999). Vulvodynia has features which are characteristic of other chronic neuropathic pain conditions. These include the persistent and burning quality of the pain, the allodynia and hyperpathia, the absence of physical findings other than erythema on examination, and the patient’s obsession with the pain (Ben-David and Friedman 1999). Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 Experimental diabetes mellitus can be induced by most commonly employed diabetogenic agents, i.e. alloxan and streptozotocin. However, streptozotocin-induced diabetes mellitus is used as a model for hyperalgesia (Bishnoi et al. 2011). Therefore, in this study, a diabetic neuropathic pain model of allodynia and vulvodynia was employed involving a single injection of streptozotocin (50 mg/kg) in female rats. Since streptozotocin exhibits a stability problem (Rakieten et al. 1963), fresh solutions were prepared at the time of administration, and diabetes was confirmed by resultant degenerative changes in the pancreatic islets of Langerhans (Fig. 1). Early neuropathic symptoms in streptozotocin-induced diabetic rodents have been reported including allodynia and decreased nerve function velocity as well as axonal dystrophy following electron microscopical analysis (Cameron et al. 1994, 1999; Li et al. 2004; Weiss et al. 1995). Moreover, later symptoms of streptozotocin neuropathy, some of which are similarly insulin reversible, have been well documented and consist of hypoalgesia, large sensory nerve fibre pathology comprising demyelination, degeneration and Schwann cell damage and regeneration (Calcutt et al. 2004; Muller et al. 2008; Shaikh and Somani 2011; Vasconcelos et al. 2011; Weiss et al. 1995). Furthermore, in streptozotocin-treated rats, bladder afferent pudendal neurons located in the L6 and S1 dorsal root ganglia (McKenna and Nadelhaft 1986) have been reported to be smaller in cross-sectional area in streptozotocininduced diabetics than in normal animals (Steers et al. 1994) indicating axonopathy (Chen et al. 2013). The current management approach to vulvodynia is redolent of neuropathic pain treatment, usually involving adjunctive therapy since monotherapy often fails to yield an effective response. Pharmacological and non-pharmacological interventions include oral and topical medications, local and regional injections, cognitive therapy, physical therapy and surgery (Andrews 2010; Bates and Timmins 2002). Literature reports regarding evidence of efficacy for pain reduction by pharmacological intervention of generalized unprovoked vulvodynia vary from fair to poor. The majority of these reports lack high-quality evidence (Andrews 2010). However, adequate substantiation of efficacy has been reported for xylocaine (5 %), oral pregabalin, oral or topical gabapentin, oral duloxetine and selective serotonin reuptake inhibitors (SSRIs). Moreover, there is evidence of poor efficacy for topical capsaicin, nitroglycerin, oral tricyclic antidepressants or venlafaxine, pentosan polysulfate, opioids, tramadol, carbamazepine, lamotrigine, oxcarbazepine, topiramate and valproic acid (Andrews 2010; Fischer 2004; Phillips and Bachmann 2010). Based on these promising case reports and retrospective investigations previously performed on allodynia and vulvodynia, it was decided to conduct systemic and topical gabapentin studies in order to validate streptozotocin-induced diabetes as an animal model of vulvodynia. 1137 Arising from a pilot study, the streptozotocin-administered animals were kept under strict observation for 29 days to develop robust symptoms of static and dynamic allodynia and vulvodynia in the animal model. The inclusion criteria for animals in each group are depicted in Table 1. Static allodynia and vulvodynia were present in some of the animals within 15 days following streptozotocin injection, while both dynamic allodynia and vulvodynia were detected in most subjects within 29 days post-streptozotocin injection. Overall assessment at different times indicated that there were exclusions of animals at different stages up to protocol day 29 in order to achieve the criteria for the paradigm. The streptozotocintreated animals also showed significant loss in body weight (Fig. 3) with elevated blood glucose levels (i.e. above 270 mg/ dL) as compared to saline-vehicle-treated animals (Fig. 2). In order to test for allodynia and vulvodynia, in the topical gel study, the animals were divided into treatment groups: saline control, streptozotocin pretreatment control, control gabapentin gel alone without streptozotocin pretreatment, streptozotocin pretreatment plus control gel vehicle and streptozotocin pretreatment plus gabapentin gel (10 %). In the case of systemic gabapentin (75 mg/kg) for allodynia and vulvodynia, the animals were divided into saline control, streptozotocin pretreatment control, gabapentin alone without streptozotocin pretreatment and streptozotocin pretreatment plus gabapentin treatment groups. Responses were measured by application of von Frey filaments or light brushing with cotton buds. Streptozotocin itself induced explicit static and dynamic mechanical allodynia and vulvodynia in comparison with saline-treated controls. This was evidenced by significantly reduced von Frey filament force PWT, reduced PWL, vulval FRT and FRL. Control and gabapentin gels were applied in the morning, noon and afternoon on the hind paws and anogenital (vulva) area of animals, while gabapentin (75 mg/kg) was administered intraperitoneally (i.p.) for subsequent allodynia and vulvodynia testing, respectively. The results showed that topically applied gabapentin gel was able to ameliorate static and dynamic allodynia in the streptozotocin-pretreated animals (Suppl Figs. 4a and 5a). Systemic gabapentin administration showed similar outcomes by alleviating static and dynamic allodynia as compared to streptozotocin-pretreated controls (Suppl Figs. 4b and 5b). In the case of static and dynamic vulvodynia, there was a statistically significant reversal of streptozotocin-induced vulvodynia of both types by topical gabapentin gel application (Suppl Figs. 6a and 7a). Likewise, systemic gabapentin against static and dynamic vulvodynia displayed significant differences between the combined gabapentin/streptozotocin and streptozotocin-alone group values reflecting symptom reversal (Suppl Figs. 6b and 7b). Vulvodynia is a chronic condition; however, testing in the model was performed on day 29 in the protracted protocol and not on days 5 and 15 for the purpose of avoiding intervening repeated-test stress induction. 1138 Consequently, the persistent quality of vulvodynia during this interim period was not fully ascertained. Thus, systemic administration of gabapentin has provided some evidence with respect to validation of the current animal model for allodynia and vulvodynia since this agent has exhibited efficacy not only in animals (Carlton and Zhou 1998; Field et al. 1999) but also in the clinic (Boardman et al. 2008; Gorson et al. 1999). In the gabapentin topical study, there was an apparent antiallodynia effect of the control gel seen in this investigation. This might derive from the finding that streptozotocin induces epidermal thinning in the plantar surface of the rat hind paw (Kadic et al. 2014), and the occlusive coating quality of the xanthan gum hydrocolloid constituent (Marchitto and Flock 2013) along with the polyacrylamide in the gel may have acted as a physical barrier to mechanical stimuli applied to the weakened paw surface. This is somewhat analogous to the hydrocolloid gel in dressings which give rise to patients experiencing less pain and the need for less analgesia in the management of abrasive wounds (Heffernan and Martin 1994) or following skin shave biopsy (Nemeth et al. 1991). Conversely, the lack of alleviation of static and dynamic vulvodynia by the control gel may be a high-level sensory consequence of the density of ventral and dorsal sensory nerve branches emerging from the pudendal canal, respectively, to the perineum (superficial perineal nerve) and genitalia (terminal branch) (Robert et al. 1998). Furthermore, the inability of gabapentin gel to show any significant antiallodynia activity in comparison to control gel may be attributed to restricted permeation of gabapentin from the gel formulation through the paw skin. However, it is noteworthy that gabapentin delivered from control gel does penetrate human trunk skin (PCCA 2012) although this particular skin area is likely to be much thinner than that of the rat paw. Gabapentin has been indicated for the treatment of neuropathy though the exact mechanism through which it inhibits neuropathic pain still remains unclear. Evidence indicates that gabapentin operates by increasing the level of gamma aminobutaric acid (GABA) (Kocsis and Honmou 1994), by acting as a non-N-methyl-D-aspartate (non-NMDA) receptor antagonist (Chen et al. 2000; Kaneko et al. 2000) and by inhibiting the α2δ1 subunit of voltage-gated calcium channels (Gee et al. 1996; Shimoyama et al. 2000). Thus, gabapentin could possibly reduce neuropathic signs via several potential mechanisms. The efficacy for pain reduction by pharmacological intervention of generalized unprovoked vulvodynia varies from fair to poor (Andrews 2010). In this respect, the topical formulation of gabapentin may be a useful addition to the pharmacological treatment options available for neuropathic pain syndromes. The topical gabapentin 10 % gel preparation tested in this study is likely to possess minimal adverse effects compared to those usually associated with systemic use of gabapentin such as sedation, dizziness, somnolence, asthenia, Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 ataxia, amblyopia, headache and nausea (McLean et al. 2006). This may be ascribed to the lower systemic concentration of active ingredient delivered by the topical formulation than that yielded by oral administration. Therefore, topical use of gabapentin in the form of a gel could be a better alternative for the clinical management of neuropathic pain syndromes such as allodynia and vulvodynia. Conclusions In summary, these findings on the outcomes of topical and systemic gabapentin in the diabetic neuropathy model of vulvodynia tend to validate it as a useful non-clinical paradigm. Nonetheless, the complex nature of the diabetic neuropathic pain syndrome and the unpredictable rate of absorption of gabapentin from the gel formulation warrant further research to correlate any antivulvodynia activity of gabapentin gel with gabapentin penetration after topical application. The antivulvodynia action of gabapentin gel in the current model also accords with retrospective clinical studies which concluded that topical gabapentin is well tolerated and associated with significant pain relief in women with vulvodynia (Andrews 2010; Boardman et al. 2008). Acknowledgments Gabapentin gel 10 % w/w and control gel (base minus the API) were supplied by SMPU (Cardiff, UK) under their MS. We also acknowledge the gift of gabapentin by MKB Pharmaceuticals (Pvt.) Ltd., Peshawar, Pakistan. Conflict of interest The authors have no conflicts of interest to declare. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. References Alba-Delgado C, Mico JA, Sánchez-Blázquez P, Berrocoso E (2012) Analgesic antidepressants promote the responsiveness of locus coeruleus neurons to noxious stimulation: implications for neuropathic pain. Pain 153:1438–1449 Andrews JC (2010) Vulvodynia: an evidence-based approach to medical management. J Clin Outcomes Manag 17:225–238 Apkarian AV, Baliki MN, Geha PY (2009) Towards a theory of chronic pain. Prog Neurobiol 87:81–97 Babu V, Gangadevi T, Subramoniam A (2003) Antidiabetic activity of ethanol extract of Cassia kleinii leaf in streptozotocin-induced diabetic rats and isolation of an active fraction and toxicity evaluation of the extract. Indian J Pharm Sci 35:290–296 Bachmann GA, Rosen R, Pinn VW, Utian WH, Ayers C, Basson R, Binik YM, Brown C, Foster DC, Gibbons J Jr, Goldstein I, Graziottin A, Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 Haefner HK, Harlow BL, Spadt SK, Leiblum SR, Masheb RM, Reed BD, Sobel JD, Veasley C, Wesselmann U, Witkin SS (2006) Vulvodynia: a state-of-the-art consensus on definitions, diagnosis and management. J Reprod Med 51:447–456 Backonja MM (2003) Defining neuropathic pain. Anesth Analg 97:785– 790 Backonja M, Beydoun A, Edwards KR, Schwartz SL, Fonseca V, Hes M, LaMoreaux L, Garofalo E (1998) Gabapentin for the symptomatic treatment of painful neuropathy in patients with diabetes mellitus. JAMA 280:1831–1836 Barrueto F, Green J, Howland MA, Hoffman RS, Nelson LS (2002) Gabapentin withdrawal presenting as status epilepticus. J Toxicol Clin Toxicol 40:925–928 Bates CM, Timmins DJ (2002) Vulvodynia - new and more effective approaches to therapy. Int J STD AIDS 13:210–212 Ben-David B, Friedman M (1999) Gabapentin therapy for vulvodynia. Anesth Analg 89:1459–1460 Bennett MI, Simpson KH (2004) Gabapentin in the treatment of neuropathic pain. Palliat Med 18:5–11 Bishnoi M, Bosgraaf CA, Abooj M, Zhong L, Premkumar LS (2011) Streptozotocin-induced early thermal hyperalgesia is independent of glycemic state of rats: role of transient receptor potential vanilloid 1 (TRPV1) and inflammatory mediators. Mol Pain 7:52–74 Boardman LA, Cooper AS, Blais LR, Raker CA (2008) Topical gabapentin in the treatment of localized and generalized vulvodynia. Obstet Gynecol 112:579–585 Boulton AJM, Vinik AI, Arezzo JC, Bril V, Feldman EL, Freeman R, Malik RA, Maser RE, Sosenko JM, Ziegler D (2005) Diabetic neuropathies: a statement by the American Diabetes Association. Diabetes Care 28:956–962 Calcutt NA, Chaplan SR (2009) Spinal pharmacology of tactile allodynia in diabetic rats. Br J Pharmacol 122:1478–1482 Calcutt NA, Freshwater JD, Mizisin AP (2004) Prevention of sensory disorders in diabetic Sprague-Dawley rats by aldose reductase inhibition or treatment with ciliary neurotrophic factor. Diabetologia 47: 718–724 Cameron NE, Cotter MA, Archibald V, Dines KC, Maxfield EK (1994) Anti-oxidant and pro-oxidant effect on nerve conduction velocity, endoneural blood flow and oxygen tension in non-diabetic and streptozotocin-diabetic rats. Diabetologia 37:449–459 Cameron NE, Cotter MA, Jack AM, Basso MD, Hohman TC (1999) Protein kinase C effect on nerve function, perfusion, Na(+), K(+)ATPase activity and glutathione content in diabetic rat. Diabetologia 42:1120–1130 Campbell JN, Meyer RA (2006) Mechanisms of neuropathic pain. Neuron 52:77–92 Carlton SM, Zhou S (1998) Attenuation of formalin-induced nociceptive behaviors following local peripheral injection of gabapentin. Pain 76:201–207 Cesena R, Calcutt N (1999) Gabapentin prevents hyperalgesia during the formalin test in diabetic rats. Neurosci Lett 262:101–104 Chaplan S, Bach F, Pogrel J, Chung J, Yaksh T (1994) Quantitative assessment of tactile allodynia in the rat paw. J Neurosci Methods 53:55–63 Chen SR, Eisenach JC, McCaslin PP, Pan HL (2000) Synergistic effect between intrathecal non-NMDA antagonist and gabapentin on allodynia induced by spinal nerve ligation in rats. Anesthesiology 92:500–506 Chen S-C, Lai C-H, Fan W-J, Peng C-W (2013) Pudendal neuromodulation improves voiding efficiency in diabetic rats. Neurourol Urodyn 32:293–300 Costigan M, Scholz J, Woolf CJ (2009) Neuropathic pain: a maladaptive response of the nervous system to damage. Annu Rev Neurosci 32: 1–32 Elliott K (1994) Taxonomy and mechanisms of neuropathic pain. Semin Neurol 14:195–205 1139 Falkeholm L, Grant CA, Magnusson A, Möller E (2001) Xylene-free method for histological preparation: a multicentre evaluation. Lab Investig 81:1213–1221 Field JM, McCleary S, Hughes J, Singh L (1999) Gabapentin and pregabalin, but not morphine and amitriptyline, block both static and dynamic components of mechanical allodynia induced by streptozocin in the rat. Pain 80:391–398 Fischer G (2004) Management of vulvar pain. Dermatol Ther 17:134– 149 Gee NS, Brown JP, Dissanayake VUK, Offord J, Thurlow R, Woodruff GN (1996) The novel anticonvulsant drug, gabapentin (Neurontin), binds to the subunit of a calcium channel. J Biol Chem 271:5768– 5776 Gilron I (2007) Gabapentin and pregabalin for chronic neuropathic and early postsurgical pain: current evidence and future directions. Curr Opin Anesthesiol 20:456–472 Gilron I, Bailey JM, Tu D, Holden RR, Weaver DF, Houlden RL (2005) Morphine, gabapentin, or their combination for neuropathic pain. N Engl J Med 352:1324–1334 Goodnick PJ, Mendosa L, Kumar A, Freund B, DeVan L (2000) Sertraline in diabetic neuropathy: response and biology. Psychosom Med 62:461–462 Gorson KC, Schott C, Herman R, Ropper AH, Rand WM (1999) Gabapentin in the treatment of painful diabetic neuropathy: a placebo controlled, double blind, crossover trial. J Neurol Neurosurg Psychiatry 66:251–252 Heffernan A, Martin AJ (1994) A comparison of a modified form of Granuflex® (Granuflexg Extra Thin) and a conventional dressing in the management of lacerations, abrasions and minor operation wounds in an accident and emergency department. J Accid Emerg 11:227–230 Hunter JC, Gogas KR, Hedley LR, Jacobson LO, Kassotakis L, Thompson J, Fontana DJ (1997) The effect of novel anti-epileptic drugs in rat experimental models of acute and chronic pain. Eur J Pharmacol 324:153–160 Indolfi C, Torella D, Cavuto L, Davalli AM, Coppola C, Esposito G, Carrierro MV, Rapacciuolo A, Di Lorenzo E, Stabile E, Perrino C, Chieffo A, Pardo F, Chiariello M (2001) Effects of balloon injury on neointimal hyperplasia in streptozotocin-induced diabetes and in hyperinsulinemic nondiabetic pancreatic islet-transplanted rats. Circulation 103:2980–2986 James J, Page J (1994) Painful diabetic peripheral neuropathy. A stepwise approach to treatment. J Am Podiatr Med Assoc 84:439–447 Jolivalt CG, Ramos KM, Herbetsson K, Esch FS, Calcutt NA (2006) Therapeutic efficacy of prosaposin-derived peptide on different models of allodynia. Pain 121:14–21 Kadic AJ, Boric M, Vidak M, Ferhatovik L, Puljak L (2014) Changes in epidermal thickness and cutaneous innervation during maturation in long-term diabetes. J Tissue Viability 23:7–12 Kalra B, Kalra S, Bajaj S (2013) Vulvodynia: an unrecognised neuropathic syndrome. Indian J Endocrinol Metab 17:787–789 Kaneko M, Mestre C, Sánchez EH, Hammond DL (2000) Intrathecally administered gabapentin inhibits formalin-evoked nociception and the expression of Fos-like immunoreactivity in the spinal cord of the rat. J Pharmacol Exp Ther 292:743–751 Kocsis JD, Honmou O (1994) Gabapentin increases GABA-induced depolarization in rat neonatal optic nerve. Neurosci Lett 169:181–184 Krause SJ, Backonja MM (2003) Development of a neuropathic pain questionnaire. Clin J Pain 19:306–314 Leo RJ (2013) A systematic review of the utility of anticonvulsant pharmacotherapy in the treatment of vulvodynia pain. J Sex Med 10: 2000–2008 Li F, Szabo C, Pacher P, Southan GJ, Abatan OI, Chaniauskaya T, Stevens MJ, Obrosova IG (2004) Evaluation of orally active poly(ADP-ribose) polymerase inhibitor in streptozotocin-diabetic rat model of early peripheral neuropathy. Diabetologia 47:710–717 1140 Lynch P (1986) Vulvodynia: a syndrome of unexplained vulvar pain, psychologic disability and sexual dysfunction. J Reprod Med 31: 773–780 Mao J, Chen LL (2000) Gabapentin in pain management. Anesth Analg 91:680–687 Marchitto KS, Flock ST (2013) Adhesive laminates for rapid wound occlusion. http://www.google.com/patents/US8343606 [accessed 19th Feb 2015] McKenna KE, Nadelhaft I (1986) The organization of the pudendal nerve in the male and female rat. J Comp Neurol 248:532–49 McLean MJ, Morrell MJ, Willmore LJ, Privitera MD, Faught RE, Holmes GL, Magnus‐Miller L, Bernstein P, Rose‐Legatt A (2006) Safety and tolerability of gabapentin as adjunctive therapy in a large, multicenter study. Epilepsia 40:965–972 Merskey HE (1986) Classification of chronic pain: descriptions of chronic pain syndromes and definitions of pain terms. Prepared by the International Association for the Study of Pain, Subcommittee on Taxonomy. Pain Suppl 3:S1–226 Muller KA, Ryals JM, Feldman EL, Wright DE (2008) Abnormal muscle spindle innervation and large-fiber neuropathy in diabetic mice. Diabetes 57:1693–1701 Nemeth AJ, Eaglstein WH, Taylor JR, Peerson LJ, Falanga V (1991) Faster healing and less pain in skin biopsy sites treated with an occlusive dressing. Arch Dermatol 127:1679–1683 Paavonen J (1995) Vulvodynia-a complex syndrome of vulvar pain. Acta Obstet Gynecol Scand 74:243–247 PCCA (Professional Compounding Centres of America) 2012. Evaluation of the percutaneous absorption of ketamine HCl, gabapentin, clonidine HCl and baclofen in Lipoderm® and Lipoderm® ActiveMax™, into human trunk skin, in vitro, using the Franz skin finite dose model. http://rxpro.net/wp-content/ uploads/2014/07/LIPODERM-AND-ACTIVEMAX-WHITEPAPER.pdf. Accessed 18 Feb 2015. Phillips N, Bachmann G (2010) Vulvodynia: an often-overlooked cause of dyspareunia in the menopausal population. Menopausal Med 18: S1–S5 Rakieten N, Rakieten ML, Nadkarni MV (1963) Studies on the diabetogenic action of streptozotocin (NSC-37917). Cancer Chemother Rep 29:91–98 Robert R, Prat-Pradal D, Labat JJ, Bensignor M, Raoul S, Rebai R, Leborgne J (1998) Anatomic basis of chronic perineal pain: role of the pudendal nerve. Surg Radiol Anat 20:93–98 Naunyn-Schmiedeberg's Arch Pharmacol (2015) 388:1129–1140 Said G (2007) Diabetic neuropathy - a review. Nat Clin Pract Neurol 3: 331–340 Schønemann NK, van der Burght M, Arendt-Nielsen L, Bjerring P (1992) Onset and duration of hypoalgesia of lidocaine spray applied to oral mucosa - a dose response study. Acta Anaesthesiol Scand 36:733– 735 Shaikh AS, Somani RS (2011) Animal models and biomarkers of neuropathy in diabetic rodents. Indian J Pharm 42:129–134 Shaw JE, Sicree RA, Zimmet PZ (2010) Global estimates of the prevalence of diabetes for 2010 and 2030. Diabetes Res Clin Pract 87:4– 14 Shimoyama M, Shimoyama N, Hori Y (2000) Gabapentin affects glutamatergic excitatory neurotransmission in the rat dorsal horn. Pain 85:405–414 Steers WD, Mackway-Gerardi AM, Ciambotti J, de Groat WC (1994) Alterations in neural pathways to the urinary bladder of the rat in response to streptozotocin-induced diabetes. J Auton Nerv Syst 47: 83–94 Vasconcelos CCA, Fazan SPV, Moore KC, Nessler RA, Valença MM (2011) Transmission electron microscopy studies of the vestibulocochlear nerve in chronic diabetic rats. Int J Morphol 29: 272–277 Vileikyte L, Peyrot M, Bundy C, Rubin RR, Leventhal H, Mora P, Shaw JE, Baker P, Boulton AJ (2003) The development and validation of a neuropathy-and foot ulcer-specific quality of life instrument. Diabetes Care 26:2549–2555 Vinik A, Park T, Stansberry K, Pittenger G (2000) Diabetic neuropathies. Diabetologia 43:957–973 Wei M, Ong L, Smith MT, Ross FB, Schmid K, Hoey AJ, Burstow D, Brown L (2003) The streptozotocin-diabetic rat as a model of the chronic complications of human diabetes. Heart Lung Circ 12:44– 50 Weiss J, Dimpfel W, Schröder JM (1995) Nerve conduction changes and fine structural alterations of extra- and intrafusal muscle and nerve fibers in streptozotocin diabetic rats. Muscle Nerve 18:175–184 Wiffen PJ, McQuay HJ, Edwards JE, Moore RA (2005) Gabapentin for acute and chronic pain. Cochrane Database Syst Rev 3, CD005452 Young AW Jr, Azoury R, McKay M (1984) Burning vulva syndrome: report of the ISSVD task force. J Reprod Med 29:457 Zhang J-L, Yang J-P, Zhang J-R, Li R-Q, Wang J, Jan J-J, Zhuang Q (2013) Gabapentin reduces allodynia and hyperalgesia in painful diabetic neuropathy rats by decreasing expression level of Nav1.7 and p-ERK1/2 in DRG neurons. Brain Res 1493:13–18
https://openalex.org/W2049223018	https://europepmc.org/articles/pmc4106794?pdf=render	English	null	Survivin as a Potential Mediator to Support Autoreactive Cell Survival in Myasthenia Gravis: A Human and Animal Model Study	PloS one	2,014	cc-by	9,211	Abstract They hold the patent to SVN53-67/M57-KLH, the vaccine used in the study. They are co-founders of MimiVax, LLC which was developed with the purpose of commercializing the SVN53-67/M57-KLH. This does not alter the authors’ adherence to PLOS ONE policies on sharing data and materials. Competing Interests: The authors acknowledge that the authors Michael Ciesielski and Robert Fenstermaker have a con manuscript. They hold the patent to SVN53-67/M57-KLH, the vaccine used in the study. They are co-founders of MimiVax, LLC purpose of commercializing the SVN53-67/M57-KLH. This does not alter the authors’ adherence to PLOS ONE policies on sharin * Email: lkusner@gwu.edu expressed during fetal development and is absent in normal adult tissue [4]. It interacts with the mitotic spindle and regulates cell division [5–7]. Furthermore, survivin’s role as an anti-apoptotic protein is to modulates the function of a number of terminal effector cell death proteases (caspases) leading to inhibition of apoptosis [8–10]. These and other actions may account survivin’s role in preventing the death of malignant cells [11]. Survivin has been implicated in the pathophysiology of certain autoimmune disorders [12,13]. Survivin expression is significantly higher in the synovial tissue of rheumatoid arthritis (RA) patients with destructive disease and lower among patients receiving therapy with disease-modifying drugs, [13,14] making in situ survivin expression an independent prognostic parameter for erosive RA. Circulating survivin is positively correlated with severity of juvenile RA [15]. Therefore, survivin expression appears to predict the clinical course of RA [16]; whereas, anti-survivin antibodies (a proxy for the anti-survivin immune response) are associated with less severe disease [13]. Survivin levels are also elevated in T cells and brain tissue of patients with primary progressive multiple sclerosis (MS), suggesting a distinct pathological function of Survivin as a Potential Mediator to Support Autoreactive Cell Survival in Myasthenia Gravis: A Human and Animal Model Study Linda L. Kusner1, Michael J. Ciesielski3,4, Alexander Marx5, Henry J. Kaminski2, Robert A. Fenstermaker3,4 1 Department of Pharmacology and Physiology, George Washington University, Washington, District of Columbia, United States of America, 2 Department of Neurology, George Washington University, Washington, District of Columbia, United States of America, 3 Department of Neurosurgery, Roswell Park Cancer Institute, Buffalo, New York, United States of America, 4 Center for Immunotherapy, Roswell Park Cancer Institute, Buffalo, New York, United States of America, 5 Institute of Pathology, University Medical Centre Mannheim, University of Heidelberg, Mannheim, Germany 1 Department of Pharmacology and Physiology, George Washington University, Washington, District of Columbia, United States of America, 2 Department of Neurology, George Washington University, Washington, District of Columbia, United States of America, 3 Department of Neurosurgery, Roswell Park Cancer Institute, Buffalo, New York, United States of America, 4 Center for Immunotherapy, Roswell Park Cancer Institute, Buffalo, New York, United States of America, 5 Institute of Pathology, University Medical Centre Mannheim, University of Heidelberg, Mannheim, Germany Abstract The mechanisms that underlie the development and maintenance of autoimmunity in myasthenia gravis are poorly understood. In this investigation, we evaluate the role of survivin, a member of the inhibitor of apoptosis protein family, in humans and in two animal models. We identified survivin expression in cells with B lymphocyte and plasma cells markers, and in the thymuses of patients with myasthenia gravis. A portion of survivin-expressing cells specifically bound a peptide derived from the alpha subunit of acetylcholine receptor indicating that they recognize the peptide. Thymuses of patients with myasthenia gravis had large numbers of survivin-positive cells with fewer cells in the thymuses of corticosteroid- treated patients. Application of a survivin vaccination strategy in mouse and rat models of myasthenia gravis demonstrated improved motor assessment, a reduction in acetylcholine receptor specific autoantibodies, and a retention of acetylcholine receptor at the neuromuscular junction, associated with marked reduction of survivin-expressing circulating CD20+ cells. These data strongly suggest that survivin expression in cells with lymphocyte and plasma cell markers occurs in patients with myasthenia gravis and in two animal models of myasthenia gravis. Survivin expression may be part of a mechanism that inhibits the apoptosis of autoreactive B cells in myasthenia gravis and other autoimmune disorders. Citation: Kusner LL, Ciesielski MJ, Marx A, Kaminski HJ, Fenstermaker RA (2014) Survivin as a Potential Mediator to Support Autoreactive Cell Survival in Myasthenia Gravis: A Human and Animal Model Study. PLoS ONE 9(7): e102231. doi:10.1371/journal.pone.0102231 Editor: William Phillips, University of Sydney, Australia Received March 11, 2014; Accepted June 16, 2014; Published July 22, 2014 Received March 11, 2014; Accepted June 16, 2014; Published July 22, 2014 Copyright: 2014 Kusner et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. Relevant data are included within the paper and its Supporting Information files. nding: This work was supported by Roswell Park Cancer Institute, Cancer Center Support Grant; NIH-CCSG-CA016056. The funders had n ta collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors acknowledge that the authors Michael Ciesielski and Robert Fenstermaker have a conflict of interest based on this manuscript. July 2014 \| Volume 9 \| Issue 7 \| e102231 Citation: Kusner LL, Ciesielski MJ, Marx A, Kaminski HJ, Fenstermaker RA (2014) Survivin as a Potential Mediator to Support Autoreactive Cell Survival in Myasthenia Gravis: A Human and Animal Model Study. PLoS ONE 9(7): e102231. doi:10.1371/journal.pone.0102231 Introduction A central question in the study of autoimmune disorders concerns the mechanisms that maintain the persistence of autoreactive cells that produce disease. Myasthenia gravis (MG) is a T cell dependent, B cell mediated autoimmune disease in which antibodies are directed against proteins concentrated at the neuromuscular junction (NMJ), primarily the nicotinic acetylcho- line receptor (AChR) [1]. The AChR antibody synthesis is dependent on autoantigen specific B cells. We hypothesized that certain inhibitor of apoptosis proteins (IAPs) that are expressed in neoplastic diseases might also support the pathologic survival of autoreactive immune cells. IAPs are a family of proteins that contain one or more baculovirus repeat (BIR) domains originally found to influence apoptosis and restrict the activation of caspases, preventing the cell from undergoing cell death, specifically in cancers [2]. A prime candidate of the maintenance of autoreactive immune cells is survivin, a 16.5 kDa intracellular protein that belongs to the IAP family and contains one BIR domain [3]. Survivin is highly July 2014 \| Volume 9 \| Issue 7 \| e102231 July 2014 \| Volume 9 \| Issue 7 \| e102231 1 PLOS ONE \| www.plosone.org Survivin Expression in Myasthenia Gravis surviving [17]. Mitogen-stimulated T lymphocytes from patients with active MS display high-level expression of survivin compared to patients with stable disease [18]. These observations indicate that survivin expression may be a common characteristic of autoimmunity and that it is likely to play a direct pathophysiologic role. repetitive decrement on repeated stimulation EMG, or single fiber examination demonstrating a neuromuscular defect, or presence of serum acetylcholine receptor, or muscle specific kinase antibodies; and, 4) willingness to participate and ability to provide informed consent. Exclusion criterion was limited to inability to provide informed consent. The MG Foundation of America Clinical Research Standards were applied for subject classification [28]. Control subject inclusion criteria were limited to willingness to participate and ability to provide informed consent. Control subject exclusion criteria were age #18 years of age and treatment with prednisone or any immunosuppressive or immune modulator in the previous 12 months. All participants provided written consent for inclusion in the study. The study was approved by the George Washington University Institutional Review Board.(Table 1) Deidentified thymic samples were obtained and analyzed by the Institute of Pathology, University Medical Centre Mannheim with the approval (#2013-802RA-MA) of the Ethics Committee of the University. Analysis of MG patients and controls by flow cytometry (FACS) ( ) Whole blood samples were used to ensure inclusion of all potential immune cells that may express survivin. Blood was diluted in cell staining buffer containing flow cytometry staining buffer (PBS/1% FCS, cat 420201 BioLegend, San Diego, CA) and TruStain fcX-Blocking reagent (cat 422302, BioLegend). Specific direct-labeled antibodies to cell surface antigens: CD8 (clone SFCI21Thy2D3 (T8) FITC, Beckman-Coulter, Brea, CA); CD20 (clone 2H7 PE and/or PE/Cy7, BioLegend); CD27 (clone M- T271 PE, BioLegend); CD38 (clone HIT-2 APC/Cy7, BioLe- gend); CD138 (clone DL-101 PerCP/Cy5.5, BioLegend) was added. After washes in PBS, cells were incubated in RBC lysis combination 1% paraformaldehyde fixation buffer (cat 422401 BioLegend). Following primary fixation cells were resuspended in permeabilization buffer (cat 421002 BioLegend). Survivin anti- body (clone 60.11 DyLight650, Novus Biologicals, Littleton, CO) was added. Cells were washed in permeabilization buffer (cat 421002 BioLegend) followed with a final PBS wash and resuspension in Fluorofix fixation buffer (cat 422101, BioLegend). Sample acquisition was obtained on a Core facility BD LSR Fortessa 3 laser flow cytometer with 11 color detection and separate interrogation of sample for each laser. Samples were acquired with FACSDiva software followed with analysis using FCS Express software. Analysis was based upon isolated gating of lymphocyte populations and co-localization of survivin with specific CD markers as indicated. Experiments utilizing Acetyl- choline Receptor Peptide (AChR) were performed as above with the addition of FITC-labeled recombinant AChR alpha 1 subunit (H00001134-Q01, peptide: 146–232 amino acid, Novus Biolog- icals, Littleton, CO) to the cell surface label cocktail. We chose MG as a disease to begin to address whether survivin is involved in the pathophysiology of antibody-mediated autoim- munity. MG is among the best characterized antibody-mediated diseases with the primary antigenic targets being neuromuscular junction (NMJ) proteins, primarily the nicotinic acetylcholine receptor (AChR) [1,24]. The antibodies produce a reduction of AChR number and damage the muscle endplate, leading to a failure of neuromuscular transmission with resulting muscle weakness. The AChR antibodies of patients are polyclonal and recognize a complex repertoire of epitopes that differ among individual patients. Experimentally acquired myasthenia gravis (EAMG) produced by administration of purified AChR to rodents has similar immune characteristics to the human disease [25,26]. Introduction Survivin is expressed in many cancers and a number of studies have demonstrated that cancer patients have anti-survivin antibodies as well as T cells capable of recognizing survivin epitopes via interaction with MHC class I molecules on tumor cells. Cytotoxic T lymphocytes (CTL) recognize the MHC-class I molecule in complex with survivin epitopes eliciting a cytotoxic antitumor response [19]. Anti-tumor vaccine strategies have been used successfully to target tumor cells that express such cell-surface complexes [20]. One such strategy has employed a survivin peptide mimic (SVN53-67/M57), which is naturally processed into epitopes that are presented by human MHC class I molecules, as well as by murine H2-Kb molecules (as a model of the human response) [21–23]. The SVN53-67/M57 peptide mimic contains a core CTL epitope SVN56-64 modified by substitution of methionine for cysteine at amino acid position 57. This core epitope binds to HLA-A02 molecules with higher affinity than the wild type molecule producing a potent immune response which is cross-reactive to the wild type molecule [22]. Conjugation of the peptide to Keyhole Limpet Hemocyanin (KLH) to form SVN53- 67/M57-KLH provides enhanced stimulation of the immune system in vivo. The SVN53-67/M57-KLH vaccine produces CTL-mediated killing of many different survivin-expressing cancer cell types, including glioma and B cell lymphoma [21]. SVN53- 67/M57-KLH is currently in clinical trials as a cancer immuno- therapeutic. Survivin staining in the thymus g y Eleven thymuses from MG patients were analyzed for survivin expression. All patients were AChR antibody positive. The treated group (5 with immunosuppressive agents and 2 with prednisolone alone) consisted of 6 women, 1 man, age range 18–42 years, mean 27 years). The untreated group had received no immunosuppres- sives or corticosteroids and consisted of 3 women, 1 man, age range 13–27 years, mean 22 years. Duration of symptoms prior to surgery was not documented for the majority of patients. Survivin expression was detected in neutral, buffered, formalin-fixed, paraffin-embedded human thymic tissue by immunohistochemis- try as described [29] using the following reagents and methods: antigen retrieval in Novocastra antigen retrieval solution pH 9.0 Analysis of MG patients and controls by flow cytometry (FACS) Studies that highlight the mechanisms involved in driving the autoimmune process have largely focused on the complexity of the breakdown of tolerance [27] but have not addressed in detail the mechanisms by which autoreactive B cells may escape pro- grammed cell death. In the present investigation, we assessed survivin expression in various subsets of peripheral blood mononuclear cells (PBMC) of patients with MG. Having identified survivin expression in cells that co-express certain plasma cell and B cell phenotypic markers, we explored the effects of specific survivin peptide vaccination on moderation of active EAMG in two rodent models. Methods Ethics statement for human subjects July 2014 \| Volume 9 \| Issue 7 \| e102231 July 2014 \| Volume 9 \| Issue 7 \| e102231 Ethics statement for animal use Six- to eight-week-old male C57BL/6J mice (Jackson Labora- tories, Bar Harbor, ME) and 100–124 gram female Lewis rats (Harlan, Indianapolis, IN) were used for the study. Animals were maintained in the George Washington University animal facility. The GWU animal facility follows IACUC, AAALAT, and AALAS standards concerning appropriate housing, cage cleaning proce- dure, air purity, feed, temperature, humidity, light and dark cycle. Animals were housed in isolator cages in a pathogen-free environment, and rodent chow and water were provided ad libitum. A veterinarian is on staff and observed the health of the animals throughout the study. All animal studies were conducted according to protocol approved by the George Washington University Institutional Animal Care and Use Committee (Permit No. A247). We used the reporting standards set by the NINDS for all outcome measurements [30]. Direct H2-Kb Tetramer Binding Assay MHC class I tetramers specific to survivin peptide epitopes were used to assess the presence of survivin-specific T cells generated by survivin vaccine (SurVaxM) stimulation, as described above. PBMC from vaccinated animals were isolated following immuni- zation with SurVaxM. Becton Dickinson MHC class I tetramers (Chicago, IL) were designed to recognize T cell receptors that bind the core CTL epitope of SurVaxM [21]. Survivin-specific R-PE- labeled tetramers were incubated with PBMC for 10 minutes at 25uC. Samples were incubated with anti-CD8 FITC (clone T8, Beckman Coulter, Brea, CA) for 20 minutes at 4uC. Cells were fixed for analysis by flow cytometry. CD8+ cells that were doubly labeled were considered positive for tetramer, ovalbumin loaded tetramers were utilized as a negative control. Data were analyzed using FCS Express software. Analysis was based upon gating on CD8+ T cells only. Isolation of splenic cells from SVN53-67/M57-KLH and PBS treated EAMG mice Isolation of splenic cells from SVN53-67/M57-KLH and PBS treated EAMG mice Spleens were harvested from EAMG induced animals and cells were teased from mouse spleens with sterile forceps and passed through a 70 mm filters (BD Falcon) in DMEM. Splenocytes were centrifuged and the pellet was re-suspended in RBC lysis buffer (R&D Systems). Cells were washed twice and re-suspended in complete RPMI culture medium. Splenocytes were frozen in DMEM containing 10% fetal bovine serum and 10% DMSO [33]. Ethics statement for human subjects Blood specimens were collected from patients of the Depart- ment of Neurology at George Washington University. For patients with MG, entrance criteria for participation were: 1) clinical diagnosis of MG; 2) age $18 years of age; 3) confirmatory diagnostic test including positive tensilon, or ice pack test, or PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e102231 2 Survivin Expression in Myasthenia Gravis able 1. Demographic and treatment characteristics of myasthenic patients (1–15) and controls (16–25) consisting of individuals with non-myasthenic neurologic conditions. tient Sex Age MGFA class (Diag/Draw) AChR level nmol/L Therapy Other conditions Symptom Duration F 78 2a/2a 10 prednisone - 3 yrs F 19 1/1 2.7 none - 6 mos M 62 1/1 11.7 none - ND F 42 1/1 6.9 none - 2 mos F 20 2/2 3.76 prednisone,pyridostigmine - 4 mos M 62 2b/2b 0.86 prednisone,pyridostigmine - 4 mos F 20 3/remission 30 prednisone - 4 yrs F 74 1/remission 68 prednisone - 7 mos M 61 2a/2a 5.09 prednisone, cellcept - 4 mos M 61 1/1 1.9 prednisone - 1.5 mos M 57 1/remission 22.6 prednisone - 6 mos F 53 2/remission 33/0 prednisone, cellcept Type I DM 3 yrs M 66 2b/2b 10 none - 2.5 yrs M 55 1/1 9.5 pyridostigmine - 2 mos F 57 2b/2b 0 none Rheumatoid Arthritis 9 mos M 53 ND none Uveitis F 48 ND none Hypothyroid F 38 ND none - F 34 ND none Cervical Spondylosis F 35 ND none - M 38 ND none Tremor M 54 ND none Neck Pain M 54 ND none Meningioma F 24 ND none Migraine F 70 ND none Hearing Problems :10 1371/journal pone 0102231 t001 July 2014 \| Volume 9 \| Issue 7 \| e102231 3 3 Survivin Expression in Myasthenia Gravis (Leica, RE7119); anti-survivin monoclonal rabbit antibody (clone EP2880Y) (ab76424, Abcam Inc., Cambridge, MA, USA); blocking of endogenous peroxidase (DAKO blocking solution, S2013); detection of bound anti-survivin antibodies by the immunoperoxidase/DAB-based DAKO REAL detection system (K5007, DAKO GmbH, Hamburg, Germany). (Leica, RE7119); anti-survivin monoclonal rabbit antibody (clone EP2880Y) (ab76424, Abcam Inc., Cambridge, MA, USA); blocking of endogenous peroxidase (DAKO blocking solution, S2013); detection of bound anti-survivin antibodies by the immunoperoxidase/DAB-based DAKO REAL detection system (K5007, DAKO GmbH, Hamburg, Germany). Post-treatment with SVN53-67/M57-KLH in active EAMG Post-treatment with SVN53-67/M57-KLH in active EAMG AChR was purified from the electric organs of the Torpedo californica by affinity chromatography [31]. Nineteen male C57BL/6J mice were injected subcutaneously (s.c.), two injections at the shoulders and two injections at the base of tail, with Torpedo AChR (tAChR, 40 mg) in complete Freunds’ adjuvant (Sigma) on day 0, day 24, and 52. Three s.c. injections of 100 mg SVN53-67/ M57-KLH (nine mice) or PBS (ten mice) in incomplete Freunds’ adjuvant (IFA, Sigma) were given at weekly intervals starting on day 42. The last two injections of SVN53-67/M57-KLH were given on days 59 and 73. Mice were evaluated in a blinded fashion biweekly using the generally accepted motor strength scale [32]. Clinical severity scale (1–4) was used: 0 = no weakness observed, 1 = slight weakness after exercise, 2 = initial weakness, 3 = weakness plus respiratory difficulty, and 4 = moribund. Grip strength and body weight measurements were assessed. A grip strength meter (Columbus Instruments, Columbus, OH) with digital force gauge (Ametek, Largo, FL) was used to assess the peak force by which an animal grasped a grid pull bar. Each mouse was exercised before measurements were taken. Mean grip strength (N) from five pulls for each mouse was used for analysis. Grip strength measurements were given as a mean value for each group/week. Animals were sacrificed on day 80. Analysis of splenocyte populations for survivin expression p Specific direct labeled antibody to cell surface antigen CD20 (BioLegend, San Diego, CA) was used to assess splenocyte population. Following primary fixation in Fluorofix fixation buffer (cat 422101, BioLegend) cells were resuspended in permeabiliza- tion buffer (cat 421002, BioLegend). Survivin antibody (clone 60.11 FITC, Novus Biologicals, Littleton, CO) was added. Cells were washed in permeabilization buffer (cat 421002, BioLegend) followed with a final PBS wash and resuspension in Fluorofix fixation buffer (cat 422101, BioLegend). Sample acquisition was obtained on a Core facility BD FACSCalibur flow cytometer running CellQuest software followed with analysis using FCS express software. Analysis was based upon isolated gating of lymphocyte populations and co-localization of survivin with specific CD markers as indicated. Antibody determination 96 well plates (Immulon, Fisher Scientific, Pittsburgh, PA) were coated with tAChR (10 mg/ml). Serum samples were diluted in two concentrations and added to the plate (rat serum dilution, IgG, 1,000; IgG1, 1:50,000; IgG2b, 1:1,000), (mouse serum dilution: IgG, 1:20,000; IgG1, 1:40,000; IgG2a, 1:5,000; IgG2b, 1:40,000). Washed plates were incubated with horseradish peroxidase conjugated antibodies: Anti-Rat IgG, Anti-Rat IgG1, or Anti-Rat IgG2b, Anti-Mouse IgG, Anti-Mouse IgG1, Anti- Mouse IgG2a, Anti-Mouse IgG2b (Alpha Diagnostics, San Antonio, TX). The color reaction was developed with SureBlue TMB peroxidase substrate (KPL, Gaithersburg, MD). Reaction was stopped with HCl and read at 450 nm on a Varioskan Flash microplate reader (Thermo Scientific, Waltham, MA). Samples and controls were analyzed in triplicates. Pre-treatment with SVN53-67/M57-KLH in active EAMG We evaluated pre-treatment with SVN53-67/M57-KLH for efficacy in a rat model of EAMG. Pre-treatment with SVN53-67/ M57-KLH involved three s.c. injections of 100 mg SVN53-67/ M57-KLH in incomplete Freunds’ adjuvant (IFA, Sigma) given at weekly intervals to eight rats, and eight rats received PBS in IFA. Active EAMG was produced a week after final SVN53-67/M57- KLH injection. Sixteen rats total received s.c. injection of 40 mg tAChR in complete Freunds’ adjuvant (Sigma) given in two injections at the base of the tail. Rats were evaluated in a blinded fashion using the generally accepted motor strength scale [32], and body weight measurement. Grip strength was performed on all animals with twenty pulls before five measurements were recorded. Animals were sacrificed 47 days after AChR injection. July 2014 \| Volume 9 \| Issue 7 \| e102231 PLOS ONE \| www.plosone.org 4 Survivin Expression in Myasthenia Gravis Figure 1. Survivin expression in myasthenia gravis patients (1– 15) and control (16–25) PBMCs. A,B) Representative FACS analysis of total survivin expression of (A) myasthenic patient (7) and (B) control (24) PBMC. C–H) PBMC were stained with antibodies to indicated markers. Data reflect the percentage of total cells staining positive for: C) Survivin (SVN) (p = 0.010), D) CD20+/SVN+ (p = 0.0009), E) CD38+/ SVN+ (p = 0.02), F) CD27+/SVN+ (p = 0.05), G) CD138+/SVN+ (p = 0.01) and H) CD8+/SVN+ (p = 0.4). Mean values of patients and controls were compared using two way ANOVA. doi:10.1371/journal.pone.0102231.g001 Figure 1. Survivin expression in peripheral blood mononuclear cells (PBMC) We examined survivin expression in patients with MG (n = 15) and control subjects (n = 10) with medical conditions indicated in Table 1. During the course of the investigation, no MuSK antibody patients were encountered and only one who was negative for both MuSK and AChR antibodies. No patient had a thymoma identified by imaging, and none had undergone a therapeutic thymectomy. There were no statistically significant differences in age (P = 0.28, ANOVA) or gender (P = 0.93, Chi Square) between the two groups (Table 1). MG patients had been treated with a range of medications including immunosuppressive therapies. Survivin expression in PBMC was assessed by flow cytometry (Fig. 1A, B, and C). MG patients consistently exhibited survivin-positive white blood cells (Fig. 1A and C); whereas, control patients did not (Fig. 1B and C). Cells were co-stained with anti-CD20 (Fig. 1D), CD27 (Fig. 1E), CD38 (Fig. 1F), CD138 (Fig. 1G), and CD8 (control) (Fig. 1H) markers. Survivin expression was significantly greater in PBMCs from MG patients and was strongly associated with CD20, CD27, CD38 and CD138 expression (Fig. 1C–G). In contrast there was no significant association between survivin and CD8 expression except for a single patient with both MG and rheumatoid arthritis who had high survivin levels in CD8+ cells (Fig. 1H). (Values for all datasets appear in Table S1.) Figure 1. Survivin expression in myasthenia gravis patients (1– 15) and control (16–25) PBMCs. A,B) Representative FACS analysis of total survivin expression of (A) myasthenic patient (7) and (B) control (24) PBMC. C–H) PBMC were stained with antibodies to indicated markers. Data reflect the percentage of total cells staining positive for: C) Survivin (SVN) (p = 0.010), D) CD20+/SVN+ (p = 0.0009), E) CD38+/ SVN+ (p = 0.02), F) CD27+/SVN+ (p = 0.05), G) CD138+/SVN+ (p = 0.01) and H) CD8+/SVN+ (p = 0.4). Mean values of patients and controls were compared using two way ANOVA. doi:10.1371/journal.pone.0102231.g001 Antibody determination Survivin expression in myasthenia gravis patients (1– Comparison of AChR content at the NMJ by Alexa Fluor 594 –bungarotoxin assessment Comparison of AChR content at the NMJ by Alexa Fluor 594 –bungarotoxin assessment Ten micron cryosections of diaphragms were stained with Alexa Fluor 594 -bungarotoxin (Life Technologies, Grand Island, NY). The sections were viewed with Leica DFC310 FX and digital images captured with the Leica microsystems LAS AF6000 modular systems (Wetzlar, Germany) and analyzed with Image Pro software (Media Cybernetics, Inc., Rockville, MD). NMJs determined by fluorescence were defined as areas of interest. Pixel density measurements were determined and mean values obtained for each animal (n = 8/group for mouse study and n = 5/group for rat study) [34,35]. Statistical analysis The data were analyzed and tested for statistical significance using paired t-tests and ANOVA. Results were considered significantly different when P,0.05. Thymus samples demonstrate survivin expression A FITC-labeled AChR alpha 1 subunit peptide was tested for its ability to bind to PBMC from MG patients. Peptide binding was detected by flow cytometry and was competitively blocked with unlabeled AChR peptide indicating specificity of binding (Fig. 2A and B). AChR peptide bound to PBMC that were co-labeled with several B lymphocyte lineage and plasma cell markers (CD20, CD38 and CD138) in association with survivin expression. Figure 2C–F shows representative five color FACS analysis of PBMC gated on SVN+/CD38+/CD138+/CD20+ cells that also bind FITC-labeled AChR-alpha subunit peptide. FACS detected high numbers of CD20+/CD38+/CD138/SVN+ gated cells in patients with MG (Fig. 2D–F), but not in control subject (Fig. 2C). Eighty percent of generalized myasthenia gravis patients with AChR positive antibody titers have thymic pathology [36]. We studied thymuses from eleven MG patients, 4 untreated and 7 treated with corticosteroids. The thymuses from the four untreated MG patients demonstrated lymphofollicular hyperplasia [37]. Cells in the cortex, cortico-medullary junction, and germinal centers were positive for survivin (Fig. 3A–D). The thymuses from patients that were treated with corticosteroids prior to thymectomy demonstrated fewer survivin-positive cells (Fig. 3E–H). In active germinal centers that persisted inside atrophic thymuses, a subpopulation of the germinal center cells remained survivin positive (Fig. 3F,H). July 2014 \| Volume 9 \| Issue 7 \| e102231 PLOS ONE \| www.plosone.org 5 Survivin Expression in Myasthenia Gravis Figure 2. Acetylcholine receptor alpha peptide binding to survivin positive cells in myasthenics by five color FACS analysis. A) FITC labeled AChR-alpha subunit peptide binding to total myasthenic PBMC (0.44%), B) Excess unlabeled AChR-alpha subunit peptide followed by competitive binding of FITC-AChR peptide (0.03%). C) Control PBMC, subject number 19, (SVN+/CD38+/CD138+/CD20+ gated sub-population) showing FITC-AChR-alpha subunit peptide binding (0.33% of gated cells). D–F; Myasthenic patient PBMC (SVN+/CD38+/CD138+/CD20+ gated sub- population) showing FITC-AChR-alpha subunit peptide binding (1.55% (pt#10), 6.54% (pt#15) and 2.50% (pt#14) of total gated cells; n = 3). doi:10.1371/journal.pone.0102231.g002 Figure 2. Acetylcholine receptor alpha peptide binding to survivin positive cells in myasthenics by five color FACS analysis. A) FITC labeled AChR-alpha subunit peptide binding to total myasthenic PBMC (0.44%), B) Excess unlabeled AChR-alpha subunit peptide followed by competitive binding of FITC-AChR peptide (0.03%). C) Control PBMC, subject number 19, (SVN+/CD38+/CD138+/CD20+ gated sub-population) showing FITC-AChR-alpha subunit peptide binding (0.33% of gated cells). AChR levels at the NMJ were increased in treated mice AChR levels at the NMJ were increased in treated mice To determine the effect of reduced AChR antibody levels of mice treated with anti-survivin vaccination, we analyzed the AChR present at the NMJ by labeling with Alexa Fluor 594 - bungarotoxin (Fig. 5J–L) [35] [38] [39]. The AChR at the NMJs of diaphragm in SVN53-67/M57-KLH treated mice (32.3966.30(SD) pixel density) showed a statistically significant (P,0.001, Mann-Whitney U) increase by density scan analysis compared to PBS control mice (23.5264.52(SD) pixel density). (Table S2). Thymus samples demonstrate survivin expression D–F; Myasthenic patient PBMC (SVN+/CD38+/CD138+/CD20+ gated sub- population) showing FITC-AChR-alpha subunit peptide binding (1.55% (pt#10), 6.54% (pt#15) and 2.50% (pt#14) of total gated cells; n = 3). doi:10.1371/journal.pone.0102231.g002 Survivin expression in EAMG mouse lymphocytes and effect of anti-survivin vaccine The survivin vaccine contains a number of class I CTL epitopes. Tetramer assays were performed to determine if the survivin vaccine induced production of survivin-specific CD8+ T cells. EAMG mice that were vaccinated with SVN53-67/M57-KLH exhibited an immune response that included larger numbers of survivin-specific CD8+ T cells (Fig. 5A, B). EAMG mice expressed survivin in CD20+ PBMC at high levels (range: 5.84%–43.14%; mean 26.5%) (Fig. 5C, E). In contrast, EAMG mice vaccinated with SVN53-67/M57-KLH had markedly reduced numbers of SVN+/CD20+ cells (range 0.10%–26.9%; mean 6.7%; P = 0.001) (Fig. 5D, E). AChR specific IgG subtypes were reduced in survivin vaccinated mice Active EAMG was induced in mice by injections of torpedo AChR in complete Freund’s adjuvant (CFA) on day 0, 24, and 52. Survivin vaccination was initiated on day 42, ten days prior to the last AChR injection and continued to day 73. Mice were evaluated for grip strength, weight and clinical scores weekly. SVN53-67/ M57-KLH treatment (n = 9) demonstrated greater grip strength (statistically significant, P,0.05, t-test) starting at week 10 compared to PBS treatment controls (n = 10) (Fig. 4A, Table S2). Overall, mice treated with SVN53-67/M57-KLH did not show weakness and had a clinical score of zero. In contrast, 20% of mice treated with PBS demonstrated general weakness (inability to grip wire bar and limited motility, clinical score of 2) and 40% demonstrated weakness after exercise (clinical score of 1) (Fig. 4B). ELISAs were performed to determine AChR specific total IgG, IgG1 and IgG2b (Fig. 5F–I). The EAMG SVN53-67/M57-KLH treated mice (n = 7) demonstrated significantly (P = 0.02, t-test) lower mean levels of AChR antibody (0.3260.14(SD), OD at 450 nm) compared to the mean level of AChR antibody in control EAMG PBS treated mice (n = 7; 0.4060.04(SD), OD at 450 nm) (Fig. 5F). We assessed the IgG subtypes to determine the expression level specific to AChR epitopes. The AChR specific IgG, IgG2a and IgG2b were also lower (P,0.02, t-test) (Fig. 5G–I) compared to PBS treated controls. (Table S2). Discussion Survivin-positive cells with B lymphocyte and plasma cell markers are present at significant levels in the circulation of humans with MG, but not in controls. In addition, survivin co- localizes to circulating white blood cells which bind specifically to AChR alpha subunit epitopes. Cells within the germinal centers of the thymus of myasthenia gravis patients express survivin. Collectively, these results suggest an association between survivin expression and the autoimmune state in patients with MG. Similarly, rodents with EAMG have large numbers of circulating CD20+/SVN+ (mice) and CD45ra+/SVN+ (rats; data not shown) cells suggesting that, despite fundamental differences in disease induction between MG and EAMG, survivin expression is a common feature of the human disease and these two animal models. Thus, it appears that survivin could be a marker for specific B lymphocytes and plasma cells that are involved in MG and that it may play a key role in enabling cells to resist apoptotic signals leading to sustained pathologic autoantibody production. Pre-treatment with SVN53-67/M57-KLH led to maintenance of grip strength in rats (Fig. 6A, Table S3). Blinded evaluation also revealed significantly (P,0.05, t-test) better strength in the treated group with mean clinical scores of 0.256.46(SD) (n = 8) compared to greater weakness in the PBS control group 1.2561.0(SD)(n = 8) (Fig. 6B). Survivin vaccination is protective in active induced experimentally acquired myasthenia gravis in rats AChR specific-total IgG in the PBS-treated group (n = 7, 0.6660.13(SD), OD at 450 nm) demonstrated higher levels (P, 0.001, t-test) compared to SVN53-67/M57-KLH treated rats (n = 7, 0.4660.14(SD), OD at 450 nm). The expression levels of IgG1 and IgG2b (1.1860.19(SD) and 1.2660.26(SD), OD at 450 nm, respectively) in the PBS control group were statistically higher (P,0.02) compared to SVN53-67/M57-KLH treated rats (0.7060.33(SD) and 0.8360.37(SD), respectively). Figure 3. Survivin expression in the thymus. Representitive images from an analysis of eleven thymuses from MG patients were analyzed for survivin expression (immunosuppression treated, n = 7; immnosuppression naı¨ve, n = 4). A–D) EOMG thymus from a 24-year-old woman who had clinical symptoms for 2 years and an AChR antibody level of 19.3 nmol and had never received immunosuppression or prednisolone showed a well developed cortex (C) and medulla (M), lymphofollicular hyperplasia with a germinal center (GC) close to Hassall’s corpuscle (HC). B,D) High number of survivin positive cells in the cortex (C), cortico-medullary junction and the GC. E–H) EOMG thymus from a AChR antibody positive 21-year-old female with lymphofollicular hyperplasia after long-term immunosuppression showing cortical atrophy and a slightly regressive germinal center (GC) close to a Hassall’s corpuscle (HC). F,H) Low number of survivin positive lymphocytes in the remnant cortical area (C), and the germinal center (GC) (A,B,E,F, 6100; C,D,G,H 6200). doi:10.1371/journal.pone.0102231.g003 Analysis of the NMJ demonstrates increased AChR in rats vaccinated against survivin Animals were sacrificed 47 days after AChR injection. AChR expresssion at the NMJs from diaphragms of rats that had been pre-treated with SVN53-67/M57-KLH (20.5364.04(SD) pixel density) showed significantly (P = 0.013) increased NMJ density with fluorescently labeled bungarotoxin (Fig. 6D) compared to PBS control (16.2663.49(SD) pixel density)(Fig. 6E, Table S3). Survivin vaccination is protective in active induced experimentally acquired myasthenia gravis in rats We reproduced the effect of SVN53-67/M57-KLH that had been observed in mice by using a second animal model of EAMG. Rats were treated with survivin peptide for three weeks prior to AChR injection to ensure the vaccine had adequate time to induce the production of survivin-specific cytotoxic T lymphocytes (CTL). PLOS ONE \| www.plosone.org July 2014 \| Volume 9 \| Issue 7 \| e102231 6 Survivin Expression in Myasthenia Gravis Figure 4. SVN53-67/M57-KLH post-treatment on active EAMG in the mouse model. A) Animals were injected with SVN53-67/M57- KLH at times marked by red arrows. Black arrows mark injection times of tAChR. Grip strength was monitored throughout experiment. Mice were treated with SVN53-67/M57-KLH (circles, n = 9), or PBS (squares, n = 10) (* denotes P,0.05, t-test). B) Table shows clinical scores of mice treated with SVN53-67/M57-KLH were significantly stronger than PBS treated mice (* denotes P,0.05, t-test). doi:10.1371/journal.pone.0102231.g004 Figure 3. Survivin expression in the thymus. Representitive images from an analysis of eleven thymuses from MG patients were analyzed for survivin expression (immunosuppression treated, n = 7; immnosuppression naı¨ve, n = 4). A–D) EOMG thymus from a 24-year-old woman who had clinical symptoms for 2 years and an AChR antibody level of 19.3 nmol and had never received immunosuppression or prednisolone showed a well developed cortex (C) and medulla (M), lymphofollicular hyperplasia with a germinal center (GC) close to Hassall’s corpuscle (HC). B,D) High number of survivin positive cells in the cortex (C), cortico-medullary junction and the GC. E–H) EOMG thymus from a AChR antibody positive 21-year-old female with lymphofollicular hyperplasia after long-term immunosuppression showing cortical atrophy and a slightly regressive germinal center (GC) close to a Hassall’s corpuscle (HC). F,H) Low number of survivin positive lymphocytes in the remnant cortical area (C), and the germinal center (GC) (A,B,E,F, 6100; C,D,G,H 6200). doi:10.1371/journal.pone.0102231.g003 Figure 4. SVN53-67/M57-KLH post-treatment on active EAMG Figure 4. SVN53-67/M57-KLH post-treatment on active EAMG in the mouse model. A) Animals were injected with SVN53-67/M57- KLH at times marked by red arrows. Black arrows mark injection times of tAChR. Grip strength was monitored throughout experiment. Mice were treated with SVN53-67/M57-KLH (circles, n = 9), or PBS (squares, n = 10) (* denotes P,0.05, t-test). B) Table shows clinical scores of mice treated with SVN53-67/M57-KLH were significantly stronger than PBS treated mice (* denotes P,0.05, t-test). doi:10.1371/journal.pone.0102231.g004 antibodies when treated with SVN53-67/M57-KLH. July 2014 \| Volume 9 \| Issue 7 \| e102231 Reduction in the AChR-specific antibody profile is observed in SVN53-67/M57-KLH treated EAMG rats compared to PBS controls ELISAs were performed to determine AChR-specific total IgG, IgG1 and IgG2b (Fig.5C, Table S3). As with findings in the mouse model, the EAMG induced rats had reduced expression of AChR July 2014 \| Volume 9 \| Issue 7 \| e102231 July 2014 \| Volume 9 \| Issue 7 \| e102231 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org 7 Survivin Expression in Myasthenia Gravis Figure 5. Analysis of post-treatment with SVN53-67/M57-KLH in EAMG mouse model. Survivin-specific iTag tetramer analysis of survivin- targeted CD8+ T cells in EAMG mice (A) PBS and (B) SVN53-67/M57-KLH-treatment. FACS analysis of PBMC with representative flow histograms of CD20+/SVN+ cells from (C) PBS and (D) SVN53-67/M57-KLH-treated mice. E) Mean values of groups were compared using two way ANOVA (P = 0.001). F–I) AChR specific total IgG (F), IgG1 (G), IgG2a (H) and IgG2b (I) levels were determined by ELISA on an AChR coated 96-well plate for PBS (n = 7) and SVN53-67/M57-KL (n = 7)-treated animals. Levels of total IgG, IgG1, IgG2a and IgG2b were significantly higher in the PBS treated group compared to SVN53-67/M57-KLH treated animals (P = 0.02 for total IgG and IgG2b; p,0.005 for IgG1 and IgG2a). J–L) AChR expression at the NMJ of PBS treated (n = 8) and SVN53-67/M57-KLH-treated (n = 8) mice was determined by imaging of Alexa Fluor 594 –bungarotoxin pixel density at the NMJ. Two representative images of PBS treated (J) and SVN53-67/M57-KLH-treated (K) mice are shown. Mean pixel density was determined (L) (P = 0.0001, t-test). doi:10.1371/journal.pone.0102231.g005 Figure 5. Analysis of post-treatment with SVN53-67/M57-KLH in EAMG mouse model. Survivin-specific iTag tetramer analysis of survivin- targeted CD8+ T cells in EAMG mice (A) PBS and (B) SVN53-67/M57-KLH-treatment. FACS analysis of PBMC with representative flow histograms of CD20+/SVN+ cells from (C) PBS and (D) SVN53-67/M57-KLH-treated mice. E) Mean values of groups were compared using two way ANOVA (P = 0.001). F–I) AChR specific total IgG (F), IgG1 (G), IgG2a (H) and IgG2b (I) levels were determined by ELISA on an AChR coated 96-well plate for PBS (n = 7) and SVN53-67/M57-KL (n = 7)-treated animals. Levels of total IgG, IgG1, IgG2a and IgG2b were significantly higher in the PBS treated group compared to SVN53-67/M57-KLH treated animals (P = 0.02 for total IgG and IgG2b; p,0.005 for IgG1 and IgG2a). Reduction in the AChR-specific antibody profile is observed in SVN53-67/M57-KLH treated EAMG rats compared to PBS controls J–L) AChR expression at the NMJ of PBS treated (n = 8) and SVN53-67/M57-KLH-treated (n = 8) mice was determined by imaging of Alexa Fluor 594 –bungarotoxin pixel density at the NMJ. Two representative images of PBS treated (J) and SVN53-67/M57-KLH-treated (K) mice are shown. Mean pixel density was determined (L) (P = 0.0001, t-test). doi:10 1371/journal pone 0102231 g005 doi:10.1371/journal.pone.0102231.g005 doi:10.1371/journal.pone.0102231.g005 Survivin expression has been found in synovial cells from patients with RA and its expression is associated with progression of that disease to a phase involving erosive joint destruction. Similarly, survivin is expressed in PBMC and in the brains of patients with progressive multiple sclerosis. Together with our findings in MG, these observations suggest that survivin plays a pathophysiologic role in some autoimmune diseases by inhibiting the apoptotic cascade that leads to the death of cells that mediate the underlying conditions. Within the limited data set presented, there is a suggestion that survivin-expression in subset of PBMC could serve as a severity and treatment responsive marker, but extensive investigation will be required to confirm such a possibility. Reports of survivin expression in lymphoid cells of patients with MS and synovial cells in RA patients suggest that it could play a similar role in other autoimmune diseases as well. SVN53-67/M57-KLH vaccine attenuated EAMG disease severity as evidenced by improved clinical motor scores, decreased numbers of survivin-expressing CD20+ cells, reduced AChR specific antibodies, reduced antigen-specific complement-fixing antibodies, and moderation of complement deposition. The SVN53-67/M57-KLH was used as a pre-treatment in the rat model and a post-treatment in the mouse model, each demon- strating an improvement in clinical scores. The AChR specific antibody levels of the treated animals were significantly reduced in all IgG subtypes assessed. Our data suggest that survivin is a factor in the maintenance of autoreactive lymphocytes in spontaneous human disease and in induced models of the disease in rodents. Antibody-mediated autoimmunity is restricted either by modification of the specificity of autoantibodies, transition of the autoreactive B cells to a state of anergy or by the apoptotic deletion of cells that express such autoantibodies [40]. Our study of survivin vaccine in EAMG does not determine the mechanism of action but the results clearly show a reduction in the autoantibodies produced. July 2014 \| Volume 9 \| Issue 7 \| e102231 Reduction in the AChR-specific antibody profile is observed in SVN53-67/M57-KLH treated EAMG rats compared to PBS controls Serum B lymphocyte activating factor (BAFF) levels are increased in patients with autoimmune MG indicating that BAFF is involved in promoting the survival and maturation of autoreactive B cells in the disease [41]. Moreover, CD138-positive plasma cells are abundant in the thymus of MG patients with expression of BAFF and a proliferation-inducing ligand (APRIL) documented in these loci, further suggesting the presence of an environment that is conducive to B-cell survival [42]. In addition, studies have demonstrated that autoantibody-producing plasma July 2014 \| Volume 9 \| Issue 7 \| e102231 July 2014 \| Volume 9 \| Issue 7 \| e102231 PLOS ONE \| www.plosone.org 8 Survivin Expression in Myasthenia Gravis Figure 6. Therapeutic effect of pre-treatment of SVN53-67/ M57-KLH on active EAMG. Rats were pre-treated with SVN53-67/ M57-KLH (circles) or PBS (squares). A) Animals were injected with SVN53-67/M57-KLH at times marked by red arrows. Black arrows mark injection time of tAChR. Grip strength was measured throughout experiment and values are a mean of biweekly measurements. (* denotes P,0.02, t-test) B) Table of clinical scores based on rats at day 47. C) Levels of total IgG, IgG1 and IgG2b were significantly higher in the PBS control group (n = 7) compared to SVN53-67/M57-KLH treated animals (n = 7). (P,0.02 for all, t-test) D) SVN53-67/M57-KLH treated rats demonstrated statistically increased amount of AChR (red) at the NMJs compared to PBS-treated rats. E) Pixel density of the labeled AChR by Alexa Fluor 594 -bungarotoxin at the NMJ in SVN53- 67/M57-KLH treated animals (n = 5) compared to PBS treated controls (n = 5) (P = 0.013, t-test). doi:10 1371/journal pone 0102231 g006 various stages of development [44]. SVN53-67/M57-KLH (SurVaxM) is currently being evaluated in Phase I clinical trials as a cancer vaccine [23]. It induces the production of specific anti- survivin antibodies and survivin-reactive T cells by tetramer and ELISPOT analysis in humans (our unpublished data). In addition, studies of human tissues demonstrate that the peptide mimic contained in SVN53-67/M57-KLH is capable of inducing a cytotoxic T cell response capable of killing survivin-expressing B-cell lymphoma cells ex vivo [21]. Together with animal data presented here, results suggest that active specific vaccination against survivin might be effective at reducing survivin-expressing autoreactive cells. Specific survivin inhibitors have been synthesized and at least one of them (YM155) is currently in clinical trials as a cancer therapeutic. Supporting Information Table S1 Survivin expression in human PBMCs. Percent of PBMCs from patients with myasthenia gravis and controls expressing survivin; survivin and CD20; survivin and CD27; survivin and 38; survivin and CD138; and survivin and CD8. (TIF) Table S1 Survivin expression in human PBMCs. Percent of PBMCs from patients with myasthenia gravis and controls expressing survivin; survivin and CD20; survivin and CD27; survivin and 38; survivin and CD138; and survivin and CD8. (TIF) Figure 6. Therapeutic effect of pre-treatment of SVN53-67/ M57-KLH on active EAMG. Rats were pre-treated with SVN53-67/ M57-KLH (circles) or PBS (squares). A) Animals were injected with SVN53-67/M57-KLH at times marked by red arrows. Black arrows mark injection time of tAChR. Grip strength was measured throughout experiment and values are a mean of biweekly measurements. (* denotes P,0.02, t-test) B) Table of clinical scores based on rats at day 47. C) Levels of total IgG, IgG1 and IgG2b were significantly higher in the PBS control group (n = 7) compared to SVN53-67/M57-KLH treated animals (n = 7). (P,0.02 for all, t-test) D) SVN53-67/M57-KLH treated rats demonstrated statistically increased amount of AChR (red) at the NMJs compared to PBS-treated rats. E) Pixel density of the labeled AChR by Alexa Fluor 594 -bungarotoxin at the NMJ in SVN53- 67/M57-KLH treated animals (n = 5) compared to PBS treated controls (n = 5) (P = 0.013, t-test). doi:10.1371/journal.pone.0102231.g006 Table S2 Mouse EAMG data. Numerical values of grip strength, clinical scores, tAChR specific total IgG, tAChR specific IgG1, tAChR specific IgG2a, tAChR specific IgG2b, and pixel density values from the EAMG mouse model. (TIF) Table S3 Rat EAMG data. Numerical values of grip strength, clinical scores, tAChR specific total IgG, tAChR specific IgG1, tAChR specific IgG2b, and pixel density values from the EAMG rat model. (TIF) cells in the thymus of MG patients are long-lived [43], suggesting an inherent resistance to normal apoptotic signals. The sustained expression of survivin in plasma and B cells of MG patients is consistent with this possibility and provides strong evidence that maintenance of autoimmune cell viability is an important component of the immunopathologic state. Author Contributions Conceived and designed the experiments: LLK MJC AM HJK RAF. Performed the experiments: LLK MJC AM. Analyzed the data: LLK MJC Conceived and designed the experiments: LLK MJC AM HJK RAF. Performed the experiments: LLK MJC AM. Analyzed the data: LLK MJC AM HJK RAF. Contributed reagents/materials/analysis tools: MJC AM HJK RAF. Wrote the paper: LLK HJK RAF. Conceived and designed the experiments: LLK MJC AM HJK RAF. Performed the experiments: LLK MJC AM. Analyzed the data: LLK MJC Conceived and designed the experiments: LLK MJC AM HJK RAF. Performed the experiments: LLK MJC AM. Analyzed the data: LLK MJC Survivin is a specific and widely expressed cancer protein that is of great interest as a target both for immunotherapy and for specific inhibitors. Survivin targeted treatment approaches using small molecules, RNAi, oligonucleotides, and ribozymes are at AM HJK RAF. Contributed reagents/materials/analysis tools: MJC AM HJK RAF. Wrote the paper: LLK HJK RAF. AM HJK RAF. Contributed reagents/materials/analysis tools: MJC AM HJK RAF. Wrote the paper: LLK HJK RAF. Reduction in the AChR-specific antibody profile is observed in SVN53-67/M57-KLH treated EAMG rats compared to PBS controls Such inhibitors could potentially be useful if survivin levels remain high following clinical immunosuppression; however, the advantage of anti-survivin vaccine therapy rests with its potential to create an extended period of immunity through production of a memory T-cell response, as opposed to life-long medical therapy. In addition, non-specific immunosuppression has many side-effects and risks which may not be the case for active specific vaccination. Previous or concomitant immunosuppressive therapies that are commonly given to MG patients may limit the effect of anti-survivin vaccination in the clinical setting, unless a vaccine is given as initial therapy, or at least early in the course of the illness. 1. Conti-Fine BM, Milani M, Kaminski HJ (2006) Myasthenia gravis: past, present, and future. The Journal of clinical investigation 116: 2843–2854. 2. LaCasse EC, Mahoney DJ, Cheung HH, Plenchette S, Baird S, et al. (2008) IAP-targeted therapies for cancer. Oncogene 27: 6252–6275. 3. Conway EM, Pollefeyt S, Cornelissen J, DeBaere I, Steiner-Mosonyi M, et al. (2000) Three differentially expressed survivin cDNA variants encode proteins with distinct antiapoptotic functions. Blood 95: 1435–1442. 3. Conway EM, Pollefeyt S, Cornelissen J, DeBaere I, Steiner-Mosonyi M, et al. (2000) Three differentially expressed survivin cDNA variants encode proteins with distinct antiapoptotic functions. Blood 95: 1435–1442. 1. Conti-Fine BM, Milani M, Kaminski HJ (2006) Myasthenia gravis: past, present, and future. The Journal of clinical investigation 116: 2843–2854. g 2. LaCasse EC, Mahoney DJ, Cheung HH, Plenchette S, Baird S, et al. (2008) IAP-targeted therapies for cancer. Oncogene 27: 6252–6275. Acknowledgments We thank Cheryl Frank for technical assistance. July 2014 \| Volume 9 \| Issue 7 \| e102231 5. Li F, Ambrosini G, Chu EY, Plescia J, Tognin S, et al. (1998) Control of apoptosis and mitotic spindle checkpoint by survivin. Nature 396: 580–584. Survivin Expression in Myasthenia Gravis Piddlesden SJ, Jiang S, Levin JL, Vincent A, Morgan BP (1996) Soluble complement receptor 1 (sCR1) protects against experimental autoimmune myasthenia gravis. J Neuroimmunol 71: 173–177. 13. Bokarewa M, Lindblad S, Bokarew D, Tarkowski A (2005) Balance between survivin, a key member of the apoptosis inhibitor family, and its specific antibodies determines erosivity in rheumatoid arthritis. Arthritis Res Ther 7: R349–358. 33. Ciesielski MJ, Apfel L, Barone TA, Castro CA, Weiss TC, et al. (2006) Antitumor effects of a xenogeneic survivin bone marrow derived dendritic cell vaccine against murine GL261 gliomas. Cancer Immunol Immunother 55: 1491–1503. 14. Smith MD, Weedon H, Papangelis V, Walker J, Roberts-Thomson PJ, et al. (2010) Apoptosis in the rheumatoid arthritis synovial membrane: modulation by disease-modifying anti-rheumatic drug treatment. Rheumatology (Oxford) 49: 862–875. 34. Kusner LL, Halperin JA, Kaminski HJ (2013) Cell surface complement regulators moderate experimental myasthenia gravis pathology. Muscle Nerve 47: 33–40. 15. Galeotti L, Adrian K, Berg S, Tarkowski A, Bokarewa M (2008) Circulating survivin indicates severe course of juvenile idiopathic arthritis. Clin Exp Rheumatol 26: 373–378. 35. Soltys J, Kusner LL, Young A, Richmonds C, Hatala D, et al. (2009) Novel complement inhibitor limits severity of experimentally myasthenia gravis. Ann Neurol 65: 67–75. 16. Isgren A, Forslind K, Erlandsson M, Axelsson C, Andersson S, et al. (2012) High survivin levels predict poor clinical response to infliximab treatment in patients with rheumatoid arthritis. Semin Arthritis Rheum 41: 652–657. 36. Cavalcante P, Cufi P, Mantegazza R, Berrih-Aknin S, Bernasconi P, et al. (2013) Etiology of myasthenia gravis: innate immunity signature in pathological thymus. Autoimmun Rev 12: 863–874. 17. Hebb AL, Moore CS, Bhan V, Campbell T, Fisk JD, et al. (2008) Expression of the inhibitor of apoptosis protein family in multiple sclerosis reveals a potential immunomodulatory role during autoimmune mediated demyelination. Mult Scler 14: 577–594. 37. Marx A, Pfister F, Schalke B, Nix W, Strobel P (2012) Thymus pathology observed in the MGTX trial. Ann N Y Acad Sci 1275: 92–100. 38. Tuzun E, Scott BG, Goluszko E, Higgs S, Christadoss P (2003) Genetic evidence for involvement of classical complement pathway in induction of experimental autoimmune myasthenia gravis. J Immunol 171: 3847–3854. 18. Sharief MK, Noori MA, Douglas MR, Semra YK (2002) Upregulated survivin expression in activated T lymphocytes correlates with disease activity in multiple sclerosis. Eur J Neurol 9: 503–510. 39. Morgan BP, Chamberlain-Banoub J, Neal JW, Song W, Mizuno M, et al. Survivin Expression in Myasthenia Gravis 25. Christadoss P, Poussin M, Deng C (2000) Animal models of myasthenia gravis. Clin Immunol 94: 75–87. 6. Rosa J, Canovas P, Islam A, Altieri DC, Doxsey SJ (2006) Survivin modulates microtubule dynamics and nucleation throughout the cell cycle. Mol Biol Cell 17: 1483–1493. 26. Baggi F, Antozzi C, Toscani C, Cordiglieri C (2012) Acetylcholine receptor- induced experimental myasthenia gravis: what have we learned from animal models after three decades? Archivum immunologiae et therapiae experimenta- lis 60: 19–30. 7. Suzuki A, Hayashida M, Ito T, Kawano H, Nakano T, et al. (2000) Survivin initiates cell cycle entry by the competitive interaction with Cdk4/p16(INK4a) and Cdk2/cyclin E complex activation. Oncogene 19: 3225–3234. and Cdk2/cyclin E complex activation. Oncogene 19: 3225–32 8. Altieri DC (2010) Survivin and IAP proteins in cell-death mechanisms. Biochem J 430: 199–205. 27. Kyewski B, Taubert R (2008) How promiscuity promotes tolerance: the case of myasthenia gravis. Ann N Y Acad Sci 1132: 157–162. 9. Shin S, Sung BJ, Cho YS, Kim HJ, Ha NC, et al. (2001) An anti-apoptotic protein human survivin is a direct inhibitor of caspase-3 and -7. Biochemistry 40: 1117–1123. 28. Benatar M, Sanders DB, Burns TM, Cutter GR, Guptill JT, et al. (2012) Recommendations for myasthenia gravis clinical trials. Muscle & nerve 45: 909– 917. 10. Tamm I, Wang Y, Sausville E, Scudiero DA, Vigna N, et al. (1998) IAP-family protein survivin inhibits caspase activity and apoptosis induced by Fas (CD95), Bax, caspases, and anticancer drugs. Cancer Res 58: 5315–5320. 29. Simon-Keller K, Paschen A, Hombach AA, Strobel P, Coindre JM, et al. (2013) Survivin blockade sensitizes rhabdomyosarcoma cells for lysis by fetal acetylcholine receptor-redirected T cells. Am J Pathol 182: 2121–2131. 11. Mita AC, Mita MM, Nawrocki ST, Giles FJ (2008) Survivin: key regulator of mitosis and apoptosis and novel target for cancer therapeutics. Clin Cancer Res 14: 5000–5005. 30. Landis SC, Amara SG, Asadullah K, Austin CP, Blumenstein R, et al. (2012) A call for transparent reporting to optimize the predictive value of preclinical research. Nature 490: 187–191. 12. Sharief MK, Semra YK (2001) Heightened expression of survivin in activated T lymphocytes from patients with multiple sclerosis. J Neuroimmunol 119: 358– 364. 31. Wu B, Goluszko E, Christadoss P (2001) Experimental autoimmune myasthenia gravis in the mouse. Curr Protoc Immunol Chapter 15: Unit 15 18. 32. References 4. Adida C, Crotty PL, McGrath J, Berrebi D, Diebold J, et al. (1998) Developmentally regulated expression of the novel cancer anti-apoptosis gene survivin in human and mouse differentiation. Am J Pathol 152: 43–49. 1. Conti-Fine BM, Milani M, Kaminski HJ (2006) Myasthenia gravis: past, present, and future. The Journal of clinical investigation 116: 2843–2854. 2. LaCasse EC, Mahoney DJ, Cheung HH, Plenchette S, Baird S, et al. (2008) IAP-targeted therapies for cancer. Oncogene 27: 6252–6275. 5. Li F, Ambrosini G, Chu EY, Plescia J, Tognin S, et al. (1998) Control of apoptosis and mitotic spindle checkpoint by survivin. Nature 396: 580–584. 5. Li F, Ambrosini G, Chu EY, Plescia J, Tognin S, et al. (1998) Control of apoptosis and mitotic spindle checkpoint by survivin. Nature 396: 580–584. 3. Conway EM, Pollefeyt S, Cornelissen J, DeBaere I, Steiner-Mosonyi M, et al. (2000) Three differentially expressed survivin cDNA variants encode proteins with distinct antiapoptotic functions. Blood 95: 1435–1442. July 2014 \| Volume 9 \| Issue 7 \| e102231 PLOS ONE \| www.plosone.org 9 Survivin Expression in Myasthenia Gravis (2006) The membrane attack pathway of complement drives pathology in passively induced experimental autoimmune myasthenia gravis in mice. Clin Exp Immunol 146: 294–302. J 19. Andersen MH, Pedersen LO, Becker JC, Straten PT (2001) Identification of a cytotoxic T lymphocyte response to the apoptosis inhibitor protein survivin in cancer patients. Cancer Res 61: 869–872. 20. Aguilar LK, Guzik BW, Aguilar-Cordova E (2011) Cytotoxic immunotherapy strategies for cancer: mechanisms and clinical development. J Cell Biochem 112: 1969–1977. 40. Khan WN, Wright JA, Kleiman E, Boucher JC, Castro I, et al. (2013) B- lymphocyte tolerance and effector function in immunity and autoimmunity. Immunol Res 57: 335–353. 21. Ciesielski MJ, Ahluwalia MS, Munich SA, Orton M, Barone T, et al. (2010) Antitumor cytotoxic T-cell response induced by a survivin peptide mimic. Cancer Immunol Immunother 59: 1211–1221. 41. Ragheb S, Lisak R, Lewis R, Van Stavern G, Gonzales F, et al. (2008) A potential role for B-cell activating factor in the pathogenesis of autoimmune myasthenia gravis. Arch Neurol 65: 1358–1362. 22. Ciesielski MJ, Kozbor D, Castanaro CA, Barone TA, Fenstermaker RA (2008) Therapeutic effect of a T helper cell supported CTL response induced by a survivin peptide vaccine against murine cerebral glioma. Cancer Immunol Immunother 57: 1827–1835. 42. Thangarajh M, Masterman T, Helgeland L, Rot U, Jonsson MV, et al. (2006) The thymus is a source of B-cell-survival factors-APRIL and BAFF-in myasthenia gravis. J Neuroimmunol 178: 161–166. 43. Fujii Y, Hashimoto J, Monden Y, Ito T, Nakahara K, et al. (1986) Specific activation of lymphocytes against acetylcholine receptor in the thymus in myasthenia gravis. J Immunol 136: 887–891. 23. Fenstermaker RA, Ciesielski MJ (2014) Challenges in the development of a survivin vaccine (SurVaxM) for malignant glioma. Expert Rev Vaccines 13: 377–385. 44. Mobahat M, Narendran A, Riabowol K (2014) Survivin as a preferential target for cancer therapy. Int J Mol Sci 15: 2494–2516. 24. Lang B, Vincent A (2009) Autoimmune disorders of the neuromuscular junction. Curr Opin Pharmacol 9: 336–340. July 2014 \| Volume 9 \| Issue 7 \| e102231 PLOS ONE \| www.plosone.org 10
https://openalex.org/W4387114341	https://link.springer.com/content/pdf/10.1007/s10935-023-00742-2.pdf	English	null	Transporting an Evidence-based Youth Development Program to a New Country: A Narrative Description and Analysis of Pre-implementation Adaptation	Journal of prevention	2,023	cc-by	10,162	Extended author information available on the last page of the article Journal of Prevention (2023) 44:729–747 https://doi.org/10.1007/s10935-023-00742-2 Journal of Prevention (2023) 44:729–747 https://doi.org/10.1007/s10935-023-00742-2 ORIGINAL PAPER Transporting an Evidence-based Youth Development Program to a New Country: A Narrative Description and Analysis of Pre-implementation Adaptation Finlay Green1 · Nick Axford2 · Ntale Eastmond3 · Vashti Berry4 · Julia Mannes1 · Kate Allen4 · Lynne Callaghan2 · Tim Hobbs1 Finlay Green1 · Nick Axford2 · Ntale Eastmond3 · Vashti Berry4 · Julia Mannes1 · Kate Allen4 · Lynne Callaghan2 · Tim Hobbs1 Accepted: 31 July 2023 / Published online: 28 September 2023 © The Author(s) 2023 Keywords Adaptation · Implementation · Mentoring · Prevention · School · Violence Introduction There is a pressing need to prevent youth crime and violence owing to its prevalence, harms and cost to society (Kieselbach & Butchart, 2015). Much is known about the effectiveness of school-, family- and community-based interventions designed to do this (e.g., Fagan and Catalano, 2013; Farrington et al., 2017; Matjasko et al., 2012; Russell et al., 2021). Those with the strongest evidence often originate in US, yet several have struggled to produce positive effects in Europe (e.g., Baldus et al., 2016; Fonagy et al., 2018; Humayun et al., 2017; Segrott et al., 2022; Skärstrand et al., 2013; Sundell et al., 2008). Reasons given for this include poor implementation, dif ferent context, lack of developer involvement and, of most relevance here, problems with program adaptation.f Adapting existing interventions for new contexts is potentially cost-effective because it saves investing in developing and evaluating new interventions (Movsi syan et al., 2019). Adaptation refers to a process of thoughtful and deliberate altera tion of the design or delivery of an intervention to improve its fit or effectiveness in each context (Stirman et al., 2019). Whereas strict fidelity to intervention blueprints was once deemed essential to replication effectiveness, it is now recognized that real ity is more complex (Chambers & Norton, 2016) and that staying true to function may be more important that adherence to form (Movsisyan et al., 2021). Making adaptations can improve program engagement, acceptability and outcomes (Stirman et al., 2019) but it can also go wrong, especially if changes remove or dilute active ingredients and thereby nullify the intervention theory of change (Evans et al., 2019; Movsisyan et al., 2019). Equally, a lack of adaptation can be unhelpful if incoming interventions inadequately fit the local service systems and culture (Moore et al., 2021). Moreover, lack of replication effect may be due to other reasons besides too much or too little adaptation, notably inflated evidence of effectiveness in the original study (Movsisyan et al., 2019). It is generally accepted, then, that adopting an intervention in a new context requires making some changes to the intervention and context to achieve optimal ‘fit’ (Evans et al., 2019; Movsisyan et al., 2021). Done well, this can contribute to improved implementation, effectiveness and maintenance (Escoffery et al., 2018). Abstract There is a pressing need to prevent and address youth crime and violence owing to its prevalence, harms and cost to society. Interventions with proven effectiveness in doing this exist. Adopting and adapting them in new contexts is potentially cost- effective. However, more research is needed into how to make adaptations that en hance intervention implementation, effectiveness and maintenance in new settings. This article reports the pre-implementation adaptation work involved in transporting Becoming a Man (BAM) from the US to the UK. BAM is a selective school-based youth development program for 12–18 year-old boys that aims to improve school engagement and reduce interactions with the criminal justice system. We describe the nature of and rationale for adaptations and identify learning for future adapta tion efforts. An adaptation team comprising the intervention developers, new pro viders and the evaluators met weekly for 10 weeks, applying a structured, pragmatic and evidence-informed approach to adapt the BAM curriculum and implementation process. Changes were informed by documentary analysis, group-based discussions and site visits. The group agreed 27 changes to the content of 17/30 lessons, at both surface (e.g., cultural references) and deep (key mechanisms or concepts) levels. Of 28 contextual factors considered, 15 discrepancies between the US and UK were identified and resolved (e.g., differences in staffing arrangements). Strengths of the process were the blend of expertise on the adaptation team in the program and local context, and constant reference to and ongoing refinement of the program theory of change. Limitations included the lack of involvement of school staff or students. Further research is needed into potential conflicts between stakeholder perspectives during adaptation and whose views to prioritise and when. Keywords Adaptation · Implementation · Mentoring · Prevention · School · Vi l Keywords Adaptation · Implementation · Mentoring · Prevention · School Vi l Keywords Violence 1 3 3 Journal of Prevention (2023) 44:729–747 730 1 3 Introduction A pragmatic, science-informed and stepwise but iterative adaptation process can help avoid ad hoc changes or program drift (Card et al., 2011; Evans et al., 2019; Moore et al., 2021; Movsisyan et al., 2019). Although several approaches to doing this exist, there is much agreement about fundamental principles (e.g., involving diverse stake holders, agreeing a way of working, protecting the theory of change and core compo nents) and steps (e.g., understanding the intervention and new community, consulting stakeholders, agreeing and making changes) (Escoffery et al., 2018; Moore et al., 2021; Movsisyan et al., 2019). However, there are few published case examples of the process in prevention and early intervention to improve youth psychosocial out comes because adaptation is often done by practitioners dynamically during imple mentation rather than a priori. One element of preventing youth crime and violence in the UK is to adopt and adapt evidence-based programs with this focus originating in other contexts. This article describes the pre-implementation adaptation process used with one such pro 1 3 Journal of Prevention (2023) 44:729–747 731 gram that originated in the US and is now being implemented for the first time in the UK. Becoming a Man (BAM) is a selective school-based youth development program targeting 12–18 year-old boys. It is a program of Youth Guidance (YG), a Chicago (US)-based non-profit organization that provides school-based social-emo tional and mental health programs across six cities in the US. It aims to improve school engagement and reduce interactions with the criminal justice system, doing so by helping boys to internalize six core values: integrity, self-determination, posi tive anger expression, accountability, respect for womanhood and visionary goal set ting. The program comprises four core activities: BAM Circles (group sessions in school settings with 8 to 12 participants), special activities (group activities outside of school property/time), brief encounters (informal check-ins), and 1:1 support. The program is delivered by prosocial male counsellors who have QCF-6 level qualifica tions and receive 300 h of BAM training. BAM circles constitute the central element and involve 50 one-hour sessions over two years (~ 25 per year) made up of check-ins and check-outs to open and close sessions, role plays, group missions, video edu cation, lectures, stories, and homework. Connections between BAM activities and desired outcomes are captured in the BAM theory of change, which was developed and adapted during the work described in this article (see below). 1 A Level 4 rating requires evidence from at least two high-quality evaluations demonstrating positive impacts across populations and environments lasting a year or longer. https://guidebook.eif.org.uk/pro gramme/becoming-a-man. 2 https://youthendowmentfund.org.uk. Adaptation Process Before making any program adaptations, the research team articulated the BAM theory of change to help guide the adaptation process and evaluation (see Green et al., 2023). This combined evidence and theory from academic literature selected by YG with local (US, UK) stakeholder expertise to articulate high-level predictions about how BAM is supposed to work, for whom, under what circumstances and why. This included: published qualitative (Lansing et al., 2016) and quantitative (Heller et al., 2013, 2017) research about BAM; theoretical frameworks for group therapy (Yalom & Leszcz, 2005), psychotherapy (Jung, 1969), youth development (Nagaoka et al., 2015), and behavior change (Michie et al., 2014), which together form the foundations of BAM’s approach to ‘action’ and ‘reflection’; systematic reviews of community-based positive youth development interventions, which are similar in their theoretical foundations to BAM, as well as reviews of process evaluations and theories of change for these interventions (Bonell et al., 2016); reviews of program documents, including the BAM circle curriculum; the observation of BAM circles in Chicago; and workshops and interviews with staff from YG and MHF. Articulating BAM’s underlying mechanisms was an important part of this pro cess. However, evaluations based on Theory of Change can sometimes overlook or under-examine the core functions of interventions (Breuer et al., 2015). This is a risk that Realist Evaluation is well-placed to mitigate, given that it offers principles and practices that help evaluators to interrogate causal relationships (Blamey & Mack enzie, 2007; Rolfe, 2019). The research team therefore adopted Realist Evaluation’s definition of a mechanism during development of the theory of change: an explana tion for how a particular set of program resources leads to a response in stakeholders’ reasoning (Dalkin et al., 2015). Pre-implementation adaptations to BAM then proceeded using a twin-track pro cess focused on curriculum and implementation respectively. First, adaptations were made to the BAM curriculum. This is contained in a 30-lesson manual covering the content that should be delivered in the BAM circle (the main program component) over two years. An adaptation team comprising representatives from YG (n = 5; oper ational, replication and evaluation expertise), MHF (n = 6; program management, delivery and evaluation roles) and the research team (n = 1; program adaptation and evaluation expertise) was set up to work through the curriculum lesson-by-lesson and make necessary adaptations. Introduction Two randomized controlled trials in Chicago, US, have shown positive impacts on numbers of arrests (for violent/all crime) and school performance for students with a mean age of ~ 15 years living in racially segregated and deprived communities (Heller et al., 2013, 2017). BAM has the highest rating on the Early Intervention Foundation Guidebook, the main UK registry for evidence-based programs, reflecting the quality of these studies and robustness of effectiveness results.1 In 2020, the Mental Health Foundation (MHF) introduced the program to the UK in three secondary schools in a south London borough. BAM was selected, based on evidence for its effectiveness, to help address a recent local increase in serious youth violence. It is being implemented with young people in school years 8 to 10 (ages 12–15 years). Funding for the project comes from the Youth Endowment Fund (YEF), a government-funded What Works Centre charged with preventing children and young people from becoming involved in violence by finding out what works and supporting efforts to put this knowledge into practice.2 A YEF-funded feasibility study and pilot outcomes evaluation under taken by this research team (the authors of this article) aims to explore (i) its potential to improve outcomes in the UK and (ii) issues pertinent to further intervention deliv ery/development and a next-stage evaluation. The adaptation process aims to enhance cultural relevance and build local own ership (Movsisyan et al., 2019). The pre-implementation adaptation work, which formed part of the feasibility study (Green et al., 2023), was informed by concepts and processes from existing guidance and frameworks (Card et al., 2011; Escoffery et al., 2018; Evans et al., 2019; Movsisyan et al., 2019; Stirman et al., 2019). The objec tives of this article are to (i) describe narratively the pre-implementation adaptation 1 A Level 4 rating requires evidence from at least two high-quality evaluations demonstrating positive impacts across populations and environments lasting a year or longer. https://guidebook.eif.org.uk/pro gramme/becoming-a-man. 1 3 732 Journal of Prevention (2023) 44:729–747 process and the nature of and rationale for the adaptations made, and (ii) reflect criti cally on the strengths and limitations of the process and identify learning for future adaptation efforts. 1 3 Adaptation Process There were 10 weekly online adaptation sessions between July and October 2020, each lasting two hours and covering approximately three lessons. The delivery partners (MHF, YG) owned and led the curriculum adaptation pro cess, with adaptation decisions arrived at by consensus. The research team docu mented the process and results to help with evaluating adaptations. This included collecting information that would allow us to describe and categorize adaptations 1 3 Journal of Prevention (2023) 44:729–747 733 and identify factors that might influence their feasibility. To support this, we drew on existing guidance (see above) to develop a framework comprising seven areas: (i) target (what is adapted?); (ii) nature (how is content adapted?); (iii) rationale (why is it adapted?); (iv) degree (how much is it adapted?); (v) agents (who did the adapt ing?); (vi) implications (what else needs adapting?); and (vii) effect on the theory of change (is the theory of change adapted by default?) (Appendix A). and identify factors that might influence their feasibility. To support this, we drew on existing guidance (see above) to develop a framework comprising seven areas: (i) target (what is adapted?); (ii) nature (how is content adapted?); (iii) rationale (why is it adapted?); (iv) degree (how much is it adapted?); (v) agents (who did the adapt ing?); (vi) implications (what else needs adapting?); and (vii) effect on the theory of change (is the theory of change adapted by default?) (Appendix A). We created a matrix based on this framework to record each change made to the curriculum. Some elements of the framework were known in advance and therefore did not need recording, notably the agents involved in making the adaptations, and it was agreed that others, such as target and nature of changes, could be recorded post hoc based on key information about the changes made. Thus, the matrix comprised these categories: (i) lesson (1 to 30); (ii) adaptation (a description of the change agreed); (iii) rationale; (iv) whether the adaptation was surface-level or deep,3 includ ing its effect on the BAM theory of change; (v) whether other adaptations were needed because of the adaptation in question (Yes/No, explanation); and (vi) the date the adaptation was agreed. The MHF Research Manager for BAM attended all cur riculum adaptation sessions and documented changes in the matrix. The second set of adaptations concerned implementation. 3 The distinction between the two rests on whether an adaptation interferes with the core functioning of the programme. Changes to cultural references, where both language and reference change but the purpose and function of the reference remain the same. Adaptation Process The contexts of London and Chicago differ in many ways, from the challenges youth face and the nature of the school environment to the wider systems and communities within which youth and schools are embedded. Within BAM, there are a series of ‘implementation teams’ designed to ensure that these contextual factors support rather than impede the imple mentation and impact of BAM. To increase the likelihood of a good intervention- context fit, it was necessary to consider differences in these wider contextual factors and their implications for the responsibilities and activities of BAM’s implementa tion teams. The context areas we considered were drawn from the BAM theory of change and relevant literature. They cover features identified by Craig et al. (2018): epidemiological; social and economic; cultural; geographic / environmental; service / organizational; ethical; policy; legal; financial; political; historical; and external shocks / cataclysmic events. They also cover the ‘inner setting’ and ‘outer setting’ domains in the Consolidated Framework for Implementation Research (Damschro der et al., 2009, 2022). Bearing these in mind, the adaptation process for implementation involved the fol lowing: (i) describing the situation in the original (US) implementation setting(s); (ii) identifying and explaining the factors hypothesized to influence outcome variation in the original setting; (iii) providing evidence of contribution or causation to sup port this hypothesis in the original setting(s); (iv) describing the situation in the new setting(s); (v) identifying discrepancies between settings and explaining why they need resolving; (vi) describing and justifying the adaptation in response; and (vii) ensuring that adaptations align with the BAM theory of change. As with curriculum adaptation, we created a matrix to document changes. This was populated using several data collection methods: analysis of program documen tation and studies; a visit by a member of the research team (FG) to Chicago to 1 3 734 Journal of Prevention (2023) 44:729–747 observe BAM delivery in schools and meet stakeholders; and interviews and group- based discussions involving MHF and YG staff (developers, trainers, local imple menters, other relevant stakeholders). The process was iterative to allow opportunity for refinement as new information emerged. Curriculum Adaptations A total of 27 changes were made to the content of 17 out of 30 lessons. Surface adap tations are described in Table 1. Deep adaptations fell into three categories. Example(s) In a session on accountability, the counsellor shares a story with students about the actions of a travelling salesman who gets lost. This prompts a discussion about being accountable for one’s actions. In the story, distance is denoted in terms of ‘city blocks’, which in the UK version became ‘streets’. In a session about self-determination, having ‘basketball tryouts’ is cited as a reason young people might give for at tending school. In the UK version, ‘basektball’ was replaced with ‘football’. In a session on accountability, the counsellor shares a story with students about the actions of a travelling salesman who gets lost. This prompts a discussion about being accountable for one’s actions. In the story, distance is denoted in terms of ‘city blocks’, which in the UK version became ‘streets’. In a session on accountability, the counsellor shares a story with students about the actions of a travelling salesman who gets lost. This prompts a discussion about being accountable for one’s actions. In the story, distance is denoted in terms of ‘city blocks’, which in the UK version became ‘streets’. In a session about self-determination, having ‘basketball tryouts’ is cited as a reason young people might give for at tending school. In the UK version, ‘basektball’ was replaced with ‘football’. Superficial changes to language, where the reference or meaning of the word(s) remains intact. Changes to cultural references, where both language and reference change but the purpose and function of the reference remain the same. In a session about self-determination, having ‘basketball tryouts’ is cited as a reason young people might give for at tending school. In the UK version, ‘basektball’ was replaced with ‘football’. In a session about integrity, there is a role play in which one student borrow money from another student but never returns it. The amount borrowed ($10) was changed to £10. Films that may resonate more with a US audience were replaced with those deemed better suited to London youth. For example, in a session on self-determination students are shown clips from the film Miracle, focusing on characters in the US ice-hockey team who push themselves to achieve their goals. This was replaced with clips from the film Pur suit of Happyness, which is about a man who experiences homelessness before becoming a successful stockbroker. Tribal Societies BAM is designed to be a two-year rite of passage from boyhood to manhood. The concept is a spiritual one, in the sense of encouraging young people to connect to BAM is designed to be a two-year rite of passage from boyhood to manhood. The concept is a spiritual one, in the sense of encouraging young people to connect to Table 1 Surface adaptations to the BAM curriculum Nature of adaptation Example(s) Superficial changes to language, where the reference or meaning of the word(s) remains intact. In a session on a with students ab gets lost. This p f ’ i Table 1 Surface adaptations to the BAM curriculum Nature of adaptation Example(s) Table 1 Surface adaptations to the BAM curriculum Example(s) They naturally attract or repel young people and provide shortcuts, helping youth to learn and internalize a concrete manifestation of what each value is and is not, rather than offering an abstract assortment of defining features. Here, the connection to something bigger makes the journey to manhood meaningful and important, which motivates young people to stay the course and practise the core program values. Upon review, MHF and YG saw that the phrase ‘tribal societies’ is problematic, and both organizations sought to change the language. In particular, MHF felt that the term ‘tribal’ had problematic colonial and imperialist associations in the UK context. Specifically, the phrase was emblematic of the Eurocentric tendency to inappropri ately group together alternative cultures and lifestyles as ‘tribal’ and overlook the differences between them. Consequently, references to ‘tribal’ were replaced with ‘communal’. Both YG and MHF felt this would allow the concept of rites of passage to retain its historical significance, while allowing participants to bring their own histories to the idea. ‘Savage’ and ‘Warrior’ Energy The ‘savage’ and the ‘warrior’ are examples of additional archetypes used in BAM, here as an important mechanism of change. During sessions on Positive Anger Expression (one of six core values), ‘savage’ is used to conjure up images of destruc tive, uncontrolled anger that creates guilt and shame, while ‘warrior’ is associated with constructive, controlled anger that brings dignity and integrity. Upon discussion, MHF felt these concepts were problematic. First, both have vio lent connotations, which could unintentionally suggest that anger equates to violence. Second, ‘savage’ has been used historically as a derogatory term for indigenous peo ples. Third, ‘savage’ is already used as a colloquialism among youth in London, often as a compliment to imply strength. As a result, MHF and YG agreed to use the terms ‘constructive’ and ‘destructive’ as archetypes. Example(s) Amending the timing and structure of sessions on ‘academic integrity’1 to account for differences in the assessment processes between the two countries Adapting to UK COVID-19 restrictions Replacing the language used during ‘check-ins’ with terms more commonly used by London youth (every session should include at least one check-in, where the group takes turns to describe how they are feeling and why). Aligning sessions with the release of grades throughout the school year. Ensuring students use hand sanitizer prior to group sessions. Certain rituals occur during check-ins, the most important being the way the group responds to someone checking in by saying “Asé”, which means “I’m with you” or “I hear you” in Yoruban language. In London, this was replaced with “safe”, “say less” or “calm”, colloquialisms common in London and used, among other things, as greetings. 1 During academic integrity sessions, participants take turns to update the group on their most recent grades. The group then affirms those students who have passed all their classes, and challenges those who have failed all or some of their classes. 1 3 3 Journal of Prevention (2023) 44:729–747 735 something bigger than themselves. In BAM, this ‘something’ is the universal experi ence of transitioning into manhood. BAM frames rites of passage as coming from and being grounded in ‘tribal’ societies, and uses various archetypes throughout the cur riculum to highlight their timeless nature. Archetypes are models of people, behav iors or personalities which have universal meanings across cultures (Jung, 1969). They naturally attract or repel young people and provide shortcuts, helping youth to learn and internalize a concrete manifestation of what each value is and is not, rather than offering an abstract assortment of defining features. Here, the connection to something bigger makes the journey to manhood meaningful and important, which motivates young people to stay the course and practise the core program values. something bigger than themselves. In BAM, this ‘something’ is the universal experi ence of transitioning into manhood. BAM frames rites of passage as coming from and being grounded in ‘tribal’ societies, and uses various archetypes throughout the cur riculum to highlight their timeless nature. Archetypes are models of people, behav iors or personalities which have universal meanings across cultures (Jung, 1969). No Discrepancies There are two broad areas where no discrepancies were found. First, the counsellors’ profiles are largely the same in terms of ethnicity (African American/Black British), gender (male), age (no younger than the age at which most people graduate college) and the extent to which they were matched with schools; counsellors who work with more challenging youth in the US must have prior experience of working with these groups, a policy adopted in the UK for the counsellor working in a Pupil Referral Unit.6 Second, when recruiting young people, YG ensures that: (i) BAM is marketed as a social-emotional program to help all youth, not a behavioral program for those displaying behavior problems; and (ii) the cohort presents with a range of social- emotional strengths and challenges, including externalizing and internalizing con cerns. This helps to prevent both negative labelling and any sense among students that BAM somehow ‘rewards’ anti-social behavior. These policies were adopted in the UK. 4 BAM is now delivered in US sites besides Chicago, notably Boston, Los Angeles and Seattle. 5 Of these, subsequently, for five no discrepancy has been identified, at least beyond the normal range of variation in the US, for one some discrepancies have been identified and addressed, and for one the level of discrepancy remains unknown. 6 A PRU is a form of alternative educational provision for children who cannot attend mainstream school owing to behavior which would disrupt other students. The ‘Liberator’ and the ‘Oppressor’ The two most important archetypes introduced during the ‘Respect for Womanhood’ core value are ‘self-liberator’ and ‘oppressor’. The former is associated with men who share their power with women and the latter with men who use that power to weaken or subjugate women. MHF felt that the term ‘liberator’ did not challenge the fundamental issue that men hold power in the first place and risked reinforcing the idea that women are weak and only gain power when men grant it to them. As a result of this discussion, YG corrected an error in the BAM curriculum, ensuring use of the term ‘self-liberator’ – to reflect that it is not an action being ‘done to’ women but rather a change that must take place internally for men (i.e., not to ‘liberate’ women, 1 1 3 Journal of Prevention (2023) 44:729–747 736 but to liberate themselves from negative perceptions or stereotypes of women). It was also agreed to further emphasize the meaning of that term during the curriculum. Implementation Adaptations Prior to implementation, 28 contextual factors were identified, with potential discrep ancies between BAM’s US4 and London contexts documented for each (Table 2). For five of these, no discrepancies were identified. For 15, discrepancies were identi fied and resolved. For seven, it was unknown whether there would be discrepancies because the project was in its infancy.5 For the final factor, discrepancies were identi fied but not addressed pre-delivery (since partially addressed). 4 BAM is now delivered in US sites besides Chicago, notably Boston, Los Angeles and Seattle.ii Problematic in the US, but Not in the UK Five contextual factors identified as previously occurring implementation barriers for BAM in the US were considered less salient in the UK, requiring no action. First, while teacher strikes can be frequent in the US, they are less common in the UK (at least in the recent past). Second, severe winter weather, which can regularly interrupt delivery in the US, is less of a barrier to implementation in the UK. Third, 1 3 Journal of Prevention (2023) 44:729–747 737 Table 2 Categorization of adaptations to implementation Broad category1 Specific category2 No discrepancy [n=5] Counsellor ethnicity Delivery Demographic Counsellor age Delivery Demographic Counsellor gender Delivery Demographic Match counsellors to school Implementation teams Selection of counsellors (pre-implementation) Marketing to avoid negative labelling Implementation teams School implementation team Discrepancies identified and resolved [n=15] Teacher strikes Delivery External shocks Winter weather Delivery Geographical / environmental Background checks on counsellors Delivery Legal Spirituality Delivery Cultural Routine data Implementation teams EQI team Viability of brief encounters (COVID-19) Delivery Service / organizational Remote delivery (COVID-19) Implementation teams BAM Training Academy School staff energy (COVID-19) Implementation teams School implementation team Counsellor recruitment (COVID-19) Implementation teams Selection of counsellors (pre-implementation) Counsellor training/coaching (format, length, frequency) (COVID-19) Implementation teams Core team School-level agreement Implementation team School implementation team Counsellor recruitment Implementation teams Selection of counsellors (pre-implementation) External champions Delivery Political Engagement / collaboration with other services Delivery Service / organizational Parent engagement Participation Social / economic Unknown pre-implementation [n=7] Stakeholder attitudes towards masculinity Delivery Cultural Rivalry with other local youth work organizations Delivery Historical Counsellor competencies Delivery Service / organizational Counsellor turnover Delivery Service / organizational Participants – demographic profile Participation Demographic Participants – epidemiological profile Participation Epidemiological Advisory Council Implementation teams Advisory Council Identified but not addressed [n=1] Clinical supervision Implementation teams Core team 1 Delivery context; Participation context; Implementation teams 2 Type of (a) context and (b) implementation team implementation delays have occurred in the US due to difficulties with counsellors clearing background checks in particular school districts, but counsellors in London were registered and cleared to work with minimal difficulty. Fourth, stakeholder atti tudes to the role of spirituality in BAM were deemed less problematic in the UK. In the US, local government partners in some sites had concerns that the spiritual con 3 3 Journal of Prevention (2023) 44:729–747 738 notations of ‘rites of passage’ might jeopardize the separation of religion and state in school. 7 The task of supporting the team with data and related activities will become more challenging as addi tional schools are added in the UK. Problematic in the US, but Not in the UK In the UK, by contrast, it is more common for state schools to be associated with religion, indeed one participating school is faith-based. Finally, the extent to which counsellors and implementation support staff participate in interpreting and acting on routine data is variable in the US, but less of a concern in the UK. This is partly because the Client Management System in London is more flexible than that used in the US, so MHF can tailor it to counsellors’ needs. Additionally, the MHF research manager has more time to support counsellors than would be normal for her US equivalent (because currently there are only three counsellors in the UK).7 COVID-19-related Adaptations The COVID-19 global pandemic and associated lockdown restrictions (including school closures) coincided with the feasibility phase of the BAM evaluation and therefore needed to be addressed in the pre-implementation phase. The first COVID- 19-related adaptation concerned ‘brief encounters’, incidental, informal and unsched uled contacts between youth and counsellors that happen spontaneously during the school day. These usually occur between classes in a communal space in the school building, although they can include a student dropping into the BAM room to talk with a counsellor. They allow BAM participants to practise core values in different contexts and settings, while also supporting the development of the participant-coun sellor relationship. YG typically expected counsellors to reach 80% of their casel oad per week with brief encounters. Due to COVID-19, however, it was difficult for counsellors to spend long periods of time in indoor communal spaces, so they were encouraged to engage young people in brief encounters in other safe spaces (e.g., the playground). As this was not mandated during COVID-19 and no ‘minimum level’ was required of counsellors in that period, fewer brief encounters were expected. The second COVID-19-related adaptation concerned the mode of delivery. While London was delivering BAM in-person, US sites were delivering online. To facilitate this, the BAM Training Academy designed a curriculum to keep students engaged through online means during this challenging time (i.e., session plans delivered via video calls). While counsellors in London were able initially to deliver face-to-face in schools, the online curriculum was available for when students were unable to attend school for a prolonged period (3–4 months depending on the school and group). The third adaptation triggered by COVID-19 concerns schools. It is important that school leadership prioritizes and champions BAM, otherwise teachers may be less likely to let students leave class for BAM sessions or to consider BAM’s needs when making decisions that impact on BAM (e.g., regarding which activities take priority for limited classroom space). The energy needed to adapt to COVID-19 pre vented school staff from being able to prioritize BAM. MHF responded by support ing schools’ COVID-19 response. This included developing mental health resources for teachers and students and supporting the implementation of COVID-19 restric tions in school. 9 Black Thrive aims to address the inequalities that negatively impact on the mental health and well-being of Black people in the London borough where BAM is being implemented: www.blackthrive.org. Colour ful Minds aims to improve public understanding of mental health and mental illness through education in the Black and minority ethnic community: www.colourful-minds.org.uk. COVID-19-related Adaptations It was anticipated that fostering positive relationships in this way 1 3 Journal of Prevention (2023) 44:729–747 739 would demonstrate that counsellors are core members of the school community and also mitigate the impact on BAM of schools’ focus on COVID-19. would demonstrate that counsellors are core members of the school community and also mitigate the impact on BAM of schools’ focus on COVID-19. Finally, COVID-19 restrictions meant that counsellor recruitment interviews, training and coaching were moved online. Training and coaching sessions were shortened (two hours rather than a half or whole day) and held more frequently (every week rather than once a month) to accommodate this change in format. Group obser vations that are a normal part of coaching activities did not occur. 8 A contract between the school and BAM which details the responsibilities of each. Staffing Discrepancies In the US, supervisors provide day-to-day administrative, project management and co-ordination support. This includes taking an active role in multiple implementation teams, notably the school implementation team. In BAM’s replication sites (Boston, Los Angeles, Seattle), the supervisor’s role is usually supplemented by a senior staff member from the local delivery organization, who undertakes more external-facing and strategic responsibilities. For example, they are a member of the Leadership Team and often responsible for establishing and maintaining the Advisory Coun cil (see below), two important implementation teams. In London, MHF recruited a project manager as the supervisor. The commitment and skill this individual demon strated meant that he assumed more responsibility than would normally be expected of a supervisor, including being a central figure in the Leadership Team. Responding to BAM’s Lack of Profile in London In the US, BAM was publicly championed by then President Obama and has a strong evidence base. This makes it an attractive prospect for school districts, communities and parents, helping to overcome indifference or resistance. No such platform exists in the UK. Moreover, various actors already influence the issue of youth violence in London, many of which will shape or be shaped by BAM in ways that will affect delivery. MHF felt that it was essential to capitalize on this network of actors or risk antagonizing them in ways that might impede implementation. Consequently, it decided to proactively develop relationships with key stakeholders. This had three elements.i The first concerns schools. School Implementation Teams represent a formal part nership between school leadership and BAM. Their role is to ensure that each school is an enabling context for implementation and to use data to inform program-related decisions. In the UK, the School-Level Agreement8 was amended to make the lan guage more approachable and less direct (e.g., removing legalese). This was partially due to differences between countries in school governance, with schools in the UK having more delegated authority than their US counterparts to make decisions about their own provision. Additionally, schools were directly involved in recruiting BAM counsellors, something that YG had never done. Second, MHF established formal partnerships with two respected commu nity organizations in London who now support implementation. Black Thrive and Colourful Minds support MHF to develop positive relationships with participants’ local communities, including their families and other organizations and individuals in the local service network.9 The aim is to ensure that their respective efforts are complementary. Third, MHF decided to hold open parents evenings early in the school year to share information about BAM. In the US, counsellor engagement of parents has tended to be more discretionary. 3 3 Journal of Prevention (2023) 44:729–747 740 Counsellor Competencies and Retention When implementation was in its early stages, it was impossible to say whether dis crepancies exist regarding the counsellors recruited in London and those delivering BAM in the US. For example, the baseline level of competency of London BAM counsellors had not been assessed, nor was it clear how counsellor turnover might affect implementation. Since then, no discrepancy has been identified in either case; London BAM counsellors’ levels of competency (in particular youth engagement skills) are in the normally expected range in the US, albeit at the upper end of the spectrum, and there has been no counsellor turnover. Perceptions Held by Other Organizations Prior to implementing BAM in London, it was unclear whether there were discrepan cies in the perceptions of BAM held by other stakeholders in the local community (e.g., whether they endorse or object to BAM’s focus on male identity, or view BAM as a threat). These perceptions can influence implementation by reinforcing or under mining young people’s sense of belonging to BAM (an important mechanism in the theory of change). Since then, it has been established that the wider community and other institutions broadly support BAM’s focus on boys and masculinity (certainly no less than would normally be expected in the US). Moreover, counsellors have not reported rivalry with other local youth work organizations when asked, despite his torical underfunding of the sector locally. 1 3 Discrepancies Identified but Not Addressed In the US, Curriculum Specialists support counsellors to develop their competencies and deliver the curriculum with quality and fidelity. However, in the UK, clinicians must receive regular clinical supervision. In the planning stage, therefore, it was intended to add a clinical supervisor to the Core Team (the primary implementation team) in London, to supplement support provided by the Curriculum Specialist. This did not happen; a decision was made collectively that supervision from the program manager and the Curriculum Specialist together was sufficient. Advisory Council A local Advisory Council is established in each region where BAM operates. This includes local community and regional/national industry leaders. Their role is to act as a two-way liaison with local communities, promote the financial sustainability of BAM, and identify programmatic opportunities for young people. As the Advisory Council was not established in London prior to delivery, it could not be compared with other BAM sites (this remans the case). One reason for this is that it did not seem a priority in the early stages of program implementation given the limited capacity of the adaptation team, challenges related to COVID-19 and the fact that the program was already fully funded for two years. Related to that are differences in funding sources for such interventions; the US has a stronger culture of individual giving and donations from large trusts, foundations and corporates, whereas in the UK it was felt that a more sustainable and realistic funding model would be via statutory bodies responsible for planning and commissioning local health care services. It is also the case in the US that the Advisory Council is typically established once the program is running. Participants Prior to completing selection, it was not possible to determine whether discrepancies existed between London and the US regarding the demographic and epidemiological profile of BAM youth because the data did not exist. We now know that the demo graphic profile (i.e., age, ethnicity) of UK BAM students is within the range of BAM Journal of Prevention (2023) 44:729–747 741 students in the US. However, .owing to a lack of data at the time of writing, it remains unknown whether a discrepancy exists for offending and school engagement. students in the US. However, .owing to a lack of data at the time of writing, it remains unknown whether a discrepancy exists for offending and school engagement. Critical Reflection and Learning The aim of this article was to report the pre-implementation adaptation work asso ciated with transporting an evidence-based program from the US to the UK. The first objective was to describe the adaptation process and the nature of and rationale for the adaptations made. Adapting the curriculum involved a 10-week group pro cess led and owned by the purveyor and provider, with the research team recording changes, specifically what was changed, why, when, if and how it affected the theory of change and whether it required further changes. The group agreed 27 changes to the content of 17/30 lessons, at both surface (e.g., cultural references) and deep (key mechanisms or concepts) levels. Adaptations to facilitate effective implementation in a new context entailed analysis of BAM documentation and studies, a site visit to the US and interviews/discussion with the purveyor and provider. Changes were recorded, focusing on the contextual situation in the original (US) setting and its 1 3 3 Journal of Prevention (2023) 44:729–747 742 influence on outcomes, the situation in new setting, potential discrepancies in need of resolving, and the nature of and rationale for any change. Of the 28 contextual factors examined, discrepancies identified and resolved (n = 15) related to implementation barriers in the US that do not apply in the UK, COVID-19-related issues, the lack of profile for BAM in the UK and differences in staffing arrangements. For some dis crepancies (n = 7) it was too early pre-implementation to say if they would be prob lematic, namely perceptions held by other organizations, counsellor competencies and retention, participant profile and the Advisory Council. Some contextual factors were deemed not to have discrepancies between settings (n = 5) and one could not be addressed pre delivery.l The second objective, which is the focus of this section, was to reflect critically on the strengths and limitations of the process and identify learning for future adaptation efforts. The context for this is that complex psychosocial interventions are difficult to design, implement and evaluate. Transporting them from one context to another adds another layer of complexity, with many such efforts resulting in null effects when trialled. Making too many or too few changes to the intervention, or failing to appre ciate and address important contextual differences, are among the reasons given for such results (e.g., Movsisyan et al., 2021). 1 3 Critical Reflection and Learning The method was efficient and effective in identifying issues requiring attention and making requisite adaptations to the curriculum and imple mentation process. Two features of the adaptation process arguably contributed to its success (mea sured in those terms), starting with the blend of expertise on the adaptation team in the program and local context. The developer/purveyor (YG) listened respectfully to local concerns from the delivery organization (MHF) and its practitioners about possible mismatches between settings and helped to make context-sensitive adapta tions that preserved core program elements. This was facilitated by similar organi zational cultures across YG and MHF. The participatory approach to adaptation was not explicitly informed by theory, unlike some adaptation projects which have delib erately drawn on community-based participatory research principles (Movsisyan et al., 2021). Nevertheless, the work did adopt a partnership approach, with equitable involvement from all parties (developer, implementers, researchers) and a strong sense of shared decision-making and joint ownership. Other studies have reported the benefits of this for program implementation, including increased acceptability, responsiveness to local needs and likelihood of sustainability (ibid.). Collaboration is not without challenges, of course, and further research is needed into potential conflicts between stakeholder perspectives and whose views to prioritise and when (ibid.).f The other aspect of the process that supported effective adaptation was the devel opment and ongoing refinement of the BAM theory of change. This allowed adap tation conversations to be well structured, acting as a constant reference point and informing both (i) the focus and substance of discussions and (ii) decisions about the nature of and rationale for changes to the curriculum or implementation. This helped to maintain consistency with intervention functions, alongside our use of Realist Evaluation’s clear and widely-used definition of program mechanisms. By contrast, none of the studies reviewed by Movsisyan et al. (2021) reported on program theo ries of change or modifications to them, or reflected on the importance of cultural or structural factors for intervention mechanisms and how they may interact with such mechanisms to affect implementation and outcomes in new settings. Our experience did not suggest the need for significant changes to existing adaptation guidance or frameworks in this respect, although we found it helpful to consider explicitly theo ries or evidence of how implementation strategies used in the original setting con tributed to outcomes. Inevitably the pre-implementation adaptation process also had limitations. Critical Reflection and Learning Although often neglected, it is therefore important to document intervention adaptations and explore their impact on accept ability, implementation and outcomes (Chambers & Norton, 2016; Escoffery et al., 2018; Stirman et al., 2019). There is also a pressing need to document how adaptation guidance is used and reflect on its usefulness (Copeland et al., 2022). Cumulatively, such studies will further knowledge about how to adapt well and inform adaptation strategies (Chambers & Norton, 2016; Moore et al., 2021). This article contributes to this endeavour by describing the process and results of the pre-implementation adaptation phase for an evidence-based intervention originat ing in Chicago and now implemented in London. As such, it captures the complexi ties and challenges of real-world intervention adaptation. Like the majority of studies describing or evaluating adaptation projects in health services and public health looked at by Movsisyan et al. (2021), the current study focused on a micro-level intervention, but in other ways it helps to address deficiencies in the evidence identi fied in that review. Unusually, it concerns a cross-continent adaptation (most such studies are within-country, especially in the US), describes the quality of evidence that informed program selection in the new setting and is specific about the rationale for different adaptations. Further, the program theory of change was central to the adaptation effort, as was consideration of multiple aspects of context in the new set ting and how they or the program might need to be adapted to achieve an optimal fit. Also in contrast to studies in the Movsisyan et al. (2021) review, the adaptations made and described here were less about program content and more about implemen tation context. We applied a pragmatic, evidence-informed approach, borrowing concepts and steps from established adaptation guidance which, as others have found, proved use ful but too long and time-consuming to apply in full (Copeland et al., 2022). This likely explains why reported adaptation efforts in health services and public health tend not to use guidance religiously but do adopt a structured approach that is in line with guidance, underpinned by key principles and following a sequential process 1 3 Journal of Prevention (2023) 44:729–747 743 (Movsisyan et al., 2021). The method was efficient and effective in identifying issues requiring attention and making requisite adaptations to the curriculum and imple mentation process. (Movsisyan et al., 2021). Critical Reflection and Learning First, the intervention was selected prior to our involvement, so we could not undertake common pre-selection analyses of fit with local needs. Second, owing to COVID-19 restrictions, MHF staff were unable to make an intended trip to the US to observe BAM in practice prior to implementation in the UK and were therefore reliant on written and third-party accounts of the program in action. Third, COVID-19 meant that some aspects of the implementation process had not been established before delivery commenced and so possible mismatches could not be identified upfront. Fourth, the focus on adapting existing implementation approaches meant that new or alternative strategies were perhaps not considered sufficiently. Finally, school staff and young people were not involved in pre-implementation adaptation. They were, 1 3 3 744 Journal of Prevention (2023) 44:729–747 however, involved in data collection later in the feasibility phase evaluation and their insights about the experience of delivering and receiving BAM respectively contrib uted to further adaptations (Green et al., 2023). however, involved in data collection later in the feasibility phase evaluation and their insights about the experience of delivering and receiving BAM respectively contrib uted to further adaptations (Green et al., 2023). As delivery continues in London, the intervention and BAM’s implementation teams continue to change; initial adaptations are evolving and new ones emerging. We are testing intervention feasibility and potential impact on outcomes and continue to work with partners to record and evaluate adaptations, with feedback loops con tributing to ongoing adaptation. The focus is primarily on ‘deep’ rather than ‘surface’ adaptations, with the aim of establishing (i) the alignment of these adaptations with their core function, and (ii) the acceptability, appropriateness and feasibility of their form. An interesting question is whether BAM has been adapted for London or for the UK, the answer to which will only become apparent if and when it is adopted elsewhere in the UK. Further research is needed to explore whether, as we suspect, deep adaptations to interventions are likely to apply at a country level and therefore not need further significant adjustment, whereas more localized differences in cul ture, service organization, and geography will require additional surface adaptations. Supplementary Information The online version contains supplementary material available at https://do org/10.1007/s10935-023-00742-2. Critical Reflection and Learning Acknowledgements We are grateful to colleagues in Youth Guidance and the Mental Health Founda tion who were involved in the pre-implementation adaptation process, especially Chris Jaffe and Jason Story (respectively Senior Manager, Partnerships and Operations, and Senior Curriculum Developer at Youth Guidance) and Victoria Zamperoni (Research Manager for Becoming a Man (BAM) at MHF). The implementation and evaluation of BAM is funded by the Youth Endowment Fund. The time of Kate Allen, Nick Axford, Vashti Berry and Lynne Callaghan is supported by the National Institute for Health and Care Research Applied Research Collaboration South West Peninsula. The views expressed in this publication are those of the authors and not necessarily those of the National Institute for Health and Care Research or the Department of Health and Social Care. Funding The implementation and evaluation of BAM is funded by the Youth Endowment Fund. Ntale Eastmond is BAM Programme Manager for the Mental Health Foundation. Declarations Conflict of Interest Vashti Berry is an Associate Editor and Nick Axford is a member of the Editorial Board of the Journal of Prevention (these are unpaid roles). The other authors have no relevant financial or non- financial interests to disclose. Ethics Approval The evaluation of BAM was approved by the Ethics Committee of the Centre for Social Policy, Dartington (affiliated to the Dartington Service Design Lab), and the study is being performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments. Informed Consent Informed consent was not required for the work reported in this article as it did not involve collecting data from research participants. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line 1 3 3 745 Journal of Prevention (2023) 44:729–747 to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. References Allen, P. J., Thomas, K., Triggs, B., & Noam, G. G. (2017). The Holistic Student Assessment (HSA) Techni cal Report. Belmont: MA, The PEAR Institute: Partnerships in Education and Resilience. Baldus, C., Thomsen, M., Sack, P. M., Bröning, S., Arnaud, N., Daubmann, A., & Thomasius, R. (2016). Evaluation of a german version of the strengthening families Programme 10–14: A randomised con trolled trial. European Journal of Public Health, 6, 953–959. Blamey, A., & Mackenzie, M. (2007). Theories of change and realistic evaluation: Peas in a pod or apples and oranges? Evaluation, 13(4), 439–455.i Bonell, C., Dickson, K., Hinds, K., Melendez-Torres, G. J., Stansfield, C., Fletcher, A., Thomas, J., Les ter, K., Oliver, E., Murphy, S., & Campbell, R. (2016). The effects of positive Youth Development interventions on substance use, violence and inequalities: Systematic review of theories of change, processes and outcomes (p. 27253003). NIHR Journals Library. Breuer, E., Lee, L., De Silva, M., et al. (2015). Using theory of change to design and evaluate public heal interventions: A systematic review. Implementation Science, 11, 63.f Card, J., Solomon, J., & Cunningham, S. D. (2011). How to adapt effective programs for use in new con texts. Health Promotion Practice, 12(1), 25–35. Chambers, D. A., & Norton, W. E. (2016). The Adaptome: Advancing the science of intervention adapt tion. American Journal of Preventive Medicine, 51(4 Suppl 2), S124–S131. Copeland, L., Littlecott, H. J., Couturiaux, D., Hoddinott, P., Segrott, J., Murphy, S., Moore, G., & Evans, R. E. (2022). Adapting population health interventions for new contexts: Qualitative interviews understanding the experiences, practices and challenges of researchers, funders and journal editors. British Medical Journal Open, 12, e066451. Craig, P., Di Ruggiero, E., Frohlich, K. L. (2018). Taking account of context in population health inter vention research: guidance for producers, users and funders of research. Southampton (UK): NIHR Journals Library; 2018 Apr. Available from: https://www.ncbi.nlm.nih.gov/books/NBK498645/ Dalkin, S. M., Greenhalgh, J., Jones, D., et al. (2015). What’s in a mechanism? Development of a ke concept in realist evaluation. Implementation Science, 10, 49. Damschroder, L. J., Aron, D. C., Keith, R. E., Kirsh, S. R., Alexander, J. A., & Lowery, J. C. (2009). Fos tering implementation of health services research findings into practice: A consolidated framework for advancing implementation science. Implementation Science, 4, 50. Damschroder, L. J., Reardon, C. M., Widerquist, M. A. O., et al. (2022). References g, ( ) yp [ ], ume 9 (Part 1), Princeton, N.J.: Princeton University Press. g, ( ) yp [ ume 9 (Part 1), Princeton, N.J.: Princeton University Press. ( ) y Kieselbach, B., & Butchart, A. (2015). Preventing Youth Violence: An overview of the evidence. WHO. Lansing, J., & Rapoport, E. (2016). Bolstering belonging in BAM and Beyond: Youth Guidance’s becom ing a man (BAM) Program Components, experiential processes, and mechanisms. A report to Youth Guidance. Chapin Hall at the University of Chicago. Matjasko, J. L., Vivolo-Kantor, A. M., Massetti, G. M., Holland, K. M., Holt, M. K., & Cruz, D., J (2012). A systematic meta-review of evaluations of youth violence prevention programs: Common and diver gent findings from 25 years of meta-analyses and systematic reviews. Aggression and Violent Behav ior, 17(6), 540–552. Michie, S., Atkins, L., & West, R. (2014). The Behaviour Change Wheel: A Guide to Designing Interven tions. Silverback Publishing. Moore, G., Campbell, M., Copeland, L., Craig, P., Movsisyan, A., Hoddinott, P., & Evans, R. (2021). Adapting interventions to new contexts: The ADAPT guidance. Bmj, 374, n1679. Movsisyan, A., Arnold, L., Evans, R., Hallingberg, B., Moore, G., O’Cathain, A., & Rehfuess, E. (2019). Adapting evidence-informed complex population health interventions for new contexts: A systematic review of guidance. Implementation Science, 14(1), 105. Movsisyan, A., Arnold, L., Copeland, L., Evans, R., Littlecott, H., Moore, G., O’Cathain, A., Pfadenhauer, L., Segrott, J., & Rehfuess, E. (2021). Adapting evidence-informed population health interventions for new contexts: A coping review of current practice. Health Research Policy and Systems, 19, 13. p g p y y , , Nagaoka, J., Farrington, C., Ehrlich, S., & Heath, R. (2015). Foundations for Young Adult Success: A D l t l F k Th U i it f Chi C ti Chi S h l R h p g p y y Nagaoka, J., Farrington, C., Ehrlich, S., & Heath, R. (2015). Foundations for Young Adult Success: A Developmental Framework. The University of Chicago Consortium on Chicago School Research. Rolfe, S. (2019). Combining theories of change and realist evaluation in practice: Lessons from a research on evaluation study. Evaluation, 25(3), 294–316. Russell, K. (2021). What works to Prevent Youth Violence: A Summary of the evidence. Scottish Government.f Segrott, J., Gillespie, D., Lau, M., Holliday, J., Murphy, S., Foxcroft, D., & Moore, L. (2022). blisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps d institutional affiliations. References The updated Consolidated Fram work for implementation research based on user feedback. Implementation Science, 17, 75.f Escoffery, C., Lebow-Skelley, E., Haardoerfer, R., Boing, E., Udelson, H., Wood, R., & Mullen, P. D. (2018). A systematic review of adaptations of evidence-based public health interventions globally. Implementation Science, 13(1), 125. Evans, R. E., Craig, P., Hoddinott, P., Littlecott, H., Moore, L., Murphy, S., O'Cathain, A., Pfadenhauer, L., Rehfuess, E., Segrott, J., & Moore, G. (2019). When and how do ‘effective’ interventions need to be adapted and/or re-evaluated in new contexts? The need for guidance. Journal of Epidemiology and Community Health, 73(6), 481-482. Fagan, A. A., & Catalano, R. F. (2013). What works in youth violence prevention: A review of the liter ture. Research on Social Work Practice, 23(2), 141–156.fif Farrington, D. P., Gaffney, H., Lösel, F., & Ttofi, M. M. (2017). Systematic reviews of the effectiveness of developmental prevention programs in reducing delinquency, aggression, and bullying. Aggression and Violent Behavior, 33, 91–106. Fonagy, P., Butler, S., Cottrell, D., Scott, S., Pilling, S., Eisler, I., & Goodyer, I. M. (2018). Multisystemic therapy versus management as usual in the treatment of adolescent antisocial behaviour (START): A pragmatic, randomised controlled, superiority trial. Lancet Psychiatry, 5(2), 119–133. Green, F., Axford, N., Harris, J., Preece, C., Mannes, J., Allen, K., Callaghan, L., Berry, V., de Santana, E., & Woodburn, A. (2023). Becoming a man (BAM): Feasibility Study Report. Youth Endowment Fund. 1 3 Journal of Prevention (2023) 44:729–747 746 Heller, S., Pollack, H. A., Ander, R., & Ludwig, J. (2013). Preventing Youth Violence and Dropout: A Randomized Field Experiment. National Bureau of Economic Research. xperiment. National Bureau of Economic Research. Heller, S. B., Shah, A. K., Guryan, J., Ludwig, J., Mullainathan, S., & Pollack, H. A. (2017). Thinking, fast and slow? Some field experiments to reduce crime and dropout in Chicago. Quarterly Journal of Economics, 132(1), 1–54. Humayun, S., Herlitz, L., Chesnokov, M., Doolan, M., Landau, S., & Scott, S. (2017). Randomized con trolled trial of Functional Family Therapy for offending and antisocial behavior in UK youth. Journal of Child Psychology and Psychiatry, 58(9), 1023–1032. Jung, C. (1969). Archetypes and the Collective Unconscious [sic], Collected Works of C.G. Jung, Vol ume 9 (Part 1), Princeton, N.J.: Princeton University Press. Jung, C. (1969). Archetypes and the Collective Unconscious [sic ume 9 (Part 1), Princeton, N.J.: Princeton University Press. References Effective ness of the strengthening families Programme in the UK at preventing substance misuse in 10–14 year-olds: A pragmatic randomised controlled trial. British Medical Journal Open, 12, e049647. Skärstrand, E., Sundell, K., & Andréasson, S. (2013). Evaluation of a swedish version of the strengthenin families Programme. European Journal of Public Health, 24, 578–584. Stirman, S. W., Baumann, A. A., & Miller, C. J. (2019). The FRAME: An expanded framework for report ing adaptations and modifications to evidence-based interventions. Implementation Science, 14, 58. Sundell, K., Hansson, K., Löfholm, C., Olsson, T., Gustle, L. H., & Kadesjö, C. (2008). The transport ability of multisystemic therapy to Sweden: Short-term results from a randomized trial of conduct- disordered youths. Journal of Family Psychology, 22, 550–560. y f y y gy, , Yalom, I., & Leszcz, M. (2005). The theory and practice of Group Psychotherapy. Basic Books. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 1 3 747 Journal of Prevention (2023) 44:729–747 Authors and Affiliations Finlay Green1 · Nick Axford2 · Ntale Eastmond3 · Vashti Berry4 · Julia Mannes1 · Kate Allen4 · Lynne Callaghan2 · Tim Hobbs1 Finlay Green1 · Nick Axford2 · Ntale Eastmond3 · Vashti Berry4 · Julia Mannes1 · Kate Allen4 · Lynne Callaghan2 · Tim Hobbs1 Finlay Green1 · Nick Axford2 · Ntale Eastmond3 · Vashti Berry4 · Julia Mannes1 · Kate Allen4 · Lynne Callaghan2 · Tim Hobbs1 Vashti Berry v.berry@exeter.ac.uk Nick Axford nick.axford@plymouth.ac.uk Vashti Berry v.berry@exeter.ac.uk Nick Axford nick.axford@plymouth.ac.uk 1 Dartington Service Design Lab, Buckfastleigh, UK 2 University of Plymouth, Plymouth, UK 3 Mental Health Foundation, London, UK 4 University of Exeter, Exeter, UK 1 Dartington Service Design Lab, Buckfastleigh, UK 3 Mental Health Foundation, London, UK 1 3
https://openalex.org/W2955605300	http://repo.lib.semmelweis.hu//bitstream/123456789/7584/1/9758176.pdf	English	null	Effectiveness of Parameters in Quantifying Root Canal Morphology Change after Instrumentation with the Aid of a Microcomputed Tomography	BioMed research international	2,019	cc-by	5,274	Alexander Juhasz,1 Csaba Hegedus ,2 Ildiko Marton,1 Balazs Benyo ,3 Kaan Orhan ,4,5 and Csaba Dobó-Nagy 6 1Department of Restorative Dentistry, Faculty of Dentistry, University of Debrecen Medical and Health Science Center, Debrecen, Hungary g y Department of Biomaterials and Prosthetic Dentistry, Facultyof Dentistry, University of Debrecen Medical and Health Debrecen, Hungary g y Department of Biomaterials and Prosthetic Dentistry, Facultyof Dentistry, University of Debrecen Medical and Health Debrecen, Hungary g y 3Department of Control Engineering and Information Technology, Budapest University of Technology and Economics, Budapest, Hungary 3Department of Control Engineering and Information Technology, Budapest University of Technology and Economics, Budapest, Hungary p g y 4Ankara University, Faculty of Dentistry, Department of Dentomaxillofacial Radiology, Ankara, Turkey p g y 4Ankara University, Faculty of Dentistry, Department of Dentomaxillofacial Radiology, Ankara, Turkey 5OMFS IMPATH Research Group, Department of Imaging & Pathology, Faculty of Medicine, University of Le Oral & Maxillofacial Surgery, University Hospitals Leuven, Leuven, Belgium f g y y p g 6Department of Oral Diagnostics, Department of Oral Diagnostics, Faculty of Dentistry, Semmelweis University, Budapest, Hungary 6Department of Oral Diagnostics, Department of Oral Diagnostics, Faculty of Dentistry, Semmelweis University, Correspondence should be addressed to Csaba Dob´o-Nagy; dobo-nagy.csaba@dent.semmelweis-univ.hu Correspondence should be addressed to Csaba Dob´o-Nagy; dobo-nagy.csaba@dent.semmelweis-univ.hu Received 25 March 2019; Revised 29 May 2019; Accepted 11 June 2019; Published 2 July 2019 Academic Editor: Nick Silikas Academic Editor: Nick Silikas Copyright © 2019 Alexander Juhasz et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The objective of this study was to analyse the effectiveness of some parameters which characterise the change in morphology in human root canals subjected to ProTaper rotary enlargement with the help of an X-ray microfocus computed tomography (MCT) and to introduce a novel parameter that is effective in quantifying changes in root canal morphology. Ten each straight and curved root canals with mature apices chosen from extracted human upper incisor and canine teeth were scanned with MCT before and after canal shaping using ProTaper rotary instruments in order to facilitate three-dimensional digital reconstruction and quantitative gauging of relevant instrumental parameters and changes therein (surface area and volume). Root canal geometry change and the effectiveness of shaping were quantified with Structure Model Index change (ΔSMI) and surface area change to volume change ratio (ΔSA/ΔV). These two parameters were also tested on simulated canals. Hindawi BioMed Research International Volume 2019, Article ID 9758176, 6 pages https://doi.org/10.1155/2019/9758176 Alexander Juhasz,1 Csaba Hegedus ,2 Ildiko Marton,1 Balazs Benyo ,3 Kaan Orhan ,4,5 and Csaba Dobó-Nagy 6 Postinstrumentation cross-sectional changes were also analysed, but only on the plastic blocks. Statistical analysis of parameters was carried out to verify the significance of results. Analysis of cross-sectional shape of postinstrumented resin simulated canals showed statistically significant decrease in Form Factor (p<0.05) and statistically significant increase in Eccentricity (p<0.005). ΔSMI did not show significant difference between straight and curved canals. SMI values showed bidirectional change during root enlargement whichquestions the reliability of this metric in analysing instrumentation. Statistically significant (p<0.005) deviations in ΔSA/ΔV were quantified as 1.92 and 3.22 for straight and curved human canals, respectively. Instrumentation-induced canal geometry change was determined to be more pronounced in curved canals using the novel parameter ΔSA/ΔV. This has been proven as being a statistically accurate and reproducible parameter for quantitative characterisationof root canal geometry change and differentiation of preparational efficacy for both straight and curved root canals. 1. Introduction influential on the efficacy of canal preparation. Frequency and magnitude of canal aberrations (e.g., zip, elbow, perforations, and asymmetric preparation) have been proven as being more prominent in curved root canals than straight ones [2]. Root canal instrumentation is strongly affected by canal con- figuration [1] with studies showing morphology to be highly 2 BioMed Research International X-ray microfocus computed tomography (MCT) has been extensively applied as a reliable methodology for the quantitative evaluation of root canal instrumentation [3, 4]. And for the evaluations the following parameters have been used to characterise the quality of instrumentation previously: surface area change (ΔSA), volume change (ΔV), Structure Model Index change (ΔSMI) and Centre of Mass change (CM shift). The ΔSMI parameter reflects the cross- sectional change in the root canal after instrumentation, especially for oval and round cross-sectional root canals. The natural root canal has concave and convex surface irregularities. Nevertheless, a simplified equation for the SMI for root canals has been proposed as 2.2. Root Canal Preparation of Natural Human Teeth. All of the twenty chosen root canals were prepared using the ProTaper protocol with a torque-controlled, constant speed endomotor (Technika, Dentsply-Maillefer, Switzerland). The SX shaping file was used first to provide an initial flare to the orifices. An ISO 10 K-reamer (Dentsply-Maillefer, Switzerland) was used with light downwards “watch winding” motion to determine the overall working length, which was calculated by subtracting 0.5 mm from the length reading when the tip was just visible at the main apical foramina of a root. ProTaper rotary files (both shaping and finishing) were then used in “pecking” motion in the following sequences: S1 and S2 were advanced to resistance but no more than two- thirds of the canal depth; SX was then introduced into the canal in “brushing” motion to 3-5mm short of the working length followed by the use of S1 and S2 at the working length; and finally F1 and F2 finishing files were used at the working length. 𝑆𝑀𝐼= 6𝑆󸀠.𝑉 𝑆2 (1) (1) where S is the root surface area preinstrumentation; S’ is the change in the surface area caused by instrumentation; and V is the original volume of the root canal preinstrumentation [5]. g Histolith 2.5% NaOCl solution (Lege Artis, Germany) was used for irrigation with a 31-gauge needle after each file use and 30 ml in total was used for each root canal. 1. Introduction Glyde (Dentsply-Maillefer, Switzerland) was used as lubricant, with the amount enough to cover all the flute area of each file. Canal recapitulation was also performed after the use of each file. Files were regularly wiped using wet gauze to remove debris. A final flush was carried out using 5 ml sterile saline solution before drying with F2 ProTaper paper points (Dentsply-Maillefer, Switzerland). All instrumentation was performed according to the manufacturer’s instructions. In order to reduce interoperational variability all preparation was conducted by the same operator. In order to reduce contamination, one set of instruments was only used for the preparation of one canal. The study was approved by the Regional and Institutional Committee of Science and Research Ethics of Semmelweis University, Budapest, Hun- gary (IRB ID: 246/2017). Surface dilatation smooths the surface irregularities and this will reduce surface area S in equation (1) causing S’ to be negative. For root canals one can expect that the SMI value change after instrumentation reflects both oval to round cross-sectional transformation and the smoothing of the root canal walls.h The root surface area (SA) and root volume (V) have been proven as being demonstrative of the canal shape, both pre- and postinstrumentation [3, 4, 6, 7]. However, when compared with experimental groups, contradictory results have been found in the changes of these parameters, ΔSA and ΔV. In a MCT study [8] reported no signif- icant differences regarding 3D parameters of root canal preparation while on the same teeth most of 2D param- eters delivered statistically significant differences between groups. These results raise a question on usefulness of three- dimensional parameters describing canal changes during enlargement.h 2.3. Preparation of Simulated Canals. Twenty transparent resin simulated root canal blocks were used to assess the ProTaper instrumentation, amongst which ten were straight (A-ETK-0 endotrainer, Frasaco, Tettnang, Germany) and ten were curved with a curvature of 30∘(A-ETK-30 endotrainer, Frasaco, Tettnang, Germany). Therefore, the objective of this study was to quantitatively evaluate the change of root canal morphology after instru- mentation with the ratio of surface area change to volume change ΔSA/ΔV, in comparison with SMI, with the aid of MCT. Root canal preparation was carried out in the same way as was described for natural human teeth except that Glycerine was used as lubricant instead of Glyde. 3. Results and Discussion The present study focused on the quantitative characterisa- tion of the difference in canal morphology using available parameters. Despite previous studies [4, 6, 10] showing the usefulness of SMI tracking, it did not prove effective in the current work. Our observation is in line with some recently published results [11, 12] where significant differences in SMI between tested groups could not been observed. The mean values (standard deviation) of the ratio of surface area change to volume change and SMI change after the ProTaper instrumentation in straight and curved natural canals are tabulated in Table 1. It is seen that ΔSA/ΔV values are not statistically different amongst the straight canals but are statistically different amongst the curved canals (p<0.05) and are even more statistically different between the two groups (p=0.01) (power=1.00). On the contrary, ΔSMI values are not significantly different within each group nor between the groups (p=0.74) (power=0.89). Unlike ΔSA/ΔV, the SMI did not change significantly after canal instrumentation in straight nor curved human canals (Table 1). The limitation of SMI in describing root canal morphology change after instrumentation is inherent in the bidirectional change of SMI value (Figure 1). Figure 1 demonstrates an increasing movement first at F1 files and then a decreasing movement at F2 files of SMI values during instrumentation. Specifically, the cross-sectional shape change from round to irregular after eccentric instrumentation reduces the SMI value, while Figure 1: Bidirectional movement of SMI values, demonstrated by instrumentation on straight and curved resin simulated root canals. Bars represent statistical differences at ∗p<0.01 level and ∗∗p<0.001 level amongst groups of untreated, F1 prepared, and F2 prepared roots. smoothing the canal walls increases the SMI value. These opposite processes cause fluctuation in SMI values, resulting in the loss of significance in statistics. i Our experiments carried out on resin blocks helped to reduce the influence of the bidirectional change on SMI because preparation of artificial canals excluded the effect of canal wall smoothing since the canal walls prior to prepa- ration were already completely smooth. So the only effect was the round to near-oval cross-sectional shape change due to eccentric movement of the files. This finding was in line with a previous observation [13]. 2.5. Data Analysis In order to track the eccentric movements of the files, the cross-sectional shape at 14 mm coronal from the apex of the straight resin simulated canals was analysed with the CTAn software (Bruker, Kontich, Belgium). Form Factor, Roundness, and Eccentricity values were used to describe the canal cross-sectional shape changes caused by canal instrumentation. Statistical analysis was performed with one-way ANOVA and two-sample t-test. 2.5.2. Cross-Sectional Analysis. In order to track the eccentric movements of the files, the cross-sectional shape at 14 mm coronal from the apex of the straight resin simulated canals was analysed with the CTAn software (Bruker, Kontich, Belgium). Form Factor, Roundness, and Eccentricity values were used to describe the canal cross-sectional shape changes caused by canal instrumentation. Statistical analysis was performed with one-way ANOVA and two-sample t-test. F1 F2 2. Materials and Methods Bars represent statistical differences at ∗p<0.01 level and ∗∗p<0.001 level amongst groups of untreated, F1 prepared, and F2 prepared roots. 3 BioMed Research International BioMed Research International 2. Materials and Methods 2.1. Sample Selection and Pretreatment. Ten each straight and curved root canals with mature apices were selected from forty-three extracted human upper incisors and canine teeth due to severe bone loss for the study. Schneider’s classification [9] was used as the grouping criterion, where root canals with <10∘and >25∘curvature were designated as straight and curved, respectively. Teeth with mature apices having one main root canal without obliteration were included. Teeth having between 11∘and 24∘curvature were excluded. Pulp remnants were removed from each root canal using barbed broaches without any canal preparation. The teeth were then stored in 0.1% thymol solution. 2.4. MCT Scanning of the Samples. All samples, both natural human teeth and resin simulated root canal blocks, were scanned before and after instrumentation with a desktop X-ray microfocus computed tomography scanner (SkyScan 1172, Bruker, Kontich, Belgium) at 10 W, 100 kV, and 98 𝜇A, with a 0.5 mm aluminium filter, resulting in an isometric voxel size of 9 𝜇m. During data acquisition, 2D projections through 180∘of rotation were stored in digital format on an electronic media. The 3D images were obtained by filtered back-projection of a series of 2D images of adjacent cross- sections. 2.4. MCT Scanning of the Samples. All samples, both natural human teeth and resin simulated root canal blocks, were scanned before and after instrumentation with a desktop X-ray microfocus computed tomography scanner (SkyScan 1172, Bruker, Kontich, Belgium) at 10 W, 100 kV, and 98 𝜇A, with a 0.5 mm aluminium filter, resulting in an isometric voxel size of 9 𝜇m. During data acquisition, 2D projections through 180∘of rotation were stored in digital format on an electronic media. The 3D images were obtained by filtered back-projection of a series of 2D images of adjacent cross- sections. 3 Table 1: The mean values (standard deviation) of the ratio of surface area change to volume change and SMI change after the ProTaper instrumentation in straight and curved natural canals. Groups ΔSA/ΔV ΔSMI Straight 2.05ns (0.88) 0.65 ns (1.00) Curved 3.23∗ (0.30) 0.44 ns (0.47) Significance between groups ∗∗ ns ∗Statistically significant at p<0.05 level. ∗∗Statistically significant at p≤0.01 level. ns: not statistically significant (n=10). 6 5 4 3 2 1 0 SMI Untreated F1 F2 Straight Curved ∗p < 0.01, ∗∗p < 0.001 ∗ ∗ ∗∗ ∗∗ Figure 1: Bidirectional movement of SMI values, demonstrated by instrumentation on straight and curved resin simulated root canals. 2.5. Data Analysis Table 1: The mean values (standard deviation) of the ratio of surface area change to volume change and SMI change after the ProTaper instrumentation in straight and curved natural canals. 2.5. Data Analysis 2.5.1. Volumetric Analysis. CTAn software (Bruker, Kontich, Belgium) was used to measure surface area (SA) and volume (V) after manual threshold segmentation of the root canal systems. Surface area change (ΔSA) and volume change (ΔV) were then calculated by subtracting the measured values of the untreated canals from those of the treated ones. The ratio of the surface area change to volume change ΔSA/ΔV was then calculated accordingly. Triangulated data were also used to determine the Structure Model Index (SMI) of the canals. This index characterises the cross-sectional geometry of the root canal as having a plate-like shape when SMI = 0, and a rod-like shape when SMI > 3 [10]. For statistical analysis of results Mann-Whitney U test in case of natural human roots and in case of plastic blocks Kruskal-Wallis and Tukey-HSD post hoc tests were used. Power analysis was done where the case of prepared sample size was 12 (six each group); the value of the power was 0.744. Based on this calculation, it can be ascertained that, in a sample size of 20, the value of the power exceeds 0.8. Groups ΔSA/ΔV ΔSMI Straight 2.05ns (0.88) 0.65 ns (1.00) Curved 3.23∗ (0.30) 0.44 ns (0.47) Significance between groups ∗∗ ns ∗Statisticall significant at p<0 05 le el 6 5 4 3 2 1 0 SMI Untreated F1 F2 Straight Curved ∗p < 0.01, ∗∗p < 0.001 ∗ ∗ ∗∗ ∗∗ Figure 1: Bidirectional movement of SMI values, demonstrated by instrumentation on straight and curved resin simulated root canals. Bars represent statistical differences at ∗p<0.01 level and ∗∗p<0.001 level amongst groups of untreated, F1 prepared, and F2 prepared roots. 6 5 4 3 2 1 0 SMI Untreated F1 F2 Straight Curved ∗p < 0.01, ∗∗p < 0.001 ∗ ∗ ∗∗ ∗∗ Figure 1: Bidirectional movement of SMI values, demonstrated by instrumentation on straight and curved resin simulated root canals. Bars represent statistical differences at ∗p<0.01 level and ∗∗p<0.001 level amongst groups of untreated, F1 prepared, and F2 prepared roots. 6 5 4 3 2 1 0 SMI Untreated F1 F2 St i ht ∗p < 0.01, ∗∗p < 0.001 ∗ ∗ ∗∗ ∗∗ 2.5.2. Cross-Sectional Analysis. 3. Results and Discussion The mean values (standard deviation) of the ratio of surface area change to volume change and SMI change after the ProTaper instrumentation in straight and curved resin simulated root canals are tabu- lated in Table 2. ΔSA/ΔV values are statistically different both within and between the two groups (p<0.001). ΔSMI values are statistically different within the curved group (p<0.01)but are not significantly different within the straight groups nor 4 BioMed Research International Table 3: Cross-sectional shape of straight resin simulated root canals at 14 mm coronal from the apex: mean values (standard deviation) of Form Factor, Roundness, and Eccentricity pre- and postinstrumentation with the ProTaper system. Table 2: The mean values (standard deviation) of the ratio of surface area change to volume change and SMI change after the ProTaper instrumentation in straight and curved resin simulated root canals. Table 2: The mean values (standard deviation) of the ratio of surface area change to volume change and SMI change after the ProTaper instrumentation in straight and curved resin simulated root canals. Groups ΔSA/ΔV ΔSMI Straight 6.13∗∗ (1.11) 0.86 ns (0.63) Curved 7.28∗∗ (0.79) 0.74∗ (0.39) Significance between groups ∗∗ ns ∗Statistically significant at p<0.01 level. ∗∗Statistically significant at p<0.001 level. ns: not statistically significant (n=10). Figure 2: Cross-sectional images of prepared straight canals at 14 mm coronal from the apex in plastic blocks. White circles rep- resent the regular round shape and white arrows point at contours of the canals showing different levels of irregularity. Groups ΔSA/ΔV ΔSMI Straight 6.13∗∗ (1.11) 0.86 ns (0.63) Curved 7.28∗∗ (0.79) 0.74∗ (0.39) Significance between groups ∗∗ ns ∗Statistically significant at p<0.01 level. ∗∗Statistically significant at p<0.001 level. ns: not statistically significant (n=10). Groups Form Factor Roundness Eccentricity Preinstrumentation 0.898 (0.039) 0.937 (0.047) 0.08 (0.091) Postinstrumentation 0.833 (0.027) 0.901 (0.04) 0.24 (0.044) Significance between Groups ∗ ns ∗∗ ∗Statistically significant at p<0.05 level. ∗∗Statistically significant at p<0.005 level. ns: not statistically significant (n=10). Figure 2: Cross-sectional images of prepared straight canals at 14 mm coronal from the apex in plastic blocks. White circles rep- resent the regular round shape and white arrows point at contours of the canals showing different levels of irregularity. difference) in their different publications when using the same 3D analysis on natural root canals [4, 6]. Similar to bone trabecular, the natural root canals have concave and convex surface irregularities. 3. Results and Discussion Thus the SMI value increases after instrumentation if the original root canal walls are smooth and decreases if they are rough. The summation of this bidirectional change of SMI value is hardly predictable and this is why SMI value change due to root canal preparation is statistically different in some canals but not in others [8, 14]. In our study on natural human root canals statistically significant differences were not found in ΔSMI but in ΔSA/ΔV (Table 1). We have demonstrated the bidirectional change of SMI, for the first time ever to the best of our knowledge, on resin blocks where an initial increase of this parameter was followed by a decrease (Figure 1). Figure 2: Cross-sectional images of prepared straight canals at 14 mm coronal from the apex in plastic blocks. White circles rep- resent the regular round shape and white arrows point at contours of the canals showing different levels of irregularity. y g Structure Model Index (SMI) [5] was proposed to provide a quantified measure of the architectural type of cancellous bone in 1997. Since then SMI has become a widely used parameter to characterise the rod- and plate-like structure of 3D trabecular bone images. Naturally, SMI calculation is based on dilatation of the analysed bone structure, giving the surface area change relative to the dilatation (Δr), i.e., the surface area derivative (dS(r)/dr) by the dilatation distance (Δr). Accordingly, the SMI is calculated from between the groups. Therefore, SMI value changes (Table 2) of resin simulated root canals reflect solely the deformity of cross-sectional shape, while those of natural teeth reflect both the cross-sectional deformity of the canals and the smoothing of the canal walls. Cross-sectional shape change of straight resin root canals showed that even in these simplified morphology the operator usually prepares one side of the root canal wall more effectively than the others. Consequently, the cross-sectional shape of prepared root canals changed from regular to irregular, resulting in Eccentricity (Figure 2). This finding was also reflected in the trend of the determined Form Factor and Eccentricity values (Table 3). When pre- and postinstrumentation values are compared, it is seen that after the instrumentation the Form Factor decreased significantly (p<0.05), the Eccentricity increased significantly (p<0.005), and the Roundness value did not show statistical difference. between the groups. References voxel layer to the surface of the volume using the classical morphological image processing step of dilatation. However, the calculated SMI with the alternative implementations are closely correlated [16]. [1] C. D. Nagy, J. Szab´o, and J. Szab´o, “A mathematically based classification of root canal curvatures on natural human teeth,” Journal of Endodontics, vol. 21, no. 11, pp. 557–560, 1995.h Since, besides the proportions of the rod- and plate- like regions, the SMI is also strongly influenced by the pro- portions of the concave and convex surfaces, the calculated SMI may show misleading value when the proportion of the concave surface is not negligible relative to the entire bone surface. In the cases where the bone structure is modelled by rods- and plates-like elements, concave surfaces are likely to be generated in the places where these elements join. Due to this increase in the proportion of concave surface the SMI may fail to characterise trabecular architecture of the bone, as has been demonstrated by Salmon et al. [16] [2] C. D. Nagy, K. Bartha, M. Bern´ath, E. Verdes, and J. Szab´o, “The effect of root canal morphology on canal shape following instru- mentation using different techniques,” International Endodontic Journal, vol. 30, no. 2, pp. 133–140, 1997. [3] L. Bergmans, J. Van Cleynenbreugel, M. Wevers, and P. Lambrechts, “A methodology for quantitative evaluation of root canal instrumentation using microcomputedtomography,” International Endodontic Journal, vol. 34, no. 5, pp. 390–398, 2001. [4] O. A. Peters, C. I. Peters, K. Schonenberger, and F. Bar- bakow, “ProTaper rotary root canal preparation: effects of canal anatomy on final shape analysed by micro CT,” International Endodontic Journal, vol. 36, no. 2, pp. 86–92, 2003. Beside the areas where rod- and plate-like elements are joining with each other, concave surfaces can also be found in the bone structure if the resolution of the imaging modality used to study the bone is high enough to capture the surface irregularities. Consequently, the higher resolution of the applied imaging modality the higher possibility there will be for SMI to show misleading trend. This limitation of the SMI index has already been recognised by Hildebrand and R¨uegsegger where the derived values corrupted when a bone structure with a rough surface was analysed [5]. In order to eliminate this limitation, these authors then suggested smoothing the surfaces prior to the calculation of volume, surface, and surface derivative by locally averaging the vertices. References [5] T. Hildebrand and P. R¨uegsegger, “Quantification of bone microarchitecture with the structure model index,” Computer Methods in Biomechanics and Biomedical Engineering, vol. 1, no. 1, pp. 15–23, 1997. [6] O. A. Peters, K. Sch¨onenberger, and A. Laib, “Effects of four Ni-Ti preparation techniques on root canal geometry assessed by micro computed tomography,” International Endodontic Journal, vol. 34, no. 3, pp. 221–230, 2001. [7] M. S. Hartmann, F. B. Barletta, V. R. Camargo Fontanella, and J. R. Vanni, “Canal transportation after root canal instrumenta- tion: a comparative study with computed tomography,” Journal of Endodontics, vol. 33, no. 8, pp. 962–965, 2007. [8] M. F. Marceliano-Alves, M. D. Sousa-Neto, S. R. Fidel et al., “Shaping ability of single-file reciprocating and heat-treated multifile rotary systems: a micro-CT study,” International Endodontic Journal, vol. 48, no. 12, pp. 1129–1136, 2015. Hildebrand and R¨uegsegger also predicted that SMI analysis using high resolution may encounter the same problem [5]. Since the microCT modality provides high resolution image sets of root canals with uneven surfaces could be inherent in this problem. The current work proved the prediction of Hildebrand and R¨uegsegger to be true with experimental results and showed that the limitation of SMI also existed when applied to the analysis of root canal morphology. [9] S. W. Schneider, “A comparison of canal preparations in straight and curved root canals,” Oral Surgery Oral Medicine Oral Pathology, vol. 32, no. 2, pp. 271–275, 1971.h [10] O. Peters, A. Laib, P. R¨uegsegger, and F. Barbakow, “Three- dimensional analysis of root canal geometry by high-resolution computed tomography,” Journal of Dental Research, vol. 79, no. 6, pp. 1405–1409, 2016. [11] M. M. Kyaw Moe, J. H. Ha, M. U. Jin, Y. K. Kim, and S. K. Kim, “Root canal shaping effect of instruments with offset mass of rotation in the mandibular first molar: a micro–computed tomographic study,” Journal of Endodontics, vol. 44, no. 5, pp. 822–827, 2018. 3. Results and Discussion Therefore, SMI value changes (Table 2) of resin simulated root canals reflect solely the deformity of cross-sectional shape, while those of natural teeth reflect both the cross-sectional deformity of the canals and the smoothing of the canal walls. Cross-sectional shape change of straight resin root canals showed that even in these simplified morphology the operator usually prepares one side of the root canal wall more effectively than the others. Consequently, the cross-sectional shape of prepared root canals changed from regular to irregular, resulting in Eccentricity (Figure 2). This finding was also reflected in the trend of the determined Form Factor and Eccentricity values (Table 3). When pre- and postinstrumentation values are compared, it is seen that after the instrumentation the Form Factor decreased significantly (p<0.05), the Eccentricity increased significantly (p<0.005), and the Roundness value did not show statistical difference. 𝑆𝑀𝐼= 6 ⋅𝑉 𝑆2 ⋅𝑑𝑆(r) 𝑑𝑟 (2) 𝑆𝑀𝐼= lim Δ𝑟󳨀→0 6 ⋅𝑉 𝑆2 ⋅Δ𝑆(r) Δ𝑟 (3) (2) (3) where S is the surface area of the analysed structure, ΔS(r) is the surface area change as a result of dilatation by Δr, and V is the bone volume. f Limitations of 3D analysis of root canal shape changes are inherent in bidirectional changes of SMI values during the enlargement of root canals as explained above. And on the other side the sample size of natural human roots may have an importance since the ideal size of sample is suggested to determine at a lower sample size of during the experiment. In this study at n=6 the power analysis resulted in 0.74 of which calculation predicted n=10 as an acceptable final sample size. It can be interesting note that another research group used 11 and 40 sample sizes (almost four times magnitude In practical applications a triangular mesh is normally fitted to the image of the bone surface [6] so the bone surface area is approximated as the sum of the triangular mesh. Then the mesh is expanded by a short distance (Δr) away from the surface in the direction of the mesh’s vertex normals. Δ𝑆(r) is calculated as the difference between the expanded surface and the original surface. Some applications like CTAn software [15] use a voxelized representation of the bone structure and the dilatation is implemented by adding an additional BioMed Research International 5 References Data Availability The [xls files] data used to support the findings of this study were supplied by the corresponding author under license and so cannot be made freely available. Request for access to these data should be made to Csaba Dobo-Nagy, dobonagy.csaba@dent.semmelweis-univ.hu. [12] G. Yuan and G. Yang, “Comparative evaluation of the shaping ability of single-file system versus multi-file system in severely curved root canals,” Journal of Dental Sciences, vol. 13, no. 1, pp. 37–42, 2018. [13] B. Brise˜no M., L. Kremers, G. Hamm, and C. Nitsch, “Com- parison by means of a computer-supported device of the enlarging characteristics of two different instruments,” Journal of Endodontics, vol. 19, no. 6, pp. 281–287, 1993.f Conflicts of Interest The authors declare that they have no conflicts of interest. [14] G. Yang, G. Yuan, X. Yun, X. Zhou, B. Liu, and H. Wu, “Effects of two nickel-titanium instrument systems, mtwo versus protaper universal, on root canal geometry assessed by micro–computed tomography,” Journal of Endodontics, vol. 37, no. 10, pp. 1412– 1416, 2011. [16] P. L. Salmon, C. Ohlsson, S. J. Shefelbine, and M. Doube, “Structure model index does not measure rods and plates in trabecular bone,” Frontiers in Endocrinology, vol. 6, no. 162, 2015. Acknowledgments The authors thank Professor Gregory A. Chass of Queen Mary University of London for informative discussions and Dr. Krisztian Csomo for help with statistical analysis. [15] “Morphometric parameters measured by SkyscanTM CT- analyser software,” CT-Analyser software. SkyScan AB, http:// www.skyscan.be/next/CTAn03.pdf. BioMed Research International 6 [16] P. L. Salmon, C. Ohlsson, S. J. Shefelbine, and M. Doube, “Structure model index does not measure rods and plates in trabecular bone,” Frontiers in Endocrinology, vol. 6, no. 162, 2015. Acknowledgments Stem Cells International Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 MEDIATORS INFLAMMATION of Endocrinology International Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Disease Markers Hindawi www.hindawi.com Volume 2018 BioMed Research International Oncology Journal of Hindawi www.hindawi.com Volume 2013 Hindawi www.hindawi.com Volume 2018 Oxidative Medicine and Cellular Longevity Hindawi www.hindawi.com Volume 2018 PPAR Research Hindawi Publishing Corporation http://www.hindawi.com Volume 2013 Hindawi www.hindawi.com The Scientific World Journal Volume 2018 Immunology Research Hindawi www.hindawi.com Volume 2018 Journal of Obesity Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Computational and Mathematical Methods in Medicine Hindawi www.hindawi.com Volume 2018 Behavioural Neurology Ophthalmology Journal of Hindawi www.hindawi.com Volume 2018 Diabetes Research Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Research and Treatment AIDS Hindawi www.hindawi.com Volume 2018 Gastroenterology Research and Practice Hindawi www.hindawi.com Volume 2018 Parkinson’s Disease Evidence-Based Complementary and Alternative Medicine Volume 2018 Hindawi www.hindawi.com Submit your manuscripts at www.hindawi.com Hindawi www.hindawi.com Volume 2018 MEDIATORS INFLAMMATION of Endocrinology International Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Disease Markers Hindawi Publishing Corporation http://www.hindawi.com Volume 2013 Hindawi www.hindawi.com The Scientific World Journal Volume 2018 Immunology Research Hindawi www.hindawi.com Volume 2018 Journal of Diabetes Research Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Gastroenterology Research and Practice Submit your manuscripts at www.hindawi.com The Scientific World Journal Hindawi Publishing Corporation Hindawi The Scientific World Journal Diabetes Research Journal of Hindawi Endocrinology International Journal of Hindawi www.hindawi.com Volume 2018 Submit your manuscripts at www.hindawi.com Hindawi www.hindawi.com Volume 2018 BioMed Research International Submit your manuscripts at www.hindawi.com Stem Cells International Hindawi www.hindawi.com Volume 2018 Oncology Journal of Hindawi www.hindawi.com Volume 2013 Hindawi www.hindawi.com Volume 2018 Oxidative Medicine and Cellular Longevity Obesity Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Computational and Mathematical Methods in Medicine Hindawi www.hindawi.com Volume 2018 Behavioural Neurology Ophthalmology Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Research and Treatment AIDS Hindawi www.hindawi.com Volume 2018 Parkinson’s Disease Evidence-Based Complementary and Alternative Medicine Volume 2018 Hindawi www.hindawi.com Stem Cells International Hindawi www.hindawi.com Volume 2018 Oncology Journal of Hindawi www.hindawi.com Volume 2013 Hindawi www.hindawi.com Volume 2018 Oxidative Medicine and Cellular Longevity Obesity Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Computational and Mathematical Methods in Medicine Hindawi www.hindawi.com Volume 2018 Behavioural Neurology Ophthalmology Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Research and Treatment AIDS Hindawi www.hindawi.com Volume 2018 Parkinson’s Disease Evidence-Based Complementary and Alternative Medicine Volume 2018 Hindawi www.hindawi.com Obesity Journal of Hindawi www.hindawi.com Volume 2018
https://openalex.org/W4390149887	https://www.journals.vu.lt/LMR/article/download/32420/31202	Lithuanian	null	Matematika ir profesinių kompetencijų ugdymas	Lietuvos matematikos rinkinys	2,003	cc-by	2,157	Ivadas Vilniaus kolegijoje pirmieji studentai, studijuojantys pagal neuniversitetinio aukštojo mokslo studij u programas, jau yra diplomantai. Gindami diplominius projektus, studentai tur˙es irodyti igij e pasirinktos pagal studij u program a srities profesines kvaliﬁkacijas bei kompetencijas – t.y., išmanym a savo pasirinktos profesijos srityje. Tarp profesini u kvaliﬁkacij u bei kompetencij u, kurias turi igyti b¯usimasis specialistas, visose kolegijos studij u programose nurodomi tokie geb˙ejimai: geb˙eti kritiškai atrinkti, apdoroti ir analizuoti kalbin e, matematin e, kult¯uros, teis˙es, verslo, ivairi u technikos sriˇci u informacij a, geb˙eti taikyti duomen u analiz˙es metodus. Matematika d˙estoma pirmame kurse ir vienas iš svarbiausi u kolegij u tiksli uj u moksl u katedr u uždavini u yra užtikrinti permanentinio ugdymo realizavim a, suteikti ir ugdyti matematinio raštingumo bei profesinio matematinio raštingumo kompetencijas. At˙ej es studijuoti i aukšt aj a mokykl a studentas turi tik˙etis naujausi u mokymo metod u, ypaˇc in- formacini u technologij u taikymo mokymo procese. Ir tur˙et u nenusivilti. Dauguma matematikos studij u moduli u tem u yra nagrin˙ejamos ir vidurin˙eje mokyk- loje, taˇciau kolegijoje šios temos yra d˙estomos giliau ir plaˇciau, tod˙el ši u dalyk u studij u s˙ekm˙e stipriai priklauso nuo vidurin˙eje mokykloje igyt u žini u ir sugeb˙ejim u, kuri u lyg i tur˙et u atspind˙eti mokyklos pažymiai. Kolegijoje d˙estomas matematikos kursas yra taikomojo pob¯udžio ir yra orientuotas i studento pasirinkt a profesijos srit i. Šiame straipsnyje nagrin˙ejamos Vilniaus kolegijos Elektronikos ir informatikos bei Verslo vadybos fakultetuose d˙estomo matematikos kurso programos profesini u kompetencij u ugdymo poži¯uriu. 287 287 Liet. matem. rink., T. 43, spec. nr., 287–291 Liet. matem. rink., T. 43, spec. nr., 287–291  2003 Matematikos ir informatikos institutas  2003 Matematikos ir informatikos institutas Matematika ir profesini u kompetencij u ugdymas Jovita SALDAUSKIEN˙E, Vytautas VIRKUTIS (Vilniaus kolegija) el. paštas: j.saldauskiene@viko.lt Jovita SALDAUSKIEN˙E, Vytautas VIRKUTIS (Vilniaus kolegija) 1. Mokymasis ir d˙estymas kolegijoje Pasirink e neuniversitetinio aukštojo mokslo mokymo institucij a – kolegij a, studentai ti- kisi gauti daugiau taikomojo pob¯udžio žini u ir igyti praktin˙es veiklos ig¯udži u, taigi studij u programose daug laiko skirta mokomosioms praktikoms, pratyboms bei laboratoriniams darbams. Rinkos s alygomis aukštoji mokykla, orientuodamasi i savo student a, privalo ieškoti optimali u d˙estymo b¯ud u bei siekti mokymo ir mokymosi balanso. Aukštojoje mokykloje mokymasis asocijuojasi su supratim a itakojanˇciu giliu procesu: jo d˙eka atrandama, kod˙el dalykai yra tokie, kokie yra, ir iš ko jie kyla. Taigi, suprasti kod˙el yra svarbiau negu ži- noti kas. Galima sakyti, kad mokymasis aukštojoje mokykloje yra mokymasis ir kritiškas mokymosi vertinimas. 288 J. Saldauskien˙e, V. Virkutis Ivertin e s araš a geb˙ejim u, reikaling u skirting u specialybi u kolegij u studentams, maty- tume, jog mokymosi ig¯udžiai reikalingi kiekvienam, nes nuolatinis mokymasis – b¯utina bet kokios karjeros s alyga. Šiuolaikini u technologij u d˙eka žinios tampa visiems prieina- mos. Jei studentas tenkina jam keliamus reikalavimus – moka „plaukioti“ informacijos j¯uroje, naudoti turtingus palaikanˇcios aplinkos išteklius, numatyti ir strukt¯urinti savo mo- kymosi program a, iš esm˙es keiˇciasi d˙estytojo funkcijos ir paties d˙estymo samprata. D˙estymas – tai vadovavimas mokymosi procesui edukacin˙eje aplinkoje. D˙estymo ir mokymosi santykio pokytis yra ne kiekybinis – daugiau mokymosi, mažiau d˙estymo, arba atvirkšˇciai, o kokybinis – kitoks d˙estymas, kitoks ir mokymasis. Pripaž istama, kad edukacinei pažangai palankiausi yra dialoginiai santykiai. Bendroje d˙estytojo ir studento veikloje auginamas ugdymo turinys. I d˙estymo ir mokymosi santykio kait a žvelgiama kaip i vien a iš aukštojo mokslo ko- kyb˙es gerinimo galimybi u. Naujos technologijos bei inovaciniai metodai mokymo/moky- mosi procese, o taip pat šiuolaikin˙es visuomen˙es poreikiai lemia esminius mokymo/mo- kymosi pokyˇcius. Aukštoji matematika kaip atskiras modulis d˙estoma visuose Vilniaus kolegijos fa- kultetuose. Vilniaus kolegijos rengiam u specialybi u aprašuose yra nurodomi geb˙ejimai, kuri u formavim a itakoja ir lemia matematikos dalykas, svarbiausias j u – geb˙eti kritiškai atrinkti, apdoroti ir analizuoti reikalingiausi a informacij a. Kolegijos tiksli uj u moksl u katedros pos˙edžiuose aptariamos nauj u edukologini u nova- cij u matematikos d˙estyme diegimo problemos, daug dirbama informacini u technologij u– matematikos mokom uj u kompiuterini u program u isisavinimo ir taikymo srityje, kolegia- liai svarstomos specialybi u programos, si¯ulomi metodai matematikos mokymui tobulinti, svarstomi vertinimo strategijos klausimai. Matematikos svarba formuojant b¯usim uj u specialist u profesines kompetencijas, deja, kolegijoje dar n˙era suprasta reikiamu lygiu ir laukia tyrin˙ejim u bei rimtos analiz˙es. 1. Mokymasis ir d˙estymas kolegijoje Šiame straipsnyje bandysime daryti pirm asias apžvalgas. 2. Matematika kompetencij u ugdymo poži¯uriu Lygindami matematikos programas Vilniaus kolegijos Elektronikos ir informatikos (EI) ir Verslo vadybos (VV) fakultetuose, matysime, jog programos sudarytos, orientuojantis i fakultetuose vykdom u studij u program u proﬁl i. EI fakultete matematikai skirti 6 kreditai, o VV fakultete – 4 kreditai. Suprantama, jog student u, studijuojanˇci u pagal technišk uj u moksl u sriˇci u programas, igyjam u matemati- kos žini u ir kompetencij u portfelis turi b¯uti svaresnis, nei student u, studijuojanˇci u pagal ekonomini u bei vadybos moksl u sriˇci u programas. Matematikos programos sudarytos remiantis kolegijos d˙estytoj u Phare profesinio mo- kymo reformos idirbiu bei Suomijos, Danijos, Anglijos ir kit u Europos šali u koledž u pa- tirtimi. Europos ir kit u šali u autori u matematikos vadov˙eliai ir specializuoti leidiniai jau senokai tapo prieinami, j u yra kolegijos fakultet u bibliotekose bei asmenin˙ese d˙estytoj u bibliotekose. Be abejo, buvo remiamasi patirtimi, igyta lankant Europos šali u koledžus, dalyvaujant tarptautin˙ese mokslin˙ese konferencijose, dalijantis informacija su kit u Lietu- vos aukšt uj u mokykl u kolegomis. 289 Matematika ir profesini u kompetencij u ugdymas Matematika ir profesini u kompetencij u ugdymas Matematikos program u strukt¯ura pateikta 1 pav. VV fakultete studentai igyja matematin˙es analiz˙es ir tiesin˙es algebros žini u bei ﬁ- nans u matematikos ir matematini u metod u taikymo ekonomikoje bei vadybos moksluose žini u. EI fakultete studentai igyja platesni u matematin˙es analiz˙es, tiesin˙es algebros, anali- zin˙es geometrijos ir diferencialini u lygˇci u žini u bei kompleksini u skaiˇci u, Bulio algebros ir matematini u metod u taikymo elektronikos ir informacini u technologij u moksl u srityse žini u. Taigi, matematikos programos yra atitinkamai orientuotos pagal fakultetuose vyk- dom u studij u proﬁl i. Vilniaus kolegijoje Elektronikos ir informatikos fakultete matematikos paskaitose, pratybose, kontroliniuose darbuose naudojamos kompiuterin˙es programos Derive, Math- cad, Mathcad Professional 2000. Kabinete yra kompiuterizuotos darbo vietos, didesnioji student u dalis, nors dar pirmame kurse, jau turi minimalius darbo su personaliniais kom- piuteriais ig¯udžius, ir didesni u problem u, naudojant matematikos kompiuterines progra- mas, neiškyla. Verslo vadybos fakultete naudojamas nešiojamasis kompiuteris su grafo- projektoriumi. Min˙etos kompiuterin˙es programos puikiai tinka rib u, diferencial u, neapibr˙ežtini u bei apibr˙ežtini u integral u d˙estymui, vaizdžiai galima pateikti vieno ir keli u kintam uj u funk- cij u graﬁkus. Remiantis jau keli u met u patirtimi, galima teigti, jog d˙estant temas „Furj˙e eilut˙e“, „Trigonometrini u funkcij u transformavimas“, „Funkcij u graﬁkai“ kompiuterini u prog- ram u taikymas sutaupo daug laiko, pats d˙estymas tampa labai informatyvus, vaizdus ir idomus. Nagrin˙ejant tem a „Graﬁk u sud˙etis“, nubraižius kelis graﬁkus, pavyzdžiui, y = 2 sin x ir y = sin(6x), galime greitai gauti ir demonstruoti graﬁk a y = 2 sin x + sin(6x). 1 pav. Matematikos programos pagrindini u tem u išd˙estymas (valandomis). 1 pav. Matematikos programos pagrindini u tem u išd˙estymas (valandomis). J. Saldauskien˙e, V. Virkutis 290 Pavyzdžiui, Furj˙e koeﬁcient u Pavyzdžiui, Furj˙e koeﬁcient u a0 = 1 π π −π f(x) dx, an = 1 π π −π f(x) cos(nx) dx, bn = 1 π π −π f(x) sin(nx) dx, skaiˇciavimas sutaupo apie 70% laiko, nes integravimas dalimis yra gana imlus laiko at- žvilgiu. skaiˇciavimas sutaupo apie 70% laiko, nes integravimas dalimis yra gana imlus laiko at- žvilgiu. Puikiai galima atlikti atskir u apskaiˇciuot u harmonik u graﬁk u braižym a, demonstruoti harmonik u sud˙et i. Pavyzdžiui, ekrane gauname keli u harmonik u graﬁk a, demonstruo- jame j u sud˙et i. Studentai mato rezultat a ekrane ir gali palyginti duotosios funkcijos gra- ﬁk a su Furj˙e eilute gautu graﬁku. Matematika ir profesini u kompetencij u ugdymas ˇCia atsiskleidžia Furj˙e eilut˙es esm˙e. Studentai isitikina, kad Furj˙e eilute galima išskleisti bet koki a funkcij a, savarankiškai skaiˇciuoja Furj˙e koe- ﬁcientus, braižo harmonik u bei funkcij u graﬁkus. Be to, kompiuterin˙es programos puikiai tinka d˙estyti temas „Matricos ir determi- nantai“, „Tiesini u lygˇci u sistem u sprendimas“, „Finans u matematika“. Sumaniai taikant iprastin e d˙estymo metodik a bei matematikos kompiuterines mokom asias programas, ga- lima pasiekti neblog u rezultat u. Be to, kompiuterini u program u naudojimas suteikia ma- tematikos mokymui aukštesn i lygmen i, yra efektyvus, informatyvus, vaizdus, noriai stu- dent u priimamas ir labai gerai vertinamas. Galima teigti, jog aktyvi u mokymo metod u ir informacini u technologij u naudojimas matematikos mokymo procese didina student u išmanym a bei ugdo profesines kompetencijas. Išanalizavus matematikos dalyko turin i bei mokymo metodus, galima pasteb˙eti, jog reikiam u profesini u kompetencij u formavimui skiriamas pakankamas d˙emesys. Dalyko d˙estymui naudojami aktyvaus mokymo metodai, informacin˙es technologijos, kompiute- rin˙es mokomosios programos, leidžiamos kolegijos matematikos d˙estytoj u parengtos mo- komosios knygos, rengiamos savarankišk u student u studij u darb u užduotys. Matematikos dalyko vertinimo strategij a sudaro kaupiam uj u bal u sistema. Galutinis balas skaiˇciuoja- mas pagal formul e: 3 i=1 (k.d. × 0, 1)i + sav. darbai × 0, 2 + egz. × 0, 5. 3 i 1 (k.d. × 0, 1)i + sav. darbai × 0, 2 + egz. × 0, 5. Egzamino bilietuose b¯utinai itraukiami ne mažiau kaip du taikomojo pob¯udžio profe- sines kompetencijas atitinkantys uždaviniai. Taigi, programos parengtos apgalvotai, naudojami šiuolaikiniai mokymo metodai, tai- koma kaupiamojo balo vertinimo strategija, taˇciau student u pažangumo rezultatai eil e met u yra problema, kurios kol kas nepavyksta išspr esti. Galima pateikti duomenis, jog VV fakultete matematikos egzamino iš pirmo karto vidutiniškai neišlaiko ir 5 bal u nesu- kaupia apie 30% student u, o EI fakultete apie 50% student u. Ši problema pasidar˙e ypaˇc aktuali pastaraisiais metais, kai i kolegij a at˙ejo silpniau pasiruoš e vidurin˙ese mokyklose studentai. Retas kuris yra laik es matematikos valstybin i egzamin a, vyrauja laikiusieji mo- kyklin i matematikos egzamin a „B“ lygiu. VV fakultete tarp pirmakursi u pasitaiko toki u, Matematika ir profesini u kompetencij u ugdymas 291 kurie matematik a mok˙esi tik pagrindin˙ese mokyklose, o v˙eliau tiesiog šio dalyko ne- pasirinko. Išvados Matematikos dalyko turinys turi b¯uti taikomojo pob¯udžio ir atitikti konkreˇcios specia- lyb˙es studij u programos profesines kvaliﬁkacijas bei kompetencijas. Matematika ir profesini u kompetencij u ugdymas Tokie studentai pirmakursiai iš karto susiduria ne tik su didžiule adaptacijos problema, bet ir supranta, jog vidurin˙eje mokykloje ne igijo pakankamo žini u bagažo ma- tematikos studijoms kolegijoje. Silpnesniems studentams dažnai nepakanka nei valios, nei ryžto patiems stengtis iveikti atsilikim a, o d˙estytojas, dirbdamas su 60 student u au- ditorija, negali jiems skirti pakankamai d˙emesio. Žinoma, pratyb u metu, kai dirbama tik su viena grupe, d˙estytojas, taikydamas diferencijuoto mokymo metodus, gali padr asinti silpnesnius studentus, taˇciau to aiškiai nepakanka. Nemažas procentas pirmo kurso stu- dent u, ypaˇc EI fakultete, jau po poros m˙enesi u palieka kolegij a, motyvuodami tuo, jog mokytis nesugeb˙es. Sunku vienareikšmiai atsakyti i klausim a, kaip spr esti student u matematikos dalyko pažangumo, taigi matematikos dalyko kompetencij u igijimo problem a. Vienas iš galim u keli u b¯ut u pirmakursiams studentams organizuoti išlyginam asias vidurin˙es mokyklos ma- tematikos kurso studijas. Taip pat tiksli uj u moksl u katedra, matematikos d˙estytojai tur˙et u ieškoti b¯ud u student u matematikos mokymosi motyvacijai skatinti. Student u supratimas, jog matematika padeda jiems igyti ir b¯usimos profesijos kompetencijas, tikrai b¯ut u ne- menkas s˙ekming u studij u paskatas. Vis d˙elto matematikos žini u s alygojam u kompetencij u ugdymo aspektai liko paliesti tik paviršutiniškai. Kai kolegijos jau tur˙es pirm aj a absolvent u, studijavusi u pagal aukštojo neuniversitetinio mokslo studij u programas, laid a, bus galima atlikti išsamesnius moksli- nius tyrimus. Literat¯ura [1] M. Jotautien˙e, B. Jani¯unait˙e, N. Veˇckien˙e, PHARE programos diegimo, modernizuojant Lietuvos profesinio rengimo sistem a, metodologiniai aspektai, Socialiniai mokslai, 3 (1999). [2] P. Juceviˇcien˙e, Edukologijos studijos Lietuvos mokyklai, Technologija, Kaunas (1998). [3] H. Hjorth, The global portfolio, in Tarptautin˙es mokslin˙es konferencijos, ivykusios 2003 m. Minske, prane- šimo medžiaga (2003). [4] R. Laužackas, Mokymo turinio projektavimas. Standartai ir programos profesiniame rengime, Kaunas (2000). [5] J. Saldauskien˙e, V. Virkutis, Informacini u technologij u taikymas Vilniaus kolegijoje matematikos d˙estyme, in: Tarptautin˙es konferencijos Pažangios informacin˙es technologijos ir j u taikymas studij u procese, ivykusios 2002 m. Alytuje, medžiaga (2002). [5] J. Saldauskien˙e, V. Virkutis, Informacini u technologij u taikymas Vilniaus kolegijoje matematikos d˙estyme, in: Tarptautin˙es konferencijos Pažangios informacin˙es technologijos ir j u taikymas studij u procese, ivykusios 2002 m. Alytuje, medžiaga (2002). Mathematics and development of professional competencies J. Saldauskien˙e, V. Virkutis J. Saldauskien˙e, V. Virkutis This work contains the material about forms and contents of teaching of mathematics. It con- tains clear description of the aims, terms, techniques and methods of teaching as well as aspects of development of students’ professional competencies. There were made the following conclusion. Mathematics teaching module in college in higher education is expedient to use emphasizing on practically applied subject. The module has to meet the speciﬁc professional competencies of student’s specialty.
https://openalex.org/W3089630073	https://jbaar.journals.ekb.eg/article_116133_bc06b7db22c8551f7a3d2b32ee7c6cda.pdf	English	null	Effect of Polyethylene and Butylated Hydroxytoluene on the histological structure and some enzymes of rat liver	Journal of Bioscience and Applied Research/Journal of Bioscience and Applied Research	2,020	cc-by-sa	6,415	Hassan Al-Harbia, Zaki Al-Hasawib, Yaser Binnaserc, Reem Al-Hasawid , Faiz Al-Solamia and Abdullah Al- Ghamdia arbia, Zaki Al-Hasawib, Yaser Binnaserc, Reem Al-Hasawid , Faiz Al-Solamia and Abdullah Al- Ghamdia aPoison Control and Medical Forensic Toxicology Center, Jeddah, Saudi Arabia bDepartment of Biological Sciences, Faculty of Science, King Abdul Aziz University, Jeddah, Saudi Arabia cDepartment of Biology, College of Sciences, Taibah University, Al-Medina Al-Munawara, Saudia Arabia dDepartment of Chemistry Sciences, Faculty of Science and Arts, King Abdul Aziz University, Rabig, Saudi Arabia * Corresponding authors: Hassan A. Al-Harbi; Email: dfhhh25@gmail.com Received: May 10, 2020. Accepted: May 30, 2020. Published: June 11, 2020 DOI: 10.21608/jbaar.2020.116133 aPoison Control and Medical Forensic Toxicology Center, Jeddah, Saudi Arabia bDepartment of Biological Sciences, Faculty of Science, King Abdul Aziz University, Jeddah, Saudi Arabia cDepartment of Biology, College of Sciences, Taibah University, Al-Medina Al-Munawara, Saudia Arabia dDepartment of Chemistry Sciences, Faculty of Science and Arts, King Abdul Aziz University, Rabig, Saudi Arabia * Corresponding authors: Hassan A. Al-Harbi; Email: dfhhh25@gmail.com Received: May 10, 2020. Accepted: May 30, 2020. Published: June 11, 2020 DOI: 10.21608/jbaar.2020.116133 aPoison Control and Medical Forensic Toxicology Center, Jeddah, Saudi Arabia bDepartment of Biological Sciences, Faculty of Science, King Abdul Aziz University, Jeddah, Saudi Arabia cDepartment of Biology, College of Sciences, Taibah University, Al-Medina Al-Munawara, Saudia Arabia dDepartment of Chemistry Sciences, Faculty of Science and Arts, King Abdul Aziz University, Rabig, Saudi Arabia * Corresponding authors: Hassan A. Al-Harbi; Email: dfhhh25@gmail.com Received: May 10, 2020. Accepted: May 30, 2020. Published: June 11, 2020 DOI: 10.21608/jbaar.2020.116133 68 68 pISSN: 2356-9174, eISSN: 2356-9182 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 BioBacta Keywords: Albino mice, Polyethylene, Butylated Hydroxytoluene Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 69 tract and the skin and distributed within the body. It is highly soluble in fat and its volume of distribution is estimated at 2.9 l/kg (Verhagen et al. 1989). Metabolism is also in the liver and part of the BHT is metabolized with metabolites in urine and bile (Takahashi 1991). The liver is the organ in which gastrointestinal food is stored and stored for use by the rest of the body. It is the mediator between the digestive system and blood (Hamoud 2003) emerges. Feeding test mice during breast-feeding doses of BHT up to 500 mg / kg (body weight) led to a change in body weight and growth parameters in infants (Meyer and Hansen, 1980). The use of 250 mg/kg of BHT resulted in pneumococcus type I, (Yamamoto et al. 1988). BHT use of 500 mg / kg orally for experimental rats with glutathione-depleted mice caused extensive necrosis in the center of the Centrilobular liver within 24 hours (Mizutani et al. 1987). BHT between 0.01-0.5 % for 12 days led to increased liver weight in white mice, and increased dose to 800 mg / kg liver destruction occurred. Polyethelene and BHT films used in packing Fillet fishes for 120 days under -25⁰C showed tissue damage and loss in firmness (Torres-Arreola et al. 2007). Introduction PE is one of the polymers used in plastic synthesis. Previous studies indicated that PE is an inert material. Recent reports about the possible hazardous effects of plastic sheets on the environment, on animal and plant lives. Several additives are added to PE for several purposes. BHT is an antioxidant that is added to PE to prevent oxidation. Several reports attributed the toxicological effects of PE to the migration of additives from PE to the surrounding environment. (Chang et al. 1977). With today's civilized progress and rapid technological development, the plastic industry has emerged with new and easy-to-use plastic products such as bags, cups, and plastic dishes, which have been used in very large quantities to preserve meat, vegetables and most food products of all kinds. But their use extends to some types of daily food used by vendors to fill what they sell foods such as beans, yogurt, juices, and others (Shehata 1999). Studies have shown that people exposed to this substance are infected with vesicles in the bones of the limbs and inflammation of the nerves and narrowing of blood vessels (Shehata 1999). High-density PE is used in the manufacture of milk containers, thick pipes that bear pressure and water transfer, and to manufacture very thin bags used in stores (Pebsworth 1996). The use of BHT at 200 ppm caused the significant increase in the activities of enzymes relevant to the functions of liver and kidney, in contrast, caused marked pathological changes in liver and kidney of rats (El-Anany and Ali 2013) The antioxidative effects of different level (200, 400,800 and 1,200 ppm) of pomposia juice and extracts as natural antioxidant compared to BHT (200 ppm) on the stability of fried sunflower oil was investigated by Ali 2010 who found that the phenolic compounds of pomposia extract at levels 800 and 1,200 ppm induced powerful antioxidant effects which were almost equal to the synthetic antioxidant (BHT) at level 200 ppm. The aim of this research was to study the effect of PE, BHT, PE+ BHT, and the thermally-treated PE on the liver of adult male mice. Abstract This research studied the effect of polyethylene (PE), Butylated Hydroxytoluene (BHT), PE with BHT, and the thermally-treated PE on the histopathological examination and some biochemical activities of the liver of adult male mice.105 mice were divided into 5 groups. The (PE group) supplied with 20% crushed PE, the (BHT group) supplied 400mg/kg of BHT dissolved in corn oil, the (PE + BHT group) were fed with 20% crushed PE and 400mg/kg of BHT dissolved in corn oil, the (thermally treated PE group) supplied food with 20% of crushed thermally-treated PE, and the control group. The weight of the mice was recorded before the study and after 6 weeks. Blood samples were collected to determine the liver enzymes aspartate aminotransferase (AST) and alanine aminotransferase (ALT). The animals were sacrificed; the liver was weighed before and after treatments and examined microscopically. Liver specimens were stained and examined by a light microscope. The results showed that the mean weights of the mice in PE, BHT, and PE + BHT groups were statistically reduced and the mean weights of their livers significantly (p≥0.05) increased and the enzymes AST and ALT in the blood significantly (p≥0.05) increased in comparison to the control group. The histopathological examination of the liver specimens showed significant changes with all PE and BHT treatments. The results of the current investigation suggest preventing the use of plastic in food preservation and replacing it with safe grade substance like wood materials pISSN: 2356-9174, eISSN: 2356-9182 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 Toxicity Tests: 1- Control: Two groups each 5 rats one fed with corn oil, the other kept with no feeding. 1- Control: Two groups each 5 rats one fed with corn oil, the other kept with no feeding. 2 – PE group: 6 groups 10 rats each, daily fed with PE+specially mixed food, each given different rate 0.5, 1.0, 1.5, 2.0, 2.5 mg/kg body weight. 2 – PE group: 6 groups 10 rats each, daily fed with PE+specially mixed food, each given different rate 0.5, 1.0, 1.5, 2.0, 2.5 mg/kg body weight. 3 – BTH group: Divided into 4 groups each 10 rats, and fed through mouth daily different rates of BTH+specially prepared diet 400, 650, 780, 850 mg/kg body weight throughout the period of the experiment. 3 – BTH group: Divided into 4 groups each 10 rats, and fed through mouth daily different rates of BTH+specially prepared diet 400, 650, 780, 850 mg/kg body weight throughout the period of the experiment. Test of Liver Tissue: The second group of 45 rats was divided into 5 each with 9 rats. 1 – Control: given balanced food with corn oil by mouth. Nutritional experiments A total of 105 adult male Albino Mice, 3 were placed in each cage and fed with a mixture of yellow corn, some plant fiber, vitamin A, amino acids, folic acid, calcium salts, phosphorus, iron and potassium, and supplied with normal water (12 hours light and 12 hours of darkness). The mice were left in the laboratory for two weeks before the experiments were carried out to adapt to the laboratory temperature (18- 20 m) and surrounding conditions. They were again weighed at the end of the trial period (42 days) and prior to autopsy to study variations in growth measures among group members and compare them with the control group. The animals were then divided into two groups; the first with 60 animals were used for the toxicity test and divided into 6 sub-groups (10 mice each). They were subjected to treatments with different concentrations of PE and BTH for 7 days for determination of the lethal dose (LD), half lethal (LD50), and under LD50 (sub LD). The second group composed of (45 animals) and divided into 5 groups (every 9 animals), to study the effect of PE, BHT, and the thermal PE bags on liver tissue, and also the behavioral and phenotypic changes in weights, in biochemical measurements, and histological changes that took place during the trial period (42 days). Hematoxylin and Eosin stains were used for liver histological staining. AST and ALT kits were obtained from SPINREACT CO. (ref; 41272, 001172 respectively, Spain). All chemicals used were of analytical reagent grade. AST and ALT kits were obtained from SPINREACT CO. (ref; 41272, 001172 respectively, Spain). All chemicals used were of analytical reagent grade. Experimental animals Adult male Albino Mice (MF1) with similar weights, 45 days old were used in the current examination and were obtained from the King Fahd Center for Medical BHT is an indirect additive to food, adding to PE used in food packaging (Lanigan and Yamarik 2002). The BHT is rapidly absorbed through the digestive Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 pISSN: 2356-9174, eISSN: 2356-9182 70 and Technical Research at King Abdul Aziz University, Jeddah, KSA. (feeding tube) 50 mm in length, 2.25 mm in diameter. Materials and Chemicals Kit 2 – PE group: Fed with 4 mg/kg PE mixed with balanced food according to ((Kalinin et al. 1964; Sheftel 1977; Michailets et al. 1978). PE: From Sabic Company, a mixture of PE powder with rat food was prepared at a rate of 4 gm/kg body weight, and different concentrations were prepared (2, 4, 6, 8 gm/kg). 3 – BHT group: Fed with 400 mg/kg of BTH dissolved in 1.5 ml of corn oil according to (Lanigan and Yamarik 2002). BHT: From Sabic Company, 1.5 gm were dissolved in 9 ml of corn oil, and different concentrations were prepared (0.5, 1.0, 1.5, 2.0, 2.5 gm/9gm corn oil) BHT: From Sabic Company, 1.5 gm were dissolved in 9 ml of corn oil, and different concentrations were prepared (0.5, 1.0, 1.5, 2.0, 2.5 gm/9gm corn oil) 4 – PE + BHT group: Fed with 4 mg/kg PE mixed with balanced food and 400 mg/kg of BTH dissolved in 1.5 ml of corn oil. Doses were administered to rats through a metal tube Doses were administered to rats through a metal tube pISSN: 2356-9174, eISSN: 2356-9182 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 71 Sections were put on glass slides, a drop of Canada balsam was added, and now ready for test using an ordinary optical microscope (20 magnification) for histological evaluation. 5 – Thermal PE bags group: Fed with balanced food heated to more than 100 C and put inside PE bags. 5 – Thermal PE bags group: Fed with balanced food heated to more than 100 C and put inside PE bags. Toxicity Tests: PE tests: All the animals fed with the special food mixed with PE for 7 days remained alive with all concentrations. None of the concentrations is lethal. Statistical analysis Animals were anesthetized after injection by diethyl ether and blood samples (0.5-1.0 mm) were collected directly from the cardiac heart puncture by means of the capillary tube (1–1.5 m) from each rat. The blood of rats in each group was centrifuged at 1,100 Xg for 20 min to obtain the sera are stored frozen (-20 °C) until further analysis. The serum biochemical was analyzed ALT and AST activities were measured according to the methods described by Bergmeyer and Harder 1986 and Belfield and Goldberg 1971, respectively. Data were statistically analyzed in a completely randomized design in a factorial arrangement according to the procedures outlined by Gomez and Gomez 1984 and the treatments mean were compared by least significant differences (L.S.D) and Duncan multiple ranges using SPSS program package. Results: Some behavioral changes have been observed on experimental animals. These changes vary depending on the stages of the experiment. E) Decreased appetite for all animals, especially the group that fed the PE, and therefore became severely weak. These results agree with Sheftel (1977) who observation changes in the behavior of animals when given different doses of (PE), and with Lanigan and Yamarik (2002) who noticed a reduction in the appetite of animals for food and water when treated with BHT. A) In the first phase of the experiment, the mice became very aggressive while placing them in cages. B) The animal group fed on food-mixed with PE, which was deprived of food for 24 hours, were observed refusing to eat for 30-45 minutes when they sniffed and left food after a period of eating gradually. Behavioral & Observations D) Animals of the BHT group after 21 days of oral feeding were observed to hid themselves in sawdust. Histopathological study The liver was cut into small pieces (5mg3) and immersed into formaldehyde 10%, and dehydrated with absolute alcohol100% for 2 hrs, then in xylol solution, and subjected to infiltration with melted paraffin wax. The paraffin-embedded section was placed in special boxes and cold water was poured for solidification of wax. The samples were cut into slides using a microtome (5–6 μ), and then stained using Haematoxylin and Eosin to Culling 1965, and wax was removed using xylol and alcohol in a decreasing order 100%- up to 70% for 5 minutes. BTH tests: All animals fed with food mixed with 400 mg/kg body weight of BHT for 7 days remained alive, 50% of the animals died when administered doses of 650 mg/kg, 75% animals died under dose of 780 mg/kg, and all animals (100%) died when given a dose of food mixed with BHT up to 860 mg/kg body weight. Based on these results, the lethal dose of half LD50 for BHT was 650 mg / kg (body weight), the sub-LD of BHT was 780 mg / kg, and the dose of 860 mg / kg is the LD that is toxic to BHT (Fig. 1). Table 1 Results of toxicity tests by PE and BHT Groups No. of animals Dose(mg/kg) No. died Death rate Control 10 -- -- 0% PE group 10 0.5-2.5 -- 0% BHT 1 10 400 -- 0% BHT 2 10 650 5 50% BHT 3 10 780 7 70% BHT 4 10 860 10 100% Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 pISSN: 2356-9174, eISSN: 2356-9182 72 Fig. 1: lethal doses of BHT Fig. 1: lethal doses of BHT Body Weight: C) The BHT group animals were observed to have fainted after the oral feeding, which lasts for 20- 60 seconds. The body weight results showed significant variations among the treated rat's body weights due to the different treatments (table 2). Body weight of treated animals showed a significant reduction compared to untreated animals (table 2) (Fig. 2). Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 pISSN: 2356-9174, eISSN: 2356-9182 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 73 Table 2: Means of rats body weight (gm) Groups Mean ± SD Rat body wt. before treatment (gm) Rat body wt. after treatment (gm) Body wt. reduction% Control 30.85 ± 0.508 35.57 ± 0.71 PE group 31.00 ± 0.816 25.57 ± 0.36 19 BHT 2 31.14 ± 0.508 25.85 ± 0.59 19 PE + BHT 30.142 ± 0.67 22.84 ± 0.57 27 Thermal PE 30.57 ± 0.84 33.71 ± 0.56 9 Fig (2): Body wt. of rats before and after treatments Table 2: Means of rats body weight (gm) Groups Mean ± SD Rat body wt. before treatment (gm) Rat body wt. after treatment (gm) Body wt. reduction% Control 30.85 ± 0.508 35.57 ± 0.71 PE group 31.00 ± 0.816 25.57 ± 0.36 19 BHT 2 31.14 ± 0.508 25.85 ± 0.59 19 PE + BHT 30.142 ± 0.67 22.84 ± 0.57 27 Thermal PE 30.57 ± 0.84 33.71 ± 0.56 9 Fig (2): Body wt. of rats before and after treatments Table 2: Means of rats body weight (gm) Groups Mean ± SD Rat body wt. before treatment (gm) Rat body wt. after treatment (gm) Body wt. reduction% Control 30.85 ± 0.508 35.57 ± 0.71 PE group 31.00 ± 0.816 25.57 ± 0.36 19 BHT 2 31.14 ± 0.508 25.85 ± 0.59 19 PE + BHT 30.142 ± 0.67 22.84 ± 0.57 27 Thermal PE 30.57 ± 0.84 33.71 ± 0.56 9 Fig (2): Body wt. of rats before and after treatments Fig (2): Body wt. of rats before and after treatments The most affected rats are those fed with PE + BHT with a reduction in their body weight up to about 27% seconded by those treated with PE and BHT with a reduction in wt. up to about 19%. The rats treated with thermal PE have body wt. reduction of up to 9% only. Liver weight: Table 3 shows the mean liver weighs of the animal groups after the end of the experiment. The highest mean weights of the liver are in the PE group (2.33 ± 0.064 gm), the BHT group (2.61 ± 0.13 gm), and in the PE and BHT group (0.103 ± 2.88 gm) which increased significantly (P <0.01) at percentages of Table 3 Means of rat liver weight (gm) Rat liver wt. (gm) Liver wt. increase% Control 1.62 ± 0.126 -- PE group 2.33 ± 0.064 43.83 BHT 2 2.61 ± 0.13 61.12 PE + BHT 2.88 ± 0.103 77.78 Thermal PE 1.63 ± 0.081 -- 502.14 ± 12.23 U/L in case of AST, and by the treatment PE + BHT with 280.42 ± 22.25 U/L in the case of ALT, the treatment with the third level of AST is PE + BHT with 335.42 ± 14.5, and the third level of ALT is BHT with 179.71 ± 11.213 U/L. The treatment with the lowest AST and ALT is the thermal PE with 202.14 ± 3.093 and 121.42 ± 18.481 U/L respectively. This finding does not agree with the result of Elgazar (2013) who treated rats with 0.4 and 0.8 mg/kg BHT and found a reduction in AST and ALT levels. 502.14 ± 12.23 U/L in case of AST, and by the treatment PE + BHT with 280.42 ± 22.25 U/L in the case of ALT, the treatment with the third level of AST is PE + BHT with 335.42 ± 14.5, and the third level of ALT is BHT with 179.71 ± 11.213 U/L. The treatment with the lowest AST and ALT is the thermal PE with 202.14 ± 3.093 and 121.42 ± 18.481 U/L respectively. This finding does not agree with the result of Elgazar (2013) who treated rats with 0.4 and 0.8 mg/kg BHT and found a reduction in AST and ALT levels. Body Weight: These results agree with the finding of Sheftel (1977) who observation reduction in the body weight of rats given different doses of (PE), and with Mayer The most affected rats are those fed with PE + BHT with a reduction in their body weight up to about 27% seconded by those treated with PE and BHT with a reduction in wt. up to about 19%. The rats treated with thermal PE have body wt. reduction of up to 9% only. These results agree with the finding of Sheftel (1977) who observation reduction in the body weight of rats given different doses of (PE), and with Mayer and Hansen(1980) who fed infant rats 500, mg/kg BHT, and with Lanigan and Yamarik (2002) who treated rats with BHT, but differ from findings of Kalinin et al. (1964) who fed rats 2.5 g of PE powder and found no change in body weight, and also differed from the study of Michailest et al. (1978), who gave rats 1.25 g of high density PE and observed no significant toxic effects. pISSN: 2356-9174, eISSN: 2356-9182 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 74 43.83%, 61.12%, 77.78% respectively compared to control group mice (1.62 ± 0.126 gm). The mean weight of the liver in the thermally treated PE group (1.63 ± 0.081 gm) showed no significant change from the liver in control group mice (1.62 ± 0.126 gm). These results agree with the finding of Nakagawa et al. (1984) who found a significant increase in rat liver treated with PE, and BHT. Effect on AST, ALT activities Enzymes play a vital role in the evaluation of functions of the liver so levels of the two (AST) and (ALT) were determined. Table 4 illustrates the means of the concentrations of AST and ALT under the effects of the different treatments. AST and ALT concentrations increased significantly (P <0.01) under all treatments, PE gave the highest concentrations 741.42 ± 18.958 and 295.42 ± 19.614 U/L respectively for AST, ALT, seconded by BTH Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 pISSN: 2356-9174, eISSN: 2356-9182 75 Table 4 means of the concentrations of AST and ALT under the effects of the different treatments. Groups Mean ± SD AST (U/L) ALT (U/L) Control 96.71* ± 1.409 27.85 ± 2.303 PE group 741.42* ± 18.958 295.42 ± 19.614 BHT 2 502.14* ± 12.23 179.71 ± 11.213 PE + BHT 335.42* ± 14.5 280.42 ± 22.25 Thermal PE 202.14* ± 3.097 121.42 ± 18.481 Data are expressed as mean±SD Values given represent means of six determinations. *. The mean difference is significant at the 0.05 level between groups. Table 4 means of the concentrations of AST and ALT under the effects of the diffe oncentrations of AST and ALT under the effects of the different treatments. Values given represent means of six determinations. region there is damage in the vascular endothelium with blood hemorrhage (Fig. 5). The results indicated that rats fed on a diet containing PE and BHT suffered a reduction in body weight, enlargement of the liver, a disorder in the function of the liver observed in the high concentration of AST, and ALT. Under BHT treatment the hepatic cells are darkly stained, irregular in shape with nucleic atrophy which is signs of Apoptosis and the cytoplasm is granulated with a relative expansion of the liver sinusoids (S) (Fig. 6). The beginning of liver fibrosis is clear, hepatocytes (H) are large in size, with large nuclei, and abnormal chromatin shape, and loss of regularity, compared to the control group (Fig. 7). Effects on Liver Histological Structures: Figure 3 shows a transverse section of the part of the liver of the control group near the Central Vein (CV) region. Hepatocyte cells (H) are observed arranged in rows separated by sinusoidal sinuses (S). The Hepatic cells have round nuclei and cytoplasm with base pigmentation, and hepatothelial cells (S) tend for base pigmentation. The liver structure has normal order. This is in agreement with that indicated by both Mizutani et al. (1987) and Nakagawa et al. (1984) and Torres-Arreola et al. (2007). The transverse sections of the liver of the PE + BHT group appear with many darkened cells having dark-colored nuclei with eosin dye and some damaged cells and the onset of hyperplasia. Some liver cells are malignant, with haemorrhage, some sections show the destruction of the epithelial lining, and atrophy and destruction of the hepatic cells becoming dark under dye of the eosin with areas of Under treatment with PE, the liver lost its normal cell structure, and many dark hepatic cells appeared stained with Eosin, indicating that it has been affected, also the presence of atrophy, while the rest of the cells are enlarged and have a large cellular cavity near the nucleus, (Fig. 4), and near the pylori Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 pISSN: 2356-9174, eISSN: 2356-9182 76 hemorrhage , and some shows severe portal vein tension (PV) and bile duct dysfunction (BD), severe hepatocyte dysfunction is observed with haemorrhage (Figs. 8, 9 and 10). This is in agreement with the results obtained by Takahashi and Hiraga (1978) and Torres-Arreola et al. (2007). In view of the thermal PE group, severe congestion in the veins (V) and the bile ducts (BD), with obvious rupture in the liver cells, resulting in the loss of the liver lobules of their normal form. The increase in the number of central veins (CV) in the same area for the occurrence of atrophy in the liver tissue compared to the control group (Figs. 11). the control group (Figs. 11). Fig. 3 Transverse section of liver of control group. Normal arrangement of cells. CV (Central Vein), N (nucleus), H (cell), S (sinusoid). ransverse section of liver of PE group. Most of cells lost arrangement, appeared dark In color, with sis (Δ), some cells appeared inflated and had a large cellular cavity near the nucleus, which appeared t shape ↑. Fig. Effects on Liver Histological Structures: 3 Transverse section of liver of control group. Normal arrangement of cells. CV (Central Vein), N (nucleus), H (cell), S (sinusoid). Fig. 3 Transverse section of liver of control group. Normal arrangement of cells. CV (Central Vein), N (nucleus), H (cell), S (sinusoid). Fig. 4 Transverse section of liver of PE group. Most of cells lost arrangement, appeared dark In color, with apoptosis (Δ), some cells appeared inflated and had a large cellular cavity near the nucleus, which appeared crescent shape ↑. Fig. 4 Transverse section of liver of PE group. Most of cells lost arrangement, appeared dark In color, with apoptosis (Δ), some cells appeared inflated and had a large cellular cavity near the nucleus, which appeared crescent shape ↑. pISSN: 2356-9174, eISSN: 2356-9182 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 77 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 Fig. 5 Transverse section of liver of PE group near the pyloric area. Shows damage in vascular endothelium (Δ), with hemorrhage, as well as the bile duct ↑ and the emergence of hepatic cells as abnormal and malignant. The bile duct appears surrounded by a homogeneous homotoxin. Fig. 5 Transverse section of liver of PE group near the pyloric area. Shows damage in vascular endothelium Fig. 5 Transverse section of liver of PE group near the pyloric area. Shows damage in vascular endothelium (Δ), with hemorrhage, as well as the bile duct ↑ and the emergence of hepatic cells as abnormal and malignant. The bile duct appears surrounded by a homogeneous homotoxin. Fig. 6 Transverse section of liver of BHT group. Dark liver cells with irregular shape ( Δ ) and have dark nuclei (Nicrosis). Other cells are enlarged and have granular cytoplasm, and some have large-sized nuclei (↑) and relative expansion of the sinuses (S). Fig. 6 Transverse section of liver of BHT group. Dark liver cells with irregular shape ( Δ ) and have dark nuclei (Nicrosis). Other cells are enlarged and have granular cytoplasm, and some have large-sized nuclei (↑) and relative expansion of the sinuses (S). pISSN: 2356-9174, eISSN: 2356-9182 78 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 Fig. 7 Transverse section of liver of BTH group near the pyloric region. Effects on Liver Histological Structures: Hepatocellular congestion (PV) and damage in the bile duct (BD) as well as portal vein ↑ and atrophy and necrosis of hepatic cells, which appeared unclear and have lost their normal order and also the relative enlargement of the hepatic sinuses (S) and irregularity of their shape. Fig. 7 Transverse section of liver of BTH group near the pyloric region. Hepatocellular congestion (PV) and damage in the bile duct (BD) as well as portal vein ↑ and atrophy and necrosis of hepatic cells, which appeared unclear and have lost their normal order and also the relative enlargement of the hepatic sinuses (S) and irregularity of their shape. Fig. 7 Transverse section of liver of BTH group near the pyloric region. Hepatocellular congestion (PV) and damage in the bile duct (BD) as well as portal vein ↑ and atrophy and necrosis of hepatic cells, which appeared unclear and have lost their normal order and also the relative enlargement of the hepatic sinuses (S) and irregularity of their shape. Fig. 8 Transverse section of liver of the PE + BHT group shows severe portal vein tension (PV) and bile duct dysfunction (BD) ↑, severe hepatocyte dysfunction is observed with haemorrhage (☆). Fig. 8 Transverse section of liver of the PE + BHT group shows severe portal vein tension (PV) and bile duct dysfunction (BD) ↑, severe hepatocyte dysfunction is observed with haemorrhage (☆). 79 pISSN: 2356-9174, eISSN: 2356-9182 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 Fig. 9 Transverse section of the liver of the PE + BHT group showing, severe malformation, hepatic cell cirrhosis, and the onset of cirrhosis (☆ ). The other cells appeared with atrophy and unclear border. Fig. 9 Transverse section of the liver of the PE + BHT group showing, severe malformation, hepatic cell cirrhosis, and the onset of cirrhosis (☆ ). The other cells appeared with atrophy and unclear border. Fig. 10 Transverse section of the liver of the PE + BHT group near the pylori area shows bile ducts and the portal vein (PV) destructed epithelial lining ↑, and the hepatic cells appear atrophic, dilute, and dark stained with eosin with areas of hemorrhage ( ☆ ). Fig. Effects on Liver Histological Structures: 10 Transverse section of the liver of the PE + BHT group near the pylori area shows bile ducts and the portal vein (PV) destructed epithelial lining ↑, and the hepatic cells appear atrophic, dilute, and dark stained with eosin with areas of hemorrhage ( ☆ ). pISSN: 2356-9174, eISSN: 2356-9182 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 80 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 Fig. 11 T.S. of a liver in the thermal PE group The congestion and severe expansion of the portal veins (PV) with rupture of the hepatocytes (H), which leads to the loss of hepatic lobules of their normal form, expansion of the bile ducts (D) and atrophy of the epithelial lining, is also observed. Fig. 11 T.S. of a liver in the thermal PE group The congestion and severe expansion of the portal veins (PV) with rupture of the hepatocytes (H), which leads to the loss of hepatic lobules of their normal form, expansion of the bile ducts (D) and atrophy of the epithelial lining, is also observed. Figs. 1–10 1 Microscopical examination of liver tissues of rats given only the basal diet (control) (H.E X 200). 2 Microscopical examination of kidney tissues of rats given only the basal diet (control) (H.E X 200). 3 Histological examination of liver tissues of rats administered the basal diet and BHT (200 ppm.) (H.E X 200). 4 Histological examination of kidney tissues of rats administered the basal diet and BHT (200 ppm.) (H.E X 200). 5 Cross-section of liver tissues of rats administered the basal diet and Pomposia fruit juice (200 ppm.) (H.E X 200). 6 Cross-section of kidney tissues of rats administered the basal diet and Pomposia fruit juice (200 ppm.) (H.E X 200). 7 Cross-section of liver tissues of rats administered the basal diet and Pomposia fruit juice (400 ppm.) (H.E X 200). 8 Cross-section of kidney tissues of rats administered the basal diet and Pomposia fruit juice (400 ppm.) (H.E X 200). 9 Cross-section of liver tissues of rats administered the basal diet and Pomposia) fruit juice (1,200 ppm.) (H.E X 200). 10 Cross section of kidney tissues of rats administered the basal diet and Pomposia fruit juice (1,200 ppm.) (H.E X 200) Figs. 1–10 1 Microscopical examination of liver tissues of rats given only the basal diet (control) (H.E X 200). Effects on Liver Histological Structures: 2 Microscopical examination of kidney tissues of rats given only the basal diet (control) (H.E X 200). 3 Histological examination of liver tissues of rats administered the basal diet and BHT (200 ppm.) (H.E X 200). 4 Histological examination of kidney tissues of rats administered the basal diet and BHT (200 ppm.) (H.E X 200). 5 Cross-section of liver tissues of rats administered the basal diet and Pomposia fruit juice (200 ppm.) (H.E X 200). 6 Cross-section of kidney tissues of rats administered the basal diet and Pomposia fruit juice (200 ppm.) (H.E X 200). 7 Cross-section of liver tissues of rats administered the basal diet and Pomposia fruit juice (400 ppm.) (H.E X 200). 8 Cross-section of kidney tissues of rats administered the basal diet and Pomposia fruit juice (400 ppm.) (H.E X 200). 9 Cross-section of liver tissues of rats administered the basal diet and Pomposia) fruit juice (1,200 ppm.) (H.E X 200). 10 Cross section of kidney tissues of rats administered the basal diet and Pomposia fruit juice (1,200 ppm.) (H.E X 200) Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 pISSN: 2356-9174, eISSN: 2356-9182 81 Elgazar AF Elgazar AF (2013) Effects of Butylated Hydroxytoluene and Butylated Hydroxyanisole Against Hepatotoxicity Induced by Carbon Tetrachloride in Rats. World Applied Sciences Journal 22 (1): 63-69. Hamoud MH (2003) Tissue Science. El Ahlia for Publishing and Distribution,Oman. Kalinin B, Zimnitskaya Yu, Zalesskaya LP (1964) Toxic properties of stabilized low- pressure polyethylene, in Toxicology and Hygiene of High- Molecular-Mass Compounds and of the Chemical Raw Material Used for Their Synthesis, Proc. 2nd AllUnion Conf., A. A. Letavet and S. L. Danishevsky, Eds., Khimiya, Leningrad, 110 (in Russian). Lanigan RS, Yamarik TA (2002) Final Report on the Safety Assessment of BHT1. International Journal of Toxicology 21(2):19–94. Loss of arrangement and rupture of hepatocytes, damage in vascular endothelium, with hemorrhage, and the emergence of hepatic cells as abnormal and malignant. The bile duct appears surrounded by a homogeneous homotoxin, necrosis, hepatocellular congestion (PV) and damage in the bile duct (BD) as well as the portal vein, and atrophy and necrosis of hepatic cells, severe malformation, hepatic cell cirrhosis and the onset of cirrhosis and loss of hepatic lobules of their normal form. The effects are more significant under the treatment of PE + BHT compared to PE alone or BHT alone. Thermal PE is the least effective treatment of rats. Meyer O, Hansen E (1980) Behavioral developmental effects of BHT dosed to rats in utero and in the lactation period. Toxicology 16:247-58. Meyer O, Hansen E (1980) Behavioral developmental effects of BHT dosed to rats in utero and in the lactation period. Toxicology 16:247-58. Michailets B, Sukhareva LV, Yevsyukov VI (1978) Hygienic characteristics of high- density polyethylene made with help of vanadium catalysts, in Environmental Protection in Plastics Production and Hygiene Aspects of Their Use, CONCLUSION Bergmeyer HU, Harder M (1986) A colorimetric method of determination of serum glutamic oxaloacetic and pyruvic transaminase. Clin Biochem 24:28–34. Treatment of adult male rates with PE, BHT, PE + BHT, thermal PE resulted in changes in behavior and changes in body weight and liver function and histology. Rats became aggressive and tend to be isolated, accompanied by a lack of appetite. The treatments resulted in a decrease in body weight, an increase in liver weight, an increase in the enzymes AST, ALT, changes in liver histology. Examination under the PE group revealed a loss of lobular architecture with the appearance of deeply stained liver cells and atrophic nuclei (Apoptotic changes), and under BHT group showed congestion of the central and portal veins with destruction in the epithelial covering layer of the biliary canaliculi and portal veins, and under the treatment PE + BHT the effects are more significant with atrophy of the liver cells that appeared as small deeply stained cells with small atrophic nuclei with the appearance of liver fibrosis and loss of lobular architecture with severe congestion of portal veins and bile canaleculi and atrophy of the epithelial lining layer of these veins and canaliculi , with a significant increase in AST and ALT levels in blood in comparison to the control group. REFERENCES pISSN: 2356-9174, eISSN: 2356-9182 82 Journal of Bioscience and Applied Research, 2020, Vol.6, No. 2, P.68 -82 hydroxytoluene on lipid oxidation and protein quality of Sierra fish (Scomberomorus sierra) muscle during frozen storage. J Agric Food Chem 55(15):6140-6. Mizutani T, Nomura H, Nakanishi K, Fujita S (1987) Hepatotoxicity of butylated hydroxytoluene and its analogs in mice depleted of hepatic glutathione. Toxicol. Appl. Pharmacol 87: 166–176. Verhagen, H, Beckers HHG, Comuth PAWV, Maas LM (1989) Disposition of single oral doses of BHT in man and rat. Food Chem.Toxicol 27:765-772. Nakagawa Y, Hiraga K, Suga T (1984) Preventive effect of cysteine effect of cysteine on butylated hydroxytoluene-induced pulmonary toxicity in mice. Biochem. Pharmacol .33:502. Yamamoto K, Tajima K, Okino N, Mizutani T (1988) Enhanced lung toxicity of butylated hydroxytoluen e in mice by coadministration of butylated hydroxyanisole. Res. Commun. Chem. Pathol. Pharmacol 59: 219–231. Sheftel VO (1977) Hygiene Aspects of the Use of Polymeric Materials in the Water Supply, Thesis Diss., All-Union Research Institute of Hygiene and Toxicology of Pesticides, Polymers and Plastic Materials, Kiyv, 61 (in Russian). Zitting A. and Savolainen H. (1979): Fire Mater; 3:80-83 Shehata HA (1998) Environmental pollution, the problem of the virus, its causes and its solution. College of Science. Al Azhar University. Takahashi O (1991) Some properties of rat platelet aggregation and effects of butylated hydroxytoluene, warfarin and aspirin. Food Chem. Toxicol 29:173– 183. Takahashi O, Hiraga K (1978) Dose-response study of hemorrhagic death by dietary butylated hydroxyl-toluene (BHT) in male rats. Toxicol. Appl. Pharmacol 43:399–406. Torres-Arreola W1, Soto-Valdez H, Peralta E, Cardenas-López JL, Ezquerra- Brauer JM (2007) Effect of a low-density polyethylene film containing butylated
https://openalex.org/W2790153657	https://europepmc.org/articles/pmc5831759?pdf=render	English	null	Yeast Cells Exposed to Exogenous Palmitoleic Acid Either Adapt to Stress and Survive or Commit to Regulated Liponecrosis and Die	Oxidative medicine and cellular longevity	2,018	cc-by	10,914	Received 2 September 2017; Revised 27 November 2017; Accepted 20 December 2017; Published 31 January 2018 Academic Editor: Paula Ludovico Copyright © 2018 Karamat Mohammad et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. A disturbed homeostasis of cellular lipids and the resulting lipotoxicity are considered to be key contributors to many human pathologies, including obesity, metabolic syndrome, type 2 diabetes, cardiovascular diseases, and cancer. The yeast Saccharomyces cerevisiae has been successfully used for uncovering molecular mechanisms through which impaired lipid metabolism causes lipotoxicity and elicits diﬀerent forms of regulated cell death. Here, we discuss mechanisms of the “liponecrotic” mode of regulated cell death in S. cerevisiae. This mode of regulated cell death can be initiated in response to a brief treatment of yeast with exogenous palmitoleic acid. Such treatment prompts the incorporation of exogenously added palmitoleic acid into phospholipids and neutral lipids. This orchestrates a global remodeling of lipid metabolism and transfer in the endoplasmic reticulum, mitochondria, lipid droplets, and the plasma membrane. Certain features of such remodeling play essential roles either in committing yeast to liponecrosis or in executing this mode of regulated cell death. We also outline four processes through which yeast cells actively resist liponecrosis by adapting to the cellular stress imposed by palmitoleic acid and maintaining viability. These prosurvival cellular processes are conﬁned in the endoplasmic reticulum, lipid droplets, peroxisomes, autophagosomes, vacuoles, and the cytosol. Hindawi Oxidative Medicine and Cellular Longevity Volume 2018, Article ID 3074769, 11 pages https://doi.org/10.1155/2018/3074769 Hindawi Oxidative Medicine and Cellular Longevity Volume 2018, Article ID 3074769, 11 pages https://doi.org/10.1155/2018/3074769 Karamat Mohammad, Paméla Dakik, Younes Medkour, Mélissa McAuley, Darya Mitrofanova, and Vladimir I. Titorenko Concordia University, Department of Biology, Montreal, QC, Canada H4B 1R6 Correspondence should be addressed to Vladimir I. Titorenko; vladimir.titorenko@concordia.ca Received 2 September 2017; Revised 27 November 2017; Accepted 20 December 2017; Published 31 January 2 1. Introduction a certain kind of exogenous and/or endogenous stress are excessively activated, thereby generating products of these processes in concentrations that are lethal to the cell [1, 3, 7]. Some forms of cell death are classiﬁed as “programmed” cell death subroutines; they involve molecular machineries dedi- cated to commit cellular “suicide” that is aimed at providing certain beneﬁts for development and/or survival of the entire organism [1–6]. Other forms of cell death are actively driven by molecular machineries that attempt to protect cells against certain stresses (without providing beneﬁts for organismal development and/or survival); these forms are known as “regulated” cell death (RCD) subroutines [1, 3]. Cells commit RCD executed by a discrete molecular machinery because (1) the capacity of a molecular machinery dedicated to cell pro- tection against a certain kind of stress is not suﬃcient to maintain cell viability if the intensity of such extracellular and/or intracellular stress exceeds a threshold and/or (2) molecular machineries driving some cellular processes that (directly or indirectly) contribute to cell protection against S. cerevisiae is a model organism most commonly and productively used for studying diﬀerent forms of RCD elicited by perturbations in lipid metabolism [8–37]. The detailed knowledge of mechanisms underlying the molecular pathways of various modes of lipotoxic RCD in this yeast is therefore instrumental to our understanding of many human pathologies that are causally linked to dysregulated lipid metabolism, unbalanced lipid homeostasis, lipotoxicity, and lipid-induced cell death [31, 34, 37–41]. Among these human pathologies are obesity, metabolic syndrome, type 2 diabetes, insulin resistance, cardiovascular diseases, hepatic steatosis, liver cirrhosis, and cancer [34, 38–56]. The scope of this review is to analyze mechanisms underlying one of the modes of lipotoxic RCD. It has been discovered in the yeast S. cerevisiae and called “liponecrosis.” Oxidative Medicine and Cellular Longevity 2 Liponecrotic RCD can be elicited by a short-term (for 2 h) treatment of yeast cells with exogenous palmitoleic acid (POA), a 16-carbon monounsaturated fatty acid (16 : 1 n-7) [57–59]. We describe diﬀerent cellular processes that yeast cells exposed to POA use for stress adaptation and viability maintenance. We critically evaluate mechanisms (including POA-induced oxidative stress) through which yeast cells that are exposed to POA die of liponecrosis if the capacities of cellular processes for protection against POA-imposed stress become insuﬃcient to maintain cell viability. We outline the most important unanswered questions and suggest directions for future research. 2. How Do Yeast Cells Die If Treated with POA and How Do They Mount a Protective Stress Response to Survive Such Treatment? p g The bulk quantities of POA-containing phospholipids initially synthesized in the ER of yeast cells that are commit- ted to liponecrotic RCD accumulate not only in the PM but also in both membranes enclosing mitochondria (Figure 1) [34, 59]. This buildup of POA-containing phospholipids in the IMM and OMM markedly weakens mitochondrial respi- ratory capacity, uncouples mitochondria respiratory chain from ATP synthesis, and lowers the electrochemical potential across the IMM (Figure 1) [34, 59]. The resulting decline in mitochondrial functionality plays an essential role in com- mitting yeast to POA-induced liponecrotic RCD, likely because these dysfunctional mitochondria cannot produce enough ATP to support the energy-demanding, prosurvival process of incorporating exogenous POA into neutral lipids (see text below for discussion of this prosurvival process) (Figure 1) [34, 59]. A model for the mechanism of liponecrotic RCD elicited by a short-term treatment of yeast with POA and for the mecha- nism protecting yeast from such RCD is schematically depicted in Figure 1. p g Yeast cells that are brieﬂy exposed to exogenous POA use the lipid-synthesizing and lipid-transporting enzymatic machineries of the endoplasmic reticulum (ER), mitochon- dria, lipid droplets (LDs), and the plasma membrane (PM) to incorporate this fatty acid into copious amounts of two classes of lipids [34, 59]. One of these POA-containing classes are the so-called “neutral” (uncharged) lipids triacylglycerols (TAGs) and ergosteryl esters (EEs), both of which are ﬁrst produced in the ER and then deposited in LDs (Figure 1) [37, 40, 60, 61]. The other class are POA-containing phos- pholipids (Figure 1); they include (1) phosphatidic acid (PA), phosphatidylserine (PS), phosphatidylcholine (PC), and phosphatidylinositol (PI), all of which are synthesized only in the ER and then transferred to mitochondria through mitochondria-ER junctions and to the PM through PM- ER junctions [62–70]; (2) phosphatidylethanolamine (PE), which is produced from ER-derived PS in the inner and outer mitochondrial membranes (IMM and OMM, resp.) and then transferred to the ER through mitochondria-ER junctions and from the ER to the PM through PM-ER junctions [62, 63, 65, 69, 71–75]; and (3) cardiolipin (CL), a signature mitochondrial phospholipid which is generated from ER- derived PA in a series of reactions conﬁned to the IMM and OMM [71, 74, 76–78]. 1. Introduction ultimately promoting transcription of the nuclear RSB1 gene [79–87]. A protein product of this gene, Rsb1, is known to regulate the bidirectional active transport of PE across the PM bilayer; speciﬁcally, Rsb1 stimulates the Lem3-driven transport of PE from the outer monolayer of the PM to its inner monolayer and also slows down the Yor1-driven trans- port of PE in the opposite direction [86, 88–91]. These eﬀects of Rsb1 elicit a depletion of PE in the outer monolayer of the PM, thereby markedly rising the permeability of the PM to small molecules (Figure 1) [34, 59]. Such increase in the per- meability of the PM to small molecules has been shown to play an essential role in committing yeast to POA-induced liponecrotic RCD (Figure 1) [34, 59]. 2. How Do Yeast Cells Die If Treated with POA and How Do They Mount a Protective Stress Response to Survive Such Treatment? These pro-death cellular processes include the following: (1) a buildup of POA-containing phospholipids (PLs) in the PM and the ensuing increase in the permeability of the PM to small molecules; (2) the accumulation of POA-containing PLs in both mitochondrial membranes and the resulting decline in mitochondrial functionality, which is needed to support the prosurvival process of incorporating exogenous POA into neutral lipids; (3) a ROS-inﬂicted oxidative damage to mitochondria and other cellular organelles, which stimulates a nonselective macroautophagic degradation of many kinds of organelles; and (4) a ROS-imposed oxidative damage to cytosolic proteins, which impairs cellular proteostasis because it promotes an accumulation of oxidatively damaged, dysfunctional, unfolded, and aggregated proteins in the cytosol. Arrows and names displayed in red color denote pro-death processes, metabolites, and proteins that commit yeast to POA-induced RCD or execute this RCD subroutine. The up or down arrows in red color denote processes or metabolites whose intensities or concentrations are increased or decreased (resp.) in yeast cells brieﬂy exposed to exogenous POA. See text for more details. ETC, mitochondrial electron transport chain; PE, phosphatidylethanolamine; PLs, phospholipids; PM, plasma membrane; ROS, reactive oxygen species; ΔΨ, electrochemical potential across the inner mitochondrial membrane. Unfolded/aggregated proteins in the cytosol Oxidatively damaged cellular organelles Dysfunctional MIT Figure 1: A model for how yeast cells exposed to exogenous palmitoleic acid (POA) either mount a protective stress response and survive or commit to POA-induced regulated liponecrosis and die. Yeast cells brieﬂy exposed to POA can employ four diﬀerent prosurvival processes to cope with the POA-induced cellular stress and maintain viability. These prosurvival cellular processes include the following: (1) an assimilation of POA into neutral lipids (triacylglycerols (TAGs) and ergosteryl esters (EEs)), in the endoplasmic reticulum (ER) and the subsequent deposition of these neutral lipids in lipid droplets (LD); (2) POA oxidation in peroxisomes (PER); (3) a macroautophagic degradation of dysfunctional or damaged mitochondria (MIT); and (4) a proteolytic degradation of oxidatively damaged, dysfunctional, unfolded, and aggregated proteins that accumulate in the cytosol of yeast cells. Arrows and names displayed in blue color denote prosurvival processes, metabolites, and proteins that protect yeast from POA-induced liponecrotic regulated cell death (RCD). Yeast cells brieﬂy treated with POA can use four diﬀerent pro-death processes to commit to POA-induced liponecrotic RCD and to execute this RCD subroutine. 2. How Do Yeast Cells Die If Treated with POA and How Do They Mount a Protective Stress Response to Survive Such Treatment? Oxidative Medicine and Cellular Longevity 3 3 Atg1 Atg11 Atg17 POA PLs TAGs EEs POA-CoA ER PM POA-containing PLs Rim101 signaling Rsb1 Lem3 Yor1 Internalization of PE Increased permeability for small molecules TAGs EEs LD Functional MIT ATP ROS ATP ETC ΔΨ Oxidatively damaged proteins in the cytosol Unfolded/aggregated proteins in the cytosol Cellular proteostasis Oxidatively damaged cellular organelles Dysfunctional cellular organelles Non-selective macroautophagic degradation PER Oxidation of POA Atg32 Macromitophagy Proteolytic degradation Yca1 Nma111 PE PE PE PE PE PE PE Dysfunctional MIT Atg1 Atg11 Atg17 POA PLs TAGs EEs POA-CoA ER PM POA-containing PLs Rim101 signaling Rsb1 Lem3 Yor1 Internalization of PE Increased permeability for small molecules TAGs EEs LD Functional MIT ATP ROS ATP ETC ΔΨ Oxidatively damaged proteins in the cytosol Unfolded/aggregated proteins in the cytosol Cellular proteostasis Oxidatively damaged cellular organelles Dysfunctional cellular organelles Non-selective macroautophagic degradation PER Oxidation of POA Atg32 Macromitophagy Proteolytic degradation Yca1 Nma111 PE PE PE PE PE PE PE Dysfunctional MIT Figure 1: A model for how yeast cells exposed to exogenous palmitoleic acid (POA) either mount a protective stress response and survive or commit to POA-induced regulated liponecrosis and die. Yeast cells brieﬂy exposed to POA can employ four diﬀerent prosurvival processes to cope with the POA-induced cellular stress and maintain viability. These prosurvival cellular processes include the following: (1) an assimilation of POA into neutral lipids (triacylglycerols (TAGs) and ergosteryl esters (EEs)), in the endoplasmic reticulum (ER) and the subsequent deposition of these neutral lipids in lipid droplets (LD); (2) POA oxidation in peroxisomes (PER); (3) a macroautophagic degradation of dysfunctional or damaged mitochondria (MIT); and (4) a proteolytic degradation of oxidatively damaged, dysfunctional, unfolded, and aggregated proteins that accumulate in the cytosol of yeast cells. Arrows and names displayed in blue color denote prosurvival processes, metabolites, and proteins that protect yeast from POA-induced liponecrotic regulated cell death (RCD). Yeast cells brieﬂy treated with POA can use four diﬀerent pro-death processes to commit to POA-induced liponecrotic RCD and to execute this RCD subroutine. 2. How Do Yeast Cells Die If Treated with POA and How Do They Mount a Protective Stress Response to Survive Such Treatment? It needs to be emphasized that genetic interventions weakening the incorporation of exoge- nously added POA into POA-containing phospholipids within the ER have been shown to increase cell resistance to POA-induced liponecrotic RCD [34, 59]. Thus, such incor- poration is a pro-death process essential for the commitment of yeast to liponecrotic RCD in response to treatment with exogenous POA. g The buildup of POA-containing phospholipids in the IMM and OMM of yeast committed to liponecrotic RCD not only impairs mitochondrial functionality but also consid- erably increases the intracellular concentration of reactive oxygen species (ROS) that are produced in mitochondria as by-products of respiration (Figure 1) [34, 59]. This rise of ROS concentrations elicits an oxidative damage to diﬀerent types of molecules in two cellular locations, namely, to (1) protein and lipid components of mitochondria and other cellular organelles and (2) proteins in the cytosol, thereby causing their unfolding and aggregation (Figure 1) [34, 59]. Both these types of cellular oxidative damage are essential contributing factors either to the commitment of yeast to POA-induced liponecrotic RCD or to an execution of this RCD subroutine. Speciﬁcally, a massive breakdown of numerous oxidatively damaged and dysfunctional organelles through a nonselective macroautophagic degradation (which is choreographed by the phagophore assembly-speciﬁc ser- ine/threonine protein kinase Atg1, adapter protein Atg11, and scaﬀold protein Atg17 [58, 59, 92–94]) plays a crucial role in executing POA-induced liponecrotic RCD (Figure 1) [34, 58, 59]. Moreover, the buildup of oxidatively damaged, dysfunctional, unfolded, and aggregated proteins in the cyto- sol of yeast cells treated with POA impairs cellular proteosta- sis, thus committing these cells to POA-induced liponecrotic RCD (Figure 1) [34, 59]. After being synthesized in the ER, the bulk quantities of POA-containing phospholipids in yeast cells committed to liponecrotic RCD amass in the PM (Figure 1) [34, 59]. Such accumulation of POA-containing phospholipids in the PM activates the alkaline pH- and lipid asymmetry- responsive Rim101 signaling pathway, which orchestrates a series of endocytic internalization and traﬃc events If the stress imposed by an exposure to POA does not exceed a toxic threshold, yeast cells can use at least four diﬀerent processes to cope with this stress and maintain viability (Figure 1). 3. What Are the Relations among Different Processes Involved in Cell Death or Cell Adaptation in Yeast Treated with POA and How Is a Balance between Pro-death and Prosurvival Processes Regulated? One of these prosurvival cellular processes is an assimi- lation of POA into neutral lipids (TAGs and EEs), which occurs in the ER and is followed by a buildup of POA- containing neutral lipids in LDs (Figure 1) [34, 58, 59]. This process lowers the extreme cellular stress caused by the accu- mulation of POA-containing phospholipids in the PM, IMM, and OMM because it attenuates the ﬂow of POA into the pathways for the synthesis of POA-containing phospho- lipids [34, 58, 59]. The assimilation of POA into neutral lipids is essential for protecting yeast from POA-induced liponecrotic RCD, as demonstrated by the ﬁnding that genetic interventions weakening the incorporation of exoge- nously added POA into POA-containing neutral lipids increase the susceptibility of yeast to this subroutine of RCD [58]. 3. What Are the Relations among Different Processes Involved in Cell Death or Cell Adaptation in Yeast Treated with POA and How Is a Balance between Pro-death and Prosurvival Processes Regulated? As outlined in the previous section, pro-death cellular pro- cesses in yeast treated with POA are direct or indirect due to the initial incorporation of this fatty acid into bulk quanti- ties of POA-containing phospholipids. Two direct pro-death processes include the following: (1) the buildup of POA- containing phospholipids in the PM and the ensuing increase in the permeability of the PM to small molecules and (2) the accumulation of POA-containing phospholipids in both mitochondrial membranes and the resulting decline in mito- chondrial functionality, which is needed to support the pro- survival process of incorporating exogenous POA into neutral lipids (Figure 1). Two other pro-death processes only indirectly caused the buildup of POA-containing phospho- lipids in both mitochondrial membranes because such buildup elicits a rise in the intracellular concentration of ROS initially produced in mitochondria. These indirect pro-death processes are as follows: (1) the ROS-inﬂicted oxi- dative damage to mitochondria and other cellular organelles, which stimulates a nonselective macroautophagic degrada- tion of many kinds of organelles and (2) the ROS-imposed oxidative damage to cytosolic proteins, which impairs cellu- lar proteostasis because it promotes the accumulation of oxi- datively damaged, dysfunctional, unfolded, and aggregated proteins in the cytosol (Figure 1). Thus, the four pro-death processes relate because they all are initiated in response to the buildup of POA-containing phospholipids. 3. What Are the Relations among Different Processes Involved in Cell Death or Cell Adaptation in Yeast Treated with POA and How Is a Balance between Pro-death and Prosurvival Processes Regulated? We hypothe- size that (1) the direct pro-death processes may precede in time the indirect ones and (2) the relative contribution of each direct or indirect pro-death process into POA-induced liponecrotic RCD may be deﬁned by the relative rates with which POA-containing phospholipids are transferred from the ER to the PM and mitochondria, mitochondria generate ROS, mitochondria and other cellular organelles undergo ROS-inﬂicted oxidative damage, oxidatively damaged cellu- lar organelles are subjected to nonselective macroautophagic degradation, cytosolic proteins are oxidatively damaged by mitochondrially produced ROS, and oxidatively damaged cytosolic proteins unfold and aggregate. In our hypothesis, none of the pro-death processes may be considered as an individual pro-death pathway. In contrast, our hypothesis posits that all four pro-death processes are likely to be nodes of a branched subnetwork that integrates the ﬂow of POA-containing phospholipids from the ER to the PM and mitochondria, mitochondrial ROS formation, the ROS-imposed oxidative damage to organelles and cytosolic proteins, the nonselective macroautophagic breakdown of diﬀerent kinds of oxidatively damaged organelles, and the unfolding and aggregation of oxidatively damaged proteins in the cytosol. Our hypothesis further suggests that prosurvival pro- Another prosurvival cellular process is the β-oxidation of POA in peroxisomes of yeast exposed to this monounsatu- rated free fatty acid (Figure 1) [34, 57–59]. Peroxisomal oxi- dation of POA mitigates POA-induced liponecrotic RCD because it weakens the incorporation of POA into phospho- lipids, thereby relieving the excessive cellular stress instigated by the buildup of POA-containing phospholipids in the PM, IMM, and OMM [34, 58, 59]. In support of an essential role of peroxisomal oxidation of POA in the protection of yeast from POA-induced liponecrotic RCD, yeast strains that carry the single-gene-deletion mutations pex5Δ and fox1Δ attenu- ating oxidative degradation of POA in peroxisomes are more susceptible to this mode of RCD than an otherwise isogenic wild-type strain [34, 57–59]. Macromitophagy, a macroautophagic degradation of dysfunctional or damaged mitochondria, is also a prosurvival process that allows yeast to cope with the POA-induced cel- lular stress [34, 58, 59]. Macromitophagy protects yeast from POA-induced liponecrotic RCD because, by selectively degrading dysfunctional mitochondria, it helps to maintain a population of functionally active mitochondria that are needed to generate enough ATP to support the prosurvival process of assimilating POA into neutral lipids (Figure 1) [34, 58, 59]. 2. How Do Yeast Cells Die If Treated with POA and How Do They Mount a Protective Stress Response to Survive Such Treatment? These pro-death cellular processes include the following: (1) a buildup of POA-containing phospholipids (PLs) in the PM and the ensuing increase in the permeability of the PM to small molecules; (2) the accumulation of POA-containing PLs in both mitochondrial membranes and the resulting decline in mitochondrial functionality, which is needed to support the prosurvival process of incorporating exogenous POA into neutral lipids; (3) a ROS-inﬂicted oxidative damage to mitochondria and other cellular organelles, which stimulates a nonselective macroautophagic degradation of many kinds of organelles; and (4) a ROS-imposed oxidative damage to cytosolic proteins, which impairs cellular proteostasis because it promotes an accumulation of oxidatively damaged, dysfunctional, unfolded, and aggregated proteins in the cytosol. Arrows and names displayed in red color denote pro-death processes, metabolites, and proteins that commit yeast to POA-induced RCD or execute this RCD subroutine. The up or down arrows in red color denote processes or metabolites whose intensities or concentrations are increased or decreased (resp.) in yeast cells brieﬂy exposed to exogenous POA. See text for more details. ETC, mitochondrial electron transport chain; PE, phosphatidylethanolamine; PLs, phospholipids; PM, plasma membrane; ROS, reactive oxygen species; ΔΨ, electrochemical potential across the inner mitochondrial membrane. 4 Oxidative Medicine and Cellular Longevity 4. Is the Subnetwork of Liponecrotic RCD Integrated into a Signaling Network Orchestrating Different RCD Scenarios in Yeast Cells? Two indirect prosurvival processes are activated to lower the extreme cellular stress created by the buildup of POA- containing phospholipids in both mitochondrial membranes and by the resulting decline in mitochondrial functionality and rise in mitochondrially produced ROS; they are as fol- lows: (1) the selective macroautophagic degradation of oxida- tively damaged and dysfunctional mitochondria, which helps to maintain a population of functionally active mitochondria generating suﬃcient quantities of ATP and producing ROS in nontoxic concentrations and (2) the Yca1- and Nma111- driven proteolysis of oxidatively damaged and aggregated cytosolic proteins, which allows to sustain eﬃcient cellular proteostasis (Figure 1). Akin to pro-death processes, the four prosurvival processes relate because they all are stimulated in attempt to relieve the extreme cellular stress that is generated (directly or indirectly) by the initial incorporation of POA into POA-containing phospholipids. Our hypothesis posits that (1) the direct prosurvival processes may occur earlier than the indirect ones and (2) the relative contribution of each direct or indirect prosurvival process into cell protection from POA-induced liponecrosis may depend on the relative rates with which POA is assimilated into neutral lipids in the ER, POA-containing neutral lipids are transferred from the ER to LDs, POA is oxidized in peroxisomes, oxidatively damaged and dysfunctional mitochondria are subjected to selective macroautophagic degradation, and oxidatively damaged and aggregated cytosolic proteins undergo proteo- lytic degradation. Yeast cells undergoing POA-induced liponecrotic RCD exhibit characteristic morphological and biochemical traits [34, 58, 59]. Some of these traits are unique to liponecrotic RCD, whereas other traits are shared by this and certain other modes of RCD (Table 1). While yeast cells committed to POA-induced lipone- crotic RCD do not display such characteristic traits of apo- ptotic RCD as nuclear fragmentation and PS externalization within the PM bilayer, the metacaspase Yca1 and serine pro- tease Nma111 play essential roles in both liponecrotic and caspase-dependent apoptotic modes of RCD [34, 59]. How- ever, the roles Yca1 and Nma111 play in each of these two RCD modes are quite diﬀerent (Table 1). As mentioned above, the Yca1- and Nma111-dependent proteolysis of oxi- datively damaged, dysfunctional, unfolded, and aggregated proteins in the cytosol of yeast cells is a prosurvival process in POA-induced liponecrotic RCD [34, 59]. Such prosurvival role of Yca1 in sustaining eﬃcient cellular proteostasis is well known [98–105]. Oxidative Medicine and Cellular Longevity Oxidative Medicine and Cellular Longevity 5 Table 1: Some of the morphological and biochemical traits characteristic of palmitoleic acid- (POA-) induced liponecrotic regulated cell death (RCD) are unique to this mode of RCD, whereas other traits are shared by this mode and other (i.e., caspase-dependent apoptotic, autophagic, and necrotic) RCD modes. LDs, lipid droplets; PM, plasma membrane; PS, phosphatidylserine. Table 1: Some of the morphological and biochemical traits characteristic of palmitoleic acid- (POA-) induced liponecrotic regulated cell death (RCD) are unique to this mode of RCD, whereas other traits are shared by this mode and other (i.e., caspase-dependent apoptotic, autophagic, and necrotic) RCD modes. LDs, lipid droplets; PM, plasma membrane; PS, phosphatidylserine. Trait Caspase-dependent apoptotic RCD [references] Autophagic RCD [references] Necrotic RCD [references] POA-induced liponecrotic RCD [references] Nuclear fragmentation + [126] − − −[58] PS externalization within the PM + [126] − − −[58] Role of Yca1 and Nma111 + (pro-death role) [106, 107] − − + (prosurvival role) [59] Excessive cytoplasmic vacuolization − + [127] − −[58] Massive degradation of various cellular organelles − + [127] − + [58] Rupture of the PM − − + [114] −[58] Permeability of the PM to small molecules − − + [114] + [59] Excessive accumulation of LDs − − − + [58] 4. Is the Subnetwork of Liponecrotic RCD Integrated into a Signaling Network Orchestrating Different RCD Scenarios in Yeast Cells? 4. Is the Subnetwork of Liponecrotic RCD Integrated into a Signaling Network Orchestrating Different RCD Scenarios in Yeast Cells? prosurvival processes relieve the extreme cellular stress by preventing the buildup of POA-containing phospholipids in the PM and mitochondria; they include the following: (1) the assimilation of POA into neutral lipids in the ER and the subsequent buildup of POA-containing neutral lipids in LDs and (2) peroxisomal oxidation of POA (Figure 1). Two indirect prosurvival processes are activated to lower the extreme cellular stress created by the buildup of POA- containing phospholipids in both mitochondrial membranes and by the resulting decline in mitochondrial functionality and rise in mitochondrially produced ROS; they are as fol- lows: (1) the selective macroautophagic degradation of oxida- tively damaged and dysfunctional mitochondria, which helps to maintain a population of functionally active mitochondria generating suﬃcient quantities of ATP and producing ROS in nontoxic concentrations and (2) the Yca1- and Nma111- driven proteolysis of oxidatively damaged and aggregated cytosolic proteins, which allows to sustain eﬃcient cellular proteostasis (Figure 1). Akin to pro-death processes, the four prosurvival processes relate because they all are stimulated in attempt to relieve the extreme cellular stress that is generated (directly or indirectly) by the initial incorporation of POA into POA-containing phospholipids. Our hypothesis posits that (1) the direct prosurvival processes may occur earlier than the indirect ones and (2) the relative contribution of each direct or indirect prosurvival process into cell protection from POA-induced liponecrosis may depend on the relative rates with which POA is assimilated into neutral lipids in the ER, POA-containing neutral lipids are transferred from the ER to LDs, POA is oxidized in peroxisomes, oxidatively damaged and dysfunctional mitochondria are subjected to selective macroautophagic degradation, and oxidatively damaged and aggregated cytosolic proteins undergo proteo- lytic degradation. prosurvival processes relieve the extreme cellular stress by preventing the buildup of POA-containing phospholipids in the PM and mitochondria; they include the following: (1) the assimilation of POA into neutral lipids in the ER and the subsequent buildup of POA-containing neutral lipids in LDs and (2) peroxisomal oxidation of POA (Figure 1). 3. What Are the Relations among Different Processes Involved in Cell Death or Cell Adaptation in Yeast Treated with POA and How Is a Balance between Pro-death and Prosurvival Processes Regulated? The Atg32Δ-dependent mutational block of macromitophagy impairs the accumulation of LD-deposited neutral lipids and sensitizes yeast to POA-induced lipone- crotic RCD [58]; thus, macromitophagy plays an essential role in protecting yeast from this subroutine of RCD. p g y The degradation of oxidatively damaged, dysfunctional, unfolded, and aggregated proteins that accumulate in the cytosol of yeast cells treated with POA is another prosurvival process in these cells; this proteolytic degradation is catalyzed by the metacaspase Yca1 and serine protease Nma111, two protein components of the caspase-dependent apoptotic RCD pathway (Figure 1) [34, 59, 95–97]. This Yca1- and Nma111-driven proteolysis of oxidatively damaged, dys- functional, unfolded, and aggregated proteins slows down a progression of POA-induced liponecrotic RCD because it allows to sustain eﬃcient cellular proteostasis, thereby weakening proteostatic cellular stress (Figure 1) [34, 59]. In support of an essential role of such proteolysis in the protection of yeast from POA-induced liponecrotic RCD, lack of Yca1 or Nma111 increases the susceptibility of yeast to this mode of RCD [34, 59]. Our hypothesis further suggests that prosurvival pro- cesses are likely to be nodes of the same branched subnet- work integrating the four pro-death processes. Two direct 4. Is the Subnetwork of Liponecrotic RCD Integrated into a Signaling Network Orchestrating Different RCD Scenarios in Yeast Cells? Moreover, in agreement with our hypothesis on only a partial overlap between liponecrotic and other pathways and subnetworks of RCD, at least one trait charac- teristic of liponecrotic RCD is unique to this mode of RCD; this trait is the accumulation of POA-containing neutral lipids in many LDs [34, 58, 59]. p y Our hypothesis on the existence of an RCD signaling net- work orchestrating diﬀerent RCD scenarios in yeast cells is reminiscent of the hypothesis on the global programmed cell death (PCD) network that has been proposed and then con- ﬁrmed for mammalian cells [117–120]. A systems biology platform has been developed for deﬁning the topology of such network operating in mammalian cells; this platform employs cell biological and computational approaches for measuring and computing the eﬀects of single and double genetic interventions on the molecular events characteristic of diﬀerent PCD modes that are integrated into the network [119]. The use of such platform, possibly in combination with powerful tools of proteomic and metabolomic analyses recently applied for molecular analyses of RCD in yeast [104, 105], will allow to test our hypothesis on the global RCD signaling network in yeast and, perhaps, to dissect the architecture of such network in the near future. p g p LCA is a geroprotective chemical compound that delays yeast chronological aging mainly under CR conditions [57]. LCA exhibits the following eﬀects on yeast susceptibility to POA-induced liponecrotic RCD: (1) it decreases such sus- ceptibility only if added to growth medium at the time of cell inoculation, during logarithmic (L) or D phase of culturing; (2) it increases such susceptibility if added during PD phase; and (3) it has no eﬀect on such susceptibility if added during ST phase [121]. Taken together, these ﬁndings suggest that liponecrotic RCD may be an essential longevity-limiting (i.e., proaging) factor in chronologically “young” yeast, may somehow contribute to longevity extension (i.e., aging delay) in chronologically “middle-aged” yeast, and may have no inﬂuence on longevity (i.e., on the pace of aging) of chrono- logically “old” yeast. Noteworthy, all these age-related varia- tions in yeast susceptibility to POA-induced liponecrotic RCD coincide with age-related changes in yeast resistance to chronic oxidative, thermal, and osmotic stresses [121]. In the future, it would be important to explore mechanisms that underlie the observed age-related coincidence between yeast susceptibility to POA-induced liponecrotic RCD and yeast resistance to long-term stresses. 4. Is the Subnetwork of Liponecrotic RCD Integrated into a Signaling Network Orchestrating Different RCD Scenarios in Yeast Cells? In contrast, the Yca1- and Nma111-driven degradation of various cellular proteins is an executing, pro- death process in several caspase-dependent modalities of apoptotic RCD in yeast exposed to certain exogenous stimuli [95–97, 106–110]. While yeast cells undergoing POA-induced liponecrotic RCD do not display such hallmark trait of autophagic RCD as extreme cytoplasmic vacuolization instigated by a buildup of double-membrane vesicles called autophagosomes [34, 58, 59], both liponecrotic and autophagic modes of RCD (1) exhibit a nonselective massive degradation of various cellular organelles and (2) depend on the phagophore assembly- speciﬁc serine/threonine protein kinase Atg1 for executing these RCD modes (Table 1) [1, 58, 59, 111–113]. In sum, the above hypothesis posits the following: (1) the balance between diﬀerent pro-death and prosurvival processes may be regulated by their relative rates and (2) these relative rates may be deﬁned by the extracellular and/ or intracellular concentrations of POA, nutrient availability, the metabolic state of a yeast cell, and the chronological age of a yeast cell. While yeast cells undergoing POA-induced liponecrotic RCD do not exhibit such hallmark feature of necrotic RCD as a severe fracture of the PM [34, 58, 59], both Oxidative Medicine and Cellular Longevity 6 liponecrotic and necrotic modes of RCD display substan- tially increased permeability of the PM to small molecules (Table 1) [31, 58, 59, 114–116]. liponecrotic and necrotic modes of RCD display substan- tially increased permeability of the PM to small molecules (Table 1) [31, 58, 59, 114–116]. can be signiﬁcantly decreased by some aging-delaying dietary and pharmacological interventions. These interventions include caloric restriction (CR) and lithocholic bile acid (LCA), each implemented at the time of cell inoculation into growth medium [57, 60, 121]. A trait which is unique to POA-induced liponecrotic RCD is a buildup of POA-containing neutral lipids in numerous LDs, a feature that has not been reported for apo- ptotic, autophagic, or necrotic subroutine of RCD (Table 1) [1, 58, 59, 96, 111–113, 115]. g [ ] Our recent unpublished ﬁndings indicate that in yeast cultured under non-CR conditions on 1% or 2% glucose, the risk of age-related death depends not only on the POA- induced liponecrotic mode of RCD but also on ROS- induced apoptotic RCD mode. Moreover, we found that the liponecrotic and apoptotic modes of RCD have diﬀerent rel- ative contributions to age-related death of non-CR yeast at diﬀerent periods of chronological lifespan (CLS). 4. Is the Subnetwork of Liponecrotic RCD Integrated into a Signaling Network Orchestrating Different RCD Scenarios in Yeast Cells? The apo- ptotic mode of RCD predominates during diauxic (D) phase, apoptotic and liponecrotic RCD modes equally increase the risk of death during post-diauxic (PD) phase, whereas the liponecrotic mode of RCD prevails during stationary (ST) phase of culturing under non-CR conditions (our unpub- lished data). The longevity-deﬁning mode of liponecrotic RCD is elicited by the accumulation of POA and other free fatty acids in chronologically aging non-CR yeast cells that progress through PD and ST phases of culturing (our unpub- lished data). In contrast, the longevity-deﬁning mode of apoptotic RCD is caused by the rapid decline of mitochon- drial functionality and rise of mitochondrially generated ROS in chronologically aging non-CR yeast cells progressing through D and PD phases of culturing (our unpublished data). CR diet, which is implemented by culturing yeast on 0.2% or 0.5% glucose, decreases the risk of age-related death by attenuating liponecrotic and apoptotic RCD modes during D, PD, and ST phases; these eﬀects of CR are due to its abilities to (1) decrease free fatty acid (including POA) concentrations during PD and ST phases of culturing and to (2) improve mitochondrial functionality and to lessen concentrations of mitochondrially generated ROS during D and PD phases of culturing (our unpublished data). Because POA-induced liponecrotic RCD has several diﬀerent traits in common with apoptotic, autophagic, and necrotic modes of RCD, we hypothesize that the molecular subnetwork of POA-induced liponecrotic RCD is integrated into a signaling network that orchestrates diﬀerent RCD scenarios in yeast cells. Other pathways and subnetworks integrated into this signaling network may include apoptotic, autophagic, and necrotic pathways and subnetworks of RCD. In our hypothesis, the molecular subnetwork of POA-induced liponecrotic RCD only partially overlaps with apoptotic, autophagic, and necrotic RCD pathways and sub- networks of the network. Our hypothesis satisfactorily explains the observed existence of several proteins that are common to liponecrotic, apoptotic, autophagic, and necrotic modes of RCD [34, 58, 59]. Furthermore, as our hypothesis suggests, some of the morphological and biochemical traits characteristic of POA-induced liponecrotic RCD are shared by this mode of RCD and other (i.e., apoptotic, autophagic, and necrotic) RCD modes integrated into the network [34, 58, 59]. 4. Is the Subnetwork of Liponecrotic RCD Integrated into a Signaling Network Orchestrating Different RCD Scenarios in Yeast Cells? Moreover, it remains to be determined if and how the concentrations of endogenously The authors declare no conﬂict of interests. The authors declare no conﬂict of interests. 5. Does Liponecrotic RCD Contribute to Yeast Chronological Aging? POA-induced liponecrotic RCD is an age-related mode of RCD, as the susceptibility of a population of yeast cells to POA-induced liponecrosis increases with the chronological age of this population [34, 58, 59, 121]. Furthermore, the sus- ceptibility of yeast cells to POA-induced liponecrotic RCD Oxidative Medicine and Cellular Longevity 7 produced free fatty acids (including POA) inﬂuence the extent of liponecrotic RCD at diﬀerent stages of yeast chro- nological aging. explore mechanisms through which such perturbations in lipid metabolism (1) modulate individual molecular path- ways and subnetworks of lipotoxic RCD and (2) orchestrate the integration of these individual pathways and subnet- works into the global signaling network of lipotoxic RCD. To address this challenge, the systems biology platform (which is discussed in Section 4) exploited for mammalian cells [119] can be used in combination with proteomic and metabolomic analyses of molecular signatures [104, 105] characteristic of diﬀerent lipotoxic RCD modes. Of note, LCA decreases yeast susceptibility to the mito- chondria-controlled, ROS-induced mode of apoptotic RCD if added to growth medium at the time of cell inoculation and during L, D, PD, or ST phase of culturing [121]. In yeast cultured under CR conditions, exogenous LCA enters cells, is sorted to mitochondria, amasses primarily in the IMM and also resides in the OMM, alters the concentrations of certain mitochondrial membrane phospholipids, elicits a major enlargement of mitochondria, signiﬁcantly decreases mitochondrial number, prompts an intramitochondrial accumulation of cristae disconnected from the IMM, triggers substantial alterations in mitochondrial proteome, decreases the frequencies of deletion and point mutations in mitochon- drial DNA, and leads to changes in vital aspects of mitochon- drial functionality [66, 68, 122, 123]. In the future, it would be important to explore how all these aging-delaying eﬀects of LCA are linked to yeast susceptibility to the mitochondria- controlled, ROS-induced mode of apoptotic RCD at diﬀerent stages of chronological aging. Acknowledgments Several exogenously added lipids [8–16], as well as diﬀerent genetic [11–13, 15–25] and pharmacological [24, 26–30] interventions that impair certain aspects of lipid metabolism, have been shown to elicit apoptotic and/or necrotic modes of lipotoxic RCD in yeast. These modes have been extensively reviewed elsewhere [31, 32, 37, 41, 125]. In brief, yeast cells committed to POA-induced liponecrotic RCD exhibit a unique combination of morphological and biochemical traits that is not characteristic of any of these other modes of lipotoxic RCD. Moreover, some of these other modes of lipotoxic RCD diﬀer from each other with respect to (1) structural and/or functional features of yeast committed to a particular mode of RCD; (2) classes of lipids whose concen- trations are altered (or are expected to be altered) in yeast committed to a particular mode of RCD; and (3) proteins that are involved in committing to and/or executing a partic- ular mode of RCD [8–32, 37, 41, 125]. The authors are grateful to other members of the Titorenko laboratory for discussions. 7. Conclusions To cope with the lipotoxic stress imposed by an exposure to POA, S. cerevisiae cells use several diﬀerent mechanisms to mount a protective stress response and maintain viability. This complex stress response consists in remodeling of at least four cellular processes. If the POA-induced lipotoxic stress exceeds a threshold, yeast cells commit suicide that is assisted by a complex molecular machinery. This molecular machinery alters the spatiotemporal dynamics of several cel- lular processes to execute a liponecrotic subroutine of RCD. The liponecrotic mode of POA-induced RCD plays an essen- tial role in deﬁning longevity of chronologically aging yeast, likely in coordination with an apoptotic mode of RCD. The molecular subnetwork of POA-induced liponecrotic RCD may be integrated into a global signaling network of partially overlapping molecular pathways and subnetworks, each exe- cuting a diﬀerent mode of lipotoxic or nonlipotoxic RCD. In sum, it is conceivable that liponecrotic and apoptotic modes of RCD may have diﬀerent eﬀects on yeast CLS at diﬀerent periods of life. This is similar to the “P” (“big P”) and “p” (“small p”) modes of death in the nematode Caenor- habditis elegans, which deﬁne lifespan earlier or later in life (resp.) [124]. The P mode of death is manifested as a substan- tial enlargement of the posterior pharyngeal bulb caused by intensiﬁed pharyngeal pumping, whereas the p mode of death is due to the complete atrophy of pharynx [124]. Oxidative Medicine and Cellular Longevity 8 the eﬀector molecules,” Science, vol. 352, no. 6281, article aaf2154, 2016. [21] E. Bener Aksam, H. Jungwirth, S. D. Kohlwein et al., “Absence of the peroxiredoxin Pmp20 causes peroxisomal protein leakage and necrotic cell death,” Free Radical Biology & Medicine, vol. 45, no. 8, pp. 1115–1124, 2008. [6] J. Shi, W. Gao, and F. Shao, “Pyroptosis: gasdermin-mediated programmed necrotic cell death,” Trends in Biochemical Sciences, vol. 42, no. 4, pp. 245–254, 2017. [22] H. Jungwirth, J. Ring, T. Mayer et al., “Loss of peroxisome function triggers necrosis,” FEBS Letters, vol. 582, no. 19, pp. 2882–2886, 2008. [7] D. R. Green and B. Victor, “The pantheon of the fallen: why are there so many forms of cell death?,” Trends in Cell Biology, vol. 22, no. 11, pp. 555-556, 2012. [23] A. D. Barbosa, H. Osório, K. J. Sims et al., “Role for Sit4p- dependent mitochondrial dysfunction in mediating the shortened chronological lifespan and oxidative stress sensi- tivity of Isc1p-deﬁcient cells,” Molecular Microbiology, vol. 81, no. 2, pp. 515–527, 2011. [8] M. Stratford and P. A. Anslow, “Comparison of the inhibitory action on Saccharomyces cerevisiae of weak-acid preserva- tives, uncouplers, and medium-chain fatty acids,” FEMS Microbiology Letters, vol. 142, no. 1, pp. 53–58, 1996. [24] X. Huang, J. Liu, and R. C. Dickson, “Down-regulating sphin- golipid synthesis increases yeast lifespan,” PLoS Genetics, vol. 8, no. 2, article e1002493, 2012. [9] K. Mitsui, D. Nakagawa, M. Nakamura, T. Okamoto, and K. Tsurugi, “Valproic acid induces apoptosis dependent of Yca1p at concentrations that mildly aﬀect the proliferation of yeast,” FEBS Letters, vol. 579, no. 3, pp. 723–727, 2005. [25] J. Tulha, F. Faria-Oliveira, C. Lucas, and C. Ferreira, “Pro- grammed cell death in Saccharomyces cerevisiae is hampered by the deletion of GUP1 gene,” BMC Microbiology, vol. 12, no. 1, p. 80, 2012. [10] Q. Sun, L. Bi, X. Su, K. Tsurugi, and K. Mitsui, “Valproate induces apoptosis by inducing accumulation of neutral lipids which was prevented by disruption of the SIR2 gene in Saccharomyces cerevisiae,” FEBS Letters, vol. 581, no. 21, pp. 3991–3995, 2007. [26] V. Zaremberg, C. Gajate, L. M. Cacharro, F. Mollinedo, and C. R. McMaster, “Cytotoxicity of an anti-cancer lysopho- spholipid through selective modiﬁcation of lipid raft compo- sition,” The Journal of Biological Chemistry, vol. 280, no. 45, pp. 38047–38058, 2005. [11] J. Garbarino, M. Padamsee, L. Oxidative Medicine and Cellular Longevity Wilcox et al., “Sterol and diacylglycerol acyltransferase deﬁciency triggers fatty acid- mediated cell death,” The Journal of Biological Chemistry, vol. 284, no. 45, pp. 30994–31005, 2009. [27] H. Zhang, C. Gajate, L. P. Yu, Y. X. Fang, and F. Mollinedo, “Mitochondrial-derived ROS in edelfosine-induced apoptosis in yeasts and tumor cells,” Acta Pharmacologica Sinica, vol. 28, no. 6, pp. 888–894, 2007. [12] J. Petschnigg, H. Wolinski, D. Kolb et al., “Good fat, essential cellular requirements for triacylglycerol synthesis to maintain membrane homeostasis in yeast,” The Journal of Biological Chemistry, vol. 284, no. 45, pp. 30981–30993, 2009. [28] V. Cerantola, I. Guillas, C. Roubaty et al., “Aureobasidin A arrests growth of yeast cells through both ceramide intoxication and deprivation of essential inositolphosphor- ylceramides,” Molecular Microbiology, vol. 71, no. 6, pp. 1523–1537, 2009. [13] P. Rockenfeller, J. Ring, V. Muschett et al., “Fatty acids trigger mitochondrion-dependent necrosis,” Cell Cycle, vol. 9, no. 14, pp. 2836–2842, 2010. [14] D. Carmona-Gutierrez, A. Reisenbichler, P. Heimbucher et al., “Ceramide triggers metacaspase-independent mito- chondrial cell death in yeast,” Cell Cycle, vol. 10, no. 22, pp. 3973–3978, 2011. [29] K. Kajiwara, T. Muneoka, Y. Watanabe, T. Karashima, H. Kitagaki, and K. Funato, “Perturbation of sphingolipid metabolism induces endoplasmic reticulum stress-mediated mitochondrial apoptosis in budding yeast,” Molecular Micro- biology, vol. 86, no. 5, pp. 1246–1261, 2012. [15] S. Fakas, Y. Qiu, J. L. Dixon et al., “Phosphatidate phos- phatase activity plays key role in protection against fatty acid-induced toxicity in yeast,” The Journal of Biological Chemistry, vol. 286, no. 33, pp. 29074–29085, 2011. [30] O. Czyz, T. Bitew, A. Cuesta-Marbán, C. R. McMaster, F. Mollinedo, and V. Zaremberg, “Alteration of plasma mem- brane organization by an anticancer lysophosphatidylcholine analogue induces intracellular acidiﬁcation and internaliza- tion of plasma membrane transporters in yeast,” The Journal of Biological Chemistry, vol. 288, no. 12, pp. 8419–8432, 2013. [16] L. Galluzzi, I. Vitale, L. Senovilla et al., “Independent tran- scriptional reprogramming and apoptosis induction by cis- platin,” Cell Cycle, vol. 11, no. 18, pp. 3472–3480, 2012. [17] M. M. Nagiec, E. E. Nagiec, J. A. Baltisberger, G. B. Wells, R. L. Lester, and R. C. Dickson, “Sphingolipid synthesis as a target for antifungal drugs. Complementation of the inositol phosphorylceramide synthase defect in a mutant strain of Saccharomyces cerevisiae by the AUR1 gene,” The Journal of Biological Chemistry, vol. 272, no. 15, pp. 9809–9817, 1997. [31] T. Eisenberg and S. References [1] L. Galluzzi, I. Vitale, J. M. Abrams et al., “Molecular deﬁni- tions of cell death subroutines: recommendations of the nomenclature committee on cell death 2012,” Cell Death and Diﬀerentiation, vol. 19, no. 1, pp. 107–120, 2012. [2] Y. Fuchs and H. Steller, “Live to die another way: modes of programmed cell death and the signals emanating from dying cells,” Nature Reviews Molecular Cell Biology, vol. 16, no. 6, pp. 329–344, 2015. [3] L. Galluzzi, J. M. Bravo-San Pedro, O. Kepp, and G. Kroemer, “Regulated cell death and adaptive stress responses,” Cellular and Molecular Life Sciences, vol. 73, no. 11-12, pp. 2405– 2410, 2016. Altogether, these ﬁndings further support our hypothesis (which is outlined in Section 4) on the possible existence of a global signaling network that integrates partially overlapping molecular pathways and subnetworks of lipotoxic RCD, each pathway and subnetwork being diﬀerently responsive to cer- tain perturbations in diverse aspects of lipid metabolism within a yeast cell. The key challenge for the future is to [4] B. Ke, M. Tian, J. Li, B. Liu, and G. He, “Targeting pro- grammed cell death using small-molecule compounds to improve potential cancer therapy,” Medicinal Research Reviews, vol. 36, no. 6, pp. 983–1035, 2016. [5] D. Wallach, T. B. Kang, C. P. Dillon, and D. R. Green, “Pro- grammed necrosis in inﬂammation: toward identiﬁcation of Oxidative Medicine and Cellular Longevity Oxidative Medicine and Cellular Longevity Büttner, “Lipids and cell death in yeast,” FEMS Yeast Research, vol. 14, no. 1, pp. 179–197, 2014. [32] A. Rego, D. Trindade, S. R. Chaves et al., “The yeast model system as a tool towards the understanding of apoptosis reg- ulation by sphingolipids,” FEMS Yeast Research, vol. 14, no. 1, pp. 160–178, 2014. [18] H. Kitagaki, L. A. Cowart, N. Matmati et al., “Isc1 regulates sphingolipid metabolism in yeast mitochondria,” Biochimica et Biophysica Acta (BBA) - Biomembranes, vol. 1768, no. 11, pp. 2849–2861, 2007. [33] K. V. Ruggles, J. Garbarino, Y. Liu et al., “A functional, genome-wide evaluation of liposensitive yeast identiﬁes the “ARE2 required for viability” (ARV1) gene product as a major component of eukaryotic fatty acid resistance,” The Journal of Biological Chemistry, vol. 289, no. 7, pp. 4417–4431, 2014. [19] A. M. Aerts, P. Zabrocki, I. E. J. A. François et al., “Ydc1p ceramidase triggers organelle fragmentation, apoptosis and accelerated ageing in yeast,” Cellular and Molecular Life Sci- ences, vol. 65, no. 12, pp. 1933–1942, 2008. [34] A. Arlia-Ciommo, V. Svistkova, S. Mohtashami, and V. I. Titorenko, “A novel approach to the discovery of anti- tumor pharmaceuticals: searching for activators of liponecro- sis,” Oncotarget, vol. 7, no. 5, pp. 5204–5225, 2016. [20] T. Almeida, M. Marques, D. Mojzita et al., “Isc1p plays a key role in hydrogen peroxide resistance and chronological life- span through modulation of iron levels and apoptosis,” Molecular Biology of the Cell, vol. 19, no. 3, pp. 865–876, 2008. [35] M. Valachovic, M. Garaiova, R. Holic, and I. Hapala, “Squa- lene is lipotoxic to yeast cells defective in lipid droplet Oxidative Medicine and Cellular Longevity 9 biogenesis,” Biochemical and Biophysical Research Communi- cations, vol. 469, no. 4, pp. 1123–1128, 2016. [55] J. Park, T. S. Morley, M. Kim, D. J. Clegg, and P. E. Scherer, “Obesity and cancer - mechanisms underlying tumour pro- gression and recurrence,” Nature Reviews Endocrinology, vol. 10, no. 8, pp. 455–465, 2014. [36] L. K. Liu, V. Choudhary, A. Toulmay, and W. A. Prinz, “An inducible ER-Golgi tether facilitates ceramide trans- port to alleviate lipotoxicity,” The Journal of Cell Biology, vol. 216, no. 1, pp. 131–147, 2017. [56] A. Piano and V. I. Titorenko, “The intricate interplay between mechanisms underlying aging and cancer,” Aging and Disease, vol. 6, no. 1, pp. 56–75, 2015. [37] D. Mitrofanova, P. Dakik, M. McAuley, Y. Medkour, K. Mohammad, and V. I. Oxidative Medicine and Cellular Longevity 281–287, 2003. [63] J. R. Friedman and G. K. Voeltz, “The ER in 3D: a multifunc- tional dynamic membrane network,” Trends in Cell Biology, vol. 21, no. 12, pp. 709–717, 2011. [45] R. H. Unger, “Longevity, lipotoxicity and leptin: the adi- pocyte defense against feasting and famine,” Biochimie, vol. 87, no. 1, pp. 57–64, 2005. [64] S. A. Henry, S. D. Kohlwein, and G. M. Carman, “Metabolism and regulation of glycerolipids in the yeast Saccharomyces cerevisiae,” Genetics, vol. 190, no. 2, pp. 317–349, 2012. [46] R. T. Brookheart, C. I. Michel, and J. E. Schaﬀer, “As a matter of fat,” Cell Metabolism, vol. 10, no. 1, pp. 9–12, 2009. [65] A. A. Rowland and G. K. Voeltz, “Endoplasmic reticulum- mitochondria contacts: function of the junction,” Nature Reviews. Molecular Cell Biology, vol. 13, no. 10, pp. 607– 625, 2012. [47] J. Garbarino and S. L. Sturley, “Saturated with fat: new per- spectives on lipotoxicity,” Current Opinion in Clinical Nutri- tion and Metabolic Care, vol. 12, no. 2, pp. 110–116, 2009. [48] C. M. Kusminski, S. Shetty, L. Orci, R. H. Unger, and P. E. Scherer, “Diabetes and apoptosis: lipotoxicity,” Apoptosis, vol. 14, no. 12, pp. 1484–1495, 2009. [66] A. Beach, V. R. Richard, A. Leonov et al., “Mitochondrial membrane lipidome deﬁnes yeast longevity,” Aging, vol. 5, no. 7, pp. 551–574, 2013. [49] S. López, B. Bermúdez, R. Abia, and F. J. Muriana, “The inﬂu- ence of major dietary fatty acids on insulin secretion and action,” Current Opinion in Lipidology, vol. 21, no. 1, pp. 15–20, 2010. [67] A. Leonov and V. I. Titorenko, “A network of interorganellar communications underlies cellular aging,” IUBMB Life, vol. 65, no. 8, pp. 665–674, 2013. [50] R. H. Unger and P. E. Scherer, “Gluttony, sloth and the met- abolic syndrome: a roadmap to lipotoxicity,” Trends in Endo- crinology and Metabolism, vol. 21, no. 6, pp. 345–352, 2010. [68] V. R. Richard, A. Leonov, A. Beach et al., “Macromitophagy is a longevity assurance process that in chronologically aging yeast limited in calorie supply sustains functional mitochon- dria and maintains cellular lipid homeostasis,” Aging, vol. 5, no. 4, pp. 234–269, 2013. [51] R. Zechner, R. Zimmermann, T. O. Eichmann et al., “FAT SIGNALS - lipases and lipolysis in lipid metabolism and sig- naling,” Cell Metabolism, vol. 15, no. 3, pp. 279–291, 2012. [69] S. Tavassoli, J. T. Chao, B. P. Oxidative Medicine and Cellular Longevity Titorenko, “Lipid metabolism and transport deﬁne longevity of the yeast Saccharomyces cerevi- siae,” Frontiers in Bioscience, vol. 23, no. 3, pp. 1166–1194, 2018. [57] A. A. Goldberg, V. R. Richard, P. Kyryakov et al., “Chemical genetic screen identiﬁes lithocholic acid as an anti-aging compound that extends yeast chronological life span in a TOR-independent manner, by modulating housekeeping longevity assurance processes,” Aging, vol. 2, no. 7, pp. 393– 414, 2010. [38] S. D. Kohlwein and J. Petschnigg, “Lipid-induced cell dys- function and cell death: lessons from yeast,” Current Hyper- tension Reports, vol. 9, no. 6, pp. 455–461, 2007. [58] S. Sheibani, V. R. Richard, A. Beach et al., “Macromitophagy, neutral lipids synthesis, and peroxisomal fatty acid oxidation protect yeast from “liponecrosis”, a previously unknown form of programmed cell death,” Cell Cycle, vol. 13, no. 1, pp. 138–147, 2014. [39] P. Rockenfeller and F. Madeo, “Ageing and eating,” Biochi- mica et Biophysica Acta, vol. 1803, no. 4, pp. 499–506, 2010. [40] S. D. Kohlwein, M. Veenhuis, and I. J. van der Klei, “Lipid droplets and peroxisomes: key players in cellular lipid homeostasis or a matter of fat - store ‘em up or burn ‘em down,” Genetics, vol. 193, no. 1, pp. 1–50, 2012. [59] V. R. Richard, A. Beach, A. Piano et al., “Mechanism of lipo- necrosis, a distinct mode of programmed cell death,” Cell Cycle, vol. 13, no. 23, pp. 3707–3726, 2014. [41] K. Natter and S. D. Kohlwein, “Yeast and cancer cells - com- mon principles in lipid metabolism,” Biochimica et Biophy- sica Acta (BBA) - Molecular and Cell Biology of Lipids, vol. 1831, no. 2, pp. 314–326, 2013. [60] A. A. Goldberg, S. D. Bourque, P. Kyryakov et al., “A novel function of lipid droplets in regulating longevity,” Biochemi- cal Society Transactions, vol. 37, no. 5, pp. 1050–1055, 2009. [42] P. G. Kopelman, “Obesity as a medical problem,” Nature, vol. 404, no. 6778, pp. 635–643, 2000. [61] L. Klug and G. Daum, “Yeast lipid metabolism at a glance,” FEMS Yeast Research, vol. 14, no. 3, pp. 369–388, 2014. [43] R. H. Unger, “Lipotoxic diseases,” Annual Review of Medi- cine, vol. 53, no. 1, pp. 319–336, 2002. [62] S. Carrasco and T. Meyer, “STIM proteins and the endoplas- mic reticulum-plasma membrane junctions,” Annual Review of Biochemistry, vol. 80, no. 1, pp. 973–1000, 2011. [44] J. E. Schaﬀer, “Lipotoxicity: when tissues overeat,” Current Opinion in Lipidology, vol. 14, no. 3, pp. Oxidative Medicine and Cellular Longevity [78] T. Tatsuta and T. Langer, “Intramitochondrial phospholipid traﬃcking,” Biochimica et Biophysica Acta, vol. 1862, no. 1, pp. 81–89, 2017. [93] Y. Feng, D. He, Z. Yao, and D. J. Klionsky, “The machinery of macroautophagy,” Cell Research, vol. 24, no. 1, pp. 24–41, 2014. [94] D. J. Klionsky, K. Abdelmohsen, A. Abe et al., “Guidelines for the use and interpretation of assays for monitoring autoph- agy (3rd edition),” Autophagy, vol. 12, no. 1, pp. 1–222, 2016. [79] M. Babst, D. J. Katzmann, E. J. Estepa-Sabal, T. Meerloo, and S. D. Emr, “Escrt-III: an endosome-associated heterooligo- meric protein complex required for MVB sorting,” Develop- mental Cell, vol. 3, no. 2, pp. 271–282, 2002. [95] A. Beach, A. Leonov, A. Arlia-Ciommo, V. Svistkova, V. Lutchman, and V. I. Titorenko, “Mechanisms by which diﬀerent functional states of mitochondria deﬁne yeast lon- gevity,” International Journal of Molecular Sciences, vol. 16, no. 3, pp. 5528–5554, 2015. [80] T. M. Lamb and A. P. Mitchell, “The transcription factor Rim101p governs ion tolerance and cell diﬀerentiation by direct repression of the regulatory genes NRG1 and SMP1 in Saccharomyces cerevisiae,” Molecular and Cellular Biology, vol. 23, no. 2, pp. 677–686, 2003. [96] D. Carmona-Gutierrez, T. Eisenberg, S. Büttner, C. Meisinger, G. Kroemer, and F. Madeo, “Apoptosis in yeast: triggers, pathways, subroutines,” Cell Death and Diﬀerentia- tion, vol. 17, no. 5, pp. 763–773, 2010. [81] W. Xu, F. J. Smith Jr, R. Subaran, and A. P. Mitchell, “Multivesicular body-ESCRT components function in pH response regulation in Saccharomyces cerevisiae and Candida albicans,” Molecular Biology of the Cell, vol. 15, no. 12, pp. 5528–5537, 2004. [97] F. Madeo, D. Carmona-Gutierrez, J. Ring, S. Büttner, T. Eisenberg, and G. Kroemer, “Caspase-dependent and caspase-independent cell death pathways in yeast,” Biochem- ical and Biophysical Research Communications, vol. 382, no. 2, pp. 227–231, 2009. [82] K. J. Barwell, J. H. Boysen, W. Xu, and A. P. Mitchell, “Rela- tionship of DFG16 to the Rim101p pH response pathway in Saccharomyces cerevisiae and Candida albicans,” Eukaryotic Cell, vol. 4, no. 5, pp. 890–899, 2005. [98] M. A. S. Khan, P. B. Chock, and E. R. Stadtman, “Knockout of caspase-like gene, YCA1, abrogates apoptosis and elevates oxidized proteins in Saccharomyces cerevisiae,” Proceedings of the National Academy of Sciences of the United States of America, vol. 102, no. 48, pp. 17326–17331, 2005. [83] M. Hayashi, T. Fukuzawa, H. Sorimachi, and T. Oxidative Medicine and Cellular Longevity 10 [87] K. Obara, H. Yamamoto, and A. Kihara, “Membrane protein Rim21 plays a central role in sensing ambient pH in Saccha- romyces cerevisiae,” The Journal of Biological Chemistry, vol. 287, no. 46, pp. 38473–38481, 2012. [72] M. J. Aaltonen, J. R. Friedman, C. Osman et al., “MICOS and phospholipid transfer by Ups2-Mdm35 organize membrane lipid synthesis in mitochondria,” The Journal of Cell Biology, vol. 213, no. 5, pp. 525–534, 2016. [88] U. Kato, K. Emoto, C. Fredriksson et al., “A novel membrane protein, Ros3p, is required for phospholipid translocation across the plasma membrane in Saccharomyces cerevisiae,” The Journal of Biological Chemistry, vol. 277, no. 40, pp. 37855–37862, 2002. [73] N. Miyata, Y. Watanabe, Y. Tamura, T. Endo, and O. Kuge, “Phosphatidylserine transport by Ups2-Mdm35 in respiration-active mitochondria,” The Journal of Cell Biology, vol. 214, no. 1, pp. 77–88, 2016. [74] K. S. Dimmer and D. Rapaport, “Mitochondrial contact sites as platforms for phospholipid exchange,” Biochimica et Bio- physica Acta (BBA) - Molecular and Cell Biology of Lipids, vol. 1862, no. 1, pp. 69–80, 2017. [89] A. Kihara and Y. Igarashi, “Identiﬁcation and characteriza- tion of a Saccharomyces cerevisiae gene, RSB1, involved in sphingoid long-chain base release,” The Journal of Biological Chemistry, vol. 277, no. 33, pp. 30048–30054, 2002. [75] C. U. Mårtensson, K. N. Doan, and T. Becker, “Eﬀects of lipids on mitochondrial functions,” Biochimica et Biophysica Acta (BBA) - Molecular and Cell Biology of Lipids, vol. 1862, no. 1, pp. 102–113, 2017. [90] A. Kihara and Y. Igarashi, “Cross talk between sphingolipids and glycerophospholipids in the establishment of plasma membrane asymmetry,” Molecular Biology of the Cell, vol. 15, no. 11, pp. 4949–4959, 2004. [76] M. Connerth, T. Tatsuta, M. Haag, T. Klecker, B. Westermann, and T. Langer, “Intramitochondrial trans- port of phosphatidic acid in yeast by a lipid transfer protein,” Science, vol. 338, no. 6108, pp. 815–818, 2012. [91] M. Ikeda, A. Kihara, A. Denpoh, and Y. Igarashi, “The Rim101 pathway is involved in Rsb1 expression induced by altered lipid asymmetry,” Molecular Biology of the Cell, vol. 19, no. 5, pp. 1922–1931, 2008. [77] M. G. Baile, Y. W. Lu, and S. M. Claypool, “The topology and regulation of cardiolipin biosynthesis and remodeling in yeast,” Chemistry and Physics of Lipids, vol. 179, pp. 25–31, 2014. [92] F. Reggiori and D. J. Klionsky, “Autophagic processes in yeast: mechanism, machinery and regulation,” Genetics, vol. 194, no. 2, pp. 341–361, 2013. Oxidative Medicine and Cellular Longevity Young et al., “Plasma membrane-endoplasmic reticulum contact sites regulate phosphatidylcholine synthesis,” EMBO Reports, vol. 14, no. 5, pp. 434–440, 2013. [52] E. Currie, A. Schulze, R. Zechner, T. C. Walther, and R. V. Farese Jr, “Cellular fatty acid metabolism and cancer,” Cell Metabolism, vol. 18, no. 2, pp. 153–161, 2013. [70] P. Dakik and V. I. Titorenko, “Communications between mitochondria, the nucleus, vacuoles, peroxisomes, the endo- plasmic reticulum, the plasma membrane, lipid droplets and the cytosol during yeast chronological aging,” Frontiers in Genetics, vol. 7, 2016. [53] L. Galluzzi, O. Kepp, M. G. Vander Heiden, and G. Kroemer, “Metabolic targets for cancer therapy,” Nature Reviews Drug Discovery, vol. 12, no. 11, pp. 829–846, 2013. [54] G. Zadra, C. Photopoulos, and M. Loda, “The fat side of prostate cancer,” Biochimica et Biophysica Acta (BBA) - Molecular and Cell Biology of Lipids, vol. 1831, no. 10, pp. 1518–1532, 2013. [71] S. E. Horvath and G. Daum, “Lipids of mitochondria,” Prog- ress in Lipid Research, vol. 52, no. 4, pp. 590–614, 2013. Oxidative Medicine and Cellular Longevity Oxidative Medicine and Cellular Longevity Zhao, A. F. Gilliat, M. Ziehm et al., “Two forms of death in ageing Caenorhabditis elegans,” Nat Commun, vol. 8, article 15458, 2017. [108] D. Walter, S. Wissing, F. Madeo, and B. Fahrenkrog, “The inhibitor-of-apoptosis protein Bir1p protects against apopto- sis in S. cerevisiae and is a substrate for the yeast homologue of Omi/HtrA2,” Journal of Cell Science, vol. 119, no. 9, pp. 1843–1851, 2006. [125] S. D. Kohlwein, “Obese and anorexic yeasts: experimental models to understand the metabolic syndrome and lipotoxi- city,” Biochimica et Biophysica Acta (BBA) - Molecular and Cell Biology of Lipids, vol. 1801, no. 3, pp. 222–229, 2010. [126] F. Madeo, E. Fröhlich, and K. U. Fröhlich, “A yeast mutant showing diagnostic markers of early and late apoptosis,” The Journal of Cell Biology, vol. 139, no. 3, pp. 729–734, 1997. [109] D. Wilkinson and M. Ramsdale, “Proteases and caspase-like activity in the yeast Saccharomyces cerevisiae,” Biochemical Society Transactions, vol. 39, no. 5, pp. 1502–1508, 2011. [110] C. Falcone and C. Mazzoni, “External and internal triggers of cell death in yeast,” Cellular and Molecular Life Sciences, vol. 73, no. 11-12, pp. 2237–2250, 2016. [127] A. Abudugupur, K. Mitsui, S. Yokota, and K. Tsurugi, “An ARL1 mutation aﬀected autophagic cell death in yeast, causing a defect in central vacuole formation,” Cell Death and Diﬀerentiation, vol. 9, no. 2, pp. 158–168, 2002. [111] N. Kourtis and N. Tavernarakis, “Autophagy and cell death in model organisms,” Cell Death and Diﬀerentiation, vol. 16, no. 1, pp. 21–30, 2008. [112] H. M. Shen and P. Codogno, “Autophagic cell death: Loch Ness monster or endangered species?,” Autophagy, vol. 7, no. 5, pp. 457–465, 2011. [113] D. Denton, S. Nicolson, and S. Kumar, “Cell death by autoph- agy: facts and apparent artefacts,” Cell Death and Diﬀerentia- tion, vol. 19, no. 1, pp. 87–95, 2012. [114] T. Eisenberg, H. Knauer, A. Schauer et al., “Induction of autophagy by spermidine promotes longevity,” Nature Cell Biology, vol. 11, no. 11, pp. 1305–1314, 2009. [115] T. Eisenberg, D. Carmona-Gutierrez, S. Büttner, N. Tavernarakis, and F. Madeo, “Necrosis in yeast,” Apopto- sis, vol. 15, no. 3, pp. 257–268, 2010. [116] P. Vandenabeele, L. Galluzzi, T. Vanden Berghe, and G. Kroemer, “Molecular mechanisms of necroptosis: an ordered cellular explosion,” Nature Reviews Molecular Cell Biology, vol. 11, no. 10, pp. 700–714, 2010. [117] S. Bialik, E. Zalckvar, Y. Ber, A. D. Rubinstein, and A. Oxidative Medicine and Cellular Longevity Oxidative Medicine and Cellular Longevity 11 [102] S. M. Hill, X. Hao, B. Liu, and T. Nyström, “Life-span exten- sion by a metacaspase in the yeast Saccharomyces cerevisiae,” Science, vol. 344, no. 6190, pp. 1389–1392, 2014. [119] E. Zalckvar, N. Yosef, S. Reef et al., “A systems level strategy for analyzing the cell death network: implication in exploring the apoptosis/autophagy connection,” Cell Death and Diﬀer- entiation, vol. 17, no. 8, pp. 1244–1253, 2010. [103] S. M. Hill and T. Nyström, “The dual role of a yeast metacas- pase: what doesn’t kill you makes you stronger,” BioEssays, vol. 37, no. 5, pp. 525–531, 2015. [120] A. D. Rubinstein and A. Kimchi, “Life in the balance - a mechanistic view of the crosstalk between autophagy and apoptosis,” Journal of Cell Science, vol. 125, no. 22, pp. 5259–5268, 2012. [104] V. Longo, M. Ždralević, N. Guaragnella, S. Giannattasio, L. Zolla, and A. M. Timperio, “Proteome and metabolome proﬁling of wild-type and YCA1-knock-out yeast cells during acetic acid-induced programmed cell death,” Journal of Pro- teomics, vol. 128, pp. 173–188, 2015. [121] M. T. Burstein, P. Kyryakov, A. Beach et al., “Lithocholic acid extends longevity of chronologically aging yeast only if added at certain critical periods of their lifespan,” Cell Cycle, vol. 11, no. 18, pp. 3443–3462, 2012. [105] M. Ždralević, V. Longo, N. Guaragnella, S. Giannattasio, A. M. Timperio, and L. Zolla, “Diﬀerential proteome- metabolome proﬁling of YCA1-knock-out and wild type cells reveals novel metabolic pathways and cellular processes dependent on the yeast metacaspase,” Molecular BioSystems, vol. 11, no. 6, pp. 1573–1583, 2015. [122] A. Beach, V. R. Richard, S. Bourque et al., “Lithocholic bile acid accumulated in yeast mitochondria orchestrates a development of an anti-aging cellular pattern by causing age-related changes in cellular proteome,” Cell Cycle, vol. 14, no. 11, pp. 1643–1656, 2015. [106] F. Madeo, E. Herker, C. Maldener et al., “A caspase-related protease regulates apoptosis in yeast,” Molecular Cell, vol. 9, no. 4, pp. 911–917, 2002. [123] A. Leonov, A. Arlia-Ciommo, S. D. Bourque et al., “Speciﬁc changes in mitochondrial lipidome alter mitochondrial pro- teome and increase the geroprotective eﬃciency of lithocholic acid in chronologically aging yeast,” Oncotarget, vol. 8, no. 19, pp. 30672–30691, 2017. [107] B. Fahrenkrog, U. Sauder, and U. Aebi, “The S. cerevisiae HtrA-like protein Nma111p is a nuclear serine protease that mediates yeast apoptosis,” Journal of Cell Science, vol. 117, no. 1, pp. 115–126, 2004. [124] Y. Oxidative Medicine and Cellular Longevity Maeda, “Constitutive activation of the pH-responsive Rim101 path- way in yeast mutants defective in late steps of the MVB/ ESCRT pathway,” Molecular and Cellular Biology, vol. 25, no. 21, pp. 9478–9490, 2005. [99] R. E. C. Lee, S. Brunette, L. G. Puente, and L. A. Megeney, “Metacaspase Yca1 is required for clearance of insoluble protein aggregates,” Proceedings of the National Academy of Sciences of the United States of America, vol. 107, no. 30, pp. 13348–13353, 2010. [84] P. Weiss, S. Huppert, and R. Kölling, “Analysis of the dual function of the ESCRT-III protein Snf7 in endocytic traﬃck- ing and in gene expression,” The Biochemical Journal, vol. 424, no. 1, pp. 89–97, 2009. [100] S. Lefevre, D. Sliwa, F. Auchère et al., “The yeast metacaspase is implicated in oxidative stress response in frataxin-deﬁcient cells,” FEBS Letters, vol. 586, no. 2, pp. 143–148, 2012. [85] A. K. Shukla, K. Xiao, and R. J. Lefkowitz, “Emerging paradigms of β-arrestin-dependent seven transmembrane receptor signaling,” Trends in Biochemical Sciences, vol. 36, no. 9, pp. 457–469, 2011. [101] A. Shrestha, L. G. Puente, S. Brunette, and L. A. Megeney, “The role of Yca1 in proteostasis. Yca1 regulates the compo- sition of the insoluble proteome,” Journal of Proteomics, vol. 81, pp. 24–30, 2013. [86] T. Maeda, “The signaling mechanism of ambient pH sensing and adaptation in yeast and fungi,” The FEBS Journal, vol. 279, no. 8, pp. 1407–1413, 2012. Oxidative Medicine and Cellular Longevity Kimchi, “Systems biology analysis of programmed cell death,” Trends in Biochemical Sciences, vol. 35, no. 10, pp. 556–564, 2010. [118] E. Zalckvar, S. Bialik, and A. Kimchi, “The road not taken: a systems level strategy for analyzing the cell death network,” Autophagy, vol. 6, no. 6, pp. 813–815, 2010.
https://openalex.org/W2938702990	https://europepmc.org/articles/pmc6522743?pdf=render	English	null	Apigenin inhibits epithelial-mesenchymal transition of human colon cancer cells through NF-κB/Snail signaling pathway	Bioscience reports	2,019	cc-by	7,901	© 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). These authors contribute equally to this work Apigenin inhibits epithelial-mesenchymal transition of human colon cancer cells through NF-κB/Snail signaling pathway Jiafeng Tong1,, Ying Shen2,, Zhenghua Zhang3, Ye Hu2, Xu Zhang1 and Li Han1 1Department of Traditional Chinese Medicine, Huadong Hospital, Shanghai 200040, China; 2Department of Gastrology, Shanghai Municipal Hospital of Traditional Chinese Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai 200071, China; 3Department of Oncology, Jing’an District Central Hospital of Shanghai (Huashan Hospital, Fudan University, Jing’an branch200040), China 1Department of Traditional Chinese Medicine, Huadong Hospital, Shanghai 200040, China; 2Department of Gastrology, Shanghai Municipal Hospital of Traditional Chinese Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai 200071, China; 3Department of Oncology, Jing’an District Central Hospital of Shanghai (Huashan Hospital, Fudan University, Jing’an branch200040), China Correspondence: Li Han (xuzhangmail123@163.com) Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4 0 (CC BY) These authors contribute equally to this work Received: 28 February 2019 Revised: 02 April 2019 Accepted: 02 April 2019 Accepted Manuscript Online: 09 April 2019 Version of Record published: 14 May 2019 Colon cancer is a leading cause of cancer-related deaths worldwide. The epithelial-mesenchymal transition (EMT) plays an important role in tumor metastasis of colon cancer. We first evaluated the effects of EMT-related transcription factors on the prognosis of colon cancer through analysis the data obtained from The Cancer Genome Atlas (TCGA). And then we screened a series of Chinese medicine monomers to find effect EMT inhibitors. First, Snail is a more important EMT transcription factors for colon cancer prognosis, compared with Twist and Slug. Then, we found that apigenin effectively inhibits the activity of Snail. Apigenin could inhibit the EMT, migration, and invasion of human colon cancer cells in vitro and in vivo through the NF-κB/Snail pathway. Snail is a key regulator of EMT in colon cancer and Snail inhibitor apigenin may be a therapeutic application for patients with colon cancer. s an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attributio Introduction Colon cancer is a leading cause of cancer-related deaths worldwide. This disease is usually diagnosed at its advanced stage [1]. The 5-year survival rate of stage IV colon cancer is less than 10%. Colon metastases readily develop through the epithelial-mesenchymal transition (EMT) of epithelial cells, a key process that supports tumor metastasis. In the process of EMT, epithelial cells lose their typical epithelial characteristics and acquire mesenchy- mal traits. Cancer cells lose their cell–cell connection, cell-matrix contact, and normal epithelial polarity while gaining mesenchymal characteristics, which enhance the migratory and invasive ability of cancer cells. EMT is an integral component of colorectal cancer progression. So targetting EMT will be beneficial for prognosis and therapy of colorectal cancer [2,3]. These authors contribute equally to this work Received: 28 February 2019 Revised: 02 April 2019 Accepted: 02 April 2019 Accepted Manuscript Online: 09 April 2019 Version of Record published: 14 May 2019 Natural products (animals, plants, and other pharmacologically active ingredients present in nature) are important sources for prevention and treatment of diseases. The drugs used in various traditional medicines in the world almost belong to natural sources. The chemical substances from natural chemi- cals are also more than 30% in modern medicine applications. And more drugs are produced by structural modification of the natural products (paclitaxel, vincristine, artemisinin, etc.). According to a review by Newman et al., about 61% of the 877 new small chemical drugs introduced in the world between 1981 and 2002 were derived from natural products or the derivatives of natural products [4]. At present, as the development of synthetic drugs becomes more and more difficult, scientists have refocussed the develop- ment of new drugs on natural products. Chinese herbal monomers, such as the flavonoid monomers, have 1 Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 extensive pharmacological activity [5] (antitumor, anti-inflammatory, immune regulation, etc.), which has attracted more and more attention. Some transcription factors, such as zinc-finger proteins of the Snail/Slug family and Twist Family BHLH Transcrip- tion Factor 1 (Twist1), have been proved to regulate the EMT [6]. In order to find effect EMT inhibitor in colorectal cancer, we first evaluated the effects of these transcription factors on the prognosis of colon cancer through analy- sis the data obtained from The Cancer Genome Atlas (TCGA). Introduction And then we screened a series of Chinese medicine monomer to find effect EMT inhibitors through detecting the promoter activity of the prognosis-related transcription factor. The antitumor effects of the screened inhibitor (apigenin) were further evaluated in vitro and in vivo. Materials and methods Chemicals and antibodies Primary antibodies including vimentin (1:1000), E-cadherin (1:1000) and GAPDH (1:5000), NF-κB (1:1000), and Snail (1:1000) were supplied by Affinty (CO, U.S.A.). Secondary antibodies included FITC-labeled goat-antimouse lgG and TRITC-labeled goat anti-rabbit IgG were obtained from EarthOx (San Francisco, CA, U.S.A.). Apigenin and other flavanoid derivates were obtained from PUSH BIO-TECHNOLOGY (Chengdu, China). Matrigel and transwell chambers were purchased from BD Biosciences (San Jose, CA, U.S.A.). Cell culture Human colon carcinoma cell lines HCT-116 and LOVO, purchased from KeyGen Biotech (Nanjing, China), were maintained according to ATCC’s recommendation. The cells were grown in DMEM (for HCT-116) or RPMI 1640 (for LOVO) medium supplemented with 10% FBS (Hyclone, U.S.A.) in a humidified atmosphere of 37◦C, 5% CO2. Wound-healing assay HCT-116 and LOVO cells were seeded in six-well plates. The next day, a sterile 200 μl pipette tip was used to draw a straight line when the whole plates were covered with cells. Then the cells were washed with PBS for three times. After that different concentration of apigenin were added to DMEM mediums for 24 h. The distance of the scratch was calculated at different time points with microscope (Olympus CX31). Immunoﬂuorescent staining g HCT-116 cells (4 × 104 cells/ml) were pretreated with 10 or 20 μM apigenin for 48 h and then the cells were seeded on the chamber slides of 24-well culture plates. After 24 h, the cells were fixed with 4% paraformaldehyde about 30 min, following permeation with 1% Trition X-100 dissolved in PBS. Then the nonspecific binding was blocked with 5% BSA in TBS and incubated with primary antibodies (E-Cadherin and vimentin were 1:100 dilution respectively) for 1 h at 4◦C overnight. Slides were washed with PBS and incubated at 37◦C for 45 min with secondary antibodies conjugated to fluorescein isothiocyanate. Then the cells were washed with PBS. DAPI were used to monitor nuclear integrity. Finally, the cells were examined using Nikon confocal microscopy. Cell invasion assays y HCT-116 and LOVO cells were cultured to near 90–100% confluence. Then the cells were digested and seeded on the upper chamber (8.0 μm) of the transwell cultured with containing different concentrations of apigenin, which were coated with matrigel. The lower compartment of the chamber was added 0.5 ml of medium supplemented with 10% FBS. After treatment for 48 h, the medium was removed and the cells were washed with PBS for three times. Subsequently, the cells were fixed with precooled methanol and then the cells were stained with crystal violet for 10 min at room temperature. Finally, the stained cells were evaluated with the microscope. © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). Immunohistochemical analysis The fixed tissues were used for immunohistochemistry staining. About 3 μm thick paraffin sections were deparaf- finized with xylene, dehydrated in ethanol in a concentration gradient, and followed by incubation with 3% hydrogen peroxide at room temperature for 10 min to quench endogeneous peroxidase. Total 10 mM citrate buffer (pH 6.0) was used to antigen retrieval in a microwave about 10 min. Nonspecific-binding sites were blocked with normal goat serum for 30 min. The paraffin slices were incubated with the primary antibodies anti-E-cadherin (rabbit polyclonal, 1;100) and antivimentin (mouse polyclonal, 1:100) in 1% BSA overnight at 4◦C. The tissues were washed in PBS for three times. Then antimouse IgG and antirabbit lgG (diluted 1:200) was added to the paraffin slices for 30 min. DAB substrate was utilized for color development. tumor volumes were measured using calipers and calculated with the formula V = length × width2/2. After treatment for 2 weeks, the mice were killed. Tumor weight was recorded and the tissues were fixed in 10% formlinata solution and then embedded by paraffin. tumor volumes were measured using calipers and calculated with the formula V = length × width2/2. After treatment for 2 weeks, the mice were killed. Tumor weight was recorded and the tissues were fixed in 10% formlinata solution and then embedded by paraffin. pigenin effectively inhibits the transcription activity of snail The wound-healing experiments showed that both 10 and 20 μM apigenin could inhibit the migration of HCT-116 and LOVO cells comparing with the control groups (Figure 2C,D). Apigenin inhibits migration and invasion in human colon cancer cells We detected the effect of apigenin on invasion and metastasis activity of cancer cells through wound-healing and transwell experiments. The wound-healing experiments showed that both 10 and 20 μM apigenin could inhibit the migration of HCT-116 and LOVO cells comparing with the control groups (Figure 2C,D). Matrigel-coated transwell chambers were used to verify the invasive ability of colon cancer cells after apigenin treat- ment. The potential metastatic capacity of tumor cells can be represented by its invasiveness. In the apigenin-treated group, the number of cells that invading the basement membrane of the transwell chamber is significantly reduced compared with the control group, and it was concentration dependent in HCT-116 and LOVO cells (Figure 2E,F). Matrigel-coated transwell chambers were used to verify the invasive ability of colon cancer cells after apigenin treat- ment. The potential metastatic capacity of tumor cells can be represented by its invasiveness. In the apigenin-treated group, the number of cells that invading the basement membrane of the transwell chamber is significantly reduced compared with the control group, and it was concentration dependent in HCT-116 and LOVO cells (Figure 2E,F). Matrigel-coated transwell chambers were used to verify the invasive ability of colon cancer cells after apigenin treat- ment. The potential metastatic capacity of tumor cells can be represented by its invasiveness. In the apigenin-treated group, the number of cells that invading the basement membrane of the transwell chamber is significantly reduced compared with the control group, and it was concentration dependent in HCT-116 and LOVO cells (Figure 2E,F). pigenin effectively inhibits the transcription activity of snail Apigenin effectively inhibits the transcription activity of snail In order to find the inhibitor of Snail, we then screened dozens of Chinese herbal monomers. The inhibitors of snail transcription activity were screened using the dual luciferase reporter assay system. We found several Chinese herbal monomers had regulatory effect on Snail and apigenin had the best effect (Figure 1C) Apigenin inhibits the proliferation of human colon cancer cells The cytotoxicity of apigenin was detected by MTT. The results showed that apigenin can inhibit the proliferation of HCT116 and LOVO in a dose-dependent manner (Figure 2A,B). The half-maximal inhibitory concentrations (IC50) of HCT-116 and LOVO were 47.33 and 34.45 μM respectively. The morphology of colon cancer cells changed sig- nificantly after treatment with apigenin, and the cells exhibited low viability after exposed to apigenin at different concentrations. Apigenin inhibits the proliferation of human colon cancer cells The cytotoxicity of apigenin was detected by MTT. The results showed that apigenin can inhibit the proliferation of HCT116 and LOVO in a dose-dependent manner (Figure 2A,B). The half-maximal inhibitory concentrations (IC50) of HCT-116 and LOVO were 47.33 and 34.45 μM respectively. The morphology of colon cancer cells changed sig- nificantly after treatment with apigenin, and the cells exhibited low viability after exposed to apigenin at different concentrations. Apigenin inhibits migration and invasion in human colon cancer cells We detected the effect of apigenin on invasion and metastasis activity of cancer cells through wound-healing and transwell experiments. The wound-healing experiments showed that both 10 and 20 μM apigenin could inhibit the i i f HCT 116 d LOVO ll i i h h l (Fi 2C D) Apigenin inhibits migration and invasion in human colon cancer cells We detected the effect of apigenin on invasion and metastasis activity of cancer cells through wound-healing and transwell experiments. The wound-healing experiments showed that both 10 and 20 μM apigenin could inhibit the migration of HCT-116 and LOVO cells comparing with the control groups (Figure 2C,D). Matrigel-coated transwell chambers were used to verify the invasive ability of colon cancer cells after apigenin treat- ment. The potential metastatic capacity of tumor cells can be represented by its invasiveness. In the apigenin-treated Apigenin inhibits migration and invasion in human colon cancer cells We detected the effect of apigenin on invasion and metastasis activity of cancer cells through wound-healing and transwell experiments. Statistical analyses Data are represented as means +−S.D. by three independent experiments. The statistical significance of different groups was analyzed by one-way ANOVA following SPSS 17.0 software. The P values <0.05 were considered signif- icant difference existence. Snail is associated with prognosis of colon cancer p g The survival analysis of EMT-related transcription factors, Snail, Slug, Twist1, and Twist2, is shown in Figure 1A). Patients with higher expression of Snail had shorter survival time (*P<0.0005) AMONGST all the factors, which showed that Snail is the key driver of EMT in colon cancer. Snail as a poor prognosis gene were highly expressed in colon cancer tissues from the results of immunohistochemical staining (Figure 1B). The patient data were all derived from the TCGA database. Animal studies The BALB/c nu/nu mice were well taken care under pathogen-free conditions in accordance with institutional guide- lines. Studies were carried out in accordance with National Institutes of Health Animal Use Guidelines and the current Chinese Regulations and Standards for the Use of Laboratory Animals. The mice were subcutaneously injected with 100 μl resuspended HCT-116 cells (2 × 106 cells) via a sterile microsyringe. After the tumor volume reached about 40 mm3, the mice were randomly assigned to control or apigenin treatment groups. Control group was treated with equal volume saline, and treatment groups were intragastrically administered with different concentrations apigenin (low dose of 200 mg/kg and high dose of 300 mg/kg). Mice weight and tumor size were recorded every 2 days. The 2 2 © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4 0 (CC BY) Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). Apigenin reversed changes in E-cadherin and vimentin expression in human colon cancer cells When EMT occurs in tumor cells, the expression of E-Cadherin is down-regulated and the expression of vimentin is up-regulated. Cells treated with apigenin displayed distinct morphological changes compared with untreated cells. Pseudopodia increased significantly after apigenin treatment. We performed immunofluorescence double staining on 3 Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 Figure 1. Snail is a key regulator of EMT in colon cancer (A) The survival curves of Snail, Slug, Twist1, and Twist2 analysis in colon cancer patients. The patient data were obtained f the TCGA database (P<0.05, *P<0.0001). (B) The immunohistochemical microarray staining of Snail in normal and colon ca tissues. The data were obtained from the TCGA database. (C) The inhibitors of Snail transcription activity were screened using dual luciferase reporter assay system. Figure 1. Snail is a key regulator of EMT in colon cancer g y g (A) The survival curves of Snail, Slug, Twist1, and Twist2 analysis in colon cancer patients. The patient data were obtained from the TCGA database (P<0.05, *P<0.0001). (B) The immunohistochemical microarray staining of Snail in normal and colon cancer tissues. The data were obtained from the TCGA database. (C) The inhibitors of Snail transcription activity were screened using the dual luciferase reporter assay system. g y g (A) The survival curves of Snail, Slug, Twist1, and Twist2 analysis in colon cancer patients. The patient data were obtained from the TCGA database (P<0.05, *P<0.0001). (B) The immunohistochemical microarray staining of Snail in normal and colon cancer tissues. The data were obtained from the TCGA database. (C) The inhibitors of Snail transcription activity were screened using the dual luciferase reporter assay system. EMT markers E-Cadherin and vimentin, and found that E-Cadherin levels increased and vimentin levels decreased in the apigenin treatment groups (Figure 3A,B). © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). Apigenin inhibits NF-κB expression and Snail transcriptional activation in colon cancer Based on the above experimental results, we learned that apigenin could inhibit EMT by inhibiting the activity of Snail in colon cancer cells. NF-κB could regulate the activity of Snail. In order to detect the effect of apigenin on NF-κB, we performed a dual luciferase reporter gene assay experiment. The results showed that apigenin could also inhibit the transcriptional activity of NF-κB (Figure 4A). We also validated the inhibit effect of apigenin on the transcriptional activity of Snail (Figure 4A). Apigenin inhibited the transcriptional activity of Snail in a dose dependent manner. 4 © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4 0 (CC BY) Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 ence Reports (2019) 39 BSR20190452 /doi.org/10.1042/BSR20190452 e 2. Effect of apigenin on cell viability, invasion, and metastasis of HCT-116 and LOVO cells CT-116 cells viability was inhibited after reincubation in medium containing 0, 10, or 20 μM apigenin for 24 h. (B) LO ty was inhibited after reincubation in medium containing 0, 10, or 20 μM apigenin for 24 h. (C,D) Effects of ap tion ability of HCT-116 and LOVO cells. Cells were treated with 0, 10, or 20 μM apigenin (P<0.05, *P<0.01). (E,F) nin on the invasion ability of HCT-116 and LOVO cells (P<0.05, *P<0.01). Cells were treated with 0, 10, or 20 μM esentative images of cells at the bottom of the membrane stained with crystal violet were visualized. Figure 2. Effect of apigenin on cell viability, invasion, and metastasis of HCT-116 and LOVO cells Figure 2. Effect of apigenin on cell viability, invasion, and metastasis of HCT-116 and LOVO cells (A) HCT-116 cells viability was inhibited after reincubation in medium containing 0, 10, or 20 μM apigenin for 24 h. (B) LOVO cells viability was inhibited after reincubation in medium containing 0, 10, or 20 μM apigenin for 24 h. (C,D) Effects of apigenin on migration ability of HCT-116 and LOVO cells. Cells were treated with 0, 10, or 20 μM apigenin (P<0.05, *P<0.01). (E,F) Effects of apigenin on the invasion ability of HCT-116 and LOVO cells (P<0.05, *P<0.01). Apigenin inhibits NF-κB expression and Snail transcriptional activation in colon cancer Cells were treated with 0, 10, or 20 μM apigenin. Representative images of cells at the bottom of the membrane stained with crystal violet were visualized. (A) HCT-116 cells viability was inhibited after reincubation in medium containing 0, 10, or 20 μM apigenin for 24 h. (B) LOVO cells viability was inhibited after reincubation in medium containing 0, 10, or 20 μM apigenin for 24 h. (C,D) Effects of apigenin on migration ability of HCT-116 and LOVO cells. Cells were treated with 0, 10, or 20 μM apigenin (P<0.05, *P<0.01). (E,F) Effects of apigenin on the invasion ability of HCT-116 and LOVO cells (P<0.05, *P<0.01). Cells were treated with 0, 10, or 20 μM apigenin. Representative images of cells at the bottom of the membrane stained with crystal violet were visualized. 5 © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 Figure 3. Apigenin reverses changes of EMT biomarkers in HCT-116 cells (A,B) In HCT-116 cells, vimentin expression decreased whereas E-cadherin expression increased after apigenin treatment. Typical images of immunofluorescent double staining for E-cadherin and vimentin in HCT-116 cells. Each experiment was performed in triplicate. Results are shown as the means of three experiments, and error bars represent the standard deviation (P<0.05, *P<0.01). Figure 3. Apigenin reverses changes of EMT biomarkers in HCT-116 cells Figure 3. Apigenin reverses changes of EMT biomarkers in HCT-116 cells (A,B) In HCT-116 cells, vimentin expression decreased whereas E-cadherin expression increased after apigenin treatment. Typical images of immunofluorescent double staining for E-cadherin and vimentin in HCT-116 cells. Each experiment was performed in triplicate. Results are shown as the means of three experiments, and error bars represent the standard deviation (P<0.05, *P<0.01). The expression of transcription factors, NF-κB and Snail which related to EMT, were significantly up-regulated in tumor cells. After apigenin treatment, the transcription activity of NF-κB and Snail (Figure 4B-D) were significantly down-regulated. Apigenin inhibits NF-κB expression and Snail transcriptional activation in colon cancer We also examined the expression of NF-κB and Snail in the nucleus, and it was proved that the change of protein level consistent with the trend of NF-κB and Snail in the cytoplasm. After apigenin treatment, the expression of both NF-κB and Snail decreased in the cytosol and nucleus. These results showed that apigenin could inhibit EMT of colon cancer cells through NF-κB/Snail pathway. Apigenin inhibits the metastasis of HCT-116 cells in xenograft model nude mice The effect of apigenin on animal body weight was not significant (Figure 5A). Our results showed that after drug treatment, the tumor volume of mice treated with apigenin was decreased compared with control group (Figure 5B,C). In vivo experiments showed that apigenin could effectively reduce the tumor volume. 6 © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution cense 4 0 (CC BY) Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 Figure 4. Apigenin inhibits NF-κB and Snail in human colon cancer cells (A,B) The effect of apigenin on NF-κB and Snail expression in HCT-116 and LOVO cells tested by dual-luciferase reporter assay (P<0.05, *P<0.01). (C,D) Effect of apigenin on NF-κB and Snail expression in HCT-116 and LOVO cells detected by western blot. Apigenin reduced the expression of NF-κB and Snail (P<0.05, *P<0.01). cancer cells Figure 4. Apigenin inhibits NF-κB and Snail in human colon cancer cells (A,B) The effect of apigenin on NF-κB and Snail expression in HCT-116 and LOVO cells tested by dual-luciferase reporter assay (P<0.05, *P<0.01). (C,D) Effect of apigenin on NF-κB and Snail expression in HCT-116 and LOVO cells detected by western blot. Apigenin reduced the expression of NF-κB and Snail (P<0.05, *P<0.01). n inhibits NF-κB and Snail in human colon cancer cells apigenin on NF-κB and Snail expression in HCT-116 and LOVO cells tested by dual-luciferase reporter assay 1). (C,D) Effect of apigenin on NF-κB and Snail expression in HCT-116 and LOVO cells detected by western blot. the expression of NF-κB and Snail (P<0.05, *P<0.01). We further performed immunohistochemical staining on the fixed tissue cells. The expression of the EMT markers in colon tissue was evaluated. Our results demonstrated that E-Cadherin and occludin expression were up-regulated after apigenin treatment, while vimentin and N-cadherin expression were down-regulated after apigenin treatment (Figure 6A,B). We then performed immunohistochemical staining analysis of EMT-related transcription factors and found that the expression of NF-κB and Snail was down-regulated after apigenin treatment, which indicated that apigenin treatment could inhibit the expression of EMT-related transcription factors. Apigenin inhibits the metastasis of HCT-116 cells in xenograft model nude mice Besides, apigenin treatment could also inhibit the expression of Ki-67 and NF-kB signaling proteins such as p65 and IKK (Figure 6C,D), suggesting that apigenin might inhibit EMT of colon cancer cells through NF-kB/Snail signaling pathways. © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). Discussion Since colon cancer is a leading cause of cancer-related deaths worldwide [7], search for effective drugs is crucial for the treatment of colon carcinoma [8–10]. EMT promotes epithelial tumor invasion and transformation toward malignant disease [11]. The phenotype of EMT is mainly embodied in the cell–cell adhesion systems down-regulation, loss of epithelial polarity, and the acquisition of a mesenchymal phenotype, which lead to migration and movement of epithelial cells [12]. Thus, targetting EMT is beneficial for inhibition of carcinoma progression [13]. So in this work we screened the EMT inhibitors for colon carcinoma. In this work, we found that the expression of Snail1 has correlation with the prognosis of colon cancer patients. And we screened a series of flavonoid Chinese herbal monomers to find 7 7 © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4 0 (CC BY) Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 Bioscience Reports (2019) 39 BSR201904 https://doi.org/10.1042/BSR201904 Figure 5. Effect of apigenin on HCT-116 xenograft model (A) Body weight changes of animals with HCT-116 xenografts (P<0.05, *P<0.01). (B,C) Changes in the tumor volume of HCT-11 xenografts after being treated with apigenin. Apigenin treatment inhibited xenograft growth in a dose-dependent manner (P<0.0 *P<0.01). The results were shown as the means of three experiments, and error bars represent standard deviation. Figure 5. Effect of apigenin on HCT-116 xenograft model (A) Body weight changes of animals with HCT-116 xenografts (P<0.05, *P<0.01). (B,C) Changes in the tumor volume of HCT-116 xenografts after being treated with apigenin. Apigenin treatment inhibited xenograft growth in a dose-dependent manner (P<0.05, *P<0.01). The results were shown as the means of three experiments, and error bars represent standard deviation. Figure 5. Effect of apigenin on HCT-116 xenograft model (A) Body weight changes of animals with HCT-116 xenografts (P<0.05, *P<0.01). (B,C) Changes in the tumor volume of HCT-116 xenografts after being treated with apigenin. Apigenin treatment inhibited xenograft growth in a dose-dependent manner (P<0.05, *P<0.01). The results were shown as the means of three experiments, and error bars represent standard deviation. Snail1 transcription activity inhibitors. We found that apigenin had good inhibitory activity for Snail1, so we further evaluated its antitumor effect on human colon cancer cells. Snail1 transcription activity inhibitors. Discussion We found that apigenin had good inhibitory activity for Snail1, so we further evaluated its antitumor effect on human colon cancer cells. Snail, Slug, and Twist1 are important transcription factors for the regulation of the EMT in tumor cells. The first discovered transcriptional repressor of E-cadherin was Snail [14,15], which plays a fundamental role in EMT. The data analysis obtained from TCGA database revealed that the expression of Snail1 has correlation with the prognosis of colon cancer patients. Patients with high expression of Snail have shorter survival. The transcriptional factors of 8 8 © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 Figure 6. Effect of apigenin on NF-κB, Snail, Ki-67, NF-κB p65 IKK, and EMT protein levels (A) Effect of apigenin on EMT biomarkers in HCT-116 xenograft model (P<0.05, *P<0.01). (B) Expression level of EMT biomarkers. (C) Effect of apigenin on expression of NF-κB, Snail, Ki-67, NF-κB p65, IKK in HCT-116 xenograft model (P<0.05, *P<0.01). (D) Expression of NF-κB, Snail, Ki-67, NF-κB p65, and IKK. Figure 6. Effect of apigenin on NF-κB, Snail, Ki-67, NF-κB p65 IKK, and EMT protein levels (A) Effect of apigenin on EMT biomarkers in HCT-116 xenograft model (P<0.05, *P<0.01). (B) Expression level of EMT biomarkers. (C) Effect of apigenin on expression of NF-κB, Snail, Ki-67, NF-κB p65, IKK in HCT-116 xenograft model (P<0.05, *P<0.01). (D) Expression of NF-κB, Snail, Ki-67, NF-κB p65, and IKK. Figure 6. Effect of apigenin on NF-κB, Snail, Ki-67, NF-κB p65 IKK, and EMT protein levels (A) Effect of apigenin on EMT biomarkers in HCT-116 xenograft model (P<0.05, *P<0.01). (B) Expression level of EMT biomarkers. (C) Effect of apigenin on expression of NF-κB, Snail, Ki-67, NF-κB p65, IKK in HCT-116 xenograft model (P<0.05, **P<0.01). (D) Expression of NF-κB, Snail, Ki-67, NF-κB p65, and IKK. Slug and Twist1 have no correlation with the prognosis of colon cancer patients. These findings reveal the importance of Snail1 in the malignant evolution of colon cancer. And Snail1 is a marker of poor prognosis of colon cancer. © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4 0 (CC BY) Discussion E-cadherin and vimentin are important markers of EMT. Studies on E-cadherin behavior provided considerable insights into carcinoma progression [16]. In addition, altered E-cadherin expression is associated with tumor stage and increased lymph node metastasis [17,18]. Vimentin acts as an intermediate filament protein, which forms the cy- toskeleton along with microtubules and actin filaments. In this work, we found that apigenin increased the expression levels of E-cadherin but decreased the expression levels of vimentin. These results suggest that apigenin suppresses EMT of colon carcinoma cells. The results of the HCT-116 xenografts in nude mice also validated the effect of api- genin on EMT. Apigenin also inhibited the migration and invasion of tumor cells, which could be promoted by EMT of tumor cells. The mechanism of apigenin in the treatment of colon cancer can be explained by EMT. Studies show that the nuclear factor-κB (NF-κB)-mediated activation of Snail transcription plays an important role in the regulation of EMT in tumor cells [19–21]. NF-κB pathway regulates Snail expression through transcriptional and post-translational mechanisms, and NF-κB increases the transcription of Snail [22]. In this work, we found that NF-κB expression and activity decreased following apigenin treatment in colon carcinoma cells. Apigenin also inhib- ited the relocation of NF-κB to the nucleus. NF-κB expression and activity decreased following apigenin treatment in colon carcinoma cells. Apigenin also inhibited the relocation of NF-κB and Snail to the nucleus. This result is consistent with the earlier finding that NF-κB/Snail is crucial to the ability of apigenin to inhibit tumor progression. Apigenin is a nontoxic, naturally occurring dietary flavonoid, which is abundant in vegetables, fruits, and beans [23]. Many researches show that apigenin has anti-inflammatory, antioxidant, and anticancer properties [24,25]. Be- cause apigenin is plant-derived natural compound, it may be an ideal and safe antitumor agent [26]. Studies show that apigenin has good antitumor activities in vitro and in vivo [27–30] and inhibits the growth of various cancer cell lines [31]. Different studies have found various molecular mechanisms underlying the anticancer effect of apigenin 9 Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 [32]. Apigenin can strongly inhibit the growth of breast cancer cells, including HER2-positive cells [33] and suppress the metastasis of human hepatocellular carcinoma by inhibiting EMT [34]. Funding g The present study was funded by Foundation of Shanghai Municipal Science and Technology Commission Project: Prevention and Treatment of Colorectal Cancer by TCM syndrome differentiation [grant number 17dz2307500] and Clinical Study on Chang Ji Tai Combined with Auricular Acupuncture in Treating Postoperative Cancer-related fatigue of Postoperative colorectal cancer, Special item of Shanghai Integrative Medicine [grant number ZY(2018-2020)-FWTX-3016]. Competing Interests The authors declare that there are no competing interests associated with the manuscript. Author Contribution J.T. was focussed on the integrity of the entire study, study concepts, manuscript editing and manuscript review. Y.S. and Z.Z. were involved in the study design, data acquisition, and statistical analysis. Y.H. and X.Z. were dedicated to the literature research and data analysis. L.H. carried out the manuscript preparation and manuscript review. All authors have read and approved this article. Discussion Although some researchers indicate that apigenin can induce apoptosis and autophagy in HCT116 colon cancer cells [35], the effect of apigenin on the EMT of colon cancer remains unclear to date. Accordingly, the present study evaluated the effect of apigenin on EMT of colon cancer cells in vitro and in vivo and clarified the mechanism of apigenin regulating EMT of colon cancer cells. We found that apigenin inhibited migration and invasion of colon carcinoma cells. Apigenin inhibited the EMT of HCT-116 and LOVO human colon cancer cells through NF-κB/Snail signaling pathway. The present study evaluated the efficacy of apigenin in the treatment of colorectal cancer at the animal level by establishing xenografts on Balbc nude mice. Actually, PDX (patient derived xenografts) model is closer to the characteristics of clinical tumor sam- ples. In the future, we will further evaluate the efficacy of apigenin in the treatment of colorectal cancer based on PDX model. In summary, we found that Snail is a more important EMT transcription factors for colon cancer prognosis, com- pared with Twist1 and Slug. After screening a series of flavonoid Chinese herbal monomers, we found that apigenin effectively inhibited the activity of Snail. And we found that apigenin suppressed the EMT, migration, and invasion of human colon cancer by inhibiting the NF-κB/Snail pathway. Our results provide a new mechanistic basis for the therapeutic application of apigenin in patients with colon cancer. Further exploration of apigenin and its mechanisms could lead to the development of a new therapeutic approach to treating colon cancer. Abbreviations EMT , epithelial-mesenchymal transition; NF-κB, nuclear factor-κB; PDX , patient derived xenograft; TCGA, The Cancer Genome Atlas. © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). References 1 Weekes, J., Lam, AK-Y, Sebesan, S. and Ho, Y.-H. (2009) Irinotecan therapy and molecular targets in colorectal cancer: a systemic review. World J. Gastroenterol. 15, 3597–3602, https://doi.org/10.3748/wjg.15.3597 2 Bates, R.C. and Mercurio, A.M. (2005) The epithelial-mesenchymal transition (EMT) and colorectal cancer progression. Cancer Biol. Ther. 4, 365–370, https://doi.org/10.4161/cbt.4.4.1655 https://doi.org/10.4161/cbt.4.4.1655 3 Lefort, E.C. and Blay, J. (2011) The dietary ﬂavonoid apigenin enhances the activities of the anti-metastatic protein CD26 on human colon carcinoma cells. Clin. Exp. Metastasis 28, 337–349, https://doi.org/10.1007/s10585-010-9364-6 4 Newman, D.J. and Cragg, G.M. (2007) Natural products as sources of new drugs over the last 25 years. J. Nat Prod. 70, 461–477, https://doi.org/10.1021/np068054v 5 Ni J Zh C D L I t l (2016) Efﬁ f t diti l Chi di i i t ti Bi d R 4 3 14 p g p 5 Nie, J., Zhao, C., Deng, L.I. et al. (2016) Efﬁcacy of traditional Chinese medicine in treating cancer. Biomed. Rep. 4, 3–14, https://doi.org/10.3892/br.2015.537 6 Yu, Q., Zhang, K., Wang, X., Liu, X. and Zhang, Z. (2010) Expression of transcription factors snail, slug, and twist in human bladder carcinoma. J. Exp. Clin. Cancer Res. 29, 119, https://doi.org/10.1186/1756-9966-29-119 7 Rashtak, S., Rego, R., Sweetser, S.R. and Sinicrope, F.A. (2017) Sessile serrated polyps and colon cancer prevention. Cancer Prev. Res. (Phila.) 10, 270–278, https://doi.org/10.1158/1940-6207.CAPR-16-0264 8 Siveen, K.S., Ahn, K.S., Ong, T.H. et al. (2014) Y-tocotrienol inhibits angiogenesis-dependent growth of human hepatocellular carcinoma through abrogation of AKT/mTOR pathway in an orthotopic mouse model. Oncotarget 5, 1897–1911, https://doi.org/10.18632/oncotarget.1876 8 Siveen, K.S., Ahn, K.S., Ong, T.H. et al. (2014) Y-tocotrienol inhibits angiogenesis-dependent growth of human hepatocellular carcinoma through abrogation of AKT/mTOR pathway in an orthotopic mouse model. Oncotarget 5, 1897–1911, https://doi.org/10.18632/oncotarget.1876 9 Cai, W., Xiong Chen, Z., Rane, G. et al. (2017) Wanted DEAD/H or alive: helicases winding up in cancers. J. Natl Cancer Inst. 109, htt //d i /10 1093/j i/dj 278 abrogation of AKT/mTOR pathway in an orthotopic mouse model. Oncotarget 5, 1897–1911, https://doi.org/10.18632/oncotarget.1876 9 Cai, W., Xiong Chen, Z., Rane, G. et al. (2017) Wanted DEAD/H or alive: helicases winding up in cancers. J. Natl Cancer Inst. 109, https://doi.org/10.1093/jnci/djw278 9 Cai, W., Xiong Chen, Z., Rane, G. et al. (2017) Wanted DEAD/H or alive: helicases winding up in cancers. J. Natl Cancer Inst. 109, https://doi.org/10.1093/jnci/djw278 10 © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 https://doi.org/10.1093/ajcn/67.2.255 p g j 26 Patel, D., Shukla, S. and Gupta, S. (2007) Apigenin and cancer chemoprevention: progress, potential and promise (review). Int. J. Oncol. 30, 233–245 27 Wang, Q.R., Yao, X.Q., Wen, G. et al. (2011) Apigenin suppresses the growth of colorectal cancer xenografts via phosphorylation and up-regulated FADD expression. Oncol. Lett. 2, 43–47, https://doi.org/10.3892/ol.2010.215 , , , p , ( ) p g p p g , p p ( ) , 27 Wang, Q.R., Yao, X.Q., Wen, G. et al. (2011) Apigenin suppresses the growth of colorectal cancer xenografts via phosphorylation and up-regulated FADD expression Oncol Lett 2 43 47 https://doi org/10 3892/ol 2010 215 27 Wang, Q.R., Yao, X.Q., Wen, G. et al. (2011) Apigenin suppresses the growth of colorectal cancer xenografts via phosphorylation and up-regulated FADD expression. Oncol. Lett. 2, 43–47, https://doi.org/10.3892/ol.2010.215 nd Pelling, J.C. (1997) Induction of p21/WAF1 and G1 cell-cycle arrest by the chemopreventive agent apigenin. Mol. Carcinog. 19, 28 Lepley, D.M. and Pelling, J.C. (1997) Induction of p21/WAF1 and G1 cell-cycle arrest by the chemopreventive agent apigenin. Mol. Carcinog. 19, 74–82, https://doi.org/10.1002/(SICI)1098-2744(199707)19:2%3c74::AID-MC2%3e3.0.CO;2-L 28 Lepley, D.M. and Pelling, J.C. (1997) Induction of p21/WAF1 and G1 cell-cycle arrest by the chemopreventive agent apigenin. 28 Lepley, D.M. and Pelling, J.C. (1997) Induction of p21/WAF1 and G1 cell-cycle arrest by the chemopreventive agent apigenin. Mol. Carcinog. 19, 74 82 https://doi org/10 1002/(SICI)1098 2744(199707)19:2%3c74::AID MC2%3e3 0 CO;2 L 28 Lepley, D.M. and Pelling, J.C. (1997) Induction of p21/WAF1 and G1 cell-cycle arrest by the chemo 74–82, https://doi.org/10.1002/(SICI)1098-2744(199707)19:2%3c74::AID-MC2%3e3.0.CO;2-L 74–82, https://doi.org/10.1002/(SICI)1098-2744(199707)19:2%3c74::AID-MC2%3e3.0.CO;2-L 29 Hu, X.-W., Meng, D. and Fang, J. (2008) Apigenin inhibited migration and invasion of human ovarian cancer A2780 cells through focal adhesion kinase. Carcinogenesis 29, 2369–2376, https://doi.org/10.1093/carcin/bgn244 30 Shukla, S. and Gupta, S. (2010) Apigenin: a promising molecule for cancer prevention. Pharm. Res. 27, 962–978, 30 Shukla, S. and Gupta, S. (2010) Apigenin: a promising molecule for cancer prevention. Pharm. Res. 27, 962–978, https://doi.org/10.1007/s11095-010-0089-7 30 Shukla, S. and Gupta, S. (2010) Apigenin: a pr https://doi.org/10.1007/s11095-010-0089-7 31 Salmani, J.M.M., Zhang, X.-P., Jacob, J.A. and Chen, B.-A. (2017) Apigenin’s anticancer properties and molecular mechanisms of a advances and future prospectives. Chin. J. Nat. Med. 15, 321–329 32 Liu, L.-Z., Fang, J., Zhou, Q., Hu, X., Shi, X. and Jiang, B.-H. (2005) Apigenin inhibits expression of vascular endothelial growth factor and angiogenesis in human lung cancer cells: implication of chemoprevention of lung cancer. Mol. Pharmacol. https://doi.org/10.1074/jbc.274.35.24819 21 Chakrabarti, O., Veeraraghavalu, K., Tergaonkar, V., Liu, Y., Androphy, E.J., Stanley, M.A. et al. (2004) Human papillomavirus type 16 E6 amino acid 83 variants enhance E6-mediated MAPK signaling and differentially regulate tumorigenesis by notch signaling and oncogenic Ras. J. Virol. 78, 5934–5945, https://doi.org/10.1128/JVI.78.11.5934-5945.2004 p g M.J., Puig, I., Dominguez, D. et al. (2004) Regulation of Snail transcription during epithelial to mesenchymal transition of tumor cells. 22 Barbera, M.J., Puig, I., Dominguez, D. et al. (2004) Regulation of Snail transcription during epithelial to mesenchymal transition of tumor cells. Oncogene 23, 7345–7354, https://doi.org/10.1038/sj.onc.1207990 23 Wang, W., VanAlstyne, P.C., Irons, K.A., Chen, S., Stewart, J.W. and Birt, D.F. (2004) Individual and interactive effects of apigenin analogs on G2/M cell-cycle arrest in human colon carcinoma cell lines. Nutr. Cancer 48, 106–114, https://doi.org/10.1207/s15327914nc4801˙14 24 Lefort, E.C. and Blay, J. (2013) Apigenin and its impact on gastrointestinal cancers. Mol. Nutr. Food Res. 57, 126–144, https://doi.org/10.1002/mnfr.201200424 24 Lefort, E.C. and Blay, J. (2013) Apigenin a https://doi.org/10.1002/mnfr.201200424 24 Lefort, E.C. and Blay, J. (2013) Apigenin and its impact on gastrointestinal cancers. Mol. Nutr. Food Res. 57, 126 144, https://doi.org/10.1002/mnfr.201200424 25 Janssen, K., Mensink, R.P., Cox, F.J., Harryvan, J.L., Hovenier, R., Hollman, P.C. et al. (1998) Effects of the ﬂavonoids quercetin and apigenin on https://doi.org/10.1002/mnfr.201200424 25 Janssen, K., Mensink, R.P., Cox, F.J., Harryvan, J.L., Hovenier, R., Hollman, P.C. et al. (1998) Effects of the ﬂavonoids quercetin and apigenin on hemostasis in healthy volunteers: results from an in vitro and a dietary supplement study. Am. J. Clin. Nutr. 67, 255–262, https://doi org/10 1093/ajcn/67 2 255 25 Janssen, K., Mensink, R.P., Cox, F.J., Harryvan, J.L., Hovenier, R., Hollman, P.C. et al. (1998) Effects of the ﬂavonoids quercetin and apigenin on hemostasis in healthy volunteers: results from an in vitro and a dietary supplement study. Am. J. Clin. Nutr. 67, 255–262, https://doi.org/10.1093/ajcn/67.2.255 https://doi.org/10.1038/sj.bjc.6605530 15 Brzozowa, M., Michalski, M., Wyrobiec, G. et al. (2015) The role of Snail1 transcription factor in colorectal cancer progression and metastasis. Contemp. Oncol. (Pozn.) 19, 265–270 16 Wijnhoven, B.P., Dinjens, W.N. and Pignatelli, M. (2000) E-cadherin-catenin cell-cell adhesion complex and human cancer. Br. J. Surg. 87, 992–1005, htt //d i /10 1046/j 1365 2168 2000 01513 https://doi.org/10.1046/j.1365-2168.2000.01513.x 17 Chao, Y.L., Shepard, C.R. and Wells, A. (2010) Breast carcinoma cells re-express E-cadherin during mesenchymal to epithelial reverting transition. Mol. Cancer 9, 179, https://doi.org/10.1186/1476-4598-9-179 18 Bae, K.-M., Parker, N.N., Dai, Y., Vieweg, J. and Siemann, D.W. (2011) E-cadherin plasticity in prostate cancer stem cell invasion. Am. J. Cancer Res. 1, 71–84 19 Julien, S., Puig, I., Caretti, E. et al. (2007) Activation of NF-kappaB by Akt upregulates Snail expression and induces epithelium mesenchyme transition. Oncogene 26, 7445–7456, https://doi.org/10.1038/sj.onc.1210546 20 Liu, Y., Tergaonkar, V., Krishna, S. and Androphy, E.J. (1999) Human papillomavirus type 16 E6-enhanced susceptibility of L929 cells to tumor necrosis factor alpha correlates with increased accumulation of reactive oxygen species. J. Biol. Chem. 274, 24819–24827, https://doi.org/10.1074/jbc.274.35.24819 Bioscience Reports (2019) 39 BSR20190452 https://doi.org/10.1042/BSR20190452 10 Siveen, K.S., Nguyen, A.H., Lee, J.H. et al. (2014) Negative regulation of signal transducer and activator of transcription-3 signalling cascade by lupeol inhibits growth and induces apoptosis in hepatocellular carcinoma cells. Br. J. Cancer 111, 1327, https://doi.org/10.1038/bjc.2014.422 11 Hay, E.D. (1995) An overview of epithelio-mesenchymal transformation. Acta Anat. (Basel) 154, 8–20, https://doi.org/10.1159/000147748 12 Bates, R.C. (2005) Colorectal cancer progression: integrin alphavbeta6 and the epithelial-mesenchymal transition (EMT). Cell Cycle 4, 1350–1352, https://doi org/10 4161/cc 4 10 2053 10 Siveen, K.S., Nguyen, A.H., Lee, J.H. et al. (2014) Negative regulation of signal transducer and activator of transcription-3 signallin inhibits growth and induces apoptosis in hepatocellular carcinoma cells. Br. J. Cancer 111, 1327, https://doi.org/10.1038/bjc.201 inhibits growth and induces apoptosis in hepatocellular carcinoma cells. Br. J. Cancer 111, 1327, https://doi.org/10.1038/bjc.2014.422 11 Hay, E.D. (1995) An overview of epithelio-mesenchymal transformation. Acta Anat. (Basel) 154, 8–20, https://doi.org/10.1159/000147748 12 Bates R C (2005) Colorectal cancer progression: integrin alphavbeta6 and the epithelial mesenchymal transition (EMT) Cell Cycle 4 1350 1352 g p p p , , p g j 11 Hay, E.D. (1995) An overview of epithelio-mesenchymal transformation. Acta Anat. (Basel) 154, 8–20, https://doi.org/10.1159/000147748 11 Hay, E.D. (1995) An overview of epithelio-mesenchymal transformation. Acta Anat. (Basel) 154, 8–20, https://doi.org/10.1159/000147748 12 Bates, R.C. (2005) Colorectal cancer progression: integrin alphavbeta6 and the epithelial-mesenchymal transition (EMT). Cell Cycle 4, 1350–1352, https://doi.org/10.4161/cc.4.10.2053 11 Hay, E.D. (1995) An overview of epithelio-mesenchymal transformation. Acta Anat. (Basel) 154, 8–20, https://doi.org/10.1159/00 12 Bates, R.C. (2005) Colorectal cancer progression: integrin alphavbeta6 and the epithelial-mesenchymal transition (EMT). Cell Cycle 4, 1350–1352, https://doi.org/10.4161/cc.4.10.2053 13 Bellovin, D.I., Bates, R.C., Muzikansky, A., Rimm, D.L. and Mercurio, A.M. (2005) Altered localization of p120 catenin during epithelial to mesenchymal transition of colon carcinoma is prognostic for aggressive disease. Cancer Res. 65, 10938–10945, https://doi.org/10.1158/0008-5472.CAN-05-1947 14 Wu, Y. and Zhou, B.P. (2010) TNF-alpha/NF-kappaB/Snail pathway in cancer cell migration and invasion. Br. J. Cancer 102, 639–644, https://doi.org/10.1038/sj.bjc.6605530 13 Bellovin, D.I., Bates, R.C., Muzikansky, A., Rimm, D.L. and Mercurio, A.M. (2005) Altered localization of p120 catenin during epithelial to mesenchymal transition of colon carcinoma is prognostic for aggressive disease. Cancer Res. 65, 10938–10945, https://doi.org/10.1158/0008-5472.CAN-05-1947 14 Wu Y and Zhou B P (2010) TNF alpha/NF kappaB/Snail pathway in cancer cell migration and invasion Br J Cancer 102 639 644 14 Wu, Y. and Zhou, B.P. (2010) TNF-alpha/NF-kappaB/Snail pathway in cancer cell migration and invasion. Br. J. Cancer 102, 63 https://doi.org/10.1038/sj.bjc.6605530 © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). https://doi.org/10.1093/ajcn/67.2.255 68, 635–643 32 Liu, L.-Z., Fang, J., Zhou, Q., Hu, X., Shi, X. and Jiang, B.-H. (2005) Apigenin inhibits expression of vascular endothelial growth factor and angiogenesis in human lung cancer cells: implication of chemoprevention of lung cancer. Mol. Pharmacol. 68, 635–643 33 Scherbakov, A.M. and Andreeva, O.E. (2015) Apigenin inhibits growth of breast cancer cells: the role of ERalpha and HER2/neu. Acta. Naturae 7, 133–139 33 Scherbakov, A.M. and Andreeva, O.E. (2015) Apigenin inhibits growth of breast cancer cells: the role of ERalpha and HER2/neu. Acta. Naturae 7, 133–139 34 Qin, Y., Zhao, D., Zhou, H.-G. et al. (2016) Apigenin inhibits NF-kappaB and snail signaling, EMT and metastasis in human hepatocellular carcinoma. Oncotarget 7, 41421–41431, https://doi.org/10.18632/oncotarget.9404 34 Qin, Y., Zhao, D., Zhou, H.-G. et al. (2016) Apigenin inhibits NF-kappaB and snail signaling, EMT and metastasis in human hepatocellular carcinoma. Oncotarget 7, 41421–41431, https://doi.org/10.18632/oncotarget.9404 35 Shao, H., Jing, K., Mahmoud, E., Huang, H., Fang, X. and Yu, C. (2013) Apigenin sensitizes colon cancer cells to antitumor activity of ABT-263. Mol. Cancer Ther. 12, 2640–2650, https://doi.org/10.1158/1535-7163.MCT-13-0066 35 Shao, H., Jing, K., Mahmoud, E., Huang, H., Fang, X. and Yu, C. (2013) Apigenin sensitizes colon cancer cells to antitumor activity of ABT-263. Mol. Cancer Ther. 12, 2640–2650, https://doi.org/10.1158/1535-7163.MCT-13-0066 35 Shao, H., Jing, K., Mahmoud, E., Huang, H., Fang, X. and Yu, C. (2013) Apigenin sens Cancer Ther. 12, 2640–2650, https://doi.org/10.1158/1535-7163.MCT-13-0066 11 © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY). © 2019 The Author(s). This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4 0 (CC BY)
https://openalex.org/W2142132792	https://www.epj-conferences.org/10.1051/epjconf/20136104011/pdf	English	null	Study of High Energy Emission from Relativistic Jets with VERITAS	EPJ web of conferences	2,013	cc-by	6,062	1 Introduction: VERITAS VERITAS (Very Energetic Radiation Imaging Telescope Array System) is an imaging atmospheric Cherenkov tele- scope array: it detects Cherenkov radiation from ener- getic particles, produced by the interaction of a gamma- ray primary in the atmosphere, using large optical reﬂec- tors on the ground. VERITAS operates at the base camp of the Fred Lawrence Whipple Observatory (FLWO) in southern Arizona. There are four 12m-diameter telescopes of Davies-Cotton design in the VERITAS array, with a camera comprising 499 photomultiplier tubes (PMTs) ar- ranged in a hexagonal pattern, at the focal plane of each telescope. The total ﬁeld of view has a diameter of 3.5◦. Gamma-rays interact in the Earth’s atmosphere be- fore reaching the ground, producing electromagnetic cas- cades of relativistic electron-positron pairs which gener- ate short-lived but bright ﬂashes of Cherenkov radiation. Cosmic rays, the principal source of background, also produce showers of charged particles and Cherenkov ra- diation but at a rate much higher than that of gamma- ray showers. Imaging Atmospheric Cherenkov Telescopes (IACTs) such as VERITAS utilize the Cherenkov light emitted from these air showers to form an image of the longitudinal development of the air shower on the camera- plane. The use of the four VERITAS telescopes in com- bination helps to increase the collection area and obtain a stereoscopic image of the particle showers to reduce cos- mic ray background. VERITAS has a sensitive energy range from 85 GeV to 30 TeV and an angular resolution of < 0◦.1 at 1 TeV. Spectral reconstruction begins at ∼150 GeV with an energy resolution ∼15% - 20%. Figure 1 is a photograph of the VERITAS observatory, showing the ar- ray of four telescopes. Details of the VERITAS telescopes are provided elsewhere [1]. Observations performed with VERITAS have led and will continue to lead to signiﬁcant Figure 1. The four VERITAS telescopes located at the base camp of the Whipple Observatory in southern Arizona. progress in our understanding of the accelerators of very high energy particles in the universe. VERITAS started four-telescope operations in 2007, and typically collects more than ∼1100 hours of data per year (including ∼200 hours data taken in moderate moon- light). A 30% improvement in sensitivity was achieved in 2009 when one of the telescopes was relocated to provide a more symmetric array layout, and a new alignment sys- tem was used which improved the instrument point spread function [2]. R. Mukherjee1,a for the VERITAS Collaboration R. Mukherjee1,a for the VERITAS Collaboration 1Department of Physics and Astronomy, Barnard College, Columbia University, NY 10027, USA epartment of Physics and Astronomy, Barnard College, Columbia University, NY 10027, USA Abstract. VERITAS is an imaging atmospheric Cherenkov telescope array carrying out observations of the gamma-ray sky at energies above 100 GeV. Located in southern Arizona, VERITAS consists of an array of four telescopes, and carries out a comprehensive observing program for Galactic and extragalactic high energy sources. Observations of astrophysical objects in the TeV band are sensitive probes of highly energetic processes occurring in these sources. The majority of the extragalactic sources detected at TeV energies are active galaxies of the blazar class, sources where we view the jet nearly along its axis. In this report we present recent results from the VERITAS extragalactic program, focusing on blazars. Figure 1. The four VERITAS telescopes located at the base camp of the Whipple Observatory in southern Arizona. DOI: 10.1051/ C ⃝Owned by the authors, published by EDP Sciences, 2013 , / 04011 (2013) 36104011 201 61 epjconf EPJ Web of Conferences DOI: 10.1051/ C ⃝Owned by the authors, published by EDP Sciences, 2013 , / 04011 (2013) 36104011 201 61 epjconf EPJ Web of Conferences DOI: 10.1051/ C ⃝Owned by the authors, published by EDP Sciences, 2013 , / 04011 (2013) 36104011 201 61 epjconf EPJ Web of Conferences DOI: 10.1051/ , / 04011 (2013) 36104011 201 61 epjconf EPJ Web of Conferences This is an Open Access article distributed under the terms of the Creative Commons Attribution License 2.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ae-mail: rmukherj@barnard.edu Article available at http://www.epj-conferences.org or http://dx.doi.org/10.1051/epjconf/20136104011 1 Introduction: VERITAS The ma- jority of the detected sources oﬀthe Galactic plane are blazars (shown in red). Source Name Class Redshift (Extragalactic) z Mrk 421 HBL 0.030 Mrk 501 HBL 0.034 1ES 2344+514 HBL 0.044 1ES 1959+650 HBL 0.047 1ES 1727+501 HBL 0.055 BL Lacertae LBL 0.069 W Comae IBL 0.102 VER J0521+211 IBL/HBL 0.108 RGB J0710+591 HBL 0.125 H 1426+428 HBL 0.129 B2 1215+303 IBL/HBL 0.130? 1ES 0806+524 HBL 0.138 1ES 0229+200 HBL 0.140 1ES 1440+122 IBL/HBL 0.162 RX J0648.7+1516 HBL 0.179 1ES 1218+304 HBL 0.182 RBS 0413 HBL 0.190 1ES 1011+496 HBL 0.212 1ES 0414+009 HBL 0.287 1ES 0502+675 HBL 0.341? 1ES 0647+250 HBL ∼0.45 PG 1553+113 HBL 0.43 < z < 0.47 3C 66A IBL 0.444? PKS 1424+240 IBL/HBL > 0.6035 M87 FR I 0.0044 M82 Starburst 3.9 Mpc Table 1: Extragalactic sources of TeV gamma-ray emission detected by VERITAS. Some redshifts are considered uncertain. Source Name Class Redshift (Extragalactic) z Mrk 421 HBL 0.030 Mrk 501 HBL 0.034 1ES 2344+514 HBL 0.044 1ES 1959+650 HBL 0.047 1ES 1727+501 HBL 0.055 BL Lacertae LBL 0.069 W Comae IBL 0.102 VER J0521+211 IBL/HBL 0.108 RGB J0710+591 HBL 0.125 H 1426+428 HBL 0.129 B2 1215+303 IBL/HBL 0.130? 1ES 0806+524 HBL 0.138 1ES 0229+200 HBL 0.140 1ES 1440+122 IBL/HBL 0.162 RX J0648.7+1516 HBL 0.179 1ES 1218+304 HBL 0.182 RBS 0413 HBL 0.190 1ES 1011+496 HBL 0.212 1ES 0414+009 HBL 0.287 1ES 0502+675 HBL 0.341? 1ES 0647+250 HBL ∼0.45 PG 1553+113 HBL 0.43 < z < 0.47 3C 66A IBL 0.444? PKS 1424+240 IBL/HBL > 0.6035 M87 FR I 0.0044 M82 Starburst 3.9 Mpc Table 1: Extragalactic sources of TeV gamma-ray emission detected by VERITAS. Some redshifts are considered uncertain. Figure 2. Photograph of one of the VERITAS cameras showing the new hQE PMTs (July 2012). Figure 2. Photograph of one of the VERITAS cameras showing the new hQE PMTs (July 2012). Figure 3. The VERITAS sky map of detected sources as of May 2013, in Galactic coordinates. The shaded area shows the region of visibility for VERITAS above 50 degree elevation. The ma- jority of the detected sources oﬀthe Galactic plane are blazars (shown in red). Figure 2. Photograph of one of the VERITAS cameras showing the new hQE PMTs (July 2012). Table 1: Extragalactic sources of TeV gamma-ray emission detected by VERITAS. Some redshifts are considered uncertain. Figure 3. 1 Introduction: VERITAS Blazars are the best extra- galactic probes of the extragalactic background light via their interaction with TeV photons traveling cosmologi- cal distances. Blazar observations have also been used to constrain the weak intergalactic magnetic ﬁelds (IGMF) (e.g. [6]), and test the validity of the Lorentz Invariance principle at high energies [7]. Blazars observations form a large part of the VERITAS program which includes a combination of planned blazar observations as part of a key science plan, long term monitoring of certain blazars, a discovery program for detecting new TeV blazars, and target-of-opportunity observations for ﬂaring sources. to eliminate background noise – at the array-level, a trig- ger requires simultaneous Cherenkov images in at least two telescopes, within a 50 ns time window, which then causes a readout of the 500 MSample/s Flash-ADC data acquisition system for each pixel [5]. To date VERITAS has detected 46 Galactic and extragalactic sources, includ- ing blazars, pulsar wind nebulae (PWNe), supernova rem- nants (SNR), binary systems, a starburst galaxy and the Crab Pulsar. Figure 3 shows all sources detected by VER- ITAS, in a galactic projection. In this article, we review some of the results from VERITAS focusing on blazars and extragalactic sources. 1 Introduction: VERITAS The VERITAS sky map of detected sources as of May 2013, in Galactic coordinates. The shaded area shows the region of visibility for VERITAS above 50 degree elevation. The ma- jority of the detected sources oﬀthe Galactic plane are blazars (shown in red). appear in stellar-mass black holes in X-ray binaries and GRBs. They are scaled-down versions of the jets seen in AGN. The observed similarity (in morphology and spec- trum) of jets from black holes of diﬀerent mass suggests that they share a common physical origin. VHE observa- tions of AGN provide the opportunity to explore a large range of scientiﬁc topics, such as the study of the proper- ties of SMBHs and their environments, indirect studies of the central engine and particle acceleration in relativistic jets. Open questions in AGN physics include the nature of particle acceleration and emission in jets, the structure and composition of jets, and with broad band blazar ob- servations, we wish to constrain models of jet formation, acceleration, and collimation. Blazars are the best extra- galactic probes of the extragalactic background light via their interaction with TeV photons traveling cosmologi- cal distances. Blazar observations have also been used to constrain the weak intergalactic magnetic ﬁelds (IGMF) (e.g. [6]), and test the validity of the Lorentz Invariance principle at high energies [7]. Blazars observations form a large part of the VERITAS program which includes a combination of planned blazar observations as part of a key science plan, long term monitoring of certain blazars, a discovery program for detecting new TeV blazars, and target-of-opportunity observations for ﬂaring sources. appear in stellar-mass black holes in X-ray binaries and GRBs. They are scaled-down versions of the jets seen in AGN. The observed similarity (in morphology and spec- trum) of jets from black holes of diﬀerent mass suggests that they share a common physical origin. VHE observa- tions of AGN provide the opportunity to explore a large range of scientiﬁc topics, such as the study of the proper- ties of SMBHs and their environments, indirect studies of the central engine and particle acceleration in relativistic jets. Open questions in AGN physics include the nature of particle acceleration and emission in jets, the structure and composition of jets, and with broad band blazar ob- servations, we wish to constrain models of jet formation, acceleration, and collimation. 1 Introduction: VERITAS In this new telescope arrangement, VERI- TAS is able to detect a source with a Crab Nebula-like spectrum and a ﬂux of 1% Crab Nebula strength in about 25 hours. In the summer of 2012 the VERITAS cameras were replaced with new high-quantum eﬃciency (hQE, QE > 32%) photomultiplier tubes [3] that increased the photon detection eﬃciency of each camera by approxi- mately 50%. This is part of a series of upgrades and im- provements that have been implemented on the VERITAS observatory since 2007 (see [4] for details on the gamma- ray detection sensitivity of the upgraded VERITAS obser- vatory). Figure 2 is a photograph of one of the VERITAS cameras showing the new hQE PMTs. VERITAS is normally operated using the full four- telescope array. A multi-level trigger system is employed EPJ Web of Conferences Figure 2. Photograph of one of the VERITAS cameras showing the new hQE PMTs (July 2012). Figure 3. The VERITAS sky map of detected sources as of May 2013, in Galactic coordinates. The shaded area shows the region of visibility for VERITAS above 50 degree elevation. The ma- jority of the detected sources oﬀthe Galactic plane are blazars (shown in red). Source Name Class Redshift (Extragalactic) z Mrk 421 HBL 0.030 Mrk 501 HBL 0.034 1ES 2344+514 HBL 0.044 1ES 1959+650 HBL 0.047 1ES 1727+501 HBL 0.055 BL Lacertae LBL 0.069 W Comae IBL 0.102 VER J0521+211 IBL/HBL 0.108 RGB J0710+591 HBL 0.125 H 1426+428 HBL 0.129 B2 1215+303 IBL/HBL 0.130? 1ES 0806+524 HBL 0.138 1ES 0229+200 HBL 0.140 1ES 1440+122 IBL/HBL 0.162 RX J0648.7+1516 HBL 0.179 1ES 1218+304 HBL 0.182 RBS 0413 HBL 0.190 1ES 1011+496 HBL 0.212 1ES 0414+009 HBL 0.287 1ES 0502+675 HBL 0.341? 1ES 0647+250 HBL ∼0.45 PG 1553+113 HBL 0.43 < z < 0.47 3C 66A IBL 0.444? PKS 1424+240 IBL/HBL > 0.6035 M87 FR I 0.0044 M82 Starburst 3.9 Mpc Table 1: Extragalactic sources of TeV gamma-ray emission detected VERITAS. Some redshifts are considered uncertain. appear in stellar-mass black holes in X-ray binaries a GRBs. They are scaled-down versions of the jets seen AGN. The observed similarity (in morphology and spe Figure 2. Photograph of one of the VERITAS cameras showing the new hQE PMTs (July 2012). Figure 3. The VERITAS sky map of detected sources as of May 2013, in Galactic coordinates. The shaded area shows the region of visibility for VERITAS above 50 degree elevation. The Innermost Regions of Relativistic Jets and Their Magnetic Fields 1ES 1218+304 (upper left) is an HBL, while 1ES 1215+303 (upper right) and W Comae are lower- frequency- peaked BL Lacs. The color scale (shown on the right) gives the signiﬁcance of detection. Note that the apparent exten- sions of the sources are due to saturation on the color scale; each blazar actually appears point-like at TeV energies. (Figure from [8]). Figure 4. Signiﬁcance map showing three blazars in a single VERITAS ﬁeld of view. 1ES 1218+304 (upper left) is an HBL, while 1ES 1215+303 (upper right) and W Comae are lower- frequency- peaked BL Lacs. The color scale (shown on the right) gives the signiﬁcance of detection. Note that the apparent exten- sions of the sources are due to saturation on the color scale; each blazar actually appears point-like at TeV energies. (Figure from [8]). Figure 4. Signiﬁcance map showing three blazars in a single VERITAS ﬁeld of view. 1ES 1218+304 (upper left) is an HBL, while 1ES 1215+303 (upper right) and W Comae are lower- frequency- peaked BL Lacs. The color scale (shown on the right) gives the signiﬁcance of detection. Note that the apparent exten- sions of the sources are due to saturation on the color scale; each blazar actually appears point-like at TeV energies. (Figure from [8]). BL Lacertae: A recent highlight from VERITAS is the detection of the eponymous blazar BL Lac during a very bright, short-lived ﬂare in 2011 June, the ﬁrst detection of minute-scale variability in a low-frequency peaked BL Lac. The source is at a redshift of 0.069 and VHE emis- sion from the source was discovered by MAGIC in 2005. VERITAS measured a ﬂux of ∼125% of the Crab Neb- ula ﬂux (> 200 GeV) during the ﬂare, and a gamma-ray spectral index of Γ = 3.8 ± 0.3. Although the rising edge of the ﬂare was not observed, during the short exposure of ∼35 min, the ﬂux was found to decay by factor of ∼3 in τ = 13 ± 4 min, placing strong constraints on the size of emission region (Figure 6a) [19]. Simultaneous changes in optical polarization, as well as the X-ray, optical and UV ﬂux were observed. The VHE ﬂare appeared during the same time as the emergence of a radio knot at 43 GHz resolved as moving downstream in the jet, and the emer- gence of the radio knot was linked to the gamma-ray ﬂare. The Innermost Regions of Relativistic Jets and Their Magnetic Fields The Innermost Regions of Relativistic Jets and Their Magnetic Fields Figure 4. Signiﬁcance map showing three blazars in a single VERITAS ﬁeld of view. 1ES 1218+304 (upper left) is an HBL, while 1ES 1215+303 (upper right) and W Comae are lower- frequency- peaked BL Lacs. The color scale (shown on the right) gives the signiﬁcance of detection. Note that the apparent exten- sions of the sources are due to saturation on the color scale; each blazar actually appears point-like at TeV energies. (Figure from [8]). sources. A large fraction of the sources are high frequency peaked objects (HBLs), in which the peaks in the spectral energy distribution (SED) lie in the X-ray and TeV bands. In fact, the TeV blazar population is largely dominated by high-frequency peaked BL Lacs, but also includes low fre- quency peaked objects, ﬂat spectrum radio quasars, and radio galaxies. In HBL objects, strongly correlated TeV and X-ray emission has been observed on many occasions, lending support to models in which the low-energy peak is explained by synchrotron emission from a population of ultra relativistic electrons, and the high-energy peak arises from inverse-Compton scattering of the synchrotron pho- tons by the same population of electrons. Given better sensitivities than earlier imaging atmospheric Cherenkov telescopes, low synchrotron peaked blazars are becoming more common in TeV catalogs and several such sources are on the VERITAS blazar list. Typically, such blazars are detected mostly during ﬂaring episodes/high ﬂux states. Figure 4 shows a sky map of three blazars detected by VERITAS in the same ﬁeld of view, illustrating the ca- pability of VERITAS to detect faint gamma-ray sources at only a few percent of the Crab Nebula ﬂux [8]. The three blazars shown in the ﬁgure are: W Comae, the ﬁrst IBL detected at TeV energies [9], the HBL 1ES 1218+304 [10] and B2 1215+30 [11]. The blazars detected by VERITAS are predominantly nearby, with the range of red shifts from 0.03 to at least 0.60 with the upper limit coming from the recently discovered redshift of PKS 1424+240 (0.6035) [12]. In the following we describe ﬁve particular results from VERITAS. A more complete review of the VERITAS blazar program may be found in some recent publications ([13], [14]). Figure 4. Signiﬁcance map showing three blazars in a single VERITAS ﬁeld of view. 2 Extragalactic TeV Gamma-ray Sources detected by VERITAS The majority of the extragalactic sources detected by VERITAS are active galaxies of the blazar class. Active galactic nuclei (AGN) are cores of galaxies believed to contain supermassive black holes (SMBHs) accreting mat- ter and generating non-thermal emission which often out- shines the host galaxy. Relativistic jets are extremely pow- erful outﬂows of collimated plasma that appear in active galactic nuclei. In addition to AGN, relativistic jets also Table 1 lists all the extragalactic sources detected by VERITAS (as of July 2013). It includes a radio galaxy (M87), and a starburst galaxy (M82), but the majority are active galaxies of the blazar class. Of the 26 sources shown in the table, ten were discovered by VERITAS as TeV 04011-p.2 The Innermost Regions of Relativistic Jets and Their Magnetic Fields MAGIC VERITAS 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Year 0 1 2 3 4 5 ΦHE ( 10−9 cm−2 s−1 ) Fermi-LAT 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 ΦX−ray (keV/s) Chandra - core 0 2 4 6 8 10 12 ΦX−ray (keV/s) Chandra - HST-1 0.0 0.2 0.4 0.6 0.8 ΦUV/O (mJy) HST STIS/F25QTZ - core HST ACS/F220W - core HST ACS/F250W - core HST STIS/F25QTZ - HST-1 HST ACS/F220W - HST-1 HST ACS/F250W - HST-1 LT - (core + HST-1) / 10 0.0 0.5 1.0 1.5 2.0 ΦRadio (Jy) VLBA 43 GHz - 1.2mas VLBA 43 GHz - peak VLBA 43 GHz - jet 1.0 1.5 2.0 2.5 3.0 3.5 ΦRadio (Jy) VLA 22GHz - core VLA 15GHz - core MOJAVE 15GHz - peak EVN 5GHz - core VLBA 2.3GHz - core VLBA 1.7GHz - core 52000 52500 53000 53500 54000 54500 55000 55500 MJD 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 ΦRadio (Jy) VLA 22GHz - HST-1 x 4.6 VLA 15GHz - HST-1 x 3.7 EVN 5GHz - HST-1 x 1.9 VLBA 2.3GHz - HST-1 x 1.2 VLBA 1.7GHz - HST-1 x 1.0 55275 55280 55285 55290 55295 55300 55305 55310 Time (MJD) -0.5 0 0.5 1 1.5 2 2.5 3 3.5 ΦVHE ( 10−11 cm−2 s−1 ) 2010 H.E.S.S. MAGIC VERITAS Fit: Φ0 e−\|t−t0\|/∆tr/d Figure 5. (a) (Top) Multiwavelength light curve of M 87 for a ten year period. The top panel shows the VHE gamma-ray ﬂux calculated above an energy threshold of 350 GeV, as detected by VERITAS, MAGIC and H.E.S.S. For details on the data analysis and references see [15]. (b) (Bottom) A bright VHE ﬂare in M 87 in 2010 recorded by VERITAS in a joint campaign with MAGIC and H.E.S.S., showing a variability timescale of ∼1 day. The solid curve is the results of the ﬁt of an exponential function to the data. (Both ﬁgures are from [15]). 0 10 20 30 40 Minutes 0 1 2 3 4 10-6 m-2 s-1 55500 55600 55700 55800 55900 Mo difie d Julia n D a te 0.0 0.5 1.0 1.5 Flux (10-6m-2s-1) E [GeV] -1 10 1 10 2 10 3 10 ] -1 s -2 [erg cm ! F ! -13 10 -12 10 -11 10 -10 10 -9 10 -8 10 VERITAS flare VERITAS pre-flare LAT simultaneous LAT 1-day LAT pre-flare Figure 6. The Innermost Regions of Relativistic Jets and Their Magnetic Fields (a) (Top) VERITAS lightcurve of BL Lac in June 2011 (> 200 GeV). The inset shows the day scale variability in BL Lac. For details see [19]. (b) (Bottom) Broadband gamma-ray spectrum of BL Lacertae. For VERITAS, the ﬂare spectrum with EBL correction is shown as blue points, with uncertainties, and the spectrum without EBL correction is shown as lighter blue points. The Fermi-LAT data shows hardening of the spectrum during the ﬂare. (Both ﬁgures are taken from [19]). ure 6. (a) (Top) VERITAS lightcurve of BL Lac in June 201 Figure 6. (a) (Top) VERITAS lightcurve of BL Lac in June 2011 (> 200 GeV). The inset shows the day scale variability in BL Lac. For details see [19]. (b) (Bottom) Broadband gamma-ray spectrum of BL Lacertae. For VERITAS, the ﬂare spectrum with EBL correction is shown as blue points, with uncertainties, and the spectrum without EBL correction is shown as lighter blue points. The Fermi-LAT data shows hardening of the spectrum during the ﬂare. (Both ﬁgures are taken from [19]). Figure 5. (a) (Top) Multiwavelength light curve of M 87 for a ten year period. The top panel shows the VHE gamma-ray ﬂux calculated above an energy threshold of 350 GeV, as detected by VERITAS, MAGIC and H.E.S.S. For details on the data analysis and references see [15]. (b) (Bottom) A bright VHE ﬂare in M 87 in 2010 recorded by VERITAS in a joint campaign with MAGIC and H.E.S.S., showing a variability timescale of ∼1 day. The solid curve is the results of the ﬁt of an exponential function to the data. (Both ﬁgures are from [15]). as VERITAS are good tools for identifying new blazars at low latitudes with the help of better localization, and higher sensitivity to ﬂux variability. Figure 7a shows a map, in Galactic coordinates, showing the arrival direc- tions of all diﬀuse class photon events with reconstructed energies greater than 30 GeV recorded by Fermi-LAT in the ﬁrst year of observations [21]. Observations of VER J0521+211 were motivated by the discovery of a cluster of > 30 GeV photons in these data. VER J0521+211 was discovered at TeV energies in 2009 and was detected again in 2012 during a ﬂare when its ﬂux was 10% of the Crab Nebula ﬂux. The Innermost Regions of Relativistic Jets and Their Magnetic Fields Such correlated multiwavelength studies during a gamma- ray ﬂare are important for locating the emission region in the blazar jet. Figure 6b shows the spectral energy distri- bution (SED) of BL Lac both during and prior to the ﬂare [19]. The TeV ﬂare occurred when the source was active and variable in GeV gamma rays as detected by the Fermi- LAT. The simultaneous VERITAS and LAT [20] spectra show that the gamma-ray SED peak lies in the ∼10 to Radio Galaxy M 87: Among non-blazar extragalactic sources, the giant radio galaxy M 87 provides an unique opportunity to study VHE emission in close proximity to a super-massive black hole. M 87 is one of only four known VHE AGN with weak-to-moderate beaming; the other three are the radio galaxies Cen A, IC 310, and NGC 1275. The source has exhibited three distinct VHE ﬂares with similarities in the VHE behaviour, but no unique multi-wavelength correlation. Figure 5a shows the mul- tiwavelength light curves of M 87 over ten years of obser- vations [15]. The exact location of the VHE gamma-ray- emitting region in M 87 remains elusive. 43 GHz VLBA radio observations of the inner jet regions indicate no in- creased ﬂux in 2010 [16], unlike the observations in 2008 where the VHE ﬂare coincided with an enhancement of the radio ﬂux of the innermost core regions. A ﬂare in M87 in 2010 was jointly observed at VHE energies by VERITAS, MAGIC [17] and H.E.S.S. [18] and triggered an exten- sive multiwavelength campaign at lower energies. Figure 5b shows results from the coordinated observation with VERITAS, H.E.S.S., and MAGIC during the bright VHE ﬂare in 2010 [15]. This TeV ﬂare was very well sampled, and the light curve shows a single isolated ﬂare character- ized by two-sided exponential decay τrise ∼1.7 days and τdecay ∼0.6 days. Such day-scale variability is rarely seen in weakly-beamed sources. 04011-p.3 EPJ Web of Conferences Time (MJD) -0.5 0 0.5 1 1.5 2 2.5 3 3.5 ΦVHE ( 10−11 cm−2 s−1 ) H.E.S.S. The Innermost Regions of Relativistic Jets and Their Magnetic Fields (a) (Top) VERITAS lightcurve of BL Lac in June 2011 (> 200 GeV). The inset shows the day scale variability in BL Lac. For details see [19]. (b) (Bottom) Broadband gamma-ray spectrum of BL Lacertae. For VERITAS, the ﬂare spectrum with EBL correction is shown as blue points, with uncertainties, and the spectrum without EBL correction is shown as lighter blue points. The Fermi-LAT data shows hardening of the spectrum during the ﬂare. (Both ﬁgures are taken from [19]). Time (MJD) -0.5 0 0.5 1 1.5 2 2.5 3 3.5 ΦVHE ( 10−11 cm−2 s−1 ) H.E.S.S. MAGIC VERITAS 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Year 0 1 2 3 4 5 ΦHE ( 10−9 cm−2 s−1 ) Fermi-LAT 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 ΦX−ray (keV/s) Chandra - core 0 2 4 6 8 10 12 ΦX−ray (keV/s) Chandra - HST-1 0.0 0.2 0.4 0.6 0.8 ΦUV/O (mJy) HST STIS/F25QTZ - core HST ACS/F220W - core HST ACS/F250W - core HST STIS/F25QTZ - HST-1 HST ACS/F220W - HST-1 HST ACS/F250W - HST-1 LT - (core + HST-1) / 10 0.0 0.5 1.0 1.5 2.0 ΦRadio (Jy) VLBA 43 GHz - 1.2mas VLBA 43 GHz - peak VLBA 43 GHz - jet 1.0 1.5 2.0 2.5 3.0 3.5 ΦRadio (Jy) VLA 22GHz - core VLA 15GHz - core MOJAVE 15GHz - peak EVN 5GHz - core VLBA 2.3GHz - core VLBA 1.7GHz - core 52000 52500 53000 53500 54000 54500 55000 55500 MJD 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 ΦRadio (Jy) VLA 22GHz - HST-1 x 4.6 VLA 15GHz - HST-1 x 3.7 EVN 5GHz - HST-1 x 1.9 VLBA 2.3GHz - HST-1 x 1.2 VLBA 1.7GHz - HST-1 x 1.0 0 10 20 30 40 Minutes 0 1 2 3 4 10-6 m-2 s-1 55500 55600 55700 55800 55900 Mo difie d Julia n D a te 0.0 0.5 1.0 1.5 Flux (10-6m-2s-1) E [GeV] -1 10 1 10 2 10 3 10 ] -1 s -2 [erg cm ! F ! -13 10 -12 10 -11 10 -10 10 -9 10 -8 10 VERITAS flare VERITAS pre-flare LAT simultaneous LAT 1-day LAT pre-flare 55275 55280 55285 55290 55295 55300 55305 55310 Time (MJD) -0.5 0 0.5 1 1.5 2 2.5 3 3.5 ΦVHE ( 10−11 cm−2 s−1 ) 2010 H.E.S.S. MAGIC VERITAS Fit: Φ0 e−\|t−t0\|/∆tr/d Figure 6. The Innermost Regions of Relativistic Jets and Their Magnetic Fields It was found to exhibit strong vari- ability from optical to TeV bands, with a peak ﬂux corre- sponding to ∼30% the steady Crab Nebula ﬂux. Recent optical spectroscopy indicates that the object is a typical BL Lac with redshift z = 0.108 [21]. Figure 7b shows the spectral energy distribution of VER J0521+211 during the VERITAS detection in 2009 [21]. The high-energy emission from VER J0521+211 peaks in the gamma-ray band, between 10 and 200 GeV. Simultaneous ﬂux mea- 100 GeV range. The LAT data show evidence for spectral hardening during the VERITAS ﬂare. VER J0521+211: The discovery of VER J0521+211 by VERITAS is an example of successfully ﬁnding new blazars behind the Galactic plane with the help of a Fermi- LAT trigger. A number of unidentiﬁed Fermi sources are expected to be blazars behind the Galactic plane. The characterization of blazars at low galactic latitudes is a diﬃcult task because of optical extinction and the diﬀuse Galactic gamma-ray background. VHE telescopes such 04011-p.4 The Innermost Regions of Relativistic Jets and Their Magnetic Fields IC 443 3C 66A RBS 0413 1ES 1218+304 VER J0521+211 Crab Vela X 1ES 1011+496 1ES 2344+514 RGB J0136+391 MAGIC J2001+435 1ES 0502+675 RX J0852.0-4622 PKS 0447-439 PKS 1424+240 PG 1553+113 PKS 2005-489 PKS 2155-304 Mrk 501 S5 0716+714 1ES 0033+595 1ES 1959+650 Mrk 421 / Hz) ν log ( 8 10 12 14 16 18 20 22 24 26 28 ) -1 s -2 / erg cm ν F ν log ( -15 -14 -13 -12 -11 -10 0.01 ± = 0.16 r α VERITAS Oct 22-30 VERITAS Nov 27 LAT Oct 22-Jan 16 XRT Oct 27-30 XRT Nov 27 MDM Oct 27 Archival data log (E / eV) -6 -4 -2 0 2 4 6 8 10 12 14 -15 -14 -13 -12 -11 -10 Figure 7. (a) (Top) Map in Galactic coordinates showing the arrival directions of all diﬀuse class photon events with zenith angle < 105◦recorded by Fermi-LAT in the ﬁrst year of obser- vations between 2008 Aug 4 and 2009 Aug 25, with energies greater than 30 GeV [21]. (b) (Bottom) Broadband spectral en- ergy distribution of VER J0521+211 during the VERITAS de- tection in 2009. The orange markers correspond to the ﬂare data. The solid blue curve represents a one-zone SSC emission model with parameters adjusted to describe the low-state data. Summary This article presents a brief summary of the VERITAS results on blazars from the past four years of observations. The population of VHE blazars detected by VERITAS is growing and VERITAS is expected to continue operations with the upgraded instrument for at least another ﬁve years with a strong focus on blazar observations. In addition, VERITAS has a coordinated eﬀort to collect multi-wavelength data on blazars. Long term monitoring of a sample of northern VHE blazars will help in the study of spectral and ﬂux variability. The observational goals of VERITAS are to catch fast ﬂares (to study the size of the emission region), measure the SEDs of the brightest blazars and increase sensitivity to weak sources (to study emission mechanisms in blazars), and detect new soft-spectrum sources that may be accessible with the upgraded camera (to increase the number count and carry out blazar population studies). Studies of relativis- tic jets in blazars will remain a high priority for VERITAS. Figure 7. (a) (Top) Map in Galactic coordinates showing the arrival directions of all diﬀuse class photon events with zenith angle < 105◦recorded by Fermi-LAT in the ﬁrst year of obser- vations between 2008 Aug 4 and 2009 Aug 25, with energies greater than 30 GeV [21]. (b) (Bottom) Broadband spectral en- ergy distribution of VER J0521+211 during the VERITAS de- tection in 2009. The orange markers correspond to the ﬂare data. The solid blue curve represents a one-zone SSC emission model with parameters adjusted to describe the low-state data. For de- tails on the broadband data and modeling results, see [21]. Both ﬁgures are taken from [21]. This research is supported by grants from the U.S. Department of Energy Oﬃce of Science, the U.S. National Science Foundation and the Smith- sonian Institution, by NSERC in Canada, by Science Foundation Ireland (SFI 10/RFP/AST2748) and by STFC in the U.K. We acknowledge the excellent work of the technical support staﬀat the Fred Lawrence Whip- ple Observatory and at the collaborating institutions in the construction and operation of the instrument. surements during a “low” state (2009 October) suggest that VER J0521+211 has synchrotron properties similar to those of the known IBLs, although the synchrotron com- ponent shows HBL-like properties during the X-ray and TeV ﬂare of 2009 November. Summary A simple one-zone lep- tonic relativistic jet model which calculates the emission as synchrotron self-Compton radiation indicates a particle- dominated jet with a bulk Lorentz factor of 30, requiring a relatively weak magnetic ﬁeld of ∼0.01 G. The Innermost Regions of Relativistic Jets and Their Magnetic Fields For de- tails on the broadband data and modeling results, see [21]. Both ﬁgures are taken from [21]. eral nights, with strong intra-night variability. Analysis of this data set is underway. 1ES 0647+250 and 1ES 1011+496: Both these blazars were detected by VERITAS after the camera upgrade in 2012 and during observations carried out in partial moon- light [13]. VERITAS observations conﬁrmed the prior dis- covery of these blazars as TeV sources by MAGIC. 1ES 0647+250 was a promising VHE target based on its hard spectrum seen in the Fermi-LAT data. VERITAS observa- tions yielded a detection at 2.9% of the Crab Nebula ﬂux (> 140 GeV). In the case of 1ES 1011+496, VERITAS observed a ﬂux ∼6.3% Crab Nebula ﬂux above 150 GeV. log (E / eV) / Hz) ν log ( 8 10 12 14 16 18 20 22 24 26 28 ) -1 s -2 / erg cm ν F ν log ( -15 -14 -13 -12 -11 -10 0.01 ± = 0.16 r α VERITAS Oct 22-30 VERITAS Nov 27 LAT Oct 22-Jan 16 XRT Oct 27-30 XRT Nov 27 MDM Oct 27 Archival data log (E / eV) -6 -4 -2 0 2 4 6 8 10 12 14 -15 -14 -13 -12 -11 -10 This research is supported by grants from the U.S. Department of Energy Oﬃce of Science, the U.S. National Science Foundation and the Smith- sonian Institution, by NSERC in Canada, by Science Foundation Ireland (SFI 10/RFP/AST2748) and by STFC in the U.K. We acknowledge the excellent work of the technical support staﬀat the Fred Lawrence Whip- ple Observatory and at the collaborating institutions in the construction and operation of the instrument. References [1] J. Holder, Proc. 32nd ICRC, Beijing, China (2011); arXiv:1111.1225. Mrk 421: The well-known blazar Mrk 421 was the ﬁrst VHE-detected extragalactic source and is typically the brightest blazar in the VHE band. The source is a sub- ject of the VERITAS long-term blazar monitoring program and major ﬂares were detected in 2008 and 2010 [22]. These initiated large multi-wavelength campaigns. Multi- ple outbursts have been observed to reach ∼10 times Crab Nebula ﬂux, allowing for studies with a high signal-to- noise ratio. In 2010 February VERITAS detected a strong, long-duration ﬂare at ∼8 Crab Nebula ﬂux units when the gamma-ray ﬂux was characterized in two-minute time bins. Recently in 2013 April, ﬂaring was detected in the source during a multi-wavelength campaign with NuSTAR [23] and Swift [24], and the TeV ﬂux was high over sev- [2] A. McCann et al., Astroparticle Physics, 32, 325 (2010). [3] A. N. Otte, Proc. 32nd ICRC, Beijing, China (2011); arXiv:1110.4702. [4] D. Kieda Proc. 33rd ICRC, Rio de Janeiro (2013); arXiv:1308.4849. [5] J. Holder et al., Astroparticle Physics, 25, 391 (2006). [6] T. Arlen & V. V. Vasiliev, 2012 Fermi Symposium proceed- ings, eConf C121028 (2012); arXiv:1303.2121. [7] N. Otte et al., White paper submitted to Snowmass 2013, CF6 Cosmic Particles and Fundamental Physics (2013); arXiv:1305.0264. [8] W. Benbow, Proc. 32nd ICRC, Beijing, China (2011); arXiv:1110.0038. [9] V. A. Acciari et al., ApJL, 684, 73 (2008). 04011-p.5 EPJ Web of Conferences EPJ Web of Conferences [10] V. A. Acciari et al., ApJL, 695, 1370 (2009). [17] MAGIC: https://magic.mpp.mpg.de/. [18] H.E.S.S.: http://www.mpi-hd.mpg.de/hfm/HESS/. [11] T. Arlen et al., ApJ submitted. [19] T. Arlen et al., ApJ, 762, 92 (2013). [12] A. Furniss et al., ApJL, 768, 31 (2013). [20] Fermi-LAT: http://www-glast.stanford.edu/. [13] J. Dumm, Proc. 33rd ICRC, Rio de Janeiro, Brazil (2013); arXiv:1308.0287. [13] J. Dumm, Proc. 33rd ICRC, Rio de Janeiro, Brazil (2013); arXiv:1308.0287. [21] S. Archambault et al., ApJ, 776, 69 (2013). [14] W. Benbow, Proc. 32nd ICRC, Beijing, China (2011); arXiv:1110.0038. [ [ [14] W. Benbow, Proc. 32nd ICRC, Beijing, China (2011); arXiv:1110.0038. [22] V. A. Acciari et al., ApJ, 738, 25 (2011). [23] NuSTAR: http://www.nustar.caltech.edu/. [15] A. Abramowski et al., ApJ, 746, 151 (2012). [24] Swift: http://swift.gsfc.nasa.gov/. [16] V. A. Acciari, et al., Science, 325, 444 (2009). Also, VLBA: http://www.vlba.nrao.edu/. [16] V. A. Acciari, et al., Science, 325, 444 (2009). Also, VLBA: http://www.vlba.nrao.edu/. 04011-p.6
https://openalex.org/W4225602995	https://epub.uni-bayreuth.de/5969/1/Adv%20Eng%20Mater%20-%202021%20-%20Ribeiro%20-%20Flexible%20and%20Porous%20Nonwoven%20SiCN%20Ceramic%20Material%20via%20Electrospinning%20of%20an%20Optimized.pdf	English	null	Flexible and Porous Nonwoven SiCN Ceramic Material via Electrospinning of an Optimized Silazane Solution	Advanced engineering materials	2,022	cc-by	4,949	Flexible and Porous Nonwoven SiCN Ceramic Material via Electrospinning of an Optimized Silazane Solution Luiz F. B. Ribeiro,* Ricardo S. Cunha, Agenor de Noni Jr, Ricardo A Günter Motz,* and Sergio Y. G. González* G. Motz The ORCID identiﬁcation number(s) for the author(s) of this article can be found under https://doi.org/10.1002/adem.202100321. © 2021 The Authors. Advanced Engineering Materials published by Wiley- VCH GmbH. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Despite several advantages, the literature on electrospun PDCs is still scarce, and the electrospinning requirements of the preceramic polymers are not fully understood yet. The main chal- lenge is to meet the needed properties for electrospinning.[23,24] 1. Introduction The processing of nonwoven porous ceramics by combining the polymer-derived ceramic (PDC) route with electrospinning offers an excellent strategy for developing new porous ceramic structures. Currently, the manufacturing of nonwoven porous materials from preceramic polymers is conducted by trial and error approaches. The necessity to predict the e-spinnability conditions from properties assessment offers a potential tool to control the manufacture and the resulting material morphology. This work assesses the relationship between the preceramic polymer solutions and the resulting electrospun nonwoven mor- phology. For this, a commercially available liquid oligosilazane (Durazane 1800) is selectively cross-linked to achieve a reliable and spinnable preceramic polymer (HTTS), which is then dissolved in tetrahydrofuran (THF). Based on the investigation of the rheological behavior of various polymer concentrations, three different polymer solution regimes (diluted, semidiluted, and concentrated) are identiﬁed and correlated with the resulting morphology of the e-spun material (spherical particles, beaded ﬁbers, and seamless ﬁbers). After the pyrolysis, the nonwoven ceramics manufactured from the solution with 65 wt% of HTTS is converted to a SiCN ceramic nonwoven with 86% of open porosity, proﬁling as a promising candidate for developing high-performance ﬁlter systems and catalytic supports in harsh environments. One of the essential advantages of the poly- mer-derived ceramics (PDC) technology is the applicability of well-established poly- mer processing techniques to manufacture ceramic components in shapes that would not be feasible via conventional powder ceramic techniques. Moreover, manipulat- ing the polymeric precursor chemistry and its molecular architecture allows designing ceramics with tailor-made properties.[1] The most common precursors (preceramic pol- ymers) are based on silicon-like siloxanes, carbosilanes, and silazanes, leading to the respective oxide or nonoxide ceramics. © 2021 The Authors. Advanced Engineering Materials published by Wiley-VCH GmbH www.aem-journal.com RESEARCH ARTICLE www.aem-journal.com 1. Introduction Due to the intrinsic properties of silicon-based nonoxide ceramics such as lightweight, high modulus, and excellent chemical and thermal stability, coupled with the versatility of being shaped in a wide variety of forms, PDCs can be poten- tially applied in several key areas related to energy, environment, transport, and defense.[1,2] Like other materials, by increasing the surface area of PDC compo- nce is improved for several applications, Due to the intrinsic properties of silicon-based nonoxide ceramics such as lightweight, high modulus, and excellent chemical and thermal stability, coupled with the versatility of being shaped in a wide variety of forms, PDCs can be poten- tially applied in several key areas related to energy, environment, transport, and defense.[1,2] Like other materials, by increasing the surface area of PDC compo- nents, its performance is improved for several applications, especially those related to catalytic supports and ﬁltration. Suitable for such applications are especially cellular materials (synergy of high surface area and the inherent properties of the ceramics)[3] and structures based on nanoﬁbers, e.g., proc- essed by combining the PDC route with electrospinning.[4] e-spun material he pyrolysis, the wt% of HTTS is ty, proﬁling as a ms and catalytic L. F. B. Ribeiro Graduate Program in Energy and Sustainability Federal University of Santa Catarina Araranguá 88906-072, Brazil E-mail: luiz.ribeiro@ufsc.br R. S. Cunha, A. de Noni, Jr, R. A. F. Machado, S. Y. G. González Graduate Program in Chemical Engineering Federal University of Santa Catarina Florianópolis 88010-970, Brazil E-mail: sergio.gomez@ufsc.br G. Motz Department of Ceramic Materials Engineering University of Bayreuth Bayreuth D-95447, Germany E-mail: guenter.motz@uni-bayreuth.de The obtained 1D nanostructure arranged in the form of mats possesses a large surface-to-volume ratio, which, combined with the aforementioned intrinsic properties of PDCs, makes these components excellent candidates for high-performance ﬁltering systems and also as catalytic sup- ports in harsh environments.[5–9] In addition, the ceramic mats can be modiﬁed or functionalized to extend their appli- cation in areas such as electrochemical devices.[10–16] Finally, the use of electrospun PDCs, in combination with different nanostructures, has been explored as promising candidates for electromagnetic wave absorption, as seen in recent literature.[17–22] © 2021 The Authors. Advanced Engineering Materials published by Wiley-VCH GmbH 2.2. Rheological Investigation of the Preceramic Polymer Solutions Commercially available preceramic polymers usually are com- posed of cyclic and linear units with a complex structure and low molecular weight, which is a drawback for the spinnability.[4] Therefore, to date, it is mandatory to use organic polymers as spinning aid and additional intermediated steps to achieve e-spinnability. However, suitable organic polymers have a mea- ger carbon yield after pyrolysis and act as sacriﬁcial ﬁller, which harms the mechanical stability. Therefore, to date, it is mandatory to use organic polymers as spinning aid and additional intermediated steps to achieve e-spinnability. However, suitable organic polymers have a mea- ger carbon yield after pyrolysis and act as sacriﬁcial ﬁller, which harms the mechanical stability. The investigation of the rheological behavior of the HTTS solu- tions provides important information about the interactions between the polymer chains in solution, which is crucial to understand and predict the polymer morphology required for e-spinning.[23,24,26] However, although this knowledge is well established for the e-spinning of organic polymers, there is no report in the literature about the rheological characterization of preceramic polymers intended for e-spinning. To avoid using an organic polymer as e-spinning additive, we used a solid polysilazane synthesized from a commercially avail- able oligosilazane[25] to achieve e-spinning conditions. Suitable e-spinning solutions had to be developed to ensure a continuous e-spinning process and the processing of regular-shaped ﬁbers. Furthermore, it was necessary to investigate the polymer to ceramic transformation by pyrolysis. This work aimed to develop an approach to establish the e-spinning process for the process- ing of nonoxide ceramic nonwoven with high interconnectivity of open porosity for applications in harsh environments, e.g., in the ﬁeld of energy, catalysis, and high-performance ﬁltration systems. The synthesized HTTS preceramic polymer was solubilized in THF with different concentrations (Table 1) to evaluate the rhe- ological behavior. The typical apparent viscosity of the prepared solutions as a function of shear rate is shown in Figure 2. tions as a function of shear rate is shown in Figure The analysis of the curves shown in Figure 2 clearly distin- guishes between two different groups regarding the rheological behavior. Samples with less than 50 wt% of the polymer have a predominantly Newtonian behavior. Therefore, increases in the shear rate have no signiﬁcant effect on the viscosity. A pro- nounced shear-thinning behavior appears for samples with higher polymer concentrations (>60 wt%). www.advancedsciencenews.com www.advancedsciencenews.com www.advancedsciencenews.com www.aem-journal.com DOI: 10.1002/adem.202100321 Adv. Eng. Mater. 2022, 24, 2100321 2100321 (1 of 7) www.aem-journal.com 2.2. Rheological Investigation of the Preceramic Polymer Solutions This typical viscoelas- tic behavior is attributed to the polymer chain alignment in the applied shear stress direction and consequently reducing the viscosity of the respective solutions with higher shear rates through disentanglement. 2.3. Manufacturing of Preceramic Polymer Fibers by e-Spinning The e-spinning process with the HTTS solutions was performed as described in the experimental procedure. The respective spin- ning results are shown in Figure 4. Figure 4 clearly shows the strong inﬂuence of the HTTS con- centration, i.e., the solution viscosity on the morphology of the resulting materials. As predicted by the rheological behavior, the e-spinning of the diluted solutions ranging from 30 to 50 wt% of HTTS yields only near-spherical polymer particles, as the Taylor cone formation was not possible. The Taylor cone formation is a precondition for the processing of regular-shaped ﬁbers via the e-spinning process.[28,29] Thus, the polymer solutions were directly ejected from the needle tip in the form of a spray jet. For this interval of concentration, the term e-spraying should be more appropriate to describe this processing scenario. In summary, for diluted solutions, the lack of chain entanglements avoids the drawn jet continuity and leads only to the deposition of small droplets at the collector. Figure 2. Dependency of the viscosity from the shear rate for HTTS sol- utions with different concentrations. Particles with a mean diameter size of 7.8 2.6, 8.4 2.0, and 9.4 1.4 μm were obtained from samples HTTS_1, HTTS_2, and HTTS_3, respectively. It is worth mentioning that the increase in polymer concentration also leads to the formation of more homogeneous particles with a narrower size (Figure 4 and 5). However, in addition to particles, the formation of small ﬁbrils was observed for sample HTTS_3, which is in agreement with the rheological results and conﬁrms the begin- ning of the transition to a semidiluted regime. polymer chains gradually start to interact through chain entanglement by increasing the polymer concentration. This change from diluted to semidiluted state allows the formation of ﬁbers, instead of beaded ﬁbers or even droplets during electrospinning.[23,27] As the polymer concentration is increased to more than 65 wt%, a second transition point occurs, characterized by a remarkable increase in the viscosity. This phenomenon is better visualized by plotting the mean viscosities at the plateau regime (η∞) as a function of the polymer concentration (Figure 3). The increase in the polymer concentration to 60 wt% (sample HTTS_4) leads to the formation of beaded ﬁbers. 2.3. Manufacturing of Preceramic Polymer Fibers by e-Spinning This morphol- ogy results from two types of instability during the e-spinning process, the Rayleigh and varicose instabilities,[30–33] which favor spheres formation because of the smallest surface area to dimin- ish the surface tension. In the sample with 65 wt% HTTS concen- tration (HTTS_5), the viscosity seems to have inhibited beads-on- string ﬁber formation, leading to seamless ﬁbers. Increasing the HTTS concentration in the solution to 70 wt% (HTTS_6), the high viscosity combined with the low solvent concentration leads to nonhomogeneous tape-like morphology with a broader size dis- tribution, here named as ﬂat-scattered ﬁbers. For better visualiza- tion, the measured average ﬁber diameters are shown in Figure 5. As will be further discussed in the e-spinning experiments section, this transition leads to seamless ﬁbers, as an increase in the solution concentration leads to a higher degree of polymer chain entanglements.[26,27] Figure 3. Dependence of viscosity of HTTS solutions with different concentrations at steady state. 2.1. Synthesis of HTTS As the Durazane 1800 used in this work, the most commercially available silazanes are liquid oligomers. In view thereof, for the processing of ﬁbers, a solid and processable polymeric precursor must be achieved. For this reason, we applied the selective cross- linking procedure used by Flores et al.[25] The simpliﬁed reaction mechanisms and their physical state before and after the cross-linking reaction are shown in Figure 1. Furthermore, the results indicate that the solution state switches from a diluted to a semidiluted regime in the range of 50–60 wt% of HTTS concentration, indicating that the Table 1. Denomination of different HTTS solutions in THF used for the following investigations. The main advantage of using selective cross-linking via dehydrocoupling reaction is to tailor the silazane molecular weight. The reaction is slow and can be easily stopped at any time using an appropriate inhibitor, such as calcium boro- hydride bis(tetrahydrofuran) (Ca(BH4)2.2THF). The resulting polysilazane HTTS has an increased molecular weight (Mw) of 8010 g mol1 in comparison with 4836 g mol1 for Durazane 1800 with a polydisperse index (Mw/Mn) of 9.4.[25] In view thereof, the previously liquid and oligomer Durazane 1800 becomes a soluble thermoplastic polymer with properties suit- able for the e-spinning process. Figure 1. Synthesis of the HTTS polysilazane from the oligosilazane Durazane 1800 by selective cross-linking. Reaction mechanism details are described elsewhere.[25] Figure 1. Synthesis of the HTTS polysilazane from the oligosilazane Durazane 1800 by selective cross-linking. Reaction elsewhere.[25] Figure 1. Synthesis of the HTTS polysilazane from the oligosilazane Durazane 1800 by selective cross-linking. Reaction mechanism details are described elsewhere.[25] © 2021 The Authors. Advanced Engineering Materials published by Wiley-VCH GmbH Adv. Eng. Mater. 2022, 24, 2100321 2100321 (2 of 7) Adv. Eng. Mater. 2022, 24, 2100321 www.aem-journal.com www.aem-journal.com d d i www.aem-journal.com Figure 2. Dependency of the viscosity from the shear rate for HTTS sol- utions with different concentrations. © 2021 The Authors. Advanced Engineering Materials published by Wiley-VCH GmbH © 2021 The Authors. Advanced Engineering Materials published by Wiley-VCH GmbH 2.4. Investigation of the Pyrolysis Behavior and NonWoven Ceramic Formation The pyrolysis of organosilazanes up to 1000 C in N2 leads to the formation of an amorphous ceramic composed of a SiCxN4x (0 ≤x ≤4) matrix and a segregated carbon phase (so-called free carbon).[34–37] For ceramic ﬁbers processed through pyrolysis, the crosslinking step is crucial to convert the thermoplastic pre- ceramic polymer to a thermoset, avoiding shape changes during the ceramization.[38] For this purpose, considering the availability of vinyl groups in HTTS, a radical initiator (AIBN) was used to perform the cross-linking step. Figure 6 shows the resulting ﬁbers pyrolyzed as described in the experimental procedure. After pyrolysis, defect-free cylindrical SiCN ﬁbers with a diameter <4 μm were obtained, which is far below the reported diameter for the same ceramic ﬁbers processed by melt spinning Figure 3. Dependence of viscosity of HTTS solutions with different concentrations at steady state. Adv. Eng. Mater. 2022, 24, 2100321 2100321 (3 2100321 (3 of 7) Adv. Eng. Mater. 2022, 24, 2100321 www.aem-journal.com www.advancedsciencenews.com Figure 4. Scanning electron microscopy (SEM) micrographs of the formed particles and ﬁbers during e-spinning in dependency on the HTTS solution concentration. microscopy (SEM) micrographs of the formed particles and ﬁbers during e-spinning in dependency on the HTTS solutio Figure 4. Scanning electron microscopy (SEM) micrographs of the formed particles and ﬁbers during e-spinning in dependency on the HTTS solution concentration. (a) (b) Figure 5. Dependence of polymer morphology and average diameter on the a) polymer concentration and b) viscosity. (a) (b) (a) gure 5. Dependence of polymer morphology and average diameter on the a) polymer concentration and b) viscosity. Figure 6. SEM micrographs of HTTS_5-derived e-spun ﬁbers pyrolyzed at 1000 C in N2 atmosphere. range of micrometers, their quality in terms of homogeneity and morphology is remarkable, as no additional organic polymer (spinning aid) was used. The decomposition gases derived from organic additives would evolve during pyrolysis, which could lead to defects and a higher shrinkage, both very deleterious for the ﬁber properties. Further optimization of the e-spinning process should lead to ﬁbers with a diameter in the range of nanometers. g Another essential characteristic of the PDCs technology is the signiﬁcant shrinkage during the pyrolysis, mainly assigned by the remarkable increase in density and the mass loss during the conversion from a polymer to ceramics. The ﬁber average diameter and the nonwoven dimensions were used to estimate the shrinkage for sample HTTS_5 during pyrolysis (Figure 7). 2.4. Investigation of the Pyrolysis Behavior and NonWoven Ceramic Formation Comparing the ﬁber diameter before and after pyrolysis, a shrinkage of 22% is noticed, as the ﬁber mean diameter decreases from 4.1 μm (before pyrolysis) to 3.2 μm (after pyroly- sis). The shrinkage is one of the drawbacks of the PDCs technol- ogy, and this phenomenon must be considered for further product designing. Regarding the ceramic nonwoven porosity, Another essential characteristic of the PDCs technology is the signiﬁcant shrinkage during the pyrolysis, mainly assigned by the remarkable increase in density and the mass loss during the conversion from a polymer to ceramics. The ﬁber average diameter and the nonwoven dimensions were used to estimate the shrinkage for sample HTTS_5 during pyrolysis (Figure 7). Comparing the ﬁber diameter before and after pyrolysis, a shrinkage of 22% is noticed, as the ﬁber mean diameter decreases from 4.1 μm (before pyrolysis) to 3.2 μm (after pyroly- sis). The shrinkage is one of the drawbacks of the PDCs technol- ogy, and this phenomenon must be considered for further product designing. Regarding the ceramic nonwoven porosity, Figure 6. SEM micrographs of HTTS_5-derived e-spun ﬁbers pyrolyzed at 1000 C in N2 atmosphere. (30–100 μm)[25] and also lower than commercially available SiC ﬁbers (10 μm).[39] Although the fabricated ﬁbers are still in the 2100321 (4 of 7) Adv. Eng. Mater. 2022, 24, 2100321 www.aem-journal.com www.advancedsciencenews.com (a) (b) (c) (d) Figure 7. SEM micrographs of HTTS_5-derived e-spun ﬁbers a) before and b) after pyrolysis. c) Histogram of polymer and ceramic ﬁber diameter distribution, and d) nonwoven porous material comparison before and after pyrolysis. (b) (a) (d) (c) Figure 7. SEM micrographs of HTTS_5-derived e-spun ﬁbers a) before and b) after pyrolysis. c) Histogram of polymer and ceramic ﬁber diameter distribution, and d) nonwoven porous material comparison before and after pyrolysis. Figure 8. Flexibility demonstration of HTTS_5-derived e-spun nonwoven after pyrolysis. organic polymers as spinning aid to enable processing via elec- trospinning. However, such organic additives act as sacriﬁcial ﬁller, which leads to a decreased ceramic yield, higher shrinkage, and the formation of pores within the formed ceramic ﬁbers during pyrolysis and results in inferior mechanical properties. The approach developed here overcomes these problems. In this work, solutions of the polysilazane HTTS in THF, syn- thesized by selective cross-linking of the oligosilazane Durazane 1800, were specially designed and comprehensively investigated for the electrospinning process. 2.4. Investigation of the Pyrolysis Behavior and NonWoven Ceramic Formation The rheological measurements of different feedstock solutions were correlated with the resulting ﬁber morphology after electrospinning, to predict the morphol- ogy of the e-spun preceramic polymers. For diluted solutions with a polymer concentration lower than 50 wt%, only near- spherical particles with diameters in the range of 7–10 μm were obtained, as no interactions of the polymer chains are present. By increasing the polymer concentration up to 60 wt%, the HTTS solution rheology reveals a shear-thinning behavior, indicating the transition to a semidiluted state due to a certain degree of chain entanglements. Fibers with beads-on-string morphology are typical for this solution regime. For concentrations higher than 65 wt%, the substantial increase in the viscosity and a vis- coelastic behavior due to a higher degree of chain entanglements lead to seamless ﬁbers during the electrospinning process. Figure 8. Flexibility demonstration of HTTS_5-derived e-spun nonwoven after pyrolysis. the percentage of total porosity was estimated at approximately 86%. The calculation was performed as described in the experi- mental section by considering the measured SiCN ceramic true density (ρT ¼ 2.416 g cm3) and the calculated nonwoven density (ρA ¼ 0.347 g cm3). Finally, it is noteworthy to mention that because of the thin ceramic ﬁbers the porous nonwoven ceramic is ﬂexible and presents sufﬁcient mechanical strength to be bent and handled without breaking or losing its structural integrity (Figure 8). The good mechanical properties, the well-known ther- mal and chemical stability, and the good shapeability open many promising applications, such as in ﬁltering systems, catalyst supports, and porous burners. The pyrolysis behavior was investigated for samples with a concentration of 65 wt% of HTTS containing AIBN as curing agent. The pyrolysis led to defect-free ceramic SiCN ﬁbers with a diameter <4 μm. Due to the resulting thin diameter, the porous ceramic nonwoven possesses excellent ﬂexibility. Finally, the ceramic nonwoven porous structure indicates a total porosity of about 86%. © 2021 The Authors. Advanced Engineering Materials published by Wiley-VCH GmbH Keywords 2014, 131. Fiber Morphology Investigation: The e-spun samples were analyzed to assess the ﬁber morphology before and after the pyrolysis by scanning electron microscopy TM3030 (Hitachi, Japan), operating at an acceleration voltage of 15 kV. The micrographs obtained were analyzed by the ImageJ software (National Institutes of Health, USA), and 200 measurements for either ﬁbers or particles were used to estimate the reported mean values. [10] A. Tolosa, M. Widmaier, B. Krüner, J. M. Grifﬁn, V. Presser, Sustain. Energy Fuels 2018, 2, 215. [11] S. A. Smith, B. P. Williams, Y. L. Joo, J. Membr. Sci. 2017, 526, 315. [12] S. A. Smith, J. H. Park, B. P. Williams, Y. L. Joo, J. Mater. Sci. 2017, 52, 3657. Investigations of the NonWoven Properties: The ceramic nonwoven porosity (ε) was estimated by using Equation 1 [13] Z. Sang, X. Yan, L. Wen, D. Su, Z. Zhao, Y. Liu, H. Ji, J. Liang, S. X. Dou, Energy Storage Mater. 2020, 25, 876. [14] Q. Chen, D. Jia, B. Liang, Z. Yang, Y. Zhou, D. Li, R. Riedel, T. Zhang, C. Gao, Ceram. Int. 2021, 47, 10958. ε ¼ 1 ρA ρT (1) C. Gao, Ceram. Int. 2021, 47, 10958. ε ¼ 1 ρA ρT (1) [15] S. Bin Mujib, R. Cuccato, S. Mukherjee, G. Franchin, P. Colombo, G. Singh, Ceram. Int. 2020, 46, 3565. [16] M. Ma, H. Wang, X. Li, K. Peng, L. Xiong, X. Du, J. Eur. Ceram. Soc. 2020, 40, 5238. where ρA is the nonwoven apparent density and ρT is the true density of the derived SiCN material. For the apparent density, the volume was obtained by measuring the nonwoven dimensions, while for the true density, the ceramic volume was determined with a gas pycnometer (Accu Pyc II 1340 – Micromeritics Instruments Corporation, USA). where ρA is the nonwoven apparent density and ρT is the true density of the derived SiCN material. For the apparent density, the volume was obtained by measuring the nonwoven dimensions, while for the true density, the ceramic volume was determined with a gas pycnometer (Accu Pyc II 1340 – Micromeritics Instruments Corporation, USA). [17] Y. Huo, Y. Tan, K. Zhao, Z. Lu, L. Zhong, Y. Tang, Chem. Phys. Lett. 2021, 763, 138230. [17] Y. Huo, Y. Tan, K. Zhao, Z. Lu, L. Zhong, 2021, 763, 138230. [18] F. Xiao, H. Sun, J. Li, X. Guo, H. Zhang, J. Lu, Z. Keywords electrospinning, nonwoven ceramics, polymer-derived ceramics, porous ceramics, SiCN, silazanes Received: March 17, 2021 Revised: June 8, 2021 Published online: July 22, 2021 Revised: June 8, 2021 E-Spinning Experiments and Pyrolysis: Solutions varying from 30 to 70 wt% of HTTS in THF were prepared for the electrospinning experiments. The respective solution was placed in a 5 mL syringe coupled with a metallic needle of 0.7 mm inner diameter clamped to a high-voltage power supply applying 10 kV and ﬁxed at 10 cm from an aluminum collector. The sample was extruded at a ﬂow rate of 1 mL h1 controlled by an infusion syringe pump arranged horizontally. [1] P. Colombo, G. Mera, R. Riedel, G. D. Sorarù, J. Am. Ceram. Soc. 2010, 93, 1805. [2] G. Mera, M. Gallei, S. Bernard, E. Ionescu, Nanomaterials 2015, 5, 468. [2] G. Mera, M. Gallei, S. Bernard, E. Ionescu, Nanomaterials 2015, 5, 468. [3] C. Vakifahmetoglu, D. Zeydanli, P. Colombo, Mater. Sci. Eng. R Rep. 2016, 106, 1. For the pyrolysis, 1.5 wt% of the curing agent AIBN, referred to HTTS amount, was added to the spinning solution. Then, the resulting polymer nonwoven was cured at 70 C for 4 h to enable cross-linking before melt- ing. Afterward, the sample was placed in a carbon crucible and heated up to 1000 C (2 K min1) under nitrogen ﬂow in a tubular furnace. [4] A. Viard, P. Miele, S. Bernard, J. Ceram. Soc. Jpn. 2016, 124, 967. [5] S. Xie, Y. Wang, Y. Lei, B. Wang, N. Wu, Y. Gou, D. Fang, RSC Adv. 2015, 5, 64911. [6] B. Wang, Y. Wang, Y. Lei, N. Wu, Y. Gou, C. Han, Mater. Manuf. Process. 2016, 31, 1357. Rheological Characterization: Parallel plate rotational rheology of the polymer solutions was performed in a Thermo Scientiﬁc Haake Mars II rheometer (Thermo Fisher Scientiﬁc, USA) using a PP60 measuring geometry (60 mm diameter). A 0.15 gap was used in the tests, and the shear rate varied from 0.1 to 250 s1 in 60 s. All measurements were performed at a temperature of 23 1C. [7] W. Nan, W. Lynn Yuqin, W. Ying-De, F. KO, J. Inorg. Mater. 2018, 33, 357. [8] S. H. Koo, S. G. Lee, H. Bong, Y.-J. Kwark, K. Cho, H. S. Lim, J. H. Cho, Polymer (Guildf). 2014, 55, 2661. [9] A. Guo, M. Roso, M. Modesti, J. Liu, P. Colombo, J. Appl. Polym. Sci. Acknowledgements ceramics should lead to specially tailored morphology, structure, and properties. The results achieved open promising PDC tech- nology applications as ﬁlter systems and catalytic supports intended for harsh environments. The authors like to thank CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) within the project PROBRAL (88881.198816/2018-01) and DAAD (German Academic Exchange Service, Project-ID 57446974) for ﬁnancial support. , j ) pp Open access funding enabled and organized by Projekt DE j ) Open access funding enabled and organized by Projekt DEAL. www.advancedsciencenews.com www.aem-journal.com Conﬂict of Interest Materials: The commercially available oligosilazane Durazane 1800 (Merck KGaA, Germany) was selected as the SiCN precursor. Tetra-n- butylammonium ﬂuoride (TBAF) (1 M in THF) was used as the catalyst, while calcium borohydride bis(tetrahydrofuran) (Ca(BH4)2·2THF) was used as an inhibitor to stop the reaction, both purchased from Sigma Aldrich (Germany). THF p.a. (Neon, Brazil) was used for the synthesis and the e-spinning experiments, and the initiator 2,2 0-azobisisobutyroni- trile (AIBN), used for curing, was supplied by Akzo Nobel (The Netherlands). The educts’ preparation and handling and all synthesis reactions were conducted in a dry argon atmosphere. The authors declare no conﬂict of interest. Data Availability Statement The data that support the ﬁndings of this study are available from the corresponding author upon reasonable request. Selective Cross-Linking of the Oligosilazane: The procedure adopted to perform the selective chemical cross-linking of the oligosilazane Durazane 1800 based on a previous work reported by Flores et al.[25] In a typical reaction, 20 g of Durazane 1800 was dissolved under vigorous stirring in 40 g of THF, and 0.25 wt% of the catalyst TBAF, concerning Durazane 1800, was added dropwise to the reaction. The solution was stirred for 90 min, and an excess of the inhibitor (Ca(BH4)2·2THF) was added to stop the reaction. The mixture was stirred for another 5 min and afterward ﬁltered off. Finally, the solvent was removed under reduced pressure at room temperature, obtaining a solid and colorless polysila- zane, hereafter called HTTS. 3. Conclusions The developed approach enables the electrospinning of precer- amic polymers without e-spinning aids and helps to conduct this process methodically. Further improvements of the electrospin- ning process combined with the functionalization of the Many preceramic polymers have a much lower molecular weight compared with organic polymers. Therefore, it is crucial to add 2100321 (5 of 7) Adv. Eng. Mater. 2022, 24, 2100321 www.aem-journal.com www.advancedsciencenews.com © 2021 The Authors. Advanced Engineering Materials published by Wiley-VCH GmbH Keywords Pan, J. Xu, Ceram. Int. 2020, 46, 12773. 2100321 (6 of 7) Adv. Eng. Mater. 2022, 24, 2100321 2100321 (6 of 7) © 2021 The Authors. Advanced Engineering Materials published by Wiley-VCH GmbH Adv. Eng. Mater. 2022, 24, 2100321 www.advancedsciencenews.com www.aem-journal.com www.aem-journal.com [30] M. M. Hohman, M. Shin, G. Rutledge, M. P. Brenner, Phys. Fluids 2001, 13, 2201. [19] X. Guo, F. Xiao, J. Li, H. Zhang, Q. Hu, G. Li, H. Sun, Ceram. Int. 2021, 47, 1184. [31] W. Zuo, M. Zhu, W. Yang, H. Yu, Y. Chen, Y. Zhang, Polym. Eng. Sci. 2005, 45, 704. [20] Y. Zhang, Y. Zhao, Q. Chen, Y. Hou, Q. Zhang, L. Cheng, L. Zheng, Ceram. Int. 2021, 47, 8123. [21] Y. Huo, K. Zhao, Z. Xu, Y. Tang, J. Alloys Compd. 2020, 815, 152458. [32] J. H. Yu, S. V. Fridrikh, G. C. Rutledge, Polymer (Guildf). 2006, 47, 4789. o, Z. Xu, Y. Tang, J. Alloys Compd. 2020, 815, 152458. [32] J. H. Yu, S. V. Fridrikh, G. C. Rutledge, Polymer (Guildf). 2006, 47, 4789. [22] Y. Hou, L. Cheng, Y. Zhang, X. Du, Y. Zhao, Z. Yang, Chem. Eng. J. 2021, 404, 126521. [33] H. Zhao, H. Chi, in Nov. Asp. Nanoﬁbers, (Eds: T. Lin) IntechOpen, London, England 2018, p. 13, Ch. 5. [34] S. Traßl, D. Suttor, G. Motz, E. Rössler, G. Ziegler, J. Eur. Ceram. Soc. 2000, 20, 215. [23] M. G. McKee, G. L. Wilkes, R. H. Colby, T. E. Long, Macromolecules 2004, 37, 1760. [24] R. Rošic, J. Pelipenko, P. Kocbek, S. Baumgartner, M. Bešter-Rogaˇc, J. Kristl, Eur. Polym. J. 2012, 48, 1374. [35] G. Ziegler, H.-J. Kleebe, G. Motz, H. Müller, S. Traßl, W. Weibelzahl, Mater. Chem. Phys. 1999, 61, 55. [36] J. Seitz, J. Bill, N. Eggerb, F. Aldinger, J. Eur. Ceram. Soc. 1996, 16, 885. [25] O. Flores, T. Schmalz, W. Krenkel, L. Heymann, G. Motz, J. Mater. Chem. A 2013, 1, 15406. , , [26] S. K. Tiwari, S. S. Venkatraman, Mater. Sci. Eng. C 2012, 32, 1037. Tiwari, S. S. Venkatraman, Mater. Sci. Eng. C 2012, 32, 103 . Gérardin, F. Taulelle, D. Bahloul, J. Mater. Chem. 1997, 7 [38] G. Motz, S. Schmidt, S. Beyer, in Ceramic Matrix Composites. (Ed: Walter Krenkel), Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany 2008, p. 165. [27] S. L. Shenoy, W. D. Bates, H. L. Frisch, G. E. Wnek, Polymer (Guildf). 2005, 46, 3372. [28] T. Subbiah, G. S. Bhat, R. W. Tock, S. Parameswaran, S. S. Ramkumar, J. Appl. Polym. Sci. 2005, 96, 557. [39] O. Flores, R. K. Bordia, D. Nestler, W. Krenkel, G. Motz, Adv. Eng. Mater. 2014, 16, 621. [29] G. © 2021 The Authors. Advanced Engineering Materials published by Wiley-VCH GmbH www.aem-journal.com Taylor, Proc. R. Soc. London. Ser. A Math. Phys. Sci. 1964, 280, 383. Adv. Eng. Mater. 2022, 24, 2100321 2100321 (7 of 7) 2100321 (7 of 7) Adv. Eng. Mater. 2022, 24, 2100321 © 2021 The Authors. Advanced Engineering Materials published by Wiley-VCH GmbH
https://openalex.org/W3208228716	https://www.bio-conferences.org/articles/bioconf/pdf/2021/10/bioconf_napd2021_00098.pdf	English	null	Some biochemical characteristics of species of the genus Monarda	Bio web of conferences/BIO web of conferences	2,021	cc-by	1,438	Some biochemical characteristics of species of the genus Monarda Anastasia Petruk1, Irina Lobanova1, and Elisaveta Bolotnik2 1Central Siberian Botanical Garden, SB RAS, 630090 Novosibirsk, Russia 2Botanical Garden, UB RAS, 620144 Ekaterinburg, Russia Anastasia Petruk1, Irina Lobanova1, and Elisaveta Bolotnik2 1Central Siberian Botanical Garden, SB RAS, 630090 Novosibirsk, Ru 2 Anastasia Petruk , Irina Lobanova , and Elisaveta Bolotnik 1Central Siberian Botanical Garden, SB RAS, 630090 Novosibirsk, Russia 2Botanical Garden, UB RAS, 620144 Ekaterinburg, Russia 1Central Siberian Botanical Garden, SB RAS, 630090 Novosibirsk, Russia 2Botanical Garden, UB RAS, 620144 Ekaterinburg, Russia Abstract. The paper presents a comparative study of the content and composition of biologically active substances of the aboveground parts of plants of some Monarda species cultivated in the Botanical Garden, Ural Branch of Russian Academy of Sciences, Ekaterinburg. The greatest quantity of flavones, tannins were found in M. citriodora Cerv. ex Lag., carotenoids, pectins, saponins - in M. fistulosa L., pectins - in M. didyma L., protopectins - in M. media Will. The component composition of the phenolic complex of the samples was studied by high-performance liquid chromatography on an analytical HPLC-system Agilent 1200. 70% ethanol extracts of plants of Monarda species contained at least 35 compounds of phenolic nature, including flavones luteolin, luteolin 7-glucoside, chlorogenic acid. *Correspondingauthor:pet.a@mail.ru BIO Web of Conferences 38, 00098 (2021) Northern Asia Plant Diversity 2021 BIO Web of Conferences 38, 00098 (2021) Northern Asia Plant Diversity 2021 https://doi.org/10.1051/bioconf/20213800098 © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). 1 Introduction The genus Monarda L. from the Lamiaceae family has up to 100 species. The homeland of representatives of this genus of North America. Under cultural conditions, species of the genus Monarda are grown in many countries of Eurasia and America. They are used as aromatic, ornamental, and medicinal plants [1-3]. Plants to 150 cm tall. Flowers are collected in inflorescences to 7 cm in diameter. The flowering of one inflorescence lasts 18- 20 days. Most species of the genus Monarda grow in dry areas [3]. Interest in the species of the genus Monarda only increases over time. Numerous research results have proven the economic value and practical importance of species of this genus. The phenolic complex of plants of the genus has been partially investigated. Essential oils have been studied in detail. Flavonoids are effective against various pathogens [2]. Essential oil and extracts of Monarda have high bactericidal, antiviral, antifungal, antimycoplasma and anthelmintic activity, immunomodulatory effect, antioxidant, radioprotective, anti-sclerotic, desensitizing, anti-cancer, anti-inflammatory, analgesic effects. They have a sedative effect on the central nervous system. Flavonoids stimulate the regeneration of damaged skin, reduce to normal the increased lipid metabolism [2,3]. The aim of the work is a comparative study of the composition and content of biologically active substances and antitoxic properties of the above-ground part of Monarda https://doi.org/10.1051/bioconf/20213800098 BIO Web of Conferences 38, 00098 (2021) Northern Asia Plant Diversity 2021 Northern Asia Plant Diversity 2021 plants cultivated in the botanical garden of the Ural Branch of the Russian Academy of Sciences, Ekaterinburg. plants cultivated in the botanical garden of the Ural Branch of the Russian Academy of Sciences, Ekaterinburg. 2 Materials and methods Five species of the genus Monarda were studied - four species of the first year of life M. didyma, M. citriodora, M. fistulosa, and M. media (collection 2019) and one species of M. menthaifolia of the second and third year of life (collection 2018, 2019). The object of the study is the above-ground part of vegetative and flowering plants, collected in the Botanical Garden of Ural Branch of RAS (Sverdlovsk Region, Ekaterinburg). The raw material was dried in well-ventilated rooms, crushed to a particle size of 2 mm, mixed and a medium sample was taken. p To extract phenolic compounds, double extraction with 70% ethanol in a water bath was performed. The quantitative content of flavones, tannins, carotenoids, and catechins was determined by the spectrophotometric method on spectrophotometer SF-56 [4-8]. The quantitative content of pectin substances was determined by the calibration curve constructed by galacturonic acid. The density of stained solutions was measured on spectrophotometer Agilent 8453 SF [9]. Saponins were determined by the weight method [10]. The component composition of the phenolic complex of the samples was studied by high-performance liquid chromatography on an analytical HPLC system consisting of an Agilent 1200 liquid chromatograph. The chromatographic separation was conducted at 25°C on a Zorbax SB-C18 Column (4.6 × 150 mm, 5 µm internal diameter) with the Agilent Guard Column Hardware Kit. Detection was performed at a wavelength of λ = 360 nm [11]. Retention times of compounds peaks on chromatograms of analyzed samples were compared with retention times of peaks of standard samples and their UV spectra. Table 1. Biologically active substances of some species of the genus Monarda (in % of the air-dry weight of raw materials). 3 Results 3,90 14,68 0,26 0,97 8,62 102,73 7,18 A k l d t W k i i d t i f f th j t f th t t i i f th Acknowledgments: Work is carried out in frames of the projects of the state mission of the ntral Siberian Botanical Garden of SB RAS № AAAA-A21-121011290025-2 Acknowledgments: Work is carried out in frames of the projects of the state mission of the Central Siberian Botanical Garden of SB RAS № AAAA-A21-121011290025-2 3 Results The greatest quantity of flavones (4.26 %) and tannins (20.65 %) was found in M. citriodora Cerv. ex Lag., carotenoids (184.07 %), pectins (1.84 %) and saponins (15.38 %) - in M. fistulosa L., pectins (1.84 %) - in M. didyma L., protopectins (12.22 %) - in M. media Willd. M. menthaifolia in the phase of mass flowering is distinguished by the increased content of catechins – 0.35 % (Table 1). The least amount of flavones (2.25%), carotenoids (87.88%) and tannins (13.39%) is typical for M. media, catechins (0.12%) - for M. fistulosa, saponins (1.86 %) and pectins (0.74 %) - M. menthaifolia (flowering, 2nd year of life, 2018), protopectins (8.14 %) - Monarda didyma . (Table 1). 3,90 % flavones, 14,68 % tannins, 0,26 % catechins, 0,97 % pectins, 8,62 % protopectins, 102,73 % carotenoids, 7,18 % saponins were found in the above-ground parts of M. menthaefolia plants (from the air-dry mass of raw materials). The component composition of the phenolic complex of the samples was studied by high-performance liquid chromatography. 70 % ethanol extracts of the aboveground parts of all studied plants of Monarda species contained at least 35 compounds of phenolic nature, of which flavones luteolin, luteolin 7-glucoside, and chlorogenic acid were identified. The results obtained can be useful in developing practical recommendations for the collection of medicinal plants in different periods of vegetation. 2 2 https://doi.org/10.1051/bioconf/20213800098 BIO Web of Conferences 38, 00098 (2021) Northern Asia Plant Diversity 2021 Species Flavones Tannins Catechins Pectins Protopectin s Carotenoid s Saponins M. didyma Vegetation. 1 year of life, 2019. 3,55 20,54 0,18 1,84 8,14 94,36 8,72 M. citriodora Vegetation. 1 year of life, 2019. 4,26 20,65 0,17 1,84 9,39 96,88 12,75 M. fistulosа Vegetation. 1 year of life, 2019. 2,87 19,01 0,12 1,78 9,44 184,07 15,38 M. media Vegetation. 1 year of life, 2019. 2,25 13,39 0,20 1,26 12,22 87,88 9,61 М. menthaifolia Flowering. 2 years of life, 2018. 2,30 17,19 0,35 0,74 9,80 125,63 1,86 М. menthaifolia Flowering. 3 years of life, 2019. References 1. S. Feskov, M. Shevchuk, Bulletin of GNBS 124 (2017) 1. S. Feskov, M. Shevchuk, Bulletin of GNBS 124 (2017) 2. G. Vysochina, Chemistry for sustainable development 2 (2020) 2. G. Vysochina, Chemistry for sustainable development 2 (2020) 3. A. Nikitina, A. Aliev, S. Feshkov, N. Nikitina, Chemistry of plant raw materials 2 (2018) 3. A. Nikitina, A. Aliev, S. Feshkov, N. Nikitina, Chemistry of plant raw materials 2 (2018) 4. T. Kukushkina, A. Zykov, L. Obukhova, Actual problems of creating new drugs of natural origin (2003) 4. T. Kukushkina, A. Zykov, L. Obukhova, Actual problems of creating new drugs of natural origin (2003) 5. V. Belikov, M. Schreiber, Pharmacy 1 (1970) 6. L. Fedoseeva, Chemistry of vegetable raw materials 2 (2005) 6. L. Fedoseeva, Chemistry of vegetable raw materials 2 (2005) 7. Methods of biochemical analysis of plants (Leninigrad, 1987) 8. Methods of biochemical analysis of plants. (L.: Publishing house of Leningrad State University, 1978) 9. V. Kriventsov, Proceedings of Nikit. Botan. Gardens 109 (1989) 10. A. Kiseleva, T. Volkhonskaya, V. Kiselev, Biologically active substances of medicin plants of Southern Siberia (Novosibirsk, 1991) 11. O. Kotsupy, N. Stepantsova, G. Vysochina, A. Petruk, Irkutsk State University Bulletin. Series: Biology. Ecology 24 (2018) 3 3
https://openalex.org/W3164168389	https://www.medrxiv.org/content/medrxiv/early/2021/05/31/2021.05.27.21257032.full.pdf	English	null	How immunity from and interaction with seasonal coronaviruses can shape SARS-CoV-2 epidemiology	medRxiv (Cold Spring Harbor Laboratory)	2,021	cc-by	11,149	. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint Corresponding Author: naomi.waterlow1@lshtm.ac.uk Keywords : coronaviruses, immunity, SARS-CoV-2, COVID-19, cross-protection Classification: Medical Sciences, Applied Mathematics Keywords : coronaviruses, immunity, SARS-CoV-2, COVID-19, cross-protection Classification: Medical Sciences, Applied Mathematics Keywords : coronaviruses, immunity, SARS-CoV-2, COVID-19, cross-protection Classification: Medical Sciences, Applied Mathematics 1 Centre for Mathematical Modeling of Infectious Disease, London School of Hygiene and Tropical Medicine, UK 1 Centre for Mathematical Modeling of Infectious Disease, London School of Hygiene and Tropical Medicine, UK 2 Statistics, Modelling and Economics Department, Public Health England, London, UK How immunity from and interaction with seasonal coronaviruses can shape SARS-CoV-2 epidemiology Authors : Naomi R Waterlow 1, , Edwin van Leeuwen 1,2 , Nicholas G. Davies 1 CMMID COVID-19 working group 1 , Stefan Flasche 1 , Rosalind M Eggo 1 . Conflicts of interest: None Funding : NRW was supported by the Medical Research Council (grant number MR/N013638/1 ). EvL was supported by the National Institute for Health Research (NIHR) Health Protection Research Unit (HPRU) in Modelling and Health Economics, a partnership between PHE, Imperial College London, and LSHTM (grant number NIHR200908). EvL was supported by the European Union's Horizon 2020 research and innovation programme - project EpiPose (101003688). SF is funded through a Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society (grant number 208812/Z/17/Z ). RME acknowledges an HDR UK Innovation Fellowship (grant: MR/S003975/1), MRC (grant: MC_PC 19065), and NIHR (grant: NIHR200908) for the Health Protection Research Unit in Modelling and Economics at LSHTM. The views expressed in this publication are those of the author(s) and not necessarily those of the NIHR or the UK Department of Health and Social Care. CMMID Working group: The following authors were part of the Centre for Mathematical Modelling of Infectious Disease COVID-19 Working Group. Each contributed in processing, cleaning and interpretation of data, interpreted findings, contributed to the manuscript, and approved the work for publication: Rachael Pung, Paul Mee, William Waites, Damien C Tully, Katherine E. Atkins, C Julian Villabona-Arenas, Graham Medley, Frank G Sandmann, Anna M Foss, Sophie R Meakin, Carl A B Pearson, Emilie Finch, Nikos I Bosse, Christopher I Jarvis, Kiesha Prem, Alicia Rosello, Kevin van Zandvoort, Rosanna C Barnard, Jiayao Lei, Yang Liu, Adam J Kucharski, Ciara V McCarthy, Sam Abbott, Emily S Nightingale, Joel Hellewell, Thibaut Jombart, David Hodgson, Gwenan M Knight, Amy Gimma, Yung-Wai Desmond Chan, Yalda Jafari, Samuel Clifford, Timothy W Russell, Fiona Yueqian Sun, Simon R Procter, Akira Endo, Oliver Brady, Kaja Abbas, Billy J Quilty, Mark Jit, Sebastian Funk, Fabienne Krauer, Matthew Quaife, Hamish P Gibbs, W John Edmunds, Mihaly Koltai, Kathleen O'Reilly, Rachel Lowe, James D Munday. CMMID Working group: The following authors were part of the Centre for Mathematical Modelling of Infectious Disease COVID-19 Working Group. Each contributed in processing, cleaning and interpretation of data, interpreted findings, contributed to the manuscript, and approved the work for publication: Rachael Pung, Paul Mee, William Waites, Damien C Tully, Katherine E. Conflicts of interest: None Atkins, C Julian Villabona-Arenas, Graham Medley, Frank G Sandmann, Anna M Foss, Sophie R Meakin, Carl A B Pearson, Emilie Finch, Nikos I Bosse, Christopher I Jarvis, Kiesha Prem, Alicia Rosello, Kevin van Zandvoort, Rosanna C Barnard, Jiayao Lei, Yang Liu, Adam J Kucharski, Ciara V McCarthy, Sam Abbott, Emily S Nightingale, Joel Hellewell, Thibaut Jombart, David Hodgson, Gwenan M Knight, Amy Gimma, Yung-Wai Desmond Chan, Yalda Jafari, Samuel Clifford, Timothy W Russell, Fiona Yueqian Sun, Simon R Procter, Akira Endo, Oliver Brady, Kaja Abbas, Billy J Quilty, Mark Jit, Sebastian Funk, Fabienne Krauer, Matthew Quaife, Hamish P Gibbs, W John Edmunds, Mihaly Koltai, Kathleen O'Reilly, Rachel Lowe, James D Munday. The following funding sources are acknowledged as providing funding for the working group authors. This research was partly funded by the Bill & Melinda Gates Foundation (INV-001754: MQ; INV-003174: JYL, KP, MJ, YL; INV-016832: SRP; NTD Modelling Consortium OPP1184344: CABP, GFM; OPP1139859: BJQ; OPP1183986: ESN; OPP1191821: KO'R). BMGF (INV-016832; OPP1157270: KA). CADDE MR/S0195/1 & FAPESP 18/14389-0 (PM). EDCTP2 NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint (RIA2020EF-2983-CSIGN: HPG). Elrha R2HC/UK FCDO/Wellcome Trust/This research was partly funded by the National Institute for Health Research (NIHR) using UK aid from the UK Government to support global health research. The views expressed in this publication are those of the author(s) and not necessarily those of the NIHR or the UK Department of Health and Social Care (KvZ). ERC Starting Grant (#757699: MQ). ERC . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Abstract We hypothesised that cross-protection from seasonal epidemics of human coronaviruses (HCoVs) could have affected SARS-CoV-2 transmission, including generating reduced susceptibility in children. To determine what the pre-pandemic distribution of immunity to HCoVs was, we fitted a mathematical model to 6 years of seasonal coronavirus surveillance data from England and Wales. We estimated a duration of immunity to seasonal HCoVs of 7.3 years (95%CI 6.8 - 7.9) and show that, while cross-protection between HCoV and SARS-CoV-2 may contribute to the age distribution, it is insufficient to explain the age pattern of SARS-CoV-2 infections in the first wave of the pandemic in England and Wales. Projections from our model illustrate how different strengths of cross-protection between circulating coronaviruses could determine the frequency and magnitude of SARS-CoV-2 epidemics over the coming decade, as well as the potential impact of cross-protection on future seasonal coronavirus transmission. Conflicts of interest: None (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint (RIA2020EF-2983-CSIGN: HPG). Elrha R2HC/UK FCDO/Wellcome Trust/This research was partly funded by the National Institute for Health Research (NIHR) using UK aid from the UK Government to support global health research. The views expressed in this publication are those of the author(s) and not necessarily those of the NIHR or the UK Department of Health and Social Care (KvZ). ERC Starting Grant (#757699: MQ). ERC (RIA2020EF-2983-CSIGN: HPG). Elrha R2HC/UK FCDO/Wellcome Trust/This research was partly funded by the National Institute for Health Research (NIHR) using UK aid from the UK Government to support global health research. The views expressed in this publication are those of the author(s) and not necessarily those of the NIHR or the UK Department of Health and Social Care (KvZ). ERC Starting Grant (#757699: MQ). ERC . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint Significance statement: Cross-protection from seasonal epidemics of human coronaviruses (HCoVs) has been hypothesised to contribute to the relative sparing of children during the early phase of the pandemic. Testing this relies on understanding the pre-pandemic age-distribution of recent HCoV infections, but little is known about their dynamics. Using England and Wales as a case study, we use a transmission model to estimate the duration of immunity to seasonal coronaviruses, and show how cross-protection could have affected the age distribution of susceptibility during the first wave, and alter SARS-CoV-2 transmission patterns over the coming decade. Introduction CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity (which was not certified by peer review) The copyright holder for this prepr this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint SARS-CoV-2 have been found in 20% 13 to 51% 14 of unexposed individuals, with evidence that these responses stem from seasonal coronavirus infection 15 . It has also been noted that these are more prevalent in children and adolescents 16 . SARS-CoV-2 have been found in 20% 13 to 51% 14 of unexposed individuals, with evidence that these responses stem from seasonal coronavirus infection 15 . It has also been noted that these are more prevalent in children and adolescents 16 . Cross-protection from seasonal HCoVs may have, therefore, partially shaped the observed epidemiology of SARS-CoV-2. In particular, this cross-protection could help to explain the relatively low SARS-CoV-2 infection rate in children 17–21 . Since children likely have a higher annual attack rate of endemic HCoVs due to their higher contact rates 22 , they may be less susceptible to SARS-CoV-2 due to cross-protection. Using England and Wales as a case study, we use dynamic models to estimate: 1) the duration of infection-induced immunity to seasonal HCoVs, 2) the ability of potential cross-protection from seasonal HCoVs to explain the age patterns in the first wave of the SARS-CoV-2 pandemic, and 3) the implications of the duration of immunity and potential cross-protection on future dynamics of SARS-CoV-2. Introduction Due to the relatively short time since SARS-CoV-2 emerged, little is yet known about the duration of infection-induced immunity. While instances of confirmed reinfection of SARS-CoV-2 have been identified 1 , these are rare, 2 indicating protection lasts for at least 6-8 months, which concurs with estimates from prospective studies 3,4 . For the success of long-term planning and design of control strategies against SARS-CoV-2, including vaccines, we must increase our understanding of the population-level immunity dynamics and interactions with other circulating viruses. In order to evaluate the impacts of cross-immunity, we first need to quantify the duration of immune protection from seasonal coronaviruses. Four coronaviruses strains from two different genera are endemic in humans: two are alphacoronaviruses (HCoV-229E, HCoV-NL63) and two are betacoronaviruses (HCoV-HKU1, HCoV-OC43); SARS-CoV-2 is a member of the latter genera as are SARS-CoV-1 and MERS-CoV. In the UK, seasonal human coronavirus (HCoV) case incidence peaks January-February each year. The first infection with seasonal HCoVs typically occurs in childhood 5 and reinfection with the same strain has been observed within a year 6,7 . However, there are also indications that immunity lasts longer, with few reinfections in a 3-year cohort study 8 and sterilising immunity to homologous strains of HCoV-229E after one year in a challenge study 9 . There may also be cross-protective immunity between seasonal HCoVs and SARS-family coronaviruses following infection. Human sera collected before the SARS-CoV-2 pandemic showed high IgG reactivity to seasonal HCoVs, but also low reactivity to SARS-CoV-2 10 , and SARS-CoV-1 infection induced antibody production against seasonal HCoVs 11 , 12 . Cross-reactive T-cells to . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . Results CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint then become infectious ( I ). They then recover (R) and via waning become fully susceptible again. At any point individuals can be infected by the other virus, although this is less likely to occur in the I and R categories, determined by the cross-protection parameter. For clarity, ageing, seasonality, births and deaths are not shown in this Figure. The initial model fit without SARS-CoV-2 only contains individuals within the compartments in the dashed box. We fitted a dynamic transmission model using England and Wales as a case study (Figure 1C, box) using only the seasonal coronavirus model. Following infection, individuals are protected against infection with any seasonal HCoVs, with reinfection possible after a period of temporary but complete immunity. Waning of immunity against circulating viruses could result from decaying protection against homotypic viruses, and/or longer-lasting immunity against homotypic viruses but evolutionary change leading to immune escape 25 . We do not track individual seasonal HCoV strains as available data are not sub-typed. We therefore assume that individual seasonal HCoV strain have the same parameter values, including duration of immunity and R 0,HCoV . Transmission is seasonally forced using a cosine function. Immunity to seasonal HCoVs is estimated to last around 7 years We fitted the model to the age group-specific seasonal HCoV data from June 09, 2014 until February 17, 2020, and estimated key seasonal HCoV parameters using parallel tempering 26 (Figure 2). We fitted the duration of infection-induced immunity, the transmissibility, age-specific reporting proportions and two seasonal forcing parameters (Table S1). We estimated that the average duration of infection-induced immunity for seasonal HCoVs was 7.3 years (95% Credible Interval (CI): 6.8 - 7.9) and that the basic reproduction number was 5.7 (95% CI 5.4 - 6.0) (Figure 2B). Further details are given in the Supplement. We fitted the model to the age group-specific seasonal HCoV data from June 09, 2014 until February 17, 2020, and estimated key seasonal HCoV parameters using parallel tempering 26 (Figure 2). We fitted the duration of infection-induced immunity, the transmissibility, age-specific reporting proportions and two seasonal forcing parameters (Table S1). Results Seasonal HCoV and SARS-CoV-2 epidemic data We extracted monthly, age group-stratified numbers of HCoV positive tests in England and Wales from the June 9, 2014 to February 17, 2020 23 and daily number of COVID-19 deaths in England and Wales during the first wave of the pandemic (March 02, 2020 to June 01, 2020) 24 (Figure 1). The timeframe for the HCoV data is from the first available date until February 2020 to avoid interference from SARS-CoV-2 transmission and reporting. Figure 1: Data and model structure. HCoV is shown in blue and SARS-CoV-2 in red. A) Time series of reported HCoV cases and COVID-19 deaths. Left axis: reported seasonal HCoV cases to Public Health England (PHE), all ages. Right axis: COVID-19 daily deaths 24 . B) Proportion of seasonal HCoV cases reported by age group. C) Model diagram for seasonal HCoV (HCoV) and SARS-CoV-2 (C19). Individuals can be Susceptible ( S ), Exposed ( E ), Infectious ( I ) or Recovered ( R ) to SARS-CoV-2 or seasonal HCoVs. Following infection, individuals enter the exposed ( E ) category and Figure 1: Data and model structure. HCoV is shown in blue and SARS-CoV-2 in red. A) Time series of reported HCoV cases and COVID-19 deaths. Left axis: reported seasonal HCoV cases to Public Health England (PHE), all ages. Right axis: COVID-19 daily deaths 24 . B) Proportion of seasonal HCoV cases reported by age group. C) Model diagram for seasonal HCoV (HCoV) and SARS-CoV-2 (C19). Individuals can be Susceptible ( S ), Exposed ( E ), Infectious ( I ) or Recovered ( R ) to SARS-CoV-2 or seasonal HCoVs. Following infection, individuals enter the exposed ( E ) category and . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . Results We estimated that the average duration of infection-induced immunity for seasonal HCoVs was 7.3 years (95% Credible Interval (CI): 6.8 - 7.9) and that the basic reproduction number was 5.7 (95% CI 5.4 - 6.0) (Figure 2B). Further details are given in the Supplement. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint Figure 2: Seasonal HCoV Fit. A) Model fit for seasonal HCoV by age. Black dots show reported HCoV cases, blue are 100 random samples from the posterior. B) Posterior distributions for the duration of waning and the R 0 of seasonal HCoV. C) Mean annual attack rate for each age group from 100 samples of the posterior and the last 5 years of the fit. Cross-protection from seasonal HCoVs is not sufficient to explain age-specific patterns of SARS-CoV-2 infection We included SARS-CoV-2 into the model, where each compartment has the state for the combined seasonal HCoVs as well as the state for SARS-CoV-2 (Figure 1C, full model). We included cross-protection that decreases susceptibility to infection by SARS-CoV-2 by an amount, σ, for individuals in the I HCoV,i or R HCoV,i states (σ = 0 is no cross-protection and 1 is full cross-protection). We assume any interaction in the opposite direction would be negligible, due to the low proportion of the population that were infected in the first SARS-CoV-2 epidemic wave. Results Using the posterior estimates of the seasonal HCoV parameters and the simulated output as initial states, we continued a simulation of epidemic seasonal HCoVs from the January 01, 2020 until June 01, 2020, including the introduction of SARS-CoV-2. Cross-protection from seasonal HCoVs and different mixing patterns (matching observed lockdown patterns, see Methods) were the only mechanisms we included that affected infection by SARS-CoV-2, so that we could evaluate the impact of cross-protection on the observed age distribution of cases. For values of the cross-protection parameter between σ=0 and σ=1, we estimated R 0,C19 and the number of introductions of SARS-CoV-2 by fitting the extended model to daily reported COVID-19 deaths (Figure 3a). We captured the national lockdown by decreasing contact rates following trends in Google mobility data 27 . Our model fits were able to closely match the reported mortality incidence for each value of the cross-protection parameter. However, the resulting R 0,C19 varied widely, reaching over 25 for the strongest cross-protection (Figure 3B). The corresponding R eff,C19 before the intervention on March 23 ranged between 2.25 and 3.75 (see Supplement). . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint Figure 3: SARS-CoV-2 simulations A) Fitted epidemic trajectories to daily SARS-CoV-2 deaths in 2020 in England and Wales 24 . Each facet is a different strength of cross-protection and contains output from 100 simulations. The red line indicates the fixed date when SARS-CoV-2 introduction starts. B) Estimated R 0 values for SARS-CoV-2 with different strengths of cross-protection. Points display the R 0, C19 and lines show the range of R eff, C19 during the simulation. C) Simulated age-specific serology rates for SARS-CoV-2 by the end of May 2020. Figure 3: SARS-CoV-2 simulations A) Fitted epidemic trajectories to daily SARS-CoV-2 deaths in 2020 in England and Wales 24 . Results Each facet is a different strength of cross-protection and contains output from 100 simulations. The red line indicates the fixed date when SARS-CoV-2 introduction starts. B) Estimated R 0 values for SARS-CoV-2 with different strengths of cross-protection. Points display the R 0, C19 and lines show the range of R eff, C19 during the simulation. C) Simulated age-specific serology rates for SARS-CoV-2 by the end of May 2020. We then evaluated the age distribution of infections that would be detected by serology by the end of May in our model, across the range of values of the cross-protection parameter (Figure 1C). In simulations with no or low cross-protection the model predicted larger proportions of children to have been infected than in older age groups, differing from observed data 17,28 . As the strength of interaction increased, the age-distribution flattened and a smaller proportion of children became infected. With complete protection, there was a higher rate in the youngest age groups, which has not been observed 19–21,28 . Future SARS-CoV2 epidemiology could be shaped by coronavirus interactions Future SARS-CoV2 epidemiology could be shaped by coronavirus interactions To determine possible long-term dynamics of interacting coronaviruses, we ran 30-year projections of our model including both HCoVs and SARS-CoV-2, with different assumptions on the strength of cross-protection and whether it acted from HCoV to SARS-CoV-2, or in both directions (Figure 4, and further samples in the supplement). In all scenarios we assumed no interventions, and used parameters estimated previously. For single-direction cross-protection, annual SARS-CoV-2 epidemics were projected to occur in scenarios with stronger cross-protection, whereas weaker / no cross-protection projected less frequent epidemics. However, strong cross-protection scenarios relied on very high and potentially unrealistic R 0 . In weaker cross-protection scenarios, interepidemic periods lasted multiple years following a pandemic. In scenarios with bi-directional cross-protection, SARS-CoV-2 infections also projected frequent epidemics, but led to the seasonal HCoV being . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint disrupted. With low levels of cross-protection, SARS-CoV-2 and seasonal HCoV epidemics alternated, but as the cross-protection increased, SARS-CoV-2 epidemics became more frequent and outcompeted seasonal HCoV: while a cross-protection of 0.6 resulted in irregular dynamics of the viruses. At higher levels of cross-protection, no seasonal HCoV transmission occurred. disrupted. With low levels of cross-protection, SARS-CoV-2 and seasonal HCoV epidemics alternated, but as the cross-protection increased, SARS-CoV-2 epidemics became more frequent and outcompeted seasonal HCoV: while a cross-protection of 0.6 resulted in irregular dynamics of the viruses. At higher levels of cross-protection, no seasonal HCoV transmission occurred. viruses. At higher levels of cross-protection, no seasonal HCoV transmission occurred. Figure 4: 10-year forward projections of seasonal HCoV and SARS-CoV-2 epidemics. Red indicates SARS-CoV-2, blue indicates Seasonal HCoVs. Future SARS-CoV2 epidemiology could be shaped by coronavirus interactions The dashed vertical line indicates a change in axis scale due to the much larger SARS-CoV-2 pandemic wave, with that to the left of the dashed line marked by the left axis and that to the right by the right axis. Cross-protection strength and estimated SARS-CoV-2 R 0 for the scenario are shown to the left of the figure. A has cross-protection from seasonal HCoV to SARS-CoV-2, and B has bidirectional cross-protection. No control measures were included. gure 4: 10-year forward projections of seasonal HCoV and SARS-CoV-2 epidemics. Red Figure 4: 10-year forward projections of seasonal HCoV and SARS-CoV-2 epidemics. Red indicates SARS-CoV-2, blue indicates Seasonal HCoVs. The dashed vertical line indicates a change in axis scale due to the much larger SARS-CoV-2 pandemic wave, with that to the left of the dashed line marked by the left axis and that to the right by the right axis. Cross-protection strength and estimated SARS-CoV-2 R 0 for the scenario are shown to the left of the figure. A has cross-protection from seasonal HCoV to SARS-CoV-2, and B has bidirectional cross-protection. No control measures were included. Figure 4: 10-year forward projections of seasonal HCoV and SARS-CoV-2 epidemics. Red indicates SARS-CoV-2, blue indicates Seasonal HCoVs. The dashed vertical line indicates a change in axis scale due to the much larger SARS-CoV-2 pandemic wave, with that to the left of the dashed line marked by the left axis and that to the right by the right axis. Cross-protection strength and estimated SARS-CoV-2 R 0 for the scenario are shown to the left of the figure. A has cross-protection from seasonal HCoV to SARS-CoV-2, and B has bidirectional cross-protection. No control measures were included. Discussion ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint Our estimates for the duration of homotypic protection following HCoV infection are comparable with other estimates, such as a cohort study where 8/216 (3%) confirmed infected individuals were reinfected over 5 years, and the median re-infection time in a study of 10 individuals 30 varied between 30 and 55 months, depending on strain. However, estimates vary, with a larger study in Michigan estimating mean strain-specific reinfection to be between 19 and 33 months 31 , 19.9% of first infections being reinfected within 6 months in Kenya 32 , and a historical study, of just one seasonal HCoV strain (229E), estimating the time until T cells could no longer neutralise new strains at 8-17 years 25 . Other coronaviruses can also give indications on the duration of immunity, with T cells to SARS-CoV-1 detectable up to 11 years post-infection 33 . Using a similar modelling approach to that presented here, others have estimated that the average duration of immunity to seasonal coronavirus could be substantially shorter, under a year 34 . However, this implies very high annual attack rates, which are not observed in surveillance data, and our model suggests that a longer period of cross-protection may be more appropriate, and should be included in the proposed range of parameters for fitting such models. Our model suggests that cross-protection between seasonal HCoVs and SARS-CoV-2 could account for some of the reduced susceptibility to infection of children in the first wave of the SARS-CoV-2 epidemic in England and Wales. Specifically, stronger cross-protection decreased the relative susceptibility to infection of children. This is in line with an American study showing that 50% of pre-pandemic donors had reactive T-cells to SARS-CoV-2 35 and serological markers for a recent seasonal HCoV infection, suggesting that immune responses to seasonal HCoV could elicit cross-protective immunity. Moreover, 48% of uninfected individuals in a cohort from Australia had cross-reactive T-cells to SARS-CoV-2 and cross-reactivity was strongly correlated with memory of a seasonal coronavirus strains 15 . Other studies among healthy individuals without SARS-CoV-2 exposure found cross-reactive T-cells targeting SARS-CoV-2 in 51% 36 , 35% 14 24% 37 , and 20% 13 of participants, suggesting a moderate amount of cross-immunity that likely stems from seasonal coronaviruses. Discussion We estimated that immunity to seasonal HCoVs lasts 7.3 years (95% CI: 6.8 - 7.9) and that the R 0 is 5.7 (95% CI: 5.4 - 6.0). We found 12.8% (95% CI: 11.9 - 13.7%) reinfection within one year, and the median reinfection time was 5.1 years (95% CI: 4.7 - 5.5 years). It was possible to match the COVID-19 mortality data with the full range of cross-protection strengths between seasonal HCoV and SARS-CoV-2, but the estimated R 0,C19 s were outside of a realistic range for very high values of cross protection, for example, a recent multi-setting study estimating the R 0,C19 to be between 3.6 and 7.3 29 . Cross-protection from seasonal HCoVs to SARS-CoV-2 did not fully explain the apparent reduced susceptibility of children to SARS-CoV-2 observed during the first wave in the UK 17,18,20,21,28 . Future projections varied in the frequency of SARS-CoV-2 epidemics, with SARS-CoV-2 epidemics every two years at low levels of cross-protection, changing to annual epidemics with increased cross-protection. In scenarios with bi-directional cross-protection, epidemics were less predictable and SARS-CoV-2 out competed seasonal HCoVs. Further elucidating possible cross-protection and potential duration of protection is therefore critical for medium-to-long-term projections of SARS-CoV-2 epidemics. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. Discussion CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint The strength and implications of cross-protection between HCoVs and SARS-CoV-2 will become increasingly evident over the coming months and years. Our projections show that, depending on the extent of cross-protection, SARS-CoV-2 could eventually cause annual epidemics (strong cross-protection) or epidemics every 2 years (little cross-protection). If bi-directional cross-protection occurs, SARS-CoV-2 also has the ability to substantially disrupt seasonal HCoV transmission. This is based on our fit of the duration of immunity and the seasonal forcing parameters of seasonal HCoVs, which are likely to differ to some extent in the case of SARS-CoV-2. These scenarios are in line with others 34,51–53 , which suggest that ongoing SARS-CoV-2 transmission is likely. Our modelled projections assumed that no interventions were implemented. However, HCoV circulation was disrupted in winter 2020–2021 40 likely due to social restrictions designed to curb the transmission of SARS-CoV-2. It is important to understand the longer-term dynamics of SARS-CoV-2, in order to minimise deaths and plan vaccination strategies. From an evolutionary perspective, cross-protection may be a strong driver for selection, so in the long run a less transmissible type with greater cross-protection against competing viruses may dominate. We modelled all seasonal HCoVs as one virus, implicitly assuming complete cross-protection between them. There is evidence for cross-protection between seasonal HCoVs, such as the presence of cross-reactive antibodies 55 and evidence from modelling studies 34 . Yet cross-protection may not be complete, or may be strain specific (alpha vs beta coronaviruses), and hence our assumption could lead to an underestimation of the true duration of protection. This is because the duration between subsequent homotypic infections would be longer than subsequent infections of any strain, for the same number of cases. Discussion We expect the assumption of modelling seasonal HCoVs to have a relatively small impact on the results of the cross-protection in the first wave of SARS-CoV-2, which uses the average cross-immunity profile at the end of the seasonal HCoV epidemic. However, the assumption may have a larger impact on the longer term dynamics, so these should be interpreted with caution. We also assumed that the strength of immunity to seasonal HCoVs is constant over repeated infections. An alternative mechanism would be that repeat infections strengthen immunity, as is hypothesised in some respiratory infections, such as Respiratory Syncytial Virus 56 , which could have led to a different estimate of immunity. The emergence of SARS-CoV-2 has highlighted our lack of knowledge on coronavirus immunity and long term dynamics. In our study, we estimate that immunity against seasonal HCoVs can last years, however by necessity we made strong assumptions about the cross-immunity between seasonal HCoV strain. Further studies exploring cross-protection between strain for seasonal coronaviruses as well as routinely subtyped surveillance data would help inform future models. Nonetheless, based on the available data our study indicates that seasonal coronavirus immunity may last multiple years, which should be considered in the planning of subsequent studies. We also conclude that cross-protection from seasonal coronaviruses is not enough to explain the age susceptibility pattern of SARS-CoV-2, indicating other mechanisms must be involved. Whilst serological data could be useful to further evaluate the extent of cross-protection, the reduction in social contacts due to government interventions against SARS-CoV-2 complicates their use. Our models relies heavily on social contact matrices, and getting an accurate understanding of social contacts in the last year comes with many challenges, such as multiple changes in public health interventions with uncertain adherence. Our study highlights the importance of understanding the background environment of coronaviruses for insights into SARS-CoV-2 pandemic progressi on. Discussion There are indications that these cross-reactive T cells are present at higher frequency in younger vs older adults, correlating with our hypothesis that this could be due to increased infection from seasonal HCoVs 38 . Antibodies have also been shown to be cross-reactive 39 , although their persistence in the body is more varied and often shorter in duration than T-cells 40 . Cross-reactive responses have also been identified in other pandemic coronaviruses 10–12,41 , with some also showing cross-protection: SARS-CoV-1 and MERS-CoV T cell epitopes where protective in mice against other human and bat coronaviruses 42 . However a longitudinal study showed that whilst cross-reactive HCoV antibodies are boosted following SARS-CoV-2 infection, this does not correlate with protection against infection or hospitalisation 43 and a lack of antibody-mediated neutralising cross-protection has been noted between sera from SARS-CoV-1 patients and SARS-CoV-2 44 . Therefore, whilst there is significant amounts of corroborating evidence that our find of some degree of cross-protection exists, the literature is not conclusive. Our results indicate that cross-protection from seasonal coronaviruses alone cannot explain reduced susceptibility to infection of children. Other factors are needed to counteract the children’s higher than average exposure probability driven by their contact behavior 22 . One mechanism for this could be due to differences in children’s immune system 45 : children can produce broadly reactive antibodies that have not been influenced by commonly circulating pathogens and have different proportions of blood cell types, such as specific subtypes of memory B cells. Genetic analysis also suggests that cross-reactivity to SARS-CoV-2 antigens can not fully be explained by seasonal coronaviruses, implying that other unknown viruses/factors may induce cross-immunity 46 . We also modelled cross-protection as only reducing susceptibility to infection, whereas there could also be a reduction in transmission and/or disease severity 47–50 . Boosting of immunity by multiple infections has also been suggested to influence cross-protection 48 , where boosting by repeat infections was hypothesised to reduce the cross-protection to SARS-CoV-2. We did not include boosting in our model due to the added complexity. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . Cross-protection Model Cross-protection Model We created a deterministic compartmental transmission model for coronavirus infections and their interactions. The population are either Susceptible (S), Exposed (E), Infectious (I) and Recovered (R) for both seasonal HCoVs and SARS-CoV-2. The subscripts used are “HCoV” for seasonal HCoVs and “C19” for SARS-CoV-2, with no differentiation between HCoV strains as the data are not sub-typed. Following infection, individuals enter the exposed category and become infected at rates and λHCoV ,i respectively, and individuals enter the infectious category at rates ν HCoV and ν C19. They then λC19,i recover and become fully susceptible again at rate . The force of infection for each virus is shown in ω equations 1 and 2. Each compartment in the model records the state for SARS-CoV-2 and seasonal HCoVs, with one for each combination of states (Figure 1) and all durations are exponentially distributed. At any point individuals can be infected by the other virus, although this is less likely to occur in the I and R categories, determined by the cross-protection parameter, . This takes into σ account both short term cross-protection from the activation of the immune system, and longer-term adaptive immunity. Both modelled viruses (HCoVs and SARS-CoV-2) are seasonally forced with a cosine function, which captures changes in seasonal human behaviour and climatic factors. ) os( ϕ) ) I λ (( HCoV ,i,t = ∑ j= N j=1 R0,HCoV A β HCoV HCoV * c 2π 52 7 * + βHCoV * αi,j * HCoV ,j (1) ) os( ϕ) ) I ) λ (( C19,i,t = ∑ j= N j=1 R0,C19 A β C19 * C19 * c 2π 52 7 * + βC19 * αi,j * C19,j (2) (1) (2) = Force of infection, i,j = age groups, N = total number of age groups, Α = seasonal amplitude, β = λ transmissibility, α = contact rates, I = number Infected, timing of seasonal forcing ϕ = = Force of infection, i,j = age groups, N = total number of age groups, Α = seasonal amplitude, β = λ transmissibility, α = contact rates, I = number Infected, timing of seasonal forcing ϕ = As the seroprevalence for SARS-CoV-2 stayed below 5% during the modelled period, we assumed that the level of cross-protection conferred by SARS-CoV-2 on HCoV is negligible during the first epidemic wave. introduction with varying strengths of cross-protection, to investigate the effect on age-specific susceptibility. The model was written in R 57 and the code is available at https://github.com/cmmid/coronavirus_immunity introduction with varying strengths of cross-protection, to investigate the effect on age-specific susceptibility. The model was written in R 57 and the code is available at https://github.com/cmmid/coronavirus_immunity introduction with varying strengths of cross-protection, to investigate the effect on age-specific susceptibility. The model was written in R 57 and the code is available at https://github.com/cmmid/coronavirus_immunity Materials and Methods We created a dynamic transmission model that includes cross-protection between seasonal HCoVs and SARS-CoV-2, using the UK as a case study (Figure 1). Initially, we fit the model without SARS-CoV-2 and estimated key seasonal HCoV parameters. Next we simulated SARS-CoV-2 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint Data We extracted the monthly, age group-stratified number of HCoV positive tests in England and Wales between June 09, 2014 and February 17, 2020, from National Health Service (NHS) and Public Health England (PHE) laboratories 21 . The reported age groups are ages 0-4, 5-14, 15-44, 45-64 and 65+. We did not use data beyond February 2020, as we wanted to estimate seasonal HCoV parameters in the absence of SARS-CoV-2. For SARS-CoV-2, we used the daily number of deaths with a confirmed SARS-CoV-2 positive test in the preceding 28 days from March 02, 2020 until May 31, 2020 reported by the Office for National Statistics (ONS) 24 . Cross-protection Model Cross-protection Model Cross-protection was the only mechanism we included for differing susceptibility to SARS-CoV-2 infection by age group, so that we could test whether it explained the observed infection pattern. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint The modelled population was stratified into 5-year bands to 75+, with constant birth rates, matching death rates and ageing in line with the population of England and Wales (see Table 1). Age-assortative mixing was modelled proportionately to patterns of conversational and physical contacts in the POLYMOD study 22 , 58 . contacts in the POLYMOD study 22 , 58 . We ran a HCoV-only model for 15 years to reach equilibrium, and a further 5 years to generate simulations to match the data on seasonal HCoV cases from June 09, 2014 to February 17, 2020. Parameter type Parameter Symbol Value Reference Seasonal HCoV Basic Reproduction number R 0,HCoV Fitted. Limits: 1-8.5 Wide range Transmission rate β HCoV Fitted Based on R 0 calculation (supplement) Latent period 1/ν HCoV 2.5 days 34,59 Duration of infectiousness 1/γ HCoV 5 days 34 Incubation period (time to symptoms) 1/δ1 HCoV 2 days 60 Reporting delay (symptom to report) 1/δ2 HCoV 3 days Based on influenza model 61 Age-specific reporting proportion μ HCoV,i Fitted. Limits 0-1. Proposed on log odds scale. Seasonal forcing amplitude Α Fitted. Limits: 0 - 2 Seasonal forcing timing ϕ Fitted. Limits: -(527)- (527) Immunity duration ω 1/ Fitted. Cross-protection Model Limits: 100 - 3000 Covers range of 100 days to over 8 years SARS-CoV-2 Basic Reproduction number R 0 , C19 Fitted Based on R 0 calculations (see supplement). Transmission rate β C19 Fitted Effective Reproduction Number R eff,C19 Fitted R 0 * proportion susceptible (see supplement) Latent period 1/ν C19 3 days 34 Duration of infectiousness 1/γ C19 5 days 62 Time between infectiousness (entering I compartment) and death 1/δ C19 22 days (split over two compartments, Erlang distributed) 63 Age-specific infection fatality proportions (age groups 0-4, 5-14, 15-44, 45-64, 65+) μ C19,i 0.00004, 0.00004, 0.00024, 0.00441, 0.06720 As in Levin 64 , weighted by model population sizes an a HCoV-only model for 15 years to reach equilibrium, and a further 5 years to generate lations to match the data on seasonal HCoV cases from June 09, 2014 to February 17, 2020. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint Table 1. Model parameters. Adult (15-64 years) introduction rate 1/η Fitted Duration of immunity 1/ω Fitted Assumed equal to HCoV waning rate Demographic Birth rate μb 640 370 per year ONS statistical bulletin 2019 65 Death rate μd 640 370 per year Equal to birth rate to maintain constant population Population size N 59 439 840 ONS 2019 population estimates for England and Wales, 5-year age bands 66 Inferring seasonal HCoV parameters (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint Poisson likelihood. We therefore created 100 simulations of HCoV and SARS-CoV-2 circulation for each strength of cross-protection. Due to the non-pharmaceutical interventions implemented in this period (“lockdown”), we adjust the contact matrices, which are split into three categories: school contacts, household contacts and all other contacts. From February 21, 2020, when Google mobility data becomes available, we adjust our “other” contacts in line with google mobility data. From February 23, 2020, we eliminate school contacts and assume that all remaining contacts are reduced to 33% of their transmission potential, due to social distancing and behavioral changes (“micro-distancing”) 68 . SARS-CoV-2 importations occur from February 15, 2020 until lockdown. See supplement for more details on the implementation of public health interventions. To look at the proportion infected during the first wave we assumed that antibodies would take 3 weeks to rise to detectable levels after infection and not wane below the detection threshold during the study period 69 . Projecting future dynamics of SARS-CoV-2 and Seasonal HCoVs We ran the model for 30 years, from January 01, 2020 without any changes in contacts, in order to project the future dynamics of SARS-CoV-2. As inputs, we used the estimated parameters from the seasonal HCoV fits, as well as the estimated transmission and introduction rates fitted for each of the samples. Projections for 10 of the samples are shown (in Figure 4 and Supplement). Inferring seasonal HCoV parameters We used reported seasonal HCoV cases from June 09, 2014 until February 17, 2020 to avoid overlap with SARS-CoV-2, where potential cross-protection could have occurred. We defined a binomial likelihood, where modelled infection incidence maps to reported cases via an age-dependent reporting proportion, . We assume equal reporting rates in age groups 5-15 and 45-65 to reduce pi the dimensions of the model, as initial fitting suggested these were very similar. The likelihood is therefore: , og(L) ～ log(p ) (n )log(1 ) l ∑ i=N i=1 ∑ X x=1 kx,i i + x,i kx,i pi (3) (3) where is the likelihood, is age group to a total of N age groups, are the reported monthly time L i x points, are the reported HCoV cases by age group, are the model estimated infections per age kx,i nx,i group and is the age-specific reporting rate. pi We fit the model to the data using parallel tempering, adapted from Vousden et al. 22 which is based on Monte Carlo Markov Chain (MCMC) inference. Unlike MCMC, multiple chains at different temperatures are run in parallel and swaps of parameter positions between chains are proposed. This allows more comprehensive exploration of the parameter space and allows the chains to move out of local maxima. We ran two sets of 16 chains and confirmed their convergence with the Gelman-Rubin statistic 67 , which was <1.1. We then combined the sample from both chains, excluding the burn-in, in order to increase sample size, resulting in 93900 samples. See supplement for more details. The percentage infected within one year and the median duration to reinfection were calculated using distribution and quantile functions from the stats R package 57 . Simulating SARS-CoV-2 with a range of strengths of cross-protection Simulating SARS-CoV-2 with a range of strengths of cross-protection We drew 100 random samples from the joint posterior distribution and simulated daily deaths reported in the first wave of the SARS-CoV-2 epidemic in England and Wales, between March 02, 2020 and May 31, 2020. We explored the full range of possible cross-protection strengths, in each case fitting the transmission and introduction rates to the death data using maximum likelihood estimation with a . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 6. Edridge, A. W. et al. Coronavirus protective immunity is short-lasting. medRxiv References References 1. Tillett, R. L. et al. Genomic evidence for reinfection with SARS-CoV-2: a case study. Lancet Infect. Dis. 0 , (2020). 2. Stokel-Walker, C. What we know about covid-19 reinfection so far. BMJ 372 , n99 (2021). 3. Engl, P. H. Past COVID-19 infection provides some immunity but people may still carry and transmit virus. GOV.UK https://www.gov.uk/government/news/past-covid-19-infection-provides-some-immunity-but -people-may-still-carry-and-transmit-virus. 4. Dan, J. M. et al. Immunological memory to SARS-CoV-2 assessed for up to 8 months after infection. Science eabf4063–eabf4063 (2021) doi:10.1126/science.abf4063. 5. Huang, A. T. et al. A systematic review of antibody mediated immunity to coronaviruses: kinetics, correlates of protection, and association with severity. Nat. Commun. 11 , 4704–4704 (2020). 6. Edridge, A. W. et al. Coronavirus protective immunity is short-lasting. medRxiv 1. Tillett, R. L. et al. Genomic evidence for reinfection with SARS-CoV-2: a case study 1. Tillett, R. L. et al. Genomic evidence for reinfection with SARS-CoV-2: a case study. Lancet Infect. Dis. 0 , (2020). 2. Stokel-Walker, C. What we know about covid-19 reinfection so far. BMJ 372 , n99 (2 2. Stokel-Walker, C. What we know about covid-19 reinfection so far. BMJ 372 , n99 (2021). 3. Engl, P. H. Past COVID-19 infection provides some immunity but people may still carry and transmit virus. GOV.UK 3. Engl, P. H. Past COVID-19 infection provides some immunity but people may still carry and transmit virus. GOV.UK https://www.gov.uk/government/news/past-covid-19-infection-provides-some-immunity-but -people-may-still-carry-and-transmit-virus. 3. Engl, P. H. Past COVID-19 infection provides some immunity but people may still carry and transmit virus. GOV.UK https://www gov uk/government/news/past-covid-19-infection-provides-some-immunity-but 3. Engl, P. H. Past COVID-19 infection provides some immunity but people may still carry and transmit virus. GOV.UK https://www.gov.uk/government/news/past-covid-19-infection-provides-some-immunity-but -people-may-still-carry-and-transmit-virus. 4. Dan, J. M. et al. Immunological memory to SARS-CoV-2 assessed for up to 8 months after infection. Science eabf4063–eabf4063 (2021) doi:10.1126/science.abf4063. 5. Huang, A. T. et al. A systematic review of antibody mediated immunity to coronaviruses: kinetics, correlates of protection, and association with severity. Nat. Commun. 11 , 4704–4704 (2020). 6. Edridge, A. W. et al. Coronavirus protective immunity is short-lasting. medRxiv . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . References CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint 2020.05.11.20086439-2020.05.11.20086439 (2020) doi:10.1101/2020.05.11.2008643 7. Galanti, M. & Shaman, J. Direct observation of repeated infections with endemic coronaviruses . 2020.04.27.20082032 (Oxford University Press (OUP), 2020). 8. Aldridge, R. W. et al. Seasonality and immunity to laboratory-confirmed seasonal coronaviruses (HCoV-NL63, HCoV-OC43, and HCoV-229E): results from the Flu Watch cohort study. Wellcome Open Res. 5 , 52 (2020). 9. Reed, S. E. The behaviour of recent isolates of human respiratory coronavirus in vitro and in volunteers: evidence of heterogeneity among 229E-related strains. J. Med. Virol. 13 , 179–192 (1984). 10. Khan, S. et al. Analysis of Serologic Cross-Reactivity Between Common Human Coronaviruses and SARS-CoV-2 Using Coronavirus Antigen Microarray. BioRxiv Prepr. Serv. Biol. (2020) doi:10.1101/2020.03.24.006544. 11. Chan, K. H. et al. Serological responses in patients with severe acute respiratory syndrome coronavirus infection and cross-reactivity with human coronaviruses 229E, OC43, and NL63. Clin. Diagn. Lab. Immunol. 12 , 1317–1321 (2005). 12. Che, X. et al. Antigenic Cross-Reactivity between Severe Acute Respiratory Syndrome–Associated Coronavirus and Human Coronaviruses 229E and OC43. J. Infect. Dis. 191 , 2033–2037 (2005). 13. Weiskopf, D. et al. Phenotype and kinetics of SARS-CoV-2–specific T cells in COVID-19 patients with acute respiratory distress syndrome. Sci. Immunol. 5 , (2020). 14. Braun, J. et al. Presence of SARS-CoV-2 reactive T cells in COVID-19 patients and healthy donors. medRxiv 2020.04.17.20061440-2020.04.17.20061440 (2020) doi:10.1101/2020.04.17.20061440. 15. Tan, H.-X. et al. Adaptive immunity to human coronaviruses is widespread but low in magnitude. medRxiv 2021.01.24.21250074 (2021) doi:10.1101/2021.01.24.21250074. 15. Tan, H.-X. et al. Adaptive immunity to human coronaviruses is widespread but low in magnitude. medRxiv 2021.01.24.21250074 (2021) doi:10.1101/2021.01.24.21250074. magnitude. medRxiv 2021.01.24.21250074 (2021) doi:10.1101/2021.01.24.21250074. 16. Ng, K. et al. Pre-existing and de novo humoral immunity to SARS-CoV-2 in humans. bioRxiv 2020.05.14.095414-2020.05.14.095414 (2020) doi:10.1101/2020.05.14.095414. 16. Ng, K. et al. Pre-existing and de novo humoral immunity to SARS-CoV-2 in humans. bioRxiv 2020.05.14.095414-2020.05.14.095414 (2020) doi:10.1101/2020.05.14.095414. 16. Ng, K. et al. Pre-existing and de novo humoral immunity to SARS-CoV-2 in humans. bioRxiv 2020.05.14.095414-2020.05.14.095414 (2020) doi:10.1101/2020.05.14.095414. 17. Viner, R. M. et al. Susceptibility to SARS-CoV-2 Infection Among Children and 17. Viner, R. M. et al. Susceptibility to SARS-CoV-2 Infection Among Children and 17. Viner, R. M. et al. Susceptibility to SARS-CoV-2 Infection Among Children and . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. 2020.05.11.20086439-2020.05.11.20086439 (2020) doi:10.1101/2020.05.11.2008643 ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint Adolescents Compared With Adults. JAMA Pediatr. (2020) doi:10.1001/jamapediatrics.2020.4573. Adolescents Compared With Adults. JAMA Pediatr. (2020) 18. Davies, N. G. et al. Age-dependent effects in the transmission and control of COVID-19 epidemics. Nat. Med. 1–7 (2020) doi:10.1038/s41591-020-0962-9. 19. Gaskell, K. M. et al. Extremely high SARS-CoV-2 seroprevalence in a strictly-Orthodox Jewish community in the UK. medRxiv 2021.02.01.21250839 (2021) doi:10.1101/2021.02.01.21250839. 20. Li, X. et al. The role of children in transmission of SARS-CoV-2: A rapid review. J. Glob. Health 10 , 011101 (2020). 21. Lai, C.-C., Wang, J.-H. & Hsueh, P.-R. Population-based seroprevalence surveys of anti-SARS-CoV-2 antibody: An up-to-date review. Int. J. Infect. Dis. IJID Off. Publ. Int. Soc. Infect. Dis. 101 , 314–322 (2020). 22. Mossong, J. et al. Social contacts and mixing patterns relevant to the spread of infectious diseases. PLoS Med. 5 , 0381–0391 (2008). 23. Respiratory infections: laboratory reports 2019 - GOV.UK. 24. Deaths registered weekly in England and Wales, provisional - Office for National Statistics. https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths /bulletins/deathsregisteredweeklyinenglandandwalesprovisional/weekending7august2020. 25. Eguia, R. et al. A human coronavirus evolves antigenically to escape antibody immunity. bioRxiv 2020.12.17.423313-2020.12.17.423313 (2020) doi:10.1101/2020.12.17.423313. 26. Vousden, W. D., Farr, W. M. & Mandel, I. Dynamic temperature selection for parallel tempering in Markov chain Monte Carlo simulations. Mon. Not. R. Astron. Soc. 455 , 1919–1937 (2016). 27. COVID-19 Community Mobility Report. COVID-19 Community Mobility Report https://www.google.com/covid19/mobility?hl=en. 28. Home \| What’s the STORY? https://whatsthestory.web.ox.ac.uk/. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint 29. Ke, R., Romero-Severson, E., Sanche, S. & Hengartner, N. Estimating the reproductive number R0 of SARS-CoV-2 in the United States and eight European countries and implications for vaccination. J. Theor. Biol. 517 , (2021). 30. Edridge, A. W. D. et al. Seasonal coronavirus protective immunity is short-lasting. Nat. Med. 26 , 1691–1693 (2020). 31. Petrie, J. G. et al. Coronavirus Occurrence in the HIVE Cohort of Michigan Households: Reinfection frequency and serologic responses to seasonal and SARS coronaviruses. J. Infect. Dis. (2021) doi:10.1093/infdis/jiab161. 32. Nyaguthii, D. M. Adolescents Compared With Adults. JAMA Pediatr. (2020) et al. Infection patterns of endemic human coronaviruses in rural households in coastal Kenya. Wellcome Open Res. 6 , 27 (2021). 33. Ng, O.-W. et al. Memory T cell responses targeting the SARS coronavirus persist up to 11 years post-infection. Vaccine 34 , 2008–2014 (2016). 34. Kissler, S. M. Projecting the transmission dynamics of SARS-CoV-2 through the postpandemic period \| Science. https://science.sciencemag.org/content/368/6493/860. 35. Grifoni, A. et al. Targets of T Cell Responses to SARS-CoV-2 Coronavirus in Humans with COVID-19 Disease and Unexposed Individuals. Cell (2020) doi:10.1016/j.cell.2020.05.015. 36. Bert, N. L. et al. Different pattern of pre-existing SARS-COV-2 specific T cell immunity in SARS-recovered and uninfected individuals. bioRxiv 2020.05.26.115832-2020.05.26.115832 (2020) doi:10.1101/2020.05.26.115832. 37. Mateus, J. et al. Selective and cross-reactive SARS-CoV-2 T cell epitopes in unexposed humans. Science eabd3871–eabd3871 (2020) doi:10.1126/science.abd3871. 37. Mateus, J. et al. Selective and cross-reactive SARS-CoV-2 T cell epitopes in unexposed humans. Science eabd3871–eabd3871 (2020) doi:10.1126/science.abd3871. 38. Saletti, G. et al. Older adults lack SARS CoV-2 cross-reactive T lymphocytes directed to human coronaviruses OC43 and NL63. Sci. Rep. 10 , 21447 (2020). 38. Saletti, G. et al. Older adults lack SARS CoV-2 cross-reactive T lymphocytes directed to human coronaviruses OC43 and NL63. Sci. Rep. 10 , 21447 (2020). 39. Hicks, J. et al. Serologic Cross-Reactivity of SARS-CoV-2 with Endemic and Seasonal Betacoronaviruses. J. Clin. Immunol. (2021) doi:10.1007/s10875-021-00997-6. 39. Hicks, J. et al. Serologic Cross-Reactivity of SARS-CoV-2 with Endemic and Seasonal Betacoronaviruses. J. Clin. Immunol. (2021) doi:10.1007/s10875-021-00997-6. 40. Chia, W. N. et al. Dynamics of SARS-CoV-2 neutralising antibody responses and duration of immunity: a longitudinal study. Lancet Microbe 0 , (2021). 40. Chia, W. N. et al. Dynamics of SARS-CoV-2 neutralising antibody responses and duration of immunity: a longitudinal study. Lancet Microbe 0 , (2021). . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint 41. Chan, K. H. et al. Adolescents Compared With Adults. JAMA Pediatr. (2020) Cross-reactive antibodies in convalescent SARS patients’ sera against the emerging novel human coronavirus EMC (2012) by both immunofluorescent and neutralizing antibody tests. J. Infect. 67 , 130–140 (2013). 42. Zhao, J. et al. Airway Memory CD4+ T Cells Mediate Protective Immunity against Emerging Respiratory Coronaviruses. Immunity 44 , 1379–1391 (2016). 43. Anderson, E. M. et al. Seasonal human coronavirus antibodies are boosted upon SARS-CoV-2 infection but not associated with protection. medRxiv 2020.11.06.20227215-2020.11.06.20227215 (2020) doi:10.1101/2020.11.06.20227215. 44. Yang, R. et al. Lack of antibody-mediated cross-protection between SARS-CoV-2 and SARS-CoV infections. EBioMedicine 58 , 102890 (2020). 45. Carsetti, R. et al. The immune system of children: the key to understanding SARS-CoV-2 susceptibility? Lancet Child Adolesc. Health 4 , 414–416 (2020). 46. Pre-existing T cell-mediated cross-reactivity to SARS-CoV-2 cannot solely be explained by prior exposure to endemic human coronaviruses \| bioRxiv. https://www.biorxiv.org/content/10.1101/2020.12.08.415703v1. 47. Yaqinuddin, A. Cross-immunity between respiratory coronaviruses may limit COVID-19 fatalities. Med. Hypotheses 144 , 110049 (2020). 48. Pinotti, F. et al. Potential impact of individual exposure histories to endemic human coronaviruses on age-dependent severity of COVID-19. BMC Med. 19 , 19 (2021). 49. Callow, K. A., Parry, H. F., Sergeant, M. & Tyrrell, D. A. J. The time course of the immune response to experimental coronavirus infection of man. Epidemiol. Infect. 105 , 435–446 (1990). 50. Dugas, M. et al. Less severe course of COVID-19 is associated with elevated levels of antibodies against seasonal human coronaviruses OC43 and HKU1 (HCoV OC43, HCoV HKU1). Int. J. Infect. Dis. 105 , 304–306 (2021). 51. Neher, R. A., Dyrdak, R., Druelle, V., Hodcroft, E. B. & Albert, J. Potential impact of seasonal forcing on a SARS-CoV-2 pandemic. Swiss Med. Wkly. 150 , w20224–w20224 (2020). . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint 52. Oberemok, V. V., Laikova, K. V., Yurchenko, K. A., Fomochkina, I. I. & Kubyshkin, A. V. Adolescents Compared With Adults. JAMA Pediatr. (2020) SARS-CoV-2 will continue to circulate in the human population: an opinion from the point of view of the virus-host relationship. Inflamm. Res. 69 , 635–640 (2020). 53. Lavine, J. S., Bjornstad, O. N. & Antia, R. Immunological characteristics govern the transition of COVID-19 to endemicity. Science 371 , 741–745 (2021). 54. Respiratory infections: laboratory reports 2020 - GOV.UK. 55. Meyer, B., Drosten, C. & Müller, M. A. Serological assays for emerging coronaviruses: Challenges and pitfalls. Virus Res. 194 , 175–183 (2014). 56. Henderson, F. W., Collier, A. M., Clyde, W. A. & Denny, F. W. Respiratory-Syncytial-Virus Infections, Reinfections and Immunity. N. Engl. J. Med. 300 , 530–534 (1979). 57. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria . (2018). 58. Funk, S. socialmixr: Social Mixing Matrices for Infectious Disease Modelling. (2018). 58. Funk, S. socialmixr: Social Mixing Matrices for Infectious Disease Modelling. (2018). 59. Liu, Z., Chu, R., Gong, L., Su, B. & Wu, J. The assessment of transmission efficiency 59. Liu, Z., Chu, R., Gong, L., Su, B. & Wu, J. The assessment of transmission efficiency and latent infection period in asymptomatic carriers of SARS-CoV-2 infection. Int. J. Infect. Dis. IJID Off. Publ. Int. Soc. Infect. Dis. 99 , 325–327 (2020). 60. Tyrrell, D. A., Cohen, S. & Schlarb, J. E. Signs and symptoms in common colds. Epidemiol. Infect. 111 , 143–156 (1993). 61. Birrell, P. J. et al. Forecasting the 2017/2018 seasonal influenza epidemic in England using multiple dynamic transmission models: a case study. BMC Public Health 20 , (2020). 62. Davies, N. G. et al. Estimated transmissibility and impact of SARS-CoV-2 lineage B.1.1.7 in England. Science (2021) doi:10.1126/science.abg3055. 63. Byrne, A. W. et al. Inferred duration of infectious period of SARS-CoV-2: rapid scoping review and analysis of available evidence for asymptomatic and symptomatic COVID-19 cases. BMJ Open 10 , e039856 (2020). 64. Levin, A. T., Cochran, K. B. & Walsh, S. P. Assessing the Age Specificity of Infection Fatality Rates for COVID-19: Meta-Analysis & Public Policy Implications. medRxiv 64. Levin, A. T., Cochran, K. B. & Walsh, S. P. Assessing the Age Specificity of Infection Fatality Rates for COVID-19: Meta-Analysis & Public Policy Implications. medRxiv . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Adolescents Compared With Adults. JAMA Pediatr. (2020) (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 31, 2021. ; https://doi.org/10.1101/2021.05.27.21257032 doi: medRxiv preprint 2020.07.23.20160895-2020.07.23.20160895 (2020) doi:10.1101/2020.07.23.2016089 5. ONS. Births in England and Wales: 2019. 66. Estimates of the population for the UK, England and Wales, Scotland and Northern Ireland - Office for National Statistics. https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populatio nestimates/datasets/populationestimatesforukenglandandwalesscotlandandnorthernirelan d. 67. Vats, D. & Knudson, C. Revisiting the Gelman-Rubin Diagnostic. (2018). 68. Health, A. G. D. of. Australian Health Protection Principal Committee (AHPPC) statement on the review of physical distancing and person density restrictions. Australian Government Department of Health https://www.health.gov.au/news/australian-health-protection-principal-committee-ahppc-st atement-on-the-review-of-physical-distancing-and-person-density-restrictions (2020). 66. Estimates of the population for the UK, England and Wales, Scotland and Northern Ireland - Office for National Statistics. 67. Vats, D. & Knudson, C. Revisiting the Gelman-Rubin Diagnostic. (2018). 69. CDC. Labs. Centers for Disease Control and Prevention https://www.cdc.gov/coronavirus/2019-ncov/lab/resources/antibody-tests-guidelines.html (2020).
https://openalex.org/W2792555860	https://europepmc.org/articles/pmc5825119?pdf=render	English	null	Work of community health agents in the Family Health Strategy: meta-synthesis	Revista de saúde pública/Revista de Saúde Pública	2,018	cc-by	7,271	Correspondence: Carolina Maria do Carmo Alonso Departamento de Terapia Ocupacional Faculdade de Medicina – UFRJ Rua Prof. Rodolpho Paulo Rocco, s/n Prédio do CCS Bloco K Sala k49 Cidade Universitária 21910-590 Rio de Janeiro, RJ, Brasil E-mail: carolina.alonso@ufrj.br http://www.rsp.fsp.usp.br/ Review Review Rev Saude Publica. 2018;52:14 Work of community health agents in the Family Health Strategy: meta-synthesis Carolina Maria do Carmo AlonsoI, Pascal Daniel BéguinII, Francisco José de Castro Moura DuarteIII I Universidade Federal do Rio de Janeiro. Faculdade de Medicina. Departamento de Terapia Ocupacional. Rio de Janeiro, RJ, Brasil J , J, II Université Lumière Lyon 2. Institut D’Etudes du Travail. UMR 5600-LabEX IMU. Lyon, France II Université Lumière Lyon 2. Institut D’Etudes du Travail. UMR 5600-LabEX IMU. Lyon, France III Universidade Federal do Rio de Janeiro. COPPE. Programa de Engenharia de Produção. Rio de Janeiro, RJ, Brasil DESCRIPTORS: Community Health Workers. Family Health Strategy. Working Conditions. Health Human Resource Evaluation. Review. ABSTRACT OBJECTIVE: To systematize and analyze the evidence from qualitative studies that address the perception of Brazilian Community Health Agents about their work. METHODS: This is a systematic review of the meta-synthesis type on the work of community health agents, carried out from the Virtual Health Library using the descriptors “Agente Comunitário de Saúde” and “Trabalho”, in Portuguese. The strategy was constructed by crossing descriptors, using the Boolean operator “AND”, and filtering Brazilian articles, published from 2004 to 2014, which resulted in 129 identified articles. We removed quantitative or quanti-qualitative research articles, essays, debates, literature reviews, reports of experiences, and research that did not include Brazilian Community Health Agents as subjects. Using these criteria, we selected and analyzed 33 studies that allowed us to identify common subjects and differences between them, to group the main conclusions, to classify subjects, and to interpret the content. RESULTS: The analysis resulted in three thematic units: characteristics of the work of community health agents, problems related to the work of community health agents, and positive aspects of the work of community health agents. On the characteristics, we could see that the work of the community health agents is permeated by the political and social dimensions of the health work with predominant use of light technologies. The main input is the knowledge that this professional obtains with the contact with families, which is developed with home visits. On the problems in the work of community health agents, we could identify the lack of limits in their attributions, poor conditions, obstacles in the relationship with the community and teams, weak professional training, and bureaucracy. The positive aspects we identified were the recognition of the work by families, resolution, bonding, work with peers, and work close to home. CONCLUSIONS: This review provided an overview of the difficulties and positive aspects that are present in the daily work of community health agents. Given this, we have raised two challenges. The first one refers to how public policy makers need to appropriation the research results and the second one refers to the need to invest in studies that are designed to generate solutions for the difficulties faced by community health agents in their work. Received: Dec 2, 2016 Approved: Mar 9, 2017 How to cite: Alonso CMC, Béguin PD, Duarte FJCM. Work of community health agents in the Family Health Strategy: meta-synthesis. Rev Saude Publica. 2018;52:14. DESCRIPTORS: Community Health Workers. How to cite: Alonso CMC, Béguin PD, Duarte FJCM. Work of community health agents in the Family Health Strategy: meta-synthesis. Rev Saude Publica. 2018;52:14. INTRODUCTION According to Tomaz38, the responsibilities of Brazilian Community Health Agents (CHA) can be summarized as the identification of risk situations, guidance for families and community, and referral of identified risk cases and situations to other members of health teams. This means that the work of CHA is to assist in the planning and implementation of health actions both locally, by forwarding information from the health territory to the Family Health Strategy (FHS), and nationally, by feeding data to information systems of the Ministry of Health29. This shows that CHA play an important role in the expansion and consolidation of primary health care (PHC) and are the subject of studies, among which we highlight, in this article, those focused on the investigation of how they perceive their work. The relevance of a study on the perspective of workers is related to how services are conceived in a sphere in which players are thought abstractly, with generic characteristics, elaborated from theoretical conceptions. Until the service starts, workers and users will have few opportunities to position themselves on the rules and to try the process to verify difficulties or problems in the operation8,16,34. In the case of public health services, such as the FHS, this situation is intensified by the distance between the front line and the regulatory agency. This distance hinders the identification of problems that occur in the operation of the service by those who could solve them, while at the same time workers are unaware of what managers expect from the actions at the front line. Such condition results in a mutual incomprehension that, in turn, reduces the chances of correcting eventual gaps in the service project8,34. Thus, research studies that have explored the perception of front-line workers, such as CHA, expose the problems of work inadequacy caused by production system projects, processes, work organization, and tasks made from simplifying stereotypes27. In this perspective, this article aims to systematize and analyze evidence from qualitative studies, carried out from 2004 to 2014, which discuss the perception of CHA about their work, offering evidence that can support the improvement of their work process. Work of CHA: meta-synthesis Alonso CMC et al. ABSTRACT Family Health Strategy. Working Conditions. Health Human Resource Evaluation. Review. Copyright: This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided that the original author and source are credited. 1 https://doi.org/10.11606/S1518-8787.2018052000395 1 https://doi.org/10.11606/S1518-8787.2018052000395 Work of CHA: meta-synthesis Alonso CMC et al. INTRODUCTION METHODS This article is a meta-synthesis review on studies on Brazilian CHA linked to the FHS, published between 2004 and 2014, which were selected from the Virtual Health Library (BIREME) that brings together Health Sciences databases, such as the Latin American and Caribbean Literature in Health Sciences (LILACS), the Scientific Electronic Library Online (SciELO), and the Medical Literature Analysis and Retrieve System Online (Medline). We chose this model because the meta-synthesis is a methodological approach used for the rigorous study of qualitative conclusions, whose interpretations and redefinitions result in the (re)conceptualization of the original conclusions11, thus meeting the objective of this article. The guiding question for this review was: what are the evidences found by qualitative studies on the work of Brazilian CHA who are linked to the FHS, from 2004 to 2014? To this end, we searched BIREME in March 2015 using the descriptors “Agente Comunitário de Saúde” and “Trabalho”, in Portuguese, crossed from the Boolean operator AND, and we filtered Brazilian articles published between 2004 and 2014. Thus, we retrieved 129 articles, being 124 indexed in the Lilacs database and five in the Medline database. This result was exported to the Mendeley reference manager, which detected six duplicate documents. Next, we started the screening process, which was carried out by the lead Work of CHA: meta-synthesis Alonso CMC et al. author with the supervision of the co-authors to minimize biases from the presence of only one evaluator. We screened the articles that presented original qualitative results and that addressed the work of CHA linked to the FHS. author with the supervision of the co-authors to minimize biases from the presence of only one evaluator. We screened the articles that presented original qualitative results and that addressed the work of CHA linked to the FHS. The exclusion criteria considered the type of methodological approach used in studies, study subjects, and nature of the article. Thus, we removed quantitative or quanti-qualitative research articles, essays, debates, literature reviews, reports of experiences, as well as those that did not include CHA as research subjects. Following these criteria, the articles were selected in three stages: analysis of the titles, analysis of abstracts, and analysis of full texts. Thus, we removed 46 articles analyzing the title, 34 analyzing the abstract, and we fully analyzed 44 of them. In this last phase, we also removed 11 articles. ANALYSIS OF RESULTS The Box summarizes the 33 articles selected and their main results. We verified the following distribution of these studies among the Brazilian regions: twenty in the Southeast, seven in the Northeast, three in the South, two in the Midwest, and one that covered the North, Midwest, and Southeast at the same time. Regarding this aspect, we highlight a discrepancy in the number of studies carried out in the Southeast region compared to the other Brazilian regions, which is consistent with an article that has shown that health research in Brazil is concentrated in the Southeast region14. The analysis of the results of the articles produced three thematic categories, presented below, that became the axis to understand the problem explored in this meta-synthesis. METHODS The Figure shows the flow of article selection, whose result was the corpus of this meta-synthesis. Characteristics of the Work of CHA The work of CHA is more permeated by the political and social dimensions of health work6,31, with predominant use of light technologies, such as communication6,7, reception and bond1,5,7,9,17, dialog4, and listening1,4,26. BIREME: Latin-American and Caribbean Center on Health Sciences Information; CHA: Community Health Agent Figure Flow of the selection process of the articles in the different phases of the meta synthesis of the 33 articles selected 80 articles removed after reading the title/abstract 5 articles removed after reading the full text 5 articles removed during data extraction 43 articles retrieved Article addressing the perception of users on the work of CHA or that focused on the work of CHA in certain fields, such as Oral Health, Occupational Health or control of endemic diseases. Articles of quantitative or quanti-qualitative research, essays, discussions, reviews of the literature, reports of experiences, as well as those that did not include CHA as research subjects. 123 non-duplicate articles BIREME 2004/2014 129 articles retrieved 33 articles selected 80 articles removed after reading the title/abstract 5 articles removed after reading the full text 5 articles removed during data extraction 43 articles retrieved Article addressing the perception of users on the work of CHA or that focused on the work of CHA in certain fields, such as Oral Health, Occupational Health or control of endemic diseases. Articles of quantitative or quanti-qualitative research, essays, discussions, reviews of the literature, reports of experiences, as well as those that did not include CHA as research subjects. 123 non-duplicate articles BIREME 2004/2014 129 articles retrieved 80 articles removed after reading the title/abstract 5 articles removed after reading the full text 5 articles removed during data extraction BIREME: Latin-American and Caribbean Center on Health Sciences Information; CHA: Community Health Agent Work of CHA: meta-synthesis Alonso CMC et al. Box. Summary of the information from selected studies for the meta-synthesis of the perception of Community Health Aagents about their work. Characteristics of the Work of CHA Authors/year Region Subjects of the research Procedure for data collection Subject Baralhas M, Pereira MAO1 (2011) Southeast CHA Interview Representation of CHA on their care practices Barbosa RHS, Menezes CAF, David HMSL, Bornstein VJ2 (2012) Southeast CHA Focus group Relationship between the gender dimension and the work of CHA Binda JB, Bianco MF, Sousa EM3 (2013) Southeast CHA Observation, discussion groups, and interview Analysis of the work processes of CHA Bornstein VJ, Stotz EM4 (2008) Southeast CHA, physicians, and nurses of FHS Interviews, documentary analysis, and participant observation Characterization of the types of mediation present in the work of CHA Carli R, Costa MC, Silva EB, Resta DG, Colomé ICS5 (2014) South CHA Interview Perceptions of CHA on the reception and bonding practices Coriolano MWL, Lima LS6 (2010) Northeast CHA Focus group Description of the work process of CHA Costa MC, Silva EB, Jahn AC, Resta DG, Colom ICS, Carli R7 (2012) South CHA Interview Analysis of the work process of CHA Ferreira VSC, Andrade CS, Franco TB, Merhy EE9 (2009) Northeast CHA Interview and focus group Production of care, work process, technologies, and restructuring of the productive process of CHA Filgueiras AS, Silva ALA10 (2011) Southeast CHA Interview Discussion of the facilitating and limiting aspects of the activities assigned to CHA Fonseca AF, Machado FRS, Bornstein VJ, Pinheiro R12 (2012) North, Midwest, and Southeast CHA, work supervisors of CHA, and service managers Interview and focus group Analysis of the processes of evaluation of the work of CHA Galavote HS, Franco TB, Lima RCD, Belizário AM13 (2013) Southeast CHA Interview and observation Work process of CHA Gomes AL, Lima Neto PJ, Silva VLA, Silva EF15 (2011) Northeast CHA Interview and observation Relationship between the organization and the work process and mental health of CHA Jardim TA, Lancman S17 (2009) Southeast CHA Group of Work Psychodynamics Discussion about living and working in the same community Lara MO, Brito MJM, Rezende LC18 (2012) Southeast CHA, nurse, physician, nursing assistant, and users Interviews Analysis of the influence of the cultural practices present in the work of CHA Lima AP, Corrêa ACP, Oliveira QC19 (2012) Midwest CHA Interview and participant observation Identification of the knowledge of CHA about PCIS instruments/records Lopes DMQ, Beck CLC, Prestes FC, Weiller TH, Colomé JS, Silva GM20 (2012) South CHA Focus group Suffering and pleasure in the work of CHA Martines WRV, Chaves EC21 (2007) Southeast CHA Interview and observation Representations of CHA about the vulnerabilities for suffering at work and the manifestations Relationship between the organization and the work process and mental health of CHA Discussion about living and working in the same community Analysis of the influence of the cultural practices present in the work of CHA Identification of the knowledge of CHA about PCIS instruments/records Suffering and pleasure in the work of CHA Representations of CHA about the vulnerabilities for suffering at work and the manifestations of this suffering in the performance of work Continue Work of CHA: meta-synthesis Alonso CMC et al. Characteristics of the Work of CHA Box. Summary of the information from selected studies for the meta-synthesis of the perception of Community Health Aagents about their work. Continuation Nascimento GM, David HMSL22 (2008) Southeast CHA Participant observation Development of an instrument for risk assessment of the work of CHA Oliveira AR, Chaves AEP, Nogueira JA, Sá LD, Collet N23 (2010) Northeast CHA Questionnaire Research on the satisfaction and limitation in the daily work of CHA Oliveira DT, Ferreira PJO, Mendonça LBA, Oliveira HS24 (2012) Northeast CHA Interview Perception of CHA about their work process Peres CRFB, Caldas Júnior AL, Silva RF, Marin MJS25 (2011) Southeast CHA Interview Analysis of the difficulties and facilities of CHA regarding teamwork. Characteristics of the Work of CHA Pinheiro RL, Guanaes-Lorenzi C26 (2014) Southeast CHA Groups of discussion Perception of CHA about their practices with social networks Queirós AAL, Lima LP28 (2012) Northeast CHA, managers, legislators, social movement, movement of CHA and popular movement, researcher Interview Analysis of the social practice of the work of CHA Rosa AJ, Bonfanti AL, Carvalho CS30 (2012) Midwest CHA and users Groups of work psychodynamics, interview, and observation Analysis of the relationship between work and psychological distress of CHA Sakata KN, Mishima SM31 (2012) Southeast CHA, dental office assistant, nursing assistant, dental surgeon, nurse, physician, and health unit manager Interview and participant observation Analysis of social relationships between CHA and family health teams Santos LFB, David HMSL32 (2011) Southeast CHA Interview Identification of the occupational stress factors reported by CHA and analysis of their relation with possible health effects Schmidt MLS, Neves TFS33 (2010) Southeast CHA Focus groups Analysis of the aspects of the implementation of the Family Health Program (FHP) as a primary public policy in the struggle with the hegemonic medical-care model in the country Sossai LCF, Pinto IC, Mello DF36 (2010) Southeast CHA and users Interviews Presentation of the work of CHA from the perspective of users and CHA Souza LJR, Freitas MC37 (2011) Northeast CHA Interview and participant observation Analysis of workplace violence in CHA Trapé CA, Soares CB39 (2007) Southeast CHA Focus groups and interviews Conceptions of health education present in the work of CHA Vilela RAG, Silva RC, Jackson Filho JM40 (2010) Southeast CHA and managers Interview and observation Analysis of the relationship between complaints of suffering and the conditions of work of CHA and proposed measures to modify them Wai MFP, Carvalho AMP41 (2009) Southeast CHA Interview Perception of CHA on events that cause overload and coping strategies used by them. Zanchetta MS, Leite LS, Perreault M, Lefebvre H42 (2005) Southeast CHA Individual and group interviews. Observation Perception of CHA on barriers in the work practice CHA: Community Health Agents; PCIS: Primary Care Information System; FHS: Family Health Strategy Box. Summary of the information from selected studies for the meta-synthesis of the perception of Community Health Aagents about their work. Continuation Box. Summary of the information from selected studies for the meta-synthesis of the perception of Community work. Continuation CHA: Community Health Agents; PCIS: Primary Care Information System; FHS: Family Health Strategy 5 https://doi.org/10.11606/S1518-8787.2018052000395 Work of CHA: meta-synthesis Alonso CMC et al. Problems Related to the Work of CHA The main challenges related to the work of CHA evidenced in the studies are described below. The main challenges related to the work of CHA evidenced in the studies are described below. Lack of limits in the work of CHA Lack of limits in the work of CHA Characteristics of the Work of CHA Some studies4,9 point out that the main input of the work of CHA is the knowledge obtained in the contact with families and, as far as this contact is concerned, research studies state that home visits are a privileged stage for its development1,4,5,10. Vilela et al.40 state that home visits are primary activities to build the relationship between CHA and users, and they also represent the main means to promote the health of the surrounding community. From the managerial point of view, home visits are valued operations that count as production of the unit. One study12 has specifically addressed the evaluation of the work of CHA, evidencing that this evaluation is based on a quantitative bias based on the achievement of goals that predominantly encompass tasks related to the biomedical field. Problems Related to the Work of CHA Lack of limits in the work of CHA Thirteen articles have addressed the lack of limits of the actions of CHA summarized in two issues: idealization of the work and lack of limits of the attributions of this professional. Regarding the idealization of the work, studies have shown that CHA perceive as their mission the solution of all the issues of the families and the community they serve, idealizing their practices and disregarding the importance of other resources to carry out their actions1,21. Studies show that CHA understand their work as a vocation based on values of friendship, solidarity, voluntariness, and charity, which are based on the perception about their attributions as elements that contribute to the difficulty in delimiting their role1,2,21,36,42. The lack of limits of the work of CHA can be understood both in studies that highlight the lack of clarity of their duties20,21 and in studies that emphasize the excess of planned functions6,15,20,22,23,30. For the excess of functions, the CHA perform tasks that do not concern them, such as: work at the reception28,36,40, scheduling of appointments1,15,28,36, organization of folders and medical records36, control of materials and warehouse and cleaning service36, delivery of referrals to specialists15,31, sending of messages from the health service to users31, and feeding of children in the absence of parents42. In addition to the tasks that are not the responsibility of CHA, other functions are gradually incorporated into their role, officially, such as the weighing of families for their re-registration in the government program Bolsa Família and actions to prevent and combat dengue fever15. This framework reinforces that CHA are seen as a multipurpose workers, given the lack of definition of the limits of their professional duties and the idealization of the role, and thus their scope of action is constantly extended. Work of CHA: meta-synthesis Alonso CMC et al. Relationship with the community As for the relationship of CHA with the community, the most reported problems were: obligation to live in the community where they work, coexistence with community problems, exposure to violence, and relationship with users. The CHA need to live in the community they serve, which is a problem as these workers have their private life exposed when sought outside working hours, on weekends, and in living spaces in the neighborhood, such as church and fairs, thus overloading them6,10,17,20,41. In this way, these professionals have an uninterrupted involvement with the users of the health service. Jardim and Lancman17 emphasize that, just as CHA enter the private world of users, the private world of these workers is also invaded by the community and its problems, as they are literally unable to keep a distance from the population for which they are responsible. The second subject of this category addresses studies that point to the fact that CHA generally work with poor populations living in peripheral regions. In these regions, social problems are more acute, increasing the emotional load related to work, both because of the difficulty in managing these issues and because CHA often share the same problems as they also live in these communities4,6,17,20,23,41. The third subject related to the relationship between CHA and the community refers to the exposure of these workers to violence. Studies carried out in large urban centers have shown that these professionals work in regions where violence is pronounced, especially that related to organized crime and drug trafficking32,37,42. Therefore, because they work and live in the same territory, CHA are more vulnerable to situations of conflict compared to the other FHS workers who do not circulate as much. In this sense, studies6,37 state that CHA generally need to build strategies to cope with these situations, corroborating the findings of Jardim and Lancman17, who report that CHA avoid reporting to the police or Child Protective Council for fear for their safety. In line with this issue, the problem of the relationship between CHA and users is manifested. According to Bornstein and Stotz4, CHA perceive that users expect a more viable access to health services, which is similar to the result of other studies10,17. Precarious working conditions Precarious working conditions Another problem related to the activities of CHA is their precarious working conditions, identified in thirteen articles: explicitness of the fragility of the employment relationship1,3,33, exposure to work hours that go beyond the opening hours of the health unit and which invades the private life1,6,7,17,40,41, care for a greater number of families than what is advocated35, exposure to unhealthy working conditions30,41, low pay and absence of social protection1,4,7,30, poor recognition of work by managers, peers, and users2,12,20,24,26,32, and precariousness of the system17,40. As a consequence of this precariousness, two studies have shown that CHA view the profession as a temporary activity, since there is no prospect of transformation of the work situations or a planned career plan20,30. The precariousness of the work of CHA is also related to the fragility of the health system, which cannot adequately meet the demand of users, with a lack of opening for appointments Work of CHA: meta-synthesis Alonso CMC et al. and examinations, materials, and medications. This situation affects the relationship between CHA and the community they serve, since many users attribute to them the administration of the lack of resources of the health unit and the system, considering their responsibility for the front line in the care of the population17,40. Relationship with the community However, when these professionals cannot respond to the demands of users for appointments, medications, exams, or access to other services, users lose their trust on CHA without considering that the lack of access is a matter of how the system works1–5,7,22,23,30,32,33,40,41. Jardim and Lancman17 have found that the credibility of CHA in the community is directly associated to the resolution of the demands of users and this credibility can be hampered by aspects related to the structuring of the service and the inoperability of the health system. y that the credibility of CHA in the community is directly associated to the resolution of the demands of users and this credibility can be hampered by aspects related to the structuring of the service and the inoperability of the health system. In this sense, another dimension that is related to this issue has been identified by Santos and David32 and refers to the violence present in the work of CHA, which emerges from the relationship of these professionals with the community. These authors have observed that the frustration of users to demands not met can also turn into verbal aggression or intense psychological pressure. Fragility in the professional qualification of CHA Bureaucratization of the work The bureaucratization of the work of CHA is another evidence highlighted in the studies included in this meta-synthesis, and it is related, in the perception of these workers, to the collection of data on the surrounding population, especially those related to diet and use of the Primary Care Information System (PCIS)19. In this sense, authors19 state that CHA have difficulties in identifying, naming, and describing the records they need to fill in to feed the PCIS, and they also do not understand the variables, terms, and pathologies that make up these instruments. As a consequence, tasks related to health surveillance, notwithstanding their importance to the FHS, focus on the work of CHA as mere statistical data collection activities, with little meaning for these workers1,32,36,40. Other aspects regarding the perception of bureaucratization were listed in the study of Nascimento and David22, which indicate the presence of the following characteristics in the work process of CHA: standardization, great number of prescriptions, and organization according to the logic of process divisions and hierarchy. In this perspective, bureaucratization starts to permeate even the tasks that are not administrative, given the predominance of the logic to count procedures, instead of assessing the quality of the care provided9,12,19. Problems in the relationship with the team Teamwork was identified as a limiting factor in the work of CHA, when there is a lack of articulation with other professionals and an inflexibility of relationships produced by the work organization that hinders the exchange between players outside the spaces established for this end7,13,41,42. Peres et al.25, when analyzing the perceptions of CHA on teamwork, indicate that these workers feel as the weakest link in the relationship with other FHS professionals. Such situation is in line with the findings of other authors3,9,40 who have identified a hierarchy of knowledge in the FHS that gives to CHA an unequal role in the planning and decision making of interventions. When analyzing the work of CHA in the interface with the teams, a dysfunction was found in the relation between CHA and teams, from the different ways of approaching the problems of users. In this sense, it is common for teams not to value what CHA raise as a priority in meeting the needs of users40. Fragility in the professional qualification of CHA The CHA consider their professional training insufficient and the main perceived gaps were: excess of standardization of contents that address predominantly technical-scientific subjects and that do not include data on the local reality, lack of focus on theoretical and practical aspects that could assist them in coping with issues of the daily work, such as Work of CHA: meta-synthesis Alonso CMC et al. the management of family and social problems, and, finally, restriction of the work hours offered for such activity1,4,7,10,19,36,39. the management of family and social problems, and, finally, restriction of the work hours offered for such activity1,4,7,10,19,36,39. Facilitating Aspects of the Work of CHA Facilitating Aspects of the Work of CHA We describe below the positive aspects that act as facilitators of the work of CHA identified by the selected studies. Bureaucratization of the work In addition, the organization of work in the FHS is guided by the achievement of goals, excess of activities, and little availability of time for the exchange between professionals, which hinders the establishment of teamwork and is negatively reflected on the work of CHA31,41,42. Other perceived difficulties are caused by “personal differences; difficulty in visualizing all actions; lack of flexibility, communication, cooperation, responsibility, and horizontality of actions” (Peres et al.25, p.908). Working together with peers Peer-to-peer activities were identified as facilitators when they were established by the positive relationships between CHA and other staff members, allowing the horizontal discussion of daily problems and allowing the sharing of work strategies9,25,31. Such a construction is favored in team meetings from the perspective of CHA24. In fact, according to Filgueiras and Silva10, the positive aspects of teamwork are the support for the actions of CHA, as the integration with the other professionals allows the collective construction of coping mechanisms for the problems identified. Formal work close to the residence Formal work close to the residence The last positive trait in the work of CHA identified in the studies analyzed concerns the formal employment relationship. This situation has been evidenced in the study by Sossai et al.36, which show the rare presence of job vacancies that comply with the norms of the Labor Law Consolidation in the region of the study. This distinguishes the work of CHA from the other workers, as CHA can combine formal employment with living close to their residence, meeting the findings of Zanchettaet al.42 In addition, two other studies2,3 have shown that work close to home is perceived as advantageous, especially for women, who are most of the CHA. This situation enables the combination of the care of the family with the professional work. Bonding with families and community Bonding with families and community Bonding with families and community Seven studies included in this meta-synthesis show that CHA perceive the bond with the user as a necessary condition for their work to happen, because, according to some authors1,5,13,31,42, this bond is intimately related to the mission of being the link between health professionals and the community. The construction of the trust and credibility of CHA among users has been identified in five studies1,5,17,18,31, as dimensions that favor the establishment of the bond. This is because they allow the worker to approach problems that extrapolate the biological dimension, despite being part of the health-disease process, such as situations of family or community conflict, domestic violence, poverty, sexual abuse, child neglect, ill-treatment of older adults, trafficking, and use of drugs. Such situations are complex and they cannot always be accessed without building a trust relationship between users and CHA. Recognition of the work by families/community and resolubility The most recurrent positive aspect in this meta-synthesis was the recognition of the families in relation to the work of CHA. Thus, several authors1,6,7,13,23,36,42 have pointed out that CHA Work of CHA: meta-synthesis Alonso CMC et al. feel satisfied with their work when they see that they are useful to the community, when there has been a change in the health conditions of users, or when families recognize their competence and commitment. Two studies2,20 reinforce these findings, adding that the recognition of the assisted population is a motivating factor for CHA, strengthening their self-esteem and contributing to the conformation of their identity. The resolubility of CHA has been identified in the study by Lopes et al.20 as “possibility to solve the problems of users and to verify that the work carried out is improving the health conditions of the community” (p.635). In this context, resolubility is identified as a positive aspect related to the work of CHA, since this characteristic is linked to the feeling of gratification and utility, thus contributing to the professional satisfaction of these workers23,36. DISCUSSION We aimed to collect and analyze evidence on the work process of CHA with this meta-synthesis. We analyzed 33 articles, which allow us to perceive the characteristics of the profession, as well as problems and positive aspects related to professional work. In this sense, we emphasize that the academic production that addresses the subject of the work of CHA is useful in pointing this evidence. 9 https://doi.org/10.11606/S1518-8787.2018052000395 Work of CHA: meta-synthesis Alonso CMC et al. The duties of CHA need to be reviewed in order to better define their role and size their actions according to the resources available, avoiding, above all, the abuse of functions. We need to overcome the precariousness of the work of CHA by building a career plan that values not only the knowledge accumulated over time but also the strategic role they play in the consolidation of the FHS. Another important issue raised in this meta-synthesis refers to the need to rethink the organization of work within the FHS, incorporating demands related to the specificities of the work of CHA. In this sense, the organization of the work of FHS teams must be reviewed to allow CHA to have a space of dialog strengthened with the other members of the team. In fact, horizontal teamwork, with integration among members, is positively reflected in the work of CHA. Regarding the organization of the work, we can observe the need to rethink the bureaucratic tasks inherent to CHA. This would be done, on the one hand, by giving a meaning to the filling of the reports used to collect population data, because as Lima et al.19 have observed, the CHA have difficulty recognizing the usefulness of these instruments. On the other hand, we can mention that although the work in the FHS and, in particular the work of CHA, is based on an expanded concept of health, in the principles of the PHC which is parametrizated qualitatively, we can observe the prevalence of the normative logic that organizes the work hierarchically and using processes. This fact is also reflected in how the work done is evaluated, which is given by the counting of procedures and not by the quality of the care. DISCUSSION Given this situation, we need to urgently rethink models that evaluated the work of CHA that are in line with the tools and values recommended for the FHS while also contemplating the specificities of the profession. Regarding the issue of living with the community problems, the fact that CHA live in the same territory where they work is a challenge. Although this exposes workers to violence and to the invasion of their private life, the most significant set of knowledge of the CHA is structured in the contact with the community. In addition, we observed that the work close to the residence also benefits these workers. Regarding the fragility in the training of the CHA, Ordinance 253, of September 25, 2015, recently established the Introductory Course for CHA, which standardized the minimum work hours for this training and defined the basic curricular components. In conclusion, regarding the distribution of research in Brazil, we verified a predominance of studies carried out in the Southeast region and in large urban centers. In this sense, new studies can be carried out in Brazilian regions with different realities, such as those that attend to riverine or rural populations, where the work of CHA assumes other characteristics. Regarding the meta-synthesis carried out, we presented an overview of the difficulties and positive aspects that are present in the daily work of CHA. Given this, this review posed two challenges for future research. The first one refers to how public policy makers need to appropriate the research results and the second one refers to the need to invest in studies that are designed to generate solutions for the difficulties faced by CHA in their work. by CHA in their work. 1. Baralhas M, Pereira MAO. Concepções dos agentes comunitários de saúde sobre suas práticas assistenciais. Physis. 2011;21(1):31-46. https://doi.org/10.1590/S0103-73312011000100003 2. Barbosa RHS, Menezes CAF, David HMSL, Bornstein VJ. Gênero e trabalho em saúde: um olhar crítico sobre o trabalho de agentes comunitárias/os de saúde. Interface (Botucatu). 2012;16(42):751-65. https://doi.org/10.1590/S1414-32832012000300013 1. Baralhas M, Pereira MAO. Concepções dos agentes comunitários de saúde sobre suas práticas assistenciais. Physis. 2011;21(1):31-46. https://doi.org/10.1590/S0103-73312011000100003 REFERENCES 10 https://doi.org/10.11606/S1518-8787.2018052000395 REFERENCES 1. Baralhas M, Pereira MAO. Concepções dos agentes comunitários de saúde sobre suas práticas assistenciais. Physis. 2011;21(1):31-46. https://doi.org/10.1590/S0103-73312011000100003 2. Barbosa RHS, Menezes CAF, David HMSL, Bornstein VJ. Gênero e trabalho em saúde: um olhar crítico sobre o trabalho de agentes comunitárias/os de saúde. Interface (Botucatu). 2012;16(42):751-65. https://doi.org/10.1590/S1414-32832012000300013 Work of CHA: meta-synthesis Alonso CMC et al. Work of CHA: meta-synthesis Alonso CMC et al. 3. Binda JB, Bianco MF, Sousa EM. O trabalho dos agentes comunitários de saúde em evidência: uma análise com foco na atividade. Saude Soc. 2013;22(2):389-402. https://doi.org/10.1590/S0104-12902013000200011 4. Bornstein VJ, Stotz EN. O trabalho dos agentes comunitários de saúde: entre a mediação convencedora e a transformadora. Trab Educ Saude. 2008;6(3):457-80. https://doi.org/10.1590/S1981-77462008000300004 5. Carli R, Costa MC, Silva EB, Resta DG, Colomé ICS. Acolhimento e vínculo nas concepções e práticas dos agentes comunitários de saúde. Texto Contexto Enferm. 2014;23(3):626-32. https://doi.org/10.1590/0104-07072014001200013 6. Coriolano MWL, Lima LS. Grupos focais com agentes comunitários de saúde: subsídios para entendimento destes atores sociais. Rev Enferm UERJ. 2010 [citado 19 out 2017];18(1):92-6. Disponível em: http://www.facenf.uerj.br/v18n1/v18n1a16.pdf 7. Costa MC, Silva EB, Jahn AC, Resta, DG, Colom ICS, Carli R. Processo de trabalho dos agentes comunitários de saúde: possibilidades e limites. Rev. Gaucha Enferm. 2012;33(3):134-40. https://doi.org/10.1590/S1983-14472012000300018 8. Derani C. Privatizações e serviços públicos: as ações do Estado na produção econômica. São Paulo: Max Limonad; 2002. 9. Ferreira VSC, Andrade CS, Franco TB, Merhy EE. Processo de trabalho do agente comunitário de saúde e a reestruturação produtiva. Cad Saude Publica. 2009;25(4):898-906. https://doi.org/10.1590/S0102-311X2009000400021 10. Filgueiras AS, Silva ALA. Agente Comunitário de Saúde: um novo ator no cenário da saúde do Brasil. Physis. 2011;21(3):899-916. https://doi.org/10.1590/S0103-73312011000300008 11. Finfgeld DL. Meta-synthesis: the state of the art - so far. Qual Health Res. 2003;13(7):893-904. https://doi.org/10.1177/1049732303253462 12. Fonseca AF, Machado FRS, Bornstein VJ, Pinheiro R. Avaliação em saúde e repercussões no trabalho do agente comunitário de saúde. Texto Contexto Enferm. 2012;21(3):519-27. https://doi.org/10.1590/S0104-07072012000300005 13. Galavote HS, Franco TB, Lima RCD, Belizário AM. Alegrias e tristezas no cotidiano de trabalho do agente comunitário de saúde: cenários de paixões e afetamentos. Interface (Botucatu). 2013;17(46):575-86. https://doi.org/10.1590/S1414-32832013005000015 14. Guimarães R. Pesquisa em saúde no Brasil: contexto e desafios. Rev Saude Publica. 2006;40 No Espec:3-10. https://doi.org/10.1590/S0034-89102006000400002 15. Gomes AL, Lima Neto PJ, Silva VLA, Silva EF. REFERENCES O elo entre o processo e a organização do trabalho e a saúde mental do agente comunitário de saúde na Estratégia Saúde da Família no município de João Pessoa – Paraíba- Brasil. Rev Bras Cienc Saude. 2011;15(3):265-76. https://doi.org/10.4034/RBCS.2011.15.03.02 16. Gomes RS. O trabalho no Programa Saúde da Família do ponto de vista da atividade: a potência, os dilemas e os riscos de ser responsável pela transformação do modelo assistencial [tese]. Rio de Janeiro: Escola Nacional de Saúde Pública Sergio Arouca; 2009. 17. Jardim TA, Lancman S. Aspectos subjetivos do morar e trabalhar na mesma comunidade: a realidade vivenciada pelo agente comunitário de saúde. Interface (Botucatu). 2009;13(28):123-35. https://doi.org/10.1590/S1414-32832009000100011 18. Lara MO, Brito MJM, Rezende LC. Aspectos culturais das práticas dos agentes comunitários de saúde em áreas rurais. Rev Esc Enferm USP. 2012;46(3):673-80. https://doi.org/10.1590/S0080-62342012000300020 19. Lima AP, Corrêa ACP, Oliveira QC. Conhecimento de agentes comunitários de saúde sobre os instrumentos de coleta de dados do SIAB. Rev Bras Enferm. 2012;65(1):121-7. https://doi.org/10.1590/S0034-71672012000100018 20. Lopes DMQ, Beck CLC, Prestes FC, Weiller TH, Colomé JS, Silva GM. Agentes comunitários de saúde e as vivências de prazer – sofrimento no trabalho: estudo qualitativo. Rev Esc Enferm USP. 2012;46(3):633-40. https://doi.org/10.1590/S0080-62342012000300015 21. Martines WRV, Chaves EC. Vulnerabilidade e sofrimento no trabalho do agente comunitário de saúde no Programa de Saúde da Família. Rev Esc Enferm USP. 2007;41(3):426-33. https://doi.org/10.1590/S0080-62342007000300012 Work of CHA: meta-synthesis Alonso CMC et al. Work of CHA: meta-synthesis Al 22. Nascimento GM, David HMSL. Avaliação de riscos no trabalho dos agentes comunitários de saúde: um processo participativo. Rev Enferm UERJ. 2008 [citado 19 out 2017];16(4):550-6. Disponível em: http://www.facenf.uerj.br/v16n4/v16n4a16.pdf 23. Oliveira AR, Chaves AEP, Nogueira JA, Sá LD, Collet N. Satisfação e limitação no cotidiano de trabalho do agente comunitário de saúde. Rev Eletr Enferm. 2010;12(1):28-36. https://doi.org/10.5216/ree.v12i1.9511 24. Oliveira DT, Ferreira PJO, Mendonça LBA, Oliveira HS. Percepções do agente comunitário de saúde sobre sua atuação na Estratégia Saúde da Família. Cogitare Enferm. 2012;17(1):132-7. https://doi.org/10.5380/ce.v17i1.26386 25. Peres CRFB, Caldas Júnior AL, Silva RF, Marin MJS. O agente comunitário de saúde frente ao processo de trabalho em equipe: facilidades e dificuldades. Rev Esc Enferm USP. 2011;45(4):905-11. https://doi.org/10.1590/S0080-62342011000400016 26. Pinheiro RL, Guanaes-Lorenzi C. Funções do agente comunitário de saúde no trabalho com redes sociais. Estud Psicol (Natal). 2014;19(1):48-57. https://doi.org/10.1590/S1413-294X2014000100007 27. Pizo CA, Menegon NL. Análise ergonômica do trabalho e o reconhecimento científico do conhecimento gerado. Produçao. 2010;20(4):657-68. https://doi.org/10.1590/S0103-65132010005000058 28. Queirós AAL, Lima LP. 41. Wai MFP, Carvalho AMP. O trabalho do agente comunitário de saúde: fatores de sobrecarga e estratégias de enfrentamento. Rev Enferm UERJ. 2009 [citado 19 out 2017];17(4):563-8. Disponível em: http://www.facenf.uerj.br/v17n4/v17n4a19.pdf 42. Zanchetta MS, Leite LS, Perreault M, Lefebvre H. Educação, crescimento e fortalecimento profissional do agente comunitário de saúde: estudo etnográfico. Online Braz J Nurs. 2005;4(3). https://doi.org/10.5935/1676-4285.200535 REFERENCES A institucionalização do trabalho do agente comunitário de saúde. Trab Educ Saude. 2012;10(2):257-81. https://doi.org/10.1590/S1981-77462012000200005 29. Rodrigues AAAO, Santos AM, Assis MMA. Agente comunitário de saúde: sujeito da prática em saúde bucal em Alagoinhas, Bahia. Cienc Saude Coletiva. 2010;15(3):907-15. https://doi.org/10.1590/S1413-81232010000300034 30. Rosa AJ, Bonfanti AL, Carvalho CS. O sofrimento psíquico de agentes comunitários de saúde e suas relações com o trabalho. Saude Soc. 2012;21(1):141-52. https://doi.org/http://dx.doi.org/10.1590/S0104-12902012000100014 31. Sakata KN, Mishima SM. Articulação das ações e interação dos agentes comunitários de saúde na equipe de Saúde da Família. Rev Esc Enferm USP. 2012;46(3):665-72. https://doi.org/10.1590/S0080-62342012000300019 32. Santos LFB, David HMSL. Percepções do estresse no trabalho pelos agentes comunitários de saúde. Rev Enferm UERJ. 2011 [citado 19 out 2017];19(1):52-7. Disponível em: http://www.facenf.uerj.br/v19n1/v19n1a09.pdf 33. Schmidt MLS, Neves TFS. O trabalho do agente comunitário de saúde e a política de atenção básica em São Paulo, Brasil. Cad Psicol Soc Trab. 2010;13(2):225-40. https://doi.org/10.11606/issn.1981-0490.v13i2p225-240 34. Silva MT, Salomão S, Alonso CMC, Matsubara S, Silva TM, Freitas ET, et. al. Transformação do modelo de atenção em saúde mental e seus efeitos no processo de trabalho. In: Lancman S, organizadora. Políticas públicas e processos de trabalho em saúde mental. Brasília (DF): Paralelo15; 2008. p. 87-128. 35. Silva TL, Dias EC, Ribeiro ECO. Saberes e práticas do agente comunitário de saúde na atenção à saúde do trabalhador. Interface (Botucatu). 2011;15(38):859-70. https://doi.org/10.1590/S1414-32832011005000035 36. Sossai LCF, Pinto IC, Mello DF. O agente comunitário de saúde (ACS) e a comunidade: percepções acerca do trabalho do ACS. Cienc Cuid Saude. 2010;9(2):228-237. https://doi.org/10.4025/cienccuidsaude.v9i2.11234 37. Souza LJR, Freitas MCS. O agente comunitário de saúde: violência e sofrimento no trabalho a céu aberto. Rev Baiana Saude Publica. 2011 [citado 19 out 2017];35(1):96-109. Disponível em: http://files.bvs.br/upload/S/0100-0233/2011/v35n1/a2100.pdf 38. Tomaz JBC. O agente comunitário de saúde não deve ser um “super-herói”. Interface (Botucatu). 2002;6(10):84-7. https://10.1590/S1414-32832002000100008 39. Trapé CA, Soares CB. A prática educativa dos agentes comunitários de saúde à luz da categoria práxis. Rev Latino-Am Enfermagem. 2007;15(1):142-9. https://doi.org/10.1590/S0104-11692007000100021 40. Vilela RAG, Silva RC, Jackson Filho JM. Poder de agir e sofrimento: estudo de caso sobre agentes comunitários de saúde. Rev Bras Saude Ocup. 2010;35(122):289-302. https://doi.org/10.1590/S0303-76572010000200011 https://doi.org/10.11606/S1518-8787.2018052000395 Funding: Coordination for the Improvement of Higher Education Personnel (CAPES – Process BEX 5217/2014-08 Program CAPES-COFECUB 702/11). Authors’ Contribution: Design, planning of the study, collection, analysis, and interpretation of the data: CMCA. Elaboration and revision of the manuscript and public responsibility for the content of the article: CMCA, PDB, FJCMD. Work of CHA: meta-synthesis Alonso CMC et al. REFERENCES Conflict of Interest: The authors declare no conflict of interest.
https://openalex.org/W1972114552	https://hal.science/hal-01557087/document	English	null	Influence of the genomic sequence on the primary structure of chromatin	Frontiers in life science	2,011	cc-by	32,590	To cite this version: Guillaume Chevereau, Alain Arnéodo, Cédric Vaillant. Influence of the genomic sequence on the primary structure of chromatin. Frontiers in Life Science, 2011, 5 (1-2), pp.29-68. 10.1080/21553769.2012.708882. hal-01557087 Distributed under a Creative Commons Attribution 4.0 International License nﬂuence of the genomic sequence on the primary structure of chromatin aUniversité de Lyon, F −69000 Lyon, France; bLaboratoire de Physique, CNRS, ENS-Lyon, 46 Allée d’Italie, 69364 Lyon Cedex 07, France As an important actor in the regulation of nuclear functions, the nucleosomal organization of the 10 nm chromatin ﬁber is the subject of increasing interest. Recent high-resolution mapping of nucleosomes along various genomes ranging from yeast to human, have revealed a patchy nucleosome landscape with alternation of depleted, well positioned and fuzzy regions. For many years, the mechanisms that control nucleosome occupancy along eukaryotic chromosomes and their coupling to transcription and replication processes have been under intense experimental and theoretical investigation. A recurrent question is to what extent the genomic sequence dictates and/or constrains nucleosome positioning and dynamics? In that context we have recently developed a simple thermodynamical model that accounts for both sequence speciﬁcity of the histone octamer and for nucleosome–nucleosome interactions. As a main issue, our modelling mimics remarkably well in vitro data showing that the sequence signaling that prevails are high energy barriers that locally inhibit nucleosome formation and condition the collective positioning of neighboring nucleosomes according to thermal equilibrium statistical ordering. When comparing to in vivo data, our physical modelling performs as well as models based on statistical learning suggesting that in vivo bulk chromatin is to a large extent controlled by the underlying genomic sequence although it is also subject to ﬁnite-range remodelling action of external factors including transcription factors and ATP-dependent chromatin remodellers. On the highly studied S. cerevisiae organism, we discuss the implications of the highlighted ‘positioning via excluding’ mechanism on the structure and function of yeast genes. The generalization of our physical modelling to human is likely to provide new insight on the isochore structure of mammalian genomes in relation with their primary nucleosomal structure. Keywords: chromatin; nucleosome; DNA sequence; physical model; statistical ordering HAL Id: hal-01557087 https://hal.science/hal-01557087v1 Submitted on 5 Jul 2017 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License Introduction 2002), including the nucleosomal array, its condensation into the 30 nm chro- matin ﬁber and the formation of chromatin loops, up to 1 periodic (‘deterministic’) distribution of bending sites, LRC likely facilitate the formation of some large-scale intrinsic curvature due to a persistent distribution of DNA curva- ture sites that predisposes DNA to make small (random) loops favoring the positioning of the histone core through- out a major part of the genome (Vaillant et al. 2005, 2006; Moukhtar et al. 2007, 2009, 2010, 2011). Furthermore, the LRC observed at scales larger than the length of DNA wrapped around the histone octamer (>150 bp) were con- jectured as contributing to the collective organization and repositioning dynamics of nucleosomes along the 10 nm chromatin ﬁber and possibly to its packing into higher- order chromatin structures (Audit et al. 2002; Vaillant et al. 2005, 2006). This conjecture was tested and veriﬁed as quite relevant when using power spectrum and correla- tion function analysis to study the nucleosome occupancy landscape obtained in the pioneering in vivo experiment of Yuan et al. (2005) on chromosome 3 of S. cerevisiae. This study (Vaillant et al. 2007) actually conﬁrmed that the spatial organization of nucleosomes is long-range cor- related with characteristics similar to the LRC imprinted in the DNA sequence. Since this pioneering experiment, the recent ﬂowering of genome-wide experimental maps of nucleosome positions for many diﬀerent organisms and cell types (Fire et al. 2006; Johnson et al. 2006; Albert et al. 2007; Lee et al. 2007; Mito et al. 2007; Ozsolak et al. 2007; Whitehouse et al. 2007; Field et al. 2008; Mavrich et al. 2008a, 2008b; Schones et al. 2008; Shivaswamy et al. 2008; Valouev et al. 2008; Kaplan et al. 2009; Zhang et al. 2009; Lantermann et al. 2010; Tsankov et al. 2010; Weiner et al. 2010; Valouev et al. 2011), has provided an unprecedented opportunity to elucidate to which extent the DNA sequence participates in the positioning of nucleosomes observed in vivo along eukaryotic chromosomes. can be induced by a stretch of regularly distributed potential wells (e.g. a stretch of highly positioning sequences), sta- tistical short-range ordering can also be observed near an energy barrier due to the interplay between boundary con- ﬁnement and rod–rod (nucleosome–nucleosome) excluded volume interaction. Introduction In the ﬁfth section, we develop a real- istic sequence-dependent model of nucleosome assembly that relies on the computation of the free-energy cost of bending a DNA fragment of a given nucleotide sequence from its natural curvature to the ﬁnal superhelical structure around the histone core. When comparing the predictions of this grand-canonical modelling at low chemical poten- tial to in vitro nucleosome occupancy data, we show that our physical model performs remarkably well conﬁrming that it accounts for both sequence speciﬁcity of the his- tone octamer and for nucleosome–nucleosome interactions. When tuning the chemical potential to higher value to repro- duce genome coverage by nucleosomes observed in vivo, we show that the collective nucleosomal organization in the in vivo bulk chromatin is to a large extent controlled by the underlying sequence. Interestingly, when some discrep- ancy is observed between the numerical predictions and the in vivo data, it actually provides very instructive informa- tion for future modelling of both transcription factor and chromatin remodeller driven ‘extrinsic’ nucleosome posi- tioning. We conclude, in the ﬁnal section, by discussing some new modelling perspectives in mammalian genomes. Introduction a full extent of condensation in metaphase chromosomes, remains controversial, at a local scale speciﬁc elements have been identiﬁed to interact with protein components of chromatin. For instance, as far as the basic unit of eukary- otic chromatin is concerned, some motifs that favor the formation and positioning of nucleosomes were found to be regularly spaced, e.g. the 10.5 bp periodicity exhibited in some dinucleotides like AA/TT/TA (Satchwell et al. 1986; Ioshikhes et al. 1996; Widom 2001). But these peri- odically distributed motifs concern only 5% of sequences that present aﬃnities for the histone octamer signiﬁcantly larger than average (Lowary and Widom 1997) and can- not account for more than ∼20% of the in vivo nucleosome positioning above what is expected by chance (Peckham et al. 2007; Yuan and Liu 2008). Alternatively, similar sequence motifs were shown to present long-range cor- relations (LRC) along the genome as the signature of the majority of nucleosomes that corresponds to 95% of bulk genomic DNA sequences having an aﬃnity for the histone octamer similar to that of random sequences (Audit et al. 2001, 2002, 2004; Arneodo et al. 2011). Hence, in con- trast to the tight histone binding obtained with an adequate The relation between the DNA primary structure and its biological function is one of the outstanding problems in modern cell biology. There are many objective reasons to believe that the functional role of the DNA sequence is not only to code for proteins (which represent less than 5% of mammalian genomes), but also to contribute to controlling the spatial structure and dynamics of DNA in chromatin. Nowadays, it is well recognized that the dynam- ics of DNA folding and unfolding within living cells plays a major role in regulating many biological processes, such as gene expression, DNA replication, recombination and repair (van Holde 1988; Wolﬀe 1998; Calladine and Drew 1999; Alberts et al. 2002). In that context, a very challeng- ing issue is to decipher to which extent the compromise between the necessity of compacting DNA in the nucleus of eukaryotic cells and the required accessibility to regulatory proteins is actually encoded in the DNA sequence. If the precise inﬂuence of the genomic sequence on the diﬀerent steps of DNA compaction (Alberts et al. In vivo and in vitro genome-wide primary structure of chromatin Nucleosome organization is generally analyzed by micro- cocal nuclease (MNase) digestion of chromatin. To perform large-scale studies, the distribution of MNase cleavage sites is determined throughout genomic regions or in the whole genome by means of either high resolution oligonucleotide tilingmicroarrays(MNase-chip)(Yuanetal.2005;Leeetal. 2007; Mito et al. 2007; Ozsolak et al. 2007; Whitehouse et al. 2007; Lantermann et al. 2010) or several diﬀerent massive DNA sequence technologies (MNase-seq) (John- son et al. 2006; Albert et al. 2007; Shivaswamy et al. 2008; Valouev et al. 2008; Kaplan et al. 2009; Tsankov et al. 2010; Weiner et al. 2010; Valouev et al. 2011). In this section, we will mainly present nucleosome occupancy proﬁle P(s) as an estimate (up to some normalization) of the probability of a base pair located at position s to be occupied by a nucleosome. As illustrated in Figures 1 to 6, a semi-logarithmic representation δY(s) = Y(s) −Y, where Y(s) = log2 (P(s)), is generally used to report the experimental MNase-chip and MNase-seq data. The paper is organized as follows. In the next section, we review in vivo and in vitro nucleosome occupancy data in various eukaryotic genomes with special focus on S. cerevisiae, S. pombe, C. elegans and human. The third section is devoted to the deﬁnition of our thermodynami- cal model of nucleosome assembly which is inspired from the well-known thermodynamics of the Tonks–Takahashi 1D ﬂuid (Tonks 1936; Takahashi 1942). As described by the Percus equation (Percus 1976), the hard-rod density (the nucleosome density along genomes) in an inhomo- geneous energetic ﬁeld (the nucleosome potential along genomes) can be determined as a function of the chemical potential (histone octamer reservoir) and the temperature. Various numerical schemes to compute the density are dis- cussed including the technical trick proposed by Vanderlick et al. (1986) to derive an explicit solution of the Percus equation that requires numerical integration. In the fourth section,forpedagogicalpurposes,weinvestigatetoyenergy landscapes involving square-like wells, square-like energy barriers, inﬁnite energy barriers, to illustrate the ‘statistical ordering’ mechanism originally proposed by Kornberg and Stryer (1988). If an array of well-positioned nucleosomes In vivo nucleosome occupancy proﬁles In Figures 1 to 6 are shown nucleosome occupancy pro- ﬁles experimentally obtained in vivo for several eukary- otic organisms, namely S. cerevisiae (Figures 1 and 2), 2 (a) (b) (c) (d) Figure 1. In vivo experimental nucleosome mapping along S. cerevisiae chromosomes obtained by Lee et al. (2007) (MNase-chip). Nucleosome occupancy proﬁle (δY(s) = Y(s) −Y, where Y(s) = log2 (P(s))) along a 10 kpb fragment of chromosomes 2 (a), 7 (b), 12 (c) and 14 (d). Symbols indicate regulatory sites: Transcription Start Sites (TSS, red dots), Transcription Termination Sites (TTS, pink circles) and Transcription Factors Binding Sites (TFBS, black triangles). (a) (b) (c) (d) Figure 1. In vivo experimental nucleosome mapping along S. cerevisiae chromosomes obtained by Lee et al. (2007) (MNase-chip). Nucleosome occupancy proﬁle (δY(s) = Y(s) −Y, where Y(s) = log2 (P(s))) along a 10 kpb fragment of chromosomes 2 (a), 7 (b), 12 (c) and 14 (d). Symbols indicate regulatory sites: Transcription Start Sites (TSS, red dots), Transcription Termination Sites (TTS, pink circles) and Transcription Factors Binding Sites (TFBS, black triangles). At a statistical level, experimental nucleosome occu- pancy ﬂuctuations can be quantiﬁed by their statistical distribution – that characterizes the variability of occu- pancy along the genome – and their auto-correlation func- tion – that characterizes the regularity of nucleosome occupancy (positioning) along the genome. As shown in Figure 7(a) for the budding yeast microarray data of Lee et al. (2007), the distribution of δY(s) is asymmet- ric with a large exponential tail at low occupancy values. The distributions obtained with the S. pombe microarray data of Lantermann et al. (2010) presents a very similar shape (Figure 7c). Interestingly, the distribution for the mutant, deﬁcient in Mit1 activity, reveals a slight enrich- ment towards lower occupancy values as compared to the wild type (WT) distribution (Figure 7c) which illustrates the global action of the remodelling factor Mit1 involved in the elimination of nucleosome depletion regions (NDRs) as recently shown by Garcia et al. (2010). The distribu- tions of MNase-seq nucleosome data for S. cerevisiae and S. pombe (Figure 7b) still present an asymmetry but are of diﬀerent shapes to those of the MNase-chip data. Let us note that S. pombe and S. cerevisiae data obtained by the same method present very similar distributions. The main S. kluyveri (Figure 3), S. pombe (Figure 4), C. elegans (Figure 5) and human (Figure 6). In vivo nucleosome occupancy proﬁles These experimental data reveal that nucleosome occupancy is globally disordered along the genomes (at chromosomal level). This conﬁrms that the primary structure of chromatin is not a regular array of nucleosomes. As exempliﬁed on the highly studied unicelullar organism S. cerevisiae whose genome is highly compact (∼75% gene coverage), the nucleosome occu- pancy proﬁle (Figure 1) reveals an alternation of (i) nucleo- some depleted regions (NDR) of typical size (100–200 bp), (ii) periodic stretches of well-positioned nucleosomes (of mean period 168 bp) and (iii) occupied but unorganized (‘fuzzy’) regions. As reported in Figure 2, a puzzling and rather annoying observation is that the nucleosome maps obtained from diﬀerent studies and diﬀerent methods (MNase-chip or MNase-seq (Zhang and Pugh 2011)) do not generally coincide. Actually they mainly diﬀer by their level of occupancy which might be due to diﬀerent tiling strate- gies between MNase-chip approaches (Yuan et al. 2005; Leeetal.2007)orbythesequencingdepthintheMNase-seq approaches. But hopefully the positions of nucleosomes and NDRs are pretty well conserved as observed in S. cerevisiae (Figure 2) and S. pombe (Figure 4). 3 (a) (b) (c) Figure 2. In vivo nucleosome occupancy proﬁle δY(s) (see Figure 1) along a 10 kbp fragment of chromosome 3 (a), 2 (b) and 2 (c) of S. cerevisiae. Comparison of Lee et al. (2007) MNase-chip data (red) with (a) Yuan et al. (2005) MNase-chip data (green), (b) Whitehouse et al. (2007) MNase-chip data (blue) and (c) Kaplan et al. (2009) MNase-seq data (violet). The symbols have the same meaning as in Figure 1. In (b), the Whitehouse et al. data correspond to a detrended hybridation proﬁle: Y(s) = Y(s) − s+a s−a Y(s) with a ∼200 bp. For the sake of comparison, we have applied the same detrending procedure to the Lee et al. data in (b). (a) (b) (c) Figure 2. In vivo nucleosome occupancy proﬁle δY(s) (see Figure 1) along a 10 kbp fragment of chromosome 3 (a), 2 (b) and 2 (c) of S. cerevisiae. Comparison of Lee et al. (2007) MNase-chip data (red) with (a) Yuan et al. (2005) MNase-chip data (green), (b) Whitehouse et al. (2007) MNase-chip data (blue) and (c) Kaplan et al. (2009) MNase-seq data (violet). The symbols have the same meaning as in Figure 1. In (b), the Whitehouse et al. data correspond to a detrended hybridation proﬁle: Y(s) = Y(s) − s+a s−a Y(s) with a ∼200 bp. In vivo nucleosome occupancy proﬁles For the sake of comparison, we have applied the same detrending procedure to the Lee et al. data in (b). (a) (b) (c) (a) (b) (c) Figure 3. (a) In vivo nucleosome occupancy proﬁle δY(s) (see Figure 1) along a 650 kbp fragment of chromosome C of S. kluyveri. MNase-seq data of Tsankov et al. (2010). (b) (resp. (c)), zoom on the 50 kbp region indicated in (a) by the red (resp. orange) colored segment, that corresponds to a high (resp. low) (G + C) content region, namely 52% (resp. 40%) as compared to the mean genome value (G+C) = 40%. Figure 3. (a) In vivo nucleosome occupancy proﬁle δY(s) (see Figure 1) along a 650 kbp fragment of chromosome C of S. kluyveri. MNase-seq data of Tsankov et al. (2010). (b) (resp. (c)), zoom on the 50 kbp region indicated in (a) by the red (resp. orange) colored segment, that corresponds to a high (resp. low) (G + C) content region, namely 52% (resp. 40%) as compared to the mean genome value (G+C) = 40%. 4 (a) (b) Figure 4. In vivo nucleosome occupancy proﬁle δY(s) (see Figure 1) along a 10 kbp fragment of chromosome 2 of S. pombe. (a) MNase-chip data of Lantermann et al. (2010): comparison between the WT (red) and the mit1-mutant (green). (b) Comparison between WT MNase-chip data of Lantermann et al. (red, see (a)) and the MNase-seq data of Tsankov et al. (2011) (cyan). (a) (b) Figure 4. In vivo nucleosome occupancy proﬁle δY(s) (see Figure 1) along a 10 kbp fragment of chromosome 2 of S. pombe. (a) MNase-chip data of Lantermann et al. (2010): comparison between the WT (red) and the mit1-mutant (green). (b) Comparison between WT MNase-chip data of Lantermann et al. (red, see (a)) and the MNase-seq data of Tsankov et al. (2011) (cyan). (a) (b) Figure 5. In vivo nucleosome occupancy proﬁle δY(s) (see Figure 1) along chromosome 2 of C. elegans (Valouev et al. 2008). (a) δY(s) versus s (green) along a 100 kbp fragment as obtained from the 5′ and 3′ ends’ tag proﬁles (black, see Zhang and Pugh 2011). (b) Zoom on a 10 kbp region. Figure 5. In vivo nucleosome occupancy proﬁle δY(s) (see Figure 1) along chromosome 2 of C. elegans (Valouev et al. 2008). In vivo nucleosome occupancy proﬁles (a) δY(s) versus s (green) along a 100 kbp fragment as obtained from the 5′ and 3′ ends’ tag proﬁles (black, see Zhang and Pugh 2011). (b) Zoom on a 10 kbp region. (a) (b) Figure 6. In vivo nucleosome (non-normalized) occupancy proﬁle P(s) along chromosome 2 of the human genome (Schones et al. 2008). (a) P(s) versus s along a 100 kbp fragment as obtained from the 5′ and 3′ ends tag proﬁles (see Zhang and Pugh 2011). (b) Zoom on a 10 kbp region. Figure 6. In vivo nucleosome (non-normalized) occupancy proﬁle P(s) along chromosome 2 of the human genome (Schones et al. 2008). (a) P(s) versus s along a 100 kbp fragment as obtained from the 5′ and 3′ ends tag proﬁles (see Zhang and Pugh 2011). (b) Zoom on a 10 kbp region. 5 5 (a) (b) (c) (d) Figure 7. Histograms of nucleosome occupation Y(s) values centered on their typical values (i.e. the maximum of the histogram is positioned at zero) for diﬀerent sets of in vivo data. (a) S. cerevisiae MNase-chip data of Lee et al. (2007). (b) S. cerevisiae MNase-seq data of Shivaswamy et al. (2008) (green) and of Kaplan et al. (2009) (violet) as compared to S. pombe MNase-seq data of Tsankov et al. (2011) (cyan). (c) S. pombe MNase-chip data of Lantermann et al. (2010): WT (red) and Mit1 mutant (green); the dashed curve corresponds to the S. cerevisiae Lee et al. data shown in (a). (d) C. elegans (green) and human (red) MNase-seq data of Valouev et al. (2008) and Schones et al. (2008), respectively. (a) (c) (b) (d) (d) Figure 7. Histograms of nucleosome occupation Y(s) values centered on their typical values (i.e. the maximum of the histogram is positioned at zero) for diﬀerent sets of in vivo data. (a) S. cerevisiae MNase-chip data of Lee et al. (2007). (b) S. cerevisiae MNase-seq data of Shivaswamy et al. (2008) (green) and of Kaplan et al. (2009) (violet) as compared to S. pombe MNase-seq data of Tsankov et al. (2011) (cyan). (c) S. pombe MNase-chip data of Lantermann et al. (2010): WT (red) and Mit1 mutant (green); the dashed curve corresponds to the S. cerevisiae Lee et al. data shown in (a). (d) C. elegans (green) and human (red) MNase-seq data of Valouev et al. (2008) and Schones et al. (2008), respectively. fold. In vivo nucleosome occupancy proﬁles These estimates indicate that the in vivo variability of nucleosome occupancy is clearly weak as compared to the 1000 fold enrichment observed in vitro between the non-natural 601 sequence, one of the sequences with the highest aﬃnity to the histone octamer, and a random sequence (Lowary and Widom 1998; Thåström et al. 1999). However, such a site-dependent variability may be suﬃ- cient to control the accessibility of protein complexes to their target sites and as such to participate in the regulation of nuclear function (Kornberg and Lorch 1999; Boeger et al. 2008; Lam et al. 2008; Wang et al. 2011a, 2011b). diﬀerence between the MNase-chip and MNase-seq data is at high occupancy values where the ﬂuctuations above the typical value are signiﬁcantly weaker for MNase-chip data (there is a more pronounced ‘bounded-like’ behavior). The origin of this discrepancy is not clear: it might be due to a lower statistical sampling in the seq experiments or to some saturation of the ﬂuorescence in the microarray experiments. For some MNase-seq data sets such as the S. cerevisiae data of Shivaswamy et al. (2008) (Figure 7b) and the human data of Schones et al. (2008) (Figure 7d), the sequencing depth (Zhang and Pugh 2011) is too low to get reliable distributions. Indeed, the high enrichment observed at low occupancy values is here artiﬁcial and is a direct consequence of unmapped genomic loci. Let us point out that for MNase-seq experiments, the C. elegans distribution (Figure 7d) is almost similar to the S. cerevisiae and S. pombe distributions (Figure 7b) apart from the slight excess of the high occupancy values which originates from excessive read enrichment at some repeated sequences. From these distributions and despite discrepancies between experiments and/or possible mapping artefacts, we can provide a rough estimate of the variability of in vivo nucle- osome occupancy between genomic sites: the occupancy ratio between highest aﬃnity sites and bulk aﬃnity sites (as deﬁned here by the sites of typical occupancy – the most probablevalue)isaround2–8fold;betweenthelowestaﬃn- ity sites and bulk sites, the depletion ratio is around 20–30 As experienced in various studies (see for a review Arneodo et al. 2011), the statistical organization of nucle- osomes along the chromatin ﬁber can be quantiﬁed by performing correlation and power-spectrum analysis. As shown in Figure 8(a), the auto-correlation function C(s) = ⟨δY(s)δY(s + s)⟩of S. cerevisiae MNase-chip data of Lee et al. In vivo nucleosome occupancy proﬁles (2007) displays a slow decrease with a periodic modulation of period l∗= 168 bp. The presence of this small-scale periodic arrangement of nucleosomes man- ifests itself in the well-deﬁned harmonic peak at f ∗= 1/l∗ in the power spectrum of the auto-correlation function (Figure 8b). Note that the same periodic modulation is also observed for the other sets of in vivo data, e.g. the White- house et al. (2007) MNase-chip data and the Shivaswamy et al. (2008) MNase-seq data (Figure 8). The local value of this period, hereafter called Nucleosome Repeat Length 6 (a) (b) Figure 8. (a) Auto-correlation function C(s) = ⟨δY(s)δY(s + s)⟩versus s. (b) Corresponding power spectrum. The diﬀer- ent colors correspond to the following S. cerevisiae data sets: MNase-chip data of Lee et al. (2007) (red) and of Whitehouse et al. (2007) (blue); MNase-seq ‘tag’ data of Shivaswamy et al. (2008) (green). (NRL), actually ﬂuctuates along the chromosomes. As reported in Figure 9, the NRL distribution is rather narrowly centered on the value 168 bp with root-mean square ﬂuc- tuations σ = 10–11 bp, in good agreement with the well- established 165 bp value for S. cerevisiae (Woodcock et al. 2006). As shown in Figure 9 (inset), the same distribution is observed for the NRL of S. kluyveri. Actually as already pointed out by Tsankov et al. (2010), the mean NRL of the Hemiascomycota fungi are relatively well conserved (around 165 bp) with the exception of C. albicans and K. lactis (∼175 bp).NotethatthischaracteristicNRLobserved in the stretches of well-ordered nucleosomes (Figure 1) is signiﬁcantly smaller than the mean in vivo NRL ∼210 bp estimated when assuming an homogeneous 75% coverage ofthe16yeastchromosomesbynucleosomes.Thissuggests the presence of some ‘conﬁning’ process. The experimen- tal two-point correlation functions in Figure 8(a) also reveal that this ‘nucleosome periodicity’ statistically appears as a modulation of a dominant slow decaying component which characterizes the large-scale disordered occupancy land- scape ﬂuctuations. In our concluding ﬁnal section, we will discuss the fact that this decay behaves as a power-law as an experimental conﬁrmation that the spatial organization of nucleosomes observed in vivo is long-range correlated with characteristics similar to the LRC imprinted in the DNA sequence (Vaillant et al. 2007). (a) Figure 8. (a) Auto-correlation function C(s) = ⟨δY(s)δY(s + s)⟩versus s. (b) Corresponding power spectrum. The diﬀer- ent colors correspond to the following S. cerevisiae data sets: MNase-chip data of Lee et al. In vivo nucleosome occupancy proﬁles (2007) (red) and of Whitehouse et al. (2007) (blue); MNase-seq ‘tag’ data of Shivaswamy et al. (2008) (green). variations of the NRL have also been measured recently in human with l∗≃205 bp in the heterochromatin HP1 domains,avaluesigniﬁcantlylargerthanintheeuchromatin domains where l∗≃178–195 bp (Valouev et al. 2011). As suggested by Woodcock et al. (2006), the control of the inter-nucleosome distance might be related in vivo to the stoichiometry of the linker histone H1 (per nucleosome) which can be species and cell-type speciﬁc: the more abun- dant the H1, the larger value the NRL (which is indeed the case for in vitro assembly (Blank and Becker 1995, see Figure 17 later). In S. cerevisiae, the H1 counterpart, Hho1 is indeed weakly present and to our knowledge no such linker histone has been found in S. pombe. However, the origin (and causality) of this relationship is not under- stood. Recent studies rather indicate that the nucleosome spacing is controlled by the action of some remodelling fac- tors such as members of the ISWI and CHD family (ACF, Mit1, Iswi1, Chd1...) (Clapier and Cairns 2009; Garcia et al. 2010; Gkikopoulos et al. 2011) which can be selectively targeted to speciﬁc genomic and/or epi-genomic loci in a developmentally regulated manner (Moshkin et al. 2012). Similar behavior of the nucleosome occupancy auto- correlation function is observed in diﬀerent eukaryotic organisms. As expected from a simple visual inspection of Figure 4, for S. pombe the periodic patterns in the occu- pancy proﬁle and consequently the periodic modulations of C(s) are more pronounced and easiest to quantify when analyzing MNase-seq data (Tsankov et al. 2011) than MNase-chip data (Lantermann et al. 2010). As shown in Figure 9, the distribution of NRL values is as wide as previ- ously obtained for S. cerevisiae but is signiﬁcantly shifted to smaller values with a mean value l∗= 151 bp, which, to our knowledge, is still unexplained. On the same Figure 9 is also reported the NRL distribution obtained for C. ele- gans data (Valouev et al. 2008); ﬂuctuations are of the samemagnitudebutthedistributionisshiftedtowardslarger NRL values with a larger mean value of 175 bp. Regional In vitro nucleosome occupancy proﬁles Only very recently, in vitro genome-wide nucleosome occupancy data have become available. One of the main Figure 9. Histograms of local NRL values computed from diﬀerent nucleosome occupancy data sets. The diﬀerent colors correspond to the following data: S. cerevisiae MNase-chip data of Lee et al. (2007) (red); S. pombe MNase-seq data of Tsankov et al. (2011) (cyan); C. elegans MNase-seq data of Valouev et al. (2008) (green). (Inset) Comparison betwen the histograms of the S. cerevisiae NRLs (Lee et al. 2007) (red) and of S. kluyveri NRLs (Tsankov et al. 2010) (orange). Figure 9. Histograms of local NRL values computed from diﬀerent nucleosome occupancy data sets. The diﬀerent colors correspond to the following data: S. cerevisiae MNase-chip data of Lee et al. (2007) (red); S. pombe MNase-seq data of Tsankov et al. (2011) (cyan); C. elegans MNase-seq data of Valouev et al. (2008) (green). (Inset) Comparison betwen the histograms of the S. cerevisiae NRLs (Lee et al. 2007) (red) and of S. kluyveri NRLs (Tsankov et al. 2010) (orange). 7 (a) (b) (c) (d) Figure 10. Nucleosome occupancy proﬁle (δY(s) = Y(s) −Y, where Y(s) = log2 (P(s))) along a 15 kpb long fragment of S. cerevisiae chromosomes 12: (a) in vivo MNase-chip data of Lee et al. (2007) (red) and in vivo MNase-seq data of Kaplan et al. (2009) (violet); (b) in vitro MNase-seq data of Kaplan et al. (2009) (orange); (c) corresponding histograms of Y(s) values centered on their typical values; (d) auto-correlation functions C(s) = ⟨δY(s)δY(s + s)⟩. (a) (b) (d) (c) Figure 10. Nucleosome occupancy proﬁle (δY(s) = Y(s) −Y, where Y(s) = log2 (P(s))) along a 15 kpb long fragment of S. cerevisiae chromosomes 12: (a) in vivo MNase-chip data of Lee et al. (2007) (red) and in vivo MNase-seq data of Kaplan et al. (2009) (violet); (b) in vitro MNase-seq data of Kaplan et al. (2009) (orange); (c) corresponding histograms of Y(s) values centered on their typical values; (d) auto-correlation functions C(s) = ⟨δY(s)δY(s + s)⟩. decay of C(s) as previously observed in vivo; this is the conﬁrmation that the long-range order observed in the collective nucleosome organization of the S. cerevisiae chromatin ﬁber likely results from the LRC that have been imprinted in the DNA sequence (at scales >200 bp) during evolution (see ﬁnal section ‘Discussion’) (Audit et al. 2001, 2002; Vaillant et al. 2007; Arneodo et al. 2011). In vitro nucleosome occupancy proﬁles diﬃculties encountered by such experiments is a dramatic limitation in the obtained genome coverage by reconsti- tuted nucleosomes. For example, Kaplan et al. (2009) have assembled chicken erythrocyte histone octamers on puriﬁed S. cerevisiae genomic DNA by salt gradient dial- ysis. They succeeded in producing in vitro occupancy proﬁles but for a nucleosome genome coverage ∼30%, which corresponds to a much smaller nucleosome density (mean inter-nucleosome distance ∼500 bp) than previously observed in vivo. As shown in Figure 10, the in vitro nucle- osome occupancy proﬁle along a 15 kbp long fragment of S. cerevisiae chromosome 12 looks much more dis- ordered as compared to the corresponding in vivo proﬁle obtained either by MNase-chip (Lee et al. 2007) or MNase- seq (Kaplan et al. 2009) experiments. In particular, if we still observe rather localized nucleosome depleted regions that remarkably coincide with some of the NDRs observed in vivo (Figure 10a), there is no longer evidence of stretches of periodically distributed nucleosomes. Thus, if the his- togram of nucleosome occupancy δY(s) in Figure 10(c) obtained in vitro almost superimposes onto the one obtained in vivo with the same MNase-seq technique, in contrast, the auto-correlation function C(s) in Figure 10(d) does not oscillateanylongerasthesignatureoftheabsenceofinvitro well-deﬁned NRL. Nevertheless, what is remarkable is the fact that we recover in vitro the same slow power-law Tonks–Takahashi ﬂuid where is the de Broglie thermal wavelength (it comes from the integration of the kinetic part of the hamiltonian). For the sake of simplicity, in the following, we will put = 1 which corresponds to a constant shift in the chemical potential μ ↔μ −β−1 ln . The one-point density distri- bution ρ(s) can be computed from the functional derivative where is the de Broglie thermal wavelength (it comes from the integration of the kinetic part of the hamiltonian). The Tonks–Takahashi ﬂuid (Tonks 1936; Takahashi 1942) is a one-dimensional ﬂuid of particles obeying the nearest neighbor two-body potential: For the sake of simplicity, in the following, we will put = 1 which corresponds to a constant shift in the chemical potential μ ↔μ −β−1 ln . The one-point density distri- bution ρ(s) can be computed from the functional derivative V(xij) = ∞, \|xij\| < l, = ψ(xij −l), l < \|xij\| < lM, = 0, \|xij\| > lM (1) (1) ρ(s) = −kT δ ln Ξ(L) δE(s) . (4) (4) where l corresponds to the hard-core length and lM to the maximal range of ψ which is an arbitrary function of xij. Lord Rayleigh (Rayleigh 1891) ﬁrst obtained the equation of state of hard-rods in an homogeneous ﬁeld in the thermodynamic limit. The distribution functions in this limit were derived later by Salsburg et al. (1953). The case of hard rods in an inhomogeneous external poten- tial was solved by Percus (1976) who derived an exact functional relationship between the local one-point den- sity ρ(s) and the residual chemical potential μ −E(s). Vanderlick et al. (1986) proposed an iterative scheme to integrate the Percus equation. Robledo and Rowlin- son (1986) investigated ﬁnite-size eﬀects on homogeneous hard-rod ﬂuids and Davis (1990) generalized these results to homogeneous Tonks–Takahashi systems. The equilibrium properties of Tonks–Takahashi ﬂuids can be derived from the conﬁguration partition function of the N-rod system: The explicit forms of this density function and of the k-body density distribution function are given by (Robledo and Rowlinson 1986; Davis 1990): ρ(s) = eβ(μ−E(s)) Ξ(s)Ξ(L −s) Ξ(L) , (5) ρ(k)(s1, . . . , sk) = eβkμ e−β k i=1 E(sk) Ξ(L) × Ξ(s1) k j=2 Ξ(sj −sj−1)Ξ(L −sk). (6) (5) ρ(s) corresponds to the probability of ﬁnding a particle at position s. ρ(2)(si, sj) corresponds to the probability of ﬁnding a pair of particles at positions si and sj. Thermodynamical model of nucleosome assembly Then, the ‘Grand- canonical’ partition function (the system equilibrates with a reservoir of particles of chemical potential μ) is given by uniform external potential, this is the well-known Tonks– Takahashi gas (Tonks 1936; Takahashi 1942). In the case of a non-uniform external potential of interest here, the prob- lem has been partly solved by Percus (1976) who derived an exact functional relationship between the residual chemical potential μ −E(s) and the hard rod density ρ(s). Ξ(L) = ∞ N=0 eβNμ N! N ZN(L), (3) (3) Thermodynamical model of nucleosome assembly Themechanismsunderlyingtheformation,thestructureand the displacement of nucleosomes are still largely unknown. As a consequence, we will attempt to provide some under- standing of the experimental observations reported in the previous section using a phenomenological approach based on simple physical arguments (Chevereau et al. 2009; Milani et al. 2009; Vaillant et al. 2010). When focusing on the dynamical assembly of histone octamers along the DNA chain, we ﬁrst assume that chromatin can be reasonnably modelled by a ﬂuid of rods of ﬁnite extension (the DNA wrapping length around the octamer), binding and moving in an external potential E(s) (the eﬀective nucleosome for- mation potential) and interacting (potential V(s, s′)) on a 1D substrate (the DNA chain) (Figure 11). The thermody- namics of such a system has been widely investigated in the literature. In the case of monodisperse hard rods on a 8 Figure 11. Grand canonical model of nucleosome assembly: bulk histones (green) may adsorb on or desorb from DNA (arrows). Barriers, such as transcription factors or other DNA binding proteins (red) can hinder nucleosome formation. The dynamics is con- trolled by a thermal bath (kT), the chemical potential of the histone reservoir (μ), the nucleosome–nucleosome interaction (V(ri, rj)) and the non-homogeneous adsorbing potential E(s). When no tridimensional degree of freedom is considered, the system reduces to a one-dimensional Tonks–Takahashi ﬂuid of hard-rods of hard core size the DNA wrapping length l (Tonks 1936; Takahashi 1942). Figure 11. Grand canonical model of nucleosome assembly: bulk histones (green) may adsorb on or desorb from DNA (arrows). Barriers, such as transcription factors or other DNA binding proteins (red) can hinder nucleosome formation. The dynamics is con- trolled by a thermal bath (kT), the chemical potential of the histone reservoir (μ), the nucleosome–nucleosome interaction (V(ri, rj)) and the non-homogeneous adsorbing potential E(s). When no tridimensional degree of freedom is considered, the system reduces to a one-dimensional Tonks–Takahashi ﬂuid of hard-rods of hard core size the DNA wrapping length l (Tonks 1936; Takahashi 1942). V(s1) and V(L −sN) assume that a particle is maintained ﬁxed at positions 0 and L, respectively. Then, the ‘Grand- canonical’ partition function (the system equilibrates with a reservoir of particles of chemical potential μ) is given by V(s1) and V(L −sN) assume that a particle is maintained ﬁxed at positions 0 and L, respectively. The homogeneous case E = Eo = cste Evaluation of the grand canonical ensemble partition func- tionΞ(L)isnoteasyingeneral.However,inthespecialcase of an homogeneous potential E = Eo, the problem is equiv- alent to the E = 0 homogeneous case when considering a residual chemical potential ˜μ = μ −Eo. To simplify the notations, we will thus treat the (E = 0, μ) case for which the canonical ensemble partition functions ZN(L) can be expressed as the N-convolution integral of the interaction Boltzmann weight e−βV(s). By denoting K(p) the Laplace transform of the interaction Boltzmann weight: βμ = βE(s) + ln ρ(s) −ln 1 − s+l s ρ(s′) ds′ + s s−l ρ(s′) 1 − s′+l s′ ρ(s′′) ds′′ ds,′ (16) (16) where K(p) = ∞ 0 e−ps−βV(s) ds, (8) (8) the partition function then writes: • s is the position along the potential; experimentally it usually corresponds to the genomic position of the nucleosome ‘dyad’ or of one of the nucleosome borders (5′ or/and 3′ extremity). ZN(L) = ZN(L −Nl) = N! 2πi i∞+τo −i∞+τ0 e(L−Nl)p[K(p)]N+1 dp, (9) (9) • l corresponds to the size of hard rod: the nucleosome wrapping length, assumed to be ﬁxed at l = 146 bp. • ρ is the density of hard rods. ( ) from which one can express Ξ(L) [Equation (3)] and the density distributions [Equations (5) and (6)] as sole functions of K(p): y • μ represents the chemical potential (i.e. the energy transferred by the bulk reservoir). Ξ(L) = ∞ N=0 eNβμ i∞+τo −i∞+τ0 e(L−Nl)p 2πi [K(p)]N+1 dp. (10) (10) • β = (kT)−1 is the reciprocal temperature. • E(s) is the free energy of nucleosome formation on a sequence at position s. At the thermodynamic limit, the pressure P [Equation (7)] of the bulk Tonks–Takahashi ﬂuid at chemical potential μ obeys the equation of state (Salsburg et al. 1953; Lieb and Mattis 1966): Remark: Note that for the homogeneous case E(s) = Eo, ρ(s) = ρ = cste and the Percus equation reduces to Equation (15). βμ = βPl −ln K(βP). (11) (11) The density–pressure equation can then be derived from this chemical potential–pressure equation using the Gibbs– Duhem equation: The density–pressure equation can then be derived from this chemical potential–pressure equation using the Gibbs– Duhem equation: Tonks–Takahashi ﬂuid This func- tion deﬁnes locally the equilibrium statistical distribution of distances between successive particles. The shape of this pair function directly depends on the interaction poten- tial V(si, sj) (Davis 1990) (see subsection ‘Homogeneous energy proﬁle E = Eo’). Classical measurement of the NRL by gel analysis of chromatin digestion products is directly related to this pair function (see Figure 17 later) (Noll and Kornberg 1977; Blank and Becker 1995; Woodcock et al. 2006). The pressure of this system can be computed directly ZN(L) = N! L 0 dsN xN 0 dsN−1 · · · s2 0 ds1 × e−β(N i>j V(si−sj)+N i E(si)) × e−β(V(s1)+V(L−sN )), ZN(L) = N! L 0 dsN xN 0 dsN−1 · · · s2 0 ds1 × e−β(N i>j V(si−sj)+N i E(si)) × e−β(V(s1)+V(L−sN )), , (2) where si correspond to the positions of the N particles; L is the size of the system and β = 1/kT. Boundary conditions 9 from the partition function (Lieb and Mattis 1966; Davis 1990): Percus equation When the energy landscape E(s) is non-uniform (with the exception of energy landscapes composed of inﬁnite energy barriers and ﬂat potential in between) the method of Laplace transform presented above cannot be applied. In 1976, Percus derived an equation that gives the density of a Tonks–Takahashi hard-rods ﬂuid in an inhomogeneous energetic ﬁeld as a function of the chemical potential and the temperature (Percus 1976): P = kT ∂ln Ξ(L) ∂L . (7) (7) The homogeneous case E = Eo = cste The homogeneous case E = Eo = cste Resolution of the Percus equation: the exact solution of Vanderlick et al. (21) (21) It is thus possible to solve Equation (20) iteratively starting from a simple initial condition for f : f (s) = 0, s ∈[−l; 0]. To compute the density let us introduce a new function b(s) such that: Figure 12. Illustration of the Vanderlick et al. exact solution Figure 12. Illustration of the Vanderlick et al. exact solution (blue) of the Percus equation [Equation (16)]. The energy land- scape (E(s)) used for the computation is shown in green. (a) f (s) ‘forward’ function; (b) b(s) ‘backward’ function; (c) the result- ing density ρ(s) = f (s)b(s). Model parameters: potential wall amplitude = +30 kT, μ = +3 kT. Figure 12. Illustration of the Vanderlick et al. exact solution (blue) of the Percus equation [Equation (16)]. The energy land- scape (E(s)) used for the computation is shown in green. (a) f (s) ‘forward’ function; (b) b(s) ‘backward’ function; (c) the result- ing density ρ(s) = f (s)b(s). Model parameters: potential wall amplitude = +30 kT, μ = +3 kT. f (s) = ρ(s) b(s) . (22) (22) From Equations (18) and (22) we deduce: From Equations (18) and (22) we deduce: b(s) = 1 − s+l s ρ(s′) ds′ = 1 − s+l s f (s′) · b(s′) ds′. (23) the system. According to Percus (1982), the density can be written as: ρ(s) = w(s)ξμ(s)ˆξμ(s) ξT , (26) The function b – for backward – depends on the ‘future’ only. When taking the derivative of Equation (23), we obtain: The function b – for backward – depends on the ‘future’ only. When taking the derivative of Equation (23), we obtain: (26) where w(s) = eβ(μ−E(s)) (27) (27) db ds (s) = f (s)b(s) −f (s + l)b(s + l), (24) (24) and and ξ(s) = ∞ N=0 eNβμQN(s), ˆξ(s) = ∞ N=0 eNβμ ˆQN(s), (28) the general solution of which writes: ξ(s) = the general solution of which writes: (28) b(s) = b(s0) exp s s0 f (s′) ds′ + s0 s exp − s′′ s f (s′) ds′ b(s′′ + l)f (s′′ + l) ds′′. (25) with QN(s) = Zleft N (0 ↔s), ˆQN(s) = Zright N (s ↔L). (29) (29) ξT is the total grand partition function. The main diﬃculty relies in the computation of Zleft N (s) (resp. Resolution of the Percus equation: the exact solution of Vanderlick et al. 1/ρ = (∂βμ)/(∂βP). (12) (12) Atechnical trick was proposed by Vanderlick et al. (1986)in order to solve the Percus equation (16) exactly. It is indeed possible to write the Percus equation as follows: In the case of hard rods, V(s) has the form of the Heavy- side distribution and K takes the simple form, K(p) = 1/p; we thus recover the equation of state derived by Rayleigh (1891): f (s) = exp βμ −βE(s) − s s−l f (s′) ds′ , (17) (17) βμ = βPl + ln(βPl) + ln(1/l), (13) βP = ρ 1 −ρl . (14) (13) where (14) f (s) ≡ ρ(s) 1 − s+l s ρ(s′) ds′ . (18) f (s) ≡ ρ(s) 1 − s+l s ρ(s′) ds′ . (18) (18) From Equations (13) and (14), we deduce the following chemical potential–density relationship: βμ = ln(ρ) −ln(1 −ρl) + ρl 1 −ρl . (15) Note that the function f – for forward – so introduced is a function that depends on the ‘past’ only. By taking the Note that the function f – for forward – so introduced is a function that depends on the ‘past’ only. By taking the (15) 10 10 10 (a) (b) (c) Figure 12. Illustration of the Vanderlick et al. exact solution (blue) of the Percus equation [Equation (16)]. The energy land- scape (E(s)) used for the computation is shown in green. (a) f (s) ‘forward’ function; (b) b(s) ‘backward’ function; (c) the result- ing density ρ(s) = f (s)b(s). Model parameters: potential wall amplitude = +30 kT, μ = +3 kT. derivative of Equation (17), we get: (a) (a) (b) (c) derivative of Equation (17), we get: (a) df ds (s) = f (s) · −β ∂E ∂s (s) + f (s −l) −f (s) . (19) df ds (s) = f (s) · −β ∂E ∂s (s) + f (s −l) −f (s) . (19) This equation has for general solution: This equation has for general solution: f (s) = u(s) exp(−βE(s0)) f (s0) + s s0 u(s′) ds′ , (20) u(s) ≡exp −βE(s) + s−l s0 f (s′) ds′ . (21) f (s) = u(s) exp(−βE(s0)) f (s0) + s s0 u(s′) ds′ , (20) where u(s) ≡exp −βE(s) + s−l s0 f (s′) ds′ . Resolution of the Percus equation: the exact solution of Vanderlick et al. Zright N ) correspond- ing to the number of conﬁgurations that accomodate N particles in the space 0 < x < s (0 ↔s) (resp. s < x < L, (s ↔L)). (25) This equation can again be solved iteratively by consider- ingtheboundaryconditions:b(s) = 1, s ∈[L; L + l].Then, once f (s) and b(s) are so computed, we use Equation (22) to get the density ρ(s) = f (s)b(s) (Figure 12). The explicit formulation of ZN [Equation (2)] (Percus 1982) is: ZN = · · · N i=2 e−βV(si−si−1) N i=1 e−βE(si) N i=1 dsi, (30) Segal et al. method (30) Recently, in the context of nucleosome positioning mod- elling, Segal et al. (2006) have proposed an alternative solu- tion for the computation of the density, using a Markovian algorithm that explicitly builds the full partition function of which is unfortunately not easily tractable. However in the particular case of hard-rods, QN(s) [Equation (29)] can be 11 by summing over all the possible accessible states: computed recursively: Ξ(L) = (1 1 . . . 1) × L s=1 Qs × ⎛ ⎜⎜⎝ 1 1 . . . 1 ⎞ ⎟⎟⎠. (34) QN(s + 1) = QN(s) + QN−1(s −l)e−βE(s+1−l), (31) (31) (34) where E(s) is the energy of a particle (nucleosome) that starts at s. Then, if we note NM the maximum number of particles that can be put in the interval [0, s], we get: Note that it is possible to compute this partition function iteratively: Ξ(L) = AL × ⎛ ⎜⎜⎝ 1 1 · · · 1 ⎞ ⎟⎟⎠, As = As−1 × Qs−1, ξμ(s + 1) = ∞ N=0 eNβμQN(s + 1), = NM N=0 eNβμQN(s) + NM N=1 eNβμQN−1(s −l) × e−βE(s+1−l), = ξμ(s) + NM −1 N=0 eNβμe+βμQN(s −l)e−βE(s+1−l), = ξμ(s) + ξμ(s −l)eβ(μ−E(s+1−l)), (32) A0 = (1 1 . . . 1). (35) (35) To determine the density distribution ρ(s) [Equation (4)], we just have to derive the partition function with respect to Ks = e−βE(s) and then to proceed iteratively consistently with Equation (35): ∂Ξ ∂Ks = ∂AN ∂Ks × ⎛ ⎜⎜⎝ 1 1 · · · 1 ⎞ ⎟⎟⎠, ∂As ∂Ks = ∂As−1 ∂Ks × Qs + As−1 × ∂Qs Ks . (36) (32) where we have explicitly used the fact that, by deﬁnition, it is possible to put no more than NM −1 particles of size l in the interval [0, s −l]. By proceeding in the same way for ˆξμ(s) [Equation (28)] we get: (36) The main advantage of the Teif and Rippe (2009) method is its adaptability to account for additional states, e.g. corresponding to other proteins or to interaction with neighboring nucleosomes. Among the disadvantages, the necessity to compute the total partition function is numeri- cally costly when working at the chromosome scales. ˆξμ(s −1) = ˆξμ(s) + ˆξμ(s + l)eβ(μ−E(s−1+l)). Statistical positioning In this section, we present and discuss the results of the numerical integration of the Percus equation (16) using the Vanderlick et al. integration scheme described in the subsection entitled ‘Resolution of the Percus equation: the exact solution of Vanderlick et al.’ for simple and illustra- tive energy landscapes made of energy barriers, traps and ﬂat regions. The nucleosome occupancy proﬁle P(s), as deﬁnedinsection‘Invivoandinvitrogenome-wideprimary structure of chromatin’, will be obtained by convolving the nucleosome density ρ(s) via the rectangular function of width 146 bp: Segal et al. method (33) (33) Then we use Equation (26) to determine the density ρ(s), where ξμ(s) and ˆξμ(s) have been estimated recursively from the left (s = 0) and the right (s = L) ends of the system according to Equations (32) and (33) respectively. As com- pared to the Vanderlick et al. algorithm described in the previous subsection, the Segal et al. method (2006) has the main inconvenience of involving sums of terms that can very quickly become very important which may result in the accumulation of numerical errors. Teif and Rippe method Teif and Rippe (2009) have recently proposed a theoretical framework for lattice models of histone–DNA interactions. Their strategy mainly consists in using the transfer matrix method to compute the grand canonical partition function (Baxter 1982). The idea is to associate a l × l matrix to each position s along the DNA sequence, where l = 146 corre- spondstothe146possiblepositionsofthenucleotides inthe nucleosome. The matrix Qs(i, j) represents the probability that the nucleotide s is in the state i knowing the fact that the nucleotide s + 1 is in the state j. Now if E(s) is the binding energy for a nucleosome starting at position s and extending up to s + l, then the matrix elements that are non-null write Qs(i, i + 1) = eβ(μ−E(s−i+1)) where 1 < i < l. The grand canonical partition [Equation (3)] can then be expressed P(s) = ρ ◦146(s). (37) (37) Low chemical potential barrier experiences a pushing force from the rest of the ﬂuid that conﬁnes it against the barrier. Positioning (as deﬁned by spatial localization) is thus the strongest at this loca- tion. The next particle is also conﬁned from one side by the ﬂuid pressure and from the other side by the (conﬁned but) moving ﬁrst particle and not by a ﬁxed barrier. So its positioning is weaker, and so on, positioning strength decreases as the distance to the barrier increases. Far from the barrier, the density proﬁle becomes uniform and equals the bulk value ρb given by Equation (15). Amplitude and range of this nonlocal-induced periodic ordering depend on the barrier height and shape (that deﬁnes the pressure exerted on the ﬂuid) and on the chemical potential (that deﬁnes internal pressure of the ﬂuid) as further discussed in the following subsections. Indeed this periodic order- ing is an internal property of dense ﬂuids; it is usually ‘revealed’ in the occupancy proﬁle by the presence of inho- mogeneities (vertical energy barriers in Figure 13) in the energy landscape. At low chemical potential, the occupancy proﬁle P(s) reﬂects the energy landscape topography: local depletion is observed at energy barriers and local enrichement at energy traps. At low chemical potential, the average density (occu- pancy) is weak and the Percus equation indeed reduces to its diluted (or non-interacting) approximation: βμ = βE(s) + ln ρ(s), (38) (38) where the nonlinear and nonlocal terms in ρ have been neglected. Then the density ρ(s) at position s only depends on the energy E(s) at this position via the simple Boltzmann relationship: ρ(s) ∼eβ(μ−E(s)). (39) (39) At regions where the energy landscape is uniform, density is constantandequalsthebulkdensity[Equation(15)]:ρ(s) = ρb ≃exp β(μ −E0) (see Figure 13, ρbl = 0.19). Illustrative energy landscape For pedagogical purposes, we show in Figure 13, the the- oretical occupancy proﬁles of nucleosomes considered as hard rods of hard-core size 146 bp obtained at diﬀerent chemical potential values in a toy energy landscape bor- dered by two inﬁnite walls and that displays a stretch of square-like wells on the right-side and two square-like energy barriers on the left side. 12 Figure 13. Hard-rod occupancy proﬁles P(s) in a non-uniform energy landscape made of discrete energy barriers and traps and bounded by two inﬁnite walls (black). The Percus equation (16) was solved using the Vanderlick et al. integration scheme (subsection ‘Resolution of the Percus equation: the exact solution of Vanderlick et al.’) at low chemical potential μ = −6 kT (orange) and high chemical potential μ = −1 kT (red). The occupancy values ρbl = 0.19 (resp. ρbl = 0.74) correspond to the bulk occupancy of the uniform system at μ = −6 kT (resp. μ = −1 kT) (see Figure 14). Figure 13. Hard-rod occupancy proﬁles P(s) in a non-uniform energy landscape made of discrete energy barriers and traps and bounded by two inﬁnite walls (black). The Percus equation (16) was solved using the Vanderlick et al. integration scheme (subsection ‘Resolution of the Percus equation: the exact solution of Vanderlick et al.’) at low chemical potential μ = −6 kT (orange) and high chemical potential μ = −1 kT (red). The occupancy values ρbl = 0.19 (resp. ρbl = 0.74) correspond to the bulk occupancy of the uniform system at μ = −6 kT (resp. μ = −1 kT) (see Figure 14). Nucleosome density where ρ(s1\|s2) is the conditional probability of having a particle at position s1 given that a particle is ﬁxed at posi- tion s2, and ρ(s2) is the particle density at position s2. In a Tonks–Takahashi ﬂuid (see subsection ‘Tonks–Takahashi ﬂuid’), ρ(s1\|s2) corresponds exactly to the density ρ(s1) of the ﬂuid when conﬁned by a wall at position s2 (imposed by the conditional ﬁxed particle) with a wall–particle interac- tion corresponding to the inter-particle interaction function V(s2, s1). The computation of the pair distribution thus only requires to compute the density ρw(s) of the semi-conﬁned Tonks–Takahashi ﬂuid. In the case of an homogeneous ﬂuid, ρ(s2) = ρb is the bulk density and ρ(s1\|s2) = ρw(r), where ρw correspond to the density of the ﬂuid at distance r = \|s1 −s2\| from the wall. As discussed in the subsec- tion ‘Tonks–Takahashi ﬂuid’, the bulk density is obtained at the thermodynamic limit by solving Equation (15). When using, Equations (3), (5) and (7), we get for ρw(r) (Davis 1990): In the ‘ideal’ case of a ﬂat potential (E = Eo = cste), the density ρ = ρb is homogeneous (the ‘bulk’ phase density) and is controlled by a single parameter, the residual chem- ical potential ˜μ = μ −Eo. As reported in the section ‘A sequence-dependent physical model of nucleosome occu- pancy’ (see Figure 27b later on, the zone F5 of 3.5 kbp), this ‘ideal’ situation can be realistically achieved in the frame of our sequence-dependent energy model at some speciﬁc genomic regions. We may also expect that, in some genomic regions, the action of remodelling factors can result in an eﬀective reduction of the sequence-dependent potential ﬂuctuations leading to an approximate local homogeniza- tion of the energy landscape (δE(s) →0)) (Zhang et al. 2011). In Figure 14, we show the evolution of the occu- pancy (ρbl) as a function of ˜μ computed from Equation (15). It has a sigmoidal shape: at very low ˜μ values the density is weak; then it slightly increases and goes through a tran- sition at ˜μc = −ln(l), where the susceptibility χ = ∂ρ/∂μ is maximum. At this point, ρ( ˜μc)l = 0.36, and the occu- pancy rapidly increases up to ρl ≃0.75. When further increasing ˜μ, it then slowly increases towards the asymp- totic limit ρ∗l = 1. Inter-nucleosomal distance As deﬁned in the section ‘In vivo and in vitro genome-wide primary structure of chromatin’, key distances characteriz- ing the primary structure of chromatin are the nucleosome repeat length (NRL) l∗(s) and the linker size d(s), i.e. the part of unwrapped DNA that joins two successive octamers. The NRL and the linker size are directly related: l∗= l + d. As we shall see later, this distance can be extracted experi- mentally by MNase digestion and gel analysis of digestion products (Blank and Becker 1995) or directly by single molecule imaging of the 10 nm ﬁber (Solis et al. 2004; Milani et al. 2009). This distance diﬀers in general from the mean inter-nucleosome distance which is simply given by lm = ρ−1 b . In the ﬂuid formalism (Hansen and McDonald 2006), the statistical properties of the inter-particle dis- tances are given by the pair density distribution ρ2(si, sj) which represents the joint probability to have particles (of ﬁxed size l) at positions s1 and s2. Using the Bayes formula, we have the following general decomposition: As deﬁned in the section ‘In vivo and in vitro genome-wide primary structure of chromatin’, key distances characteriz- ing the primary structure of chromatin are the nucleosome repeat length (NRL) l∗(s) and the linker size d(s), i.e. the part of unwrapped DNA that joins two successive octamers. The NRL and the linker size are directly related: l∗= l + d. Figure 14. Occupancy (ρl) of hard rods in a homogeneous energy landscape E(s) = Eo, as a function of the residual chem- ical potential ˜μ = μ −Eo. Theoretical curves obtained from Equation (14) (red, inﬁnite size system) and from Vanderlick numerical method (blue, large but ﬁnite size system). The dots indicate the bulk occupancy values for the chemical potential val- ues used in Figure 13: ˜μ = −6 kT (orange dot) and ˜μ = −1 kT (red dot). ρ2(s1, s2) = ρ(s1\|s2)ρ(s2), (40) Homogeneous energy proﬁle E = Eo Nucleosome density (40) High chemical potential Remark: An interesting point here is that from low den- sity (low chemical potential) proﬁle, we can extract the underlying energy landscape via the simple Boltzmann rela- tionship [Equation (39)]. Actually the Percus equation (16) provides a direct computation of the energy landscape from the density proﬁle at every chemical potential. How- ever, since this relationship is only valid for hard-core repulsion, applying it to a Tonks–Takahashi ﬂuid with an ‘a priori’ unknown interaction potential V (for exam- ple, applying it to high density in vivo data, e.g. the distribution of tags in MNase-seq experiments, see sub- section ‘In vivo nucleosome occupancy proﬁles’) may lead to an interaction-dependent bias in the energy landscape estimation. This ‘bias’ is likely to be minimized when using low density proﬁle (for example, the low density in vitro data, see subsection ‘In vitro nucleosome occupancy proﬁles’). When increasing the chemical potential, the average density increases, local depletion and enrichment are still observed at barriers and traps respectively. As expected and already observed at low μ values, an array of well-positioned nucle- osomes is observed as induced by the stretch of regularly distributed potential wells (e.g. a stretch of highly position- ing sequences like the well-known 601 sequence (Lowary andWidom1997)). Moreinterestingaretheoscillationsthat appearinP(s)inthevicinityoftheenergybarriersandofthe bordering walls. This ‘periodic’ positioning, which is not induced by any local periodically distributed energy traps, is purely entropic (Kornberg and Stryer 1988). It results from the conﬁnement of the hard rods imposed by the excluding energy barriers: the pressure imposed on each particle by the surrounding particles increases with the density; so at high enough density the ﬁrst particle next to (ﬂanking) the 13 Figure 14. Occupancy (ρl) of hard rods in a homogeneous energy landscape E(s) = Eo, as a function of the residual chem- ical potential ˜μ = μ −Eo. Theoretical curves obtained from Equation (14) (red, inﬁnite size system) and from Vanderlick numerical method (blue, large but ﬁnite size system). The dots indicate the bulk occupancy values for the chemical potential val- ues used in Figure 13: ˜μ = −6 kT (orange dot) and ˜μ = −1 kT (red dot). Inter-nucleosomal distance Nucleosome density Let us ﬁrst point out that peri- odic ordering could only be deﬁned for ˜μ > μc = ln l, which corresponds to the transition point of the ρb versus μ plot in Figure 14. At this critical point, for hard rods of size l = 146 bp, l∗= 1.45l = 212 bp and lm = 0.36−1l = 405 bp. The damping length is λ = 1.18l = 172 bp. More interestingly, for a higher chemical potential ˜μ = −1 kT corresponding to a bulk occupancy of ρbl = 0.74, we ﬁnd a period of l∗= 1.22l = 178 bp and lm = l/0.74 = 198 bp. The damping length is λ = 2.5l = 365 bp which means that at a distance ∼6 from the (wall) particle, the ordering strength (i.e. the density ‘excess’: ρ(6l) −ρb) is 10 fold less than the one observed next to the reference particle (wall). At a distance of 12l, ordering is lost (the density excess is 100 fold less than the one observed next to the refer Figure 15. Evolution of the pair function g(r) [Equation (44)] with the residual chemical potential ˜μ. Black and red curves correspond to a chemical potential ˜μ = −5 kT and +5 kT respec- tively. Gray curves correspond to intermediate values of ˜μ. The inter-particle distance r = \|s1 −s2\| is expressed in l units. Figure 15. Evolution of the pair function g(r) [Equation (44)] with the residual chemical potential ˜μ. Black and red curves correspond to a chemical potential ˜μ = −5 kT and +5 kT respec- tively. Gray curves correspond to intermediate values of ˜μ. The inter-particle distance r = \|s1 −s2\| is expressed in l units. g(s1, s2) as: g(s1, s2) = ρ2(s1, s2)/ρ(s1)ρ(s2) = ρ(s1\|s2)/ρ(s1). (43) In the case of a uniform ﬂuid: g(s1, s2) as: g(s1, s2) = ρ2(s1, s2)/ρ(s1)ρ(s2) = ρ(s1\|s2)/ρ(s1). (43) g(s1, s2) as: g(s1, s2) = ρ2(s1, s2)/ρ(s1)ρ(s2) = ρ(s1\|s2)/ρ(s1). (43) g(s1, s2) = ρ2(s1, s2)/ρ(s1)ρ(s2) = ρ(s1\|s2)/ρ(s1). (43) In the case of a uniform ﬂuid: g(r) = ρ2(r)/ρ2 b = ρw(r)/ρb. (44) (44) Let us note again that whatever the distribution func- tion, the mean inter-nucleosome distance is given by: lm = 1/L L l rg(r) dr = ρ−1 b . Nucleosome density The pair function g(r) can be com- puted from the analytical form [Equation (42)], where for short distances r only a few modes N < r/l are contribut- ing. In Figure 15, to get the complete pair function, we proceed as in Figure 13 by solving the Percus equation (16) in the semi-conﬁnement conﬁguration (vertical barrier: E = ∞, x < 0; E = 0, x > 0 (see subsection ‘Statistical order- ing near an energy barrier’). At very low chemical potential, i.e. at very low bulk density, the pair function is almost constant (g = 1), consistent with the fact that particles are almost isolated (and don’t interact) leading to a constant pair distribution ρ2(r) = ρ2 b: all inter-nucleosomal distances are equiprobable. When increasing the chemical potential, par- ticles interact resulting in oscillations in the pair function, whose amplitude (resp. period) increases (resp. decreases) with the density (Figure 15): around each particle there is preferential periodic ordering (of period l∗) of ﬂanking particles. Suﬃciently far from the reference particle, the pair function tends to the value g = 1 of the diluted (unor- ganized) ﬂuid, as well as ρw(r) tends to the bulk density ρb. The origin of the oscillations (internal ordering) is by deﬁnition the same as the one previously described in the density proﬁles for the semi-conﬁned ﬂuid near an inﬁnite wall (Figure 13). Here, each particle acts as a conﬁning wall in its reference frame. It is important to emphasize here that in this homoge- neous energy proﬁle E = Eo, the density proﬁle ρ(s) = ρb is homogeneous and doesn’t present any oscillation as observed in Figure 13 far from the conﬁning energy bar- riers. To observe a similar ordering pattern in the density proﬁle one has to introduce a vertical barrier in the energy landscape. This will be discussed in the subsection ‘Sta- tistical ordering near an energy barrier’. Experimentally, the NRL is measured through an enzymatic digestion pro- cedure followed by a gel characterization of digestion products. Restriction enzymes like MNase preferentially digest linker DNA (‘naked’ DNA) resulting in mono-, di-, tri- ... nucleosome DNA fragments whose length distribu- tion is quantiﬁed by gel migration (Blank and Becker 1995). Nucleosome density Interestingly, for very large ˜μ values, ﬁnite-size systems experience a transition towards a sat- urated ρl = 0.85 occupancy state (Figure 14, blue curve) which has been identiﬁed as a pseudo-crystalline state con- sisting of quasi regularly spaced particles ‘self-conﬁned’ inside equipartitioned regions of length equal to the aver- age length per particle lm = 1/ρ (Piasecki and Peliti 1993; Giaquinta 2008). In inhomogeneous ﬂuids, this saturation only occurs for the bulk occupancy (ρbl = l/L L 0 ρ(s) ds) as illustrated in Figure 13 where at the edges near the inﬁnite energy walls, the local occupancy ρ(s)l can reach a value close to one. ρw(r) = exp (βμ) exp [−βPr]Ξ(r −l). (41) (41) For hard-rod ﬂuids, when using Equation (10) with K(p) = 1/p, we get: ρw(r) = eβμ exp − ρb 1 −ρbl r × ∞ N=0 eNβμ N! (r −(N + 1)l)Nθ(r −(N + 1)l), N 0 (42) (42) where θ is the heavyside function (θ(r) = 0, r > 0, = 1, r < 0). Now let us introduce the pair correlation function 14 Figure 15. Evolution of the pair function g(r) [Equation (44)] with the residual chemical potential ˜μ. Black and red curves correspond to a chemical potential ˜μ = −5 kT and +5 kT respec- tively. Gray curves correspond to intermediate values of ˜μ. The inter-particle distance r = \|s1 −s2\| is expressed in l units. and this whatever the chemical potential value (Figure 15). For hard rods, this typical distance thus diﬀers from the NRL l∗which rather corresponds to the order parameter of the ﬂuid. The ordered phase of the dense uniform hard-rod ﬂuid is indeed characterized by this period l∗and the damp- ing length λ of the pair function. As shown in Figures 16(a) and (b), the attenuation of the oscillations and the relax- ation towards the asymptotic value 1 is indeed exponential with a well-deﬁned characteristic correlation length λ; these oscillations are periodic with a well-deﬁned period l∗. The computation of these characteristic lengths from the analyti- cal expression of the pair function [Equations (42) and (44)] is not obvious; so we used numerical calculations to produce theresultsshowninFigure16.InFigure16(c)isreportedthe evolution of the characteristic lengths λ and l∗, as compared to the mean inter-nucleosome distance lm, as a function of the chemical potential ˜μ. Nucleosome density (c) Evolution of the ordering range λ (black) and the mean period l∗(red) as a function of the chemical potential ˜μ. Both lengths are measured as explained in (a) and (b). In blue is reported the mean inter-nucleosome distance lm = ρ−1 b as a function of ˜μ. Figure 16. (a) Pair function g(r) of a dense (uniform, E(s) = Eo = cste) hard-rod ﬂuid. The residual chemical potential is ˜μ = 0 which gives a bulk density ρbl = 0.78. The statistical ordering of this dense phase is characterized by the period and the range of the oscillating pattern: (inset) ln (gm(i)l −1) as a function of pm(i) −l, where gm(i) is the value of the pair function at the ith extremum and pm(i) is the position of the ith extremum: maximum (red points) and minimum (green points) of the pair function. The linear regression gives a slope −λ = −1.9l, with λ deﬁning the damping length. (b) Position pm(i) of the ith maxima (red) or minima (green) as a function of i. The linear regression gives the mean period of the oscillations l∗= 1.2l. (c) Evolution of the ordering range λ (black) and the mean period l∗(red) as a function of the chemical potential ˜μ. Both lengths are measured as explained in (a) and (b). In blue is reported the mean inter-nucleosome distance lm = ρ−1 b as a function of ˜μ. Figure 17. Gradual increase of NRL by histone H1. (a) Chro- matin was assembled in histone depleted embryo extracts comple- mented with core histones and the indicated amounts of histone H1. (b) Plot proﬁle of the ﬁrst lane of each MNase digestion in (a). Peaks of mono- (M), di- (D) and tri-nucleosomes (T) are indicated. Adapted from Blank and Becker (1995). hence reveals regularly spaced gel bands (Figure 17a) cor- responding to the migration of oligomers of quantized size n × l∗. It is usually extracted from the intensity proﬁle (Figure 17b) as the period of the oscillation (by a linear regression as illustrated in Figure 16). Formally this diges- tion proﬁle is reminiscent of the pair distribution and the NRL coincides with the period l∗of this pair function (actually the MNase proﬁles are rather related to the pair function of linkers). As already mentioned, it depends on the organism and the cell type. Nucleosome density For a well-ordered primary structure, the digestion pattern It is quite interesting to note that the typical inter- nucleosome distance is actually l the size of the hard rods: the typical (most probable) conﬁguration consists in having neighboring hard rods in contact (i.e. linker size d(s) = 0), 15 (a) (b) (c) Figure 16. (a) Pair function g(r) of a dense (uniform, E(s) = Eo = cste) hard-rod ﬂuid. The residual chemical potential is ˜μ = 0 which gives a bulk density ρbl = 0.78. The statistical ordering of this dense phase is characterized by the period and the range of the oscillating pattern: (inset) ln (gm(i)l −1) as a function of pm(i) −l, where gm(i) is the value of the pair function at the ith extremum and pm(i) is the position of the ith extremum: maximum (red points) and minimum (green points) of the pair function. The linear regression gives a slope −λ = −1.9l, with λ deﬁning the damping length. (b) Position pm(i) of the ith maxima (red) or minima (green) as a function of i. The linear regression gives the mean period of the oscillations l∗= 1.2l. (c) Evolution of the ordering range λ (black) and the mean period l∗(red) as a function of the chemical potential ˜μ. Both lengths are measured as explained in (a) and (b). In blue is reported the mean inter-nucleosome distance lm = ρ−1 b as a function of ˜μ. (a) (b) (c) (a) (b) (c) Figure 16. (a) Pair function g(r) of a dense (uniform, E(s) = Eo = cste) hard-rod ﬂuid. The residual chemical potential is ˜μ = 0 which gives a bulk density ρbl = 0.78. The statistical ordering of this dense phase is characterized by the period and the range of the oscillating pattern: (inset) ln (gm(i)l −1) as a function of pm(i) −l, where gm(i) is the value of the pair function at the ith extremum and pm(i) is the position of the ith extremum: maximum (red points) and minimum (green points) of the pair function. The linear regression gives a slope −λ = −1.9l, with λ deﬁning the damping length. (b) Position pm(i) of the ith maxima (red) or minima (green) as a function of i. The linear regression gives the mean period of the oscillations l∗= 1.2l. Nucleosome density It has been shown by in vitro experiments, supported by theoretical investigations, that the NRL plays a crucial role in the formation of higher- order chromatin structure (Bednar et al. 1998; Dorigo et al. 2004; Mergell et al. 2004; Lesne and Victor 2006; Robin- son et al. 2006; Kepper et al. 2008; Depken and Schiessel 2009). Figure 17. Gradual increase of NRL by histone H1. (a) Chro- matin was assembled in histone depleted embryo extracts comple- mented with core histones and the indicated amounts of histone H1. (b) Plot proﬁle of the ﬁrst lane of each MNase digestion in (a). Peaks of mono- (M), di- (D) and tri-nucleosomes (T) are indicated. Adapted from Blank and Becker (1995). Inﬁnite energy barriers As shown in Figure 18(a), near an inﬁnite energy barrier, whenprogressivelyincreasingthechemicalpotential ˜μ,sta- tistical ordering becomes more and more pronounced and manifests as periodic oscillations in the density (and occu- pancy) proﬁle very similar to what we have observed in the pair function in Figure 15. Indeed, as explicitly used in sub- section ‘Inter-nucleosomal distance’, the density ρ(s) near an inﬁnite wall [ρw, Equation (42)] takes the same form, up to a multiplicative factor, as the pair correlation func- tion in an homogeneous landscape. Thus, the (positional) statistical ordering can again be characterized by the two characteristic lengths, the damping length λ and the period of the spatial modulation l∗. Of course there is the question of the biological rele- vancy of such periodic ordering induced by inﬁnite barriers. As a very convincing example of biological interest, we report in Figure 19(a), the experimental nucleosome posi- tioning data around CTCF sites obtained by Fu et al. (2008) in human cells. The binding protein CTCF has been extensively studied for its impact on imprinting and X-inactivation (Lee 2003). It is known to bind to insula- tor elements to prevent heterochromatin spreading and may function as a transcriptional repressor or activator (Klen- ova et al. 1993; Burcin et al. 1997; Ohlsson et al. 2010). As shown in Figure 19(b), the remarkable nucleosome order- ing observed on both sides of the CTCF bound proteins and that progressively vanishes for distances larger than 1 kbp is remarkably reproduced by the statistical position- ing predicted by numerically solving the Percus equation [Equation (16)] in a ﬂat energy landscape with an inﬁnite energy barrier of width 240 bp positioned at the CTCF site. A good agreement is actually obtained for a chemical poten- tial value ˜μ = −2 kT which yields for the NRL a value l∗= 185 bp very close to the 190 bp estimated experimen- tally. Consistently, the predicted damping length accounts quite well for the exponential decay of the nucleosome ordering observed in the data. Bistability induced by statistical conﬁning in between two energy barriers When analyzing the nucleosome occupancy proﬁles obtained in vivo in budding yeast (Figures 1 and 2), we realize that nucleosome depleted regions commonly called the Nucleosome Free Region (NFR) are distributed along the chromosomes with a mean separation distance ∼1– 2 kbp (Chevereau et al. 2009; Vaillant et al. 2010; Arneodo et al. 2011). This observation raises the issue of the statis- tical conﬁning of nucleosomes in between two inhibitory energy barriers that are separated by a distance L of a few nucleosome DNA wrapping lengths l. As illustrated in Figure 18(b) for the very simpliﬁed situation of a ﬂat energy landscape bordered by two inﬁnite energy barriers, when solving the Percus equation (16) for increasing values of the chemical potential ˜μ, we observe the establishment of a clear statistical ordering near the two bordering energy barriers, as previously shown in Figure 18(a), that progres- sively invades the system to transform into a clear periodic packing at high nucleosome (hard rod) density. The more nucleosomes are conﬁned, the more they adopt a long-range and compact periodic organization with the inter-boundary distance L as a fundamental control parameter. The problem of stacking hard rods of size l = 146 bp in a box of size L with inﬁnite wall boundaries is actully easy to solve (Chevereau 2010). For a given (rather high) value of the chemical potential ˜μ, the theoretical weighted probabil- ityofan-nucleosomeconﬁgurationisshowninFigure21(d) for various n-values as a function of the inter-barrier dis- tance L. The nucleosome occupancy proﬁle is then obtained as the weighted sum of each n = 1, 2, . . . nucleosome occupancy proﬁles (Figures 21a–c). Thus this theoretical Statistical ordering near an energy barrier As already illustrated in Figure 13, a simple way to pro- duce periodic positioning without any local ‘positioning’ signal (i.e. energy traps) is to introduce energy barriers, i.e. exclusion regions (Kornberg and Stryer 1988; Chevereau et al. 2009; Möbius and Gerland 2010; Vaillant et al. 2010). These inhibitory energy barriers can be encoded in the DNA sequence via unfavorable sequences that potentially resist to the structural distortions required for nucleosome for- mation (e.g. the presence of poly(dA:dT) (Bao et al. 2006; Field et al. 2008; Mavrich et al. 2008a; Kaplan et al.. 2009; Segal and Widom 2009b) or particular sequences that may 16 recruit transcription factors (Pusarla et al. 2007; Hartley and Madhani 2009; Kaplan et al. 2009) or/and other pro- tein complexes such as chromatin regulators (Whitehouse and Tsukiyama 2006; Clapier and Cairns 2009; Teif and Rippe 2009) that may compete with the nucleosomes. a need to quantify the eﬀect of the energy barrier char- acteristics on the statistical positioning observed nearby these obstacles. In Figure 20, we summarize the numeri- cal results obtained when solving the Percus equation with an energy barrier of variable height emerging at the center of a ﬂat energy landscape. In agreement with the experi- mental observation in budding yeast (Figure 1), we ﬁxed the barrier width to w = 180 bp (which allows a nuclosome to form for small barrier height as shown in Figure 20a). For the chemical potential value ˜μ = −1, 0, 4 and 10 kT, corresponding to high nucleosome bulk occupancy (0.74, 0.78, 0.83 and 0.84), we have computed the NRL l∗and the maximal density value ρw obtained near an energy bar- rier for diﬀerent values of the barrier height ranging from 1 to 20 kT (Figure 20b). Let us ﬁrst note that above 5 kT, l∗ and ρw reach the asymptotic values of the inﬁnite wall case. If the height of the energy barrier does not signiﬁcantly aﬀect the NRL, it has a clear inﬂuence on the maximum density ρw which increases between low (< 1 kT) and high (>5 kT) barrier height by 2, 2.5 and 3.5 fold for μ = −1, 0 and >4 kT respectively. Finite energy barrier (b) Modelling of the data in (a) obtained by solving the Percus equation (16) in a ﬂat energy landscape with an inﬁnite energy barrier centered on the CTCF site and of width 240 bp (gray area) and a chemical potential value ˜μ = −2 kT. As clearly seen on the NRL in Figure 22, as the inter- barrier distance L increases, the extent of crystal domains is expected to decrease to the beneﬁt of the bistable (or multistable) coexistence domains. At very large L, peri- odic packing is lost at the center of the system where the nucleosome positioning is no longer dictated by the long- range inﬂuence of the bordering inﬁnite walls. The NRL l∗is no more L-dependent and tends to the value of the uniform system (Figure 16c) at the chemical potential ˜μ. situation predicts the existence of ‘crystallization’ domains that are characterized by a single dominating crystal n con- ﬁguration (Figures 21a and c) with a NRL that increases with L over the range lmin < NRL ∼L/n < lmax as shown in Figure 22. Importantly, this model also predicts that in between the crystal n (Figure 21a) and n + 1 (Figure 21c) domains, there exists a coexistence domain where these two (or more) crystalline conﬁgurations contribute statistically to an apparently irregular occupancy proﬁle (Figure 21b). As clearly seen on the NRL in Figure 22, as the inter- barrier distance L increases, the extent of crystal domains is expected to decrease to the beneﬁt of the bistable (or multistable) coexistence domains. At very large L, peri- odic packing is lost at the center of the system where the nucleosome positioning is no longer dictated by the long- range inﬂuence of the bordering inﬁnite walls. The NRL l∗is no more L-dependent and tends to the value of the uniform system (Figure 16c) at the chemical potential ˜μ. situation predicts the existence of ‘crystallization’ domains that are characterized by a single dominating crystal n con- ﬁguration (Figures 21a and c) with a NRL that increases with L over the range lmin < NRL ∼L/n < lmax as shown in Figure 22. Importantly, this model also predicts that in between the crystal n (Figure 21a) and n + 1 (Figure 21c) domains, there exists a coexistence domain where these two (or more) crystalline conﬁgurations contribute statistically to an apparently irregular occupancy proﬁle (Figure 21b). (a) (b) Figure 20. Finite energy barrier Inhibitory energy barriers encountered along eukaryotic chromosomes can be of variable shape. As we will discuss in the next section ‘A sequence-dependent physical model of nucleosome occupancy’, the ones that are encoded in the mechanical properties of the DNA double helices are typically of a few kT high. Their width is at most of the order of a nucleosome DNA wrapping length. So there is 17 (b) (a) (b) Figure 18. Evolution of the density proﬁle ρ(s) with the chemical potential ˜μ from ˜μ = −5 kT (black) to +5 kT (red). Statistical conﬁnement near an inﬁnite wall (a) and in between two inﬁnite energy barriers (b). (a) Figure 18. Evolution of the density proﬁle ρ(s) with the chemical potential ˜μ from ˜μ = −5 kT (black) to +5 kT (red). Statistical conﬁnement near an inﬁnite wall (a) and in between two inﬁnite energy barriers (b). (a) (b) Figure 19. (a) Aggregation of nucleosome signals around CTCF sites from the experiment of Fu et al. (2008). The coordinate origin is set to the 5′ end position of the 20 bp-long CTCF sites. Schematic arrangement of nucleosomes (blue ovals) around a CTCF binding site (orange rectangle). Blue arrows indicate sequence tags on the same strand as the CTCF site (nucleosome 5′ extremity) and orange arrows indicate opposite-strand tags (nucleosome 3′ extremity). In green (resp. purple) are reported the 5′ (resp. 3′) extremity nucleosome counts in the absence of bound-CTCF. (b) Modelling of the data in (a) obtained by solving the Percus equation (16) in a ﬂat energy landscape with an inﬁnite energy barrier centered on the CTCF site and of width 240 bp (gray area) and a chemical potential value ˜μ = −2 kT. (a) (b) Figure 19. (a) Aggregation of nucleosome signals around CTCF sites from the experiment of Fu et al. (2008). The coordinate origin is set to the 5′ end position of the 20 bp-long CTCF sites. Schematic arrangement of nucleosomes (blue ovals) around a CTCF binding site (orange rectangle). Blue arrows indicate sequence tags on the same strand as the CTCF site (nucleosome 5′ extremity) and orange arrows indicate opposite-strand tags (nucleosome 3′ extremity). In green (resp. purple) are reported the 5′ (resp. 3′) extremity nucleosome counts in the absence of bound-CTCF. Finite energy barrier Statistical periodic ordering observed near an energy barrier as a fuction of the barrier height. (a) Density proﬁles obtained by solving the Percus equation (16) in a ﬂat energy landscape with a ﬁnite energy barrier of width l = 180 bp centered at s = 0 (gray area) and a chemical potential value ˜μ = 4 kT. The proﬁles correspond to barrier heights ranging from 1 kT (black) to 20 kT (red). (b) NRL l∗ (dots) and wall density value ρ(s = ±90) = ρw (curves) as a function of the barrier height, extracted from the density proﬁles (see (a)) at diﬀerent values of the chemical potential : μ = −1 kT (black), μ = 0 kT (red), μ = 4 kT (green) and μ = 10 kT (blue). (b) (a) Figure 20. Statistical periodic ordering observed near an energy barrier as a fuction of the barrier height. (a) Density proﬁles obtained by solving the Percus equation (16) in a ﬂat energy landscape with a ﬁnite energy barrier of width l = 180 bp centered at s = 0 (gray area) and a chemical potential value ˜μ = 4 kT. The proﬁles correspond to barrier heights ranging from 1 kT (black) to 20 kT (red). (b) NRL l∗ (dots) and wall density value ρ(s = ±90) = ρw (curves) as a function of the barrier height, extracted from the density proﬁles (see (a)) at diﬀerent values of the chemical potential : μ = −1 kT (black), μ = 0 kT (red), μ = 4 kT (green) and μ = 10 kT (blue). 18 (a) (b) (C) (d) Figure 21. Theoretical probability of nucleosome occupancy at each point of a box of size L bordered by two inﬁnite walls. (a) Box large enough to shelter ﬁve nucleosomes (green). (b) Larger box where the two dotted conﬁgurations are possible; the weighted average of the 5 and 6 nucleosome crystal-like proﬁles yields an irregular-looking average proﬁle (red). (c) Larger box where six nucleosomes can be inserted without being tightly packed. (d) Probability of crystal conﬁgurations with a ﬁxed number n of nucleosomes with respect to the box size L. Vertical colored lines correspond to the inter-barrier distances L used respectively in (a), (b) and (c). While only one conﬁguration has clearly the highest probability for (a) and (c), two conﬁgurations are equally probable in (b), which justiﬁes the superposition. Finite energy barrier The distances are expressed in nucleosome length units (hard core length l). (a) (b) (C) (d) (d) Figure 21. Theoretical probability of nucleosome occupancy at each point of a box of size L bordered by two inﬁnite walls. (a) Box large enough to shelter ﬁve nucleosomes (green). (b) Larger box where the two dotted conﬁgurations are possible; the weighted average of the 5 and 6 nucleosome crystal-like proﬁles yields an irregular-looking average proﬁle (red). (c) Larger box where six nucleosomes can be inserted without being tightly packed. (d) Probability of crystal conﬁgurations with a ﬁxed number n of nucleosomes with respect to the box size L. Vertical colored lines correspond to the inter-barrier distances L used respectively in (a), (b) and (c). While only one conﬁguration has clearly the highest probability for (a) and (c), two conﬁgurations are equally probable in (b), which justiﬁes the superposition. The distances are expressed in nucleosome length units (hard core length l). Figure 22. Theoretical NRL l∗dependency on the box size L (see Figure 21); black dotted lines correspond to a ﬁxed number n of nucleosomes and the red lines to the NRL size at a given chem- ical potential ˜μ = 4 kT. Vertical gray shaded bands correspond to the bistable domains. The distances are expressed in nucleosome length l units. distribution inside the genes. (Note that we use L as a substitute of the distance L ∼L + 188 between the 5′ and 3′ NFRs which is more diﬃcult to measure accurately because of the NFR shape variability.) Small genes (L < 1.5 kbp) present a clear periodic packing in between the two bor- dering NFRs with a well-deﬁned number n of regularly spaced nucleosomes (Figure 23c). As the interdistance L increases, these ‘crystallized’ genes cluster into L-domains with the gene having the same number of nucleosomes, for n = 2 to about 9 nucleosomes. For rather large gene sizes (L > 1.5 kbp), the nucleosome positioning appears periodic essentially at the two boundaries and fuzzy in the middle where the conﬁnement induced by both boundaries is too weaktoconstrainthepositioningofthecentralnucleosomes (Figure 23a). Finite energy barrier (a)The4554genesareorderedverticallybythedistanceLbetween the ﬁrst (5′) and last (3′) nucleosomes. The nucleosome occupancy proﬁle of each gene is ﬁgured along a horizontal line: red dots correspond to the minima of nucleosome occupancy; nucleosomes occupy the white zones; in vivo data are retrieved from Lee et al. (2007). (b) Predictions of our theoretical modelling (blue) with ﬁxed force boundary energy barriers (see text and Vaillant et al. (2010)) drawn on top of experimental data (red). Insets: mean experimental (red) and theoretical (blue) nucleosome occupancy proﬁles for crystal genes harboring 5 nucleosomes (right, top), 6 nucleosomes (right, bottom) and for bi-stable genes with 5/6 nucleosomes. (c) Zoom on the ﬁrst 2000 genes in (b); gray-shaded areas correspond to some bi-stable L-domains. In (b) and (c), the black curves indicate the 5′ and 3′ end positions of the theoretical excluding nucleosome energy barriers. Figure 23. 2D-map of nucleosomes along budding yeast genes. (a)The4554genesareorderedverticallybythedistanceLbetween the ﬁrst (5′) and last (3′) nucleosomes. The nucleosome occupancy proﬁle of each gene is ﬁgured along a horizontal line: red dots correspond to the minima of nucleosome occupancy; nucleosomes occupy the white zones; in vivo data are retrieved from Lee et al. (2007). (b) Predictions of our theoretical modelling (blue) with ﬁxed force boundary energy barriers (see text and Vaillant et al. (2010)) drawn on top of experimental data (red). Insets: mean experimental (red) and theoretical (blue) nucleosome occupancy proﬁles for crystal genes harboring 5 nucleosomes (right, top), 6 nucleosomes (right, bottom) and for bi-stable genes with 5/6 nucleosomes. (c) Zoom on the ﬁrst 2000 genes in (b); gray-shaded areas correspond to some bi-stable L-domains. In (b) and (c), the black curves indicate the 5′ and 3′ end positions of the theoretical excluding nucleosome energy barriers. g y ( ) The nucleosome structure of promoters and its implica- tions in transcription initiation has been studied in various organisms from yeast to human including the nematode and Drosophila (Bernstein et al. 2004; Lee et al. 2004; Yuan et al. 2005; Lee et al. 2007; Ozsolak et al. 2007; Mavrich et al. 2008a; Miele et al. 2008; Shivaswamy et al. 2008; Tirosh and Barkai 2008; Valouev et al. 2008; Arneodo et al. 2011). In Chevereau et al. (2009), and Vaillant et al. (2010), we have identiﬁed a new paradigm of transcriptional con- trol mediated by the stability and the level of compaction of the intra-genic chromatin architecture. Finite energy barrier To what extent these chromatin mediated regulation processes generalize to other eukaryotic species is a very challenging question for future experimental and theoretical studies. genes. Hopefully, probably thanks to chromatin remodellers that are found all over yeast chromosomes likely increasing the eﬀective temperature (Rippe et al. 2007), equilibrium statistical nucleosome ordering will be recovered along most genes with a characteristic time much shorter than the typical time separating the successive chromatin alter- ations induced by elongation. In a work under progress, we are revisiting the experimental in vitro nucleosome posi- tioning data of Zhang et al. (2011) under the scope of our theoretical modelling with the idea that the 1 pN con- ﬁning force previously found to account for the in vivo Finite energy barrier The nucleosome occupancy proﬁle of each gene is ﬁgured along a horizontal line: red dots correspond to the minima of nucleosome occupancy; nucleosomes occupy the white zones; in vivo data are retrieved from Lee et al. (2007). (b) Predictions of our theoretical modelling (blue) with ﬁxed force boundary energy barriers (see text and Vaillant et al. (2010)) drawn on top of experimental data (red). Insets: mean experimental (red) and theoretical (blue) nucleosome occupancy proﬁles for crystal genes harboring 5 nucleosomes (right, top), 6 nucleosomes (right, bottom) and for bi-stable genes with 5/6 nucleosomes. (c) Zoom on the ﬁrst 2000 genes in (b); gray-shaded areas correspond to some bi-stable L-domains. In (b) and (c), the black curves indicate the 5′ and 3′ end positions of the theoretical excluding nucleosome energy barriers. intra-genic nucleosome organization might well result from active remodelling at the 5′ end of most yeast genes. But what is remarkable in the results reported in Figure 23 (Chevereau et al. 2009; Vaillant et al. 2010) is the functional implications of the intra-genic chromatin structure of yeast genes. In agreement with the predic- tions of our thermodynamical modelling, we have been able to identify two main classes of in vivo intra-genic nucleosome organizations: crystal-like genes with regularly positioned nucleosomes and bi-stable genes with rather irregular nucleosome occupancy proﬁles resulting from the coexistence of two possible crystal-like states with diﬀerent compaction levels, a weakly compacted n nucleosome state and a highly compact n + 1 nucleosome state. As compared to crystal-like genes that present a constitutive expression level, bi-stable genes show a higher transcriptional plastic- ity and are more sensitive to chromatin regulators. Indeed, by means of a single nucleosome switching, bi-stable genes may drastically alter their expression level in response to external changes. In that context, a less intuitive result is the fact that the transcription rate tends to increase when the NRL decreases, so when the linear compaction level increases. A very similar trend has been also observed recently in human cells by Valouev et al. (2011). Several possible interpretations of this coupling between intra-genic chromatin and polymerase elongation process have been proposed (Vaillant et al. 2010) including the fact that a short linker size would rather lead to a loose 30 nm ﬁber favoring the accessibility and sequential action of components of the elongation machinery (Lesne and Victor 2006). Figure 23. 2D-map of nucleosomes along budding yeast genes. Finite energy barrier This intra-genic nucleosome organization is totally consistent with the statistical ordering mechanism induced by exclusion from the boundaries except that to quantitatively account for the in vivo data in Figures 23(b) and (c), we had to consider in our theoretical modelling ﬁnite-size linear energy barriers of height EM = 6 kT and width = 80 bp and a chemical potential value ˜μ = 1 kT so that the nucleosomes cover 75% of the yeast genome as observed in vivo (see Vaillant et al. 2010, for more tech- nical details). This peculiar linear shape of the bordering energy barriers actually amounts to imposing a constant force F = EM/ = 6 kT/27.2 nm ∼1 pN on both sides of the intra-genic nucleosome array. Note that 1 pN is com- parable and actually a little less than the few piconewton tensions generated by elongating polymerases (Wang et al. 1998; Hall et al. 2009) and helicases (Strick et al. 2003; Lionnet et al. 2006) suggesting that these enzymes can dras- tically aﬀect the nucleosome ordering observed inside yeast Figure 22. Theoretical NRL l∗dependency on the box size L (see Figure 21); black dotted lines correspond to a ﬁxed number n of nucleosomes and the red lines to the NRL size at a given chem- ical potential ˜μ = 4 kT. Vertical gray shaded bands correspond to the bistable domains. The distances are expressed in nucleosome length l units. Altogether, these theoretical results demonstrate the cru- cial role of the inter-barrier distance L as a ﬁne control of chromatin structure. As pointed out in a previous study (Chevereau et al. 2009; Vaillant et al. 2010; Arneodo et al. 2011), in S. cere- visiae, NFR are mainly observed in vivo at transcription start sites (TSS) and transcription termination sites (TTS). Whatever the origin of the underlying energy barriers, this observation suggests that the chromatin organization inside budding yeast genes is susceptible to being described by a phenomenological model as simple as the one previously studied in Figures 21 and 22. As reported in Figure 23, when ordering budding yeast genes by the distance L that sepa- rates the ﬁrst (5′) and last (3′) nucleosome, we obtain a 2D map that reveals a striking organization of the nucleosomes’ 19 Figure 23. 2D-map of nucleosomes along budding yeast genes. (a)The4554genesareorderedverticallybythedistanceLbetween the ﬁrst (5′) and last (3′) nucleosomes. A sequence-dependent physical model of nucleosome occupancy As shown previously with CTCF (subsection ‘Inﬁnite energy barriers’, Figure 19), in vivo nucleosome occupancy and positioning can be locally controlled by external fac- tors: the stable binding of proteins or protein complexes at a genomic locus act as an ‘eﬀective’ vertical barrier that induces stretches of periodically distributed nucleo- somes in agreement with a statistical ordering principle. ‘Chromatin regulators’ such as chromatin remodellers have 20 Figure 24. Our physical modelling consists of computing the energy cost to bend a DNA fragment of length lw into almost two turns of the DNA double helix which are involved in the crystallized nucleosome particle (radius R = 4.19 nm, pitch P = 2.59 nm). Adapted with permission from Richmond and Davey (2003). Copyright 2003 by Nature Publishing Group. Figure 24. Our physical modelling consists of computing the energy cost to bend a DNA fragment of length lw into almost two turns of the DNA double helix which are involved in the crystallized nucleosome particle (radius R = 4.19 nm, pitch P = 2.59 nm). Adapted with permission from Richmond and Davey (2003). Copyright 2003 by Nature Publishing Group. are described by the set of three local angles Ω1(s) (tilt), Ω2(s) (roll), Ω3(s) (twist), and (ii) the DNA chain along the nucleosome at position s is constrained to form an ideal superhelix of radius R = 4.19 nm and pitch P = 2.59 nm as observed in the X-ray crystallographic nucleosome struc- ture (Luger et al. 1997; Richmond and Davey 2003) over a total length lw which ﬁxes the distribution of angular defor- mations (Ωnuc i (u −s))i=1,2,3, u = s, . . . , s + l (Figure 24). Within linear elasticity approximation, the energy cost for nucleosome formation is given by: a fundamental role in shaping in vivo primary structure. Some remodellers, in combination with DNA binding fac- tors or/and chromatin modiﬁers, are targeted to speciﬁc genomic loci or histone epigenomic marks, to actively eject [e.g. RSC (Hartley and Madhani 2009; Wang et al. 2011b)] or maintain [e.g. RSC (Moreira and Holmberg 1999; Floer et al. 2010), Mit1 (Garcia et al. 2010)] nucleo- somes. Besides these ‘extrinsic’ regulations of nucleosome positioning/occupancy, there may also exist an ‘intrinsic’ contribution from the sequence-speciﬁcity of histone–DNA interactions. The question of the sequence-speciﬁcity and of the role it plays in vivo has been extensively studied for the last 30 years. A sequence-dependent physical model of nucleosome occupancy However there is still no consensus on it. According to recent reports (Peckham et al. 2007; Yuan and Liu 2008), in S. cerevisiae, no more than 20% of the in vivo nucleosome positioning above what is expected by chance is determined by intrinsic signals in the genomic DNA. As reported in ‘In vivo and in vitro genome-wide primary structure of chromatin’, the recent availability of in vivo and in vitro genome wide nucleosome occupancy data (Figures 1, 2 and 10) (Lee et al. 2007; Kaplan et al. 2009) has led to a renewal of interest in modelling nucleo- some organization along the 10 nm chromatin ﬁber. Thus a model of DNA-sequence dependent nucleosome position- ing based on statistical learning (Field et al. 2008; Kaplan et al. 2009) was shown to be signiﬁcantly predictive of the nucleosome organization in vivo in budding yeast as well as in other organisms like ﬂy and human. In the spirit of the physical modelling developed in an early work (Vaillant et al. 2005, 2006), we have recently proposed a model which is based on the sequence-dependent DNA bending proper- ties (Vaillant et al. 2007; Chevereau et al. 2009; Milani et al. 2009) and, as reported in this section, which performs as well as models based on statistical learning. βE(s, lw) = s+lw s 3 i=1 Ai 2 (Ωnuc i (u −s) −Ωo i (u))2 du, βE(s, lw) = s+lw s 3 i=1 Ai 2 (Ωnuc i (u −s) −Ωo i (u))2 du, (45) (45) where A1, A2 and A3 are the stiﬀnesses associated with the tilt, roll and twist deformations around their intrinsic values Ωo 1, Ωo 2 and Ωo 3, respectively. Consistently with our previ- ous works (Audit et al. 2002; Vaillant et al. 2005, 2006, 2007), we will use here the ‘Pnuc’ structural bending table (Goodsell and Dickerson 1994) which is mainly a trinu- cleotide roll coding table (Ωo 2), with zero tilt (Ωo 1 = 0) and constant twist (Ωo 3 = 2π/10.5). Since the values of this bending table were arbitrarily assigned between 0 and π/18, we have performed the following aﬃne rescaling Ω∗o 2 = γ (Ωo 2 −η) with η = 0.15 and γ a tuning param- eter that controls the ﬂuctuation range δ = ⟨(E −E)2⟩1/2 of the energy landscape. A sequence-dependent physical model of nucleosome occupancy For example, for yeast, we have ﬁxed γ = 0.4 which amounts to imposing δ = 2 kT, a value which allowed us to get comparable overall nucleosome occupancy distributions as observed both in vitro and in vivo (Vaillant et al. 2007). In Figure 25 is shown the theoretical nucleosome occu- pancy proﬁle obtained along the yeast chromosome 2 when ﬁxing the model parameters to provide a very good match, at a statistical level, with the experimental in vitro (Kaplan et al. 2009) and in vivo (Lee et al. 2007) data. The 2D map in Figure 25(a) actually represents the evolution of the nucleosome occupancy probability P(s) when increas- ing the mean residual chemical potential ˜μ. In Figure 25(b) ‘Intrinsic’ nucleosome formation energy landscape To compute the free energy landscape associated with the formation of one nucleosome at a given position s along DNA, we will assume that (Vaillant et al. 2007) (i) DNA is an unshearable elastic rod whose conformations 21 (a) (b) Figure 25. (a) 2D map representing the theoretical nucleosome occupancy probability P(s) [Equation (37)] along a 12 kbp long fragment of the budding yeast chromosome 2 as a function of the residual chemical potential ˜μ = μ −E (Chevereau et al. 2009): dark blue corresponds to low probability and red to high probability. The two white occupancy proﬁles are the theoretical proﬁles obtained for ˜μ = −6 kT and −1.3 kT that correspond to a genome nucleosome coverage of 30% and 75% as observed in vitro (Kaplan et al. 2009) and in vivo (Lee et al. 2007) respectively; the corresponding in vitro and in vivo experimental nucleosome occupancy proﬁles are shown in red for comparison. (b) The corresponding energy landscape E(s) computed with the following parameter values: δ = ⟨(E −E)2⟩1/2 = 2 kT and lw = 125 bp (see text). (a) (b) Figure 25. (a) 2D map representing the theoretical nucleosome occupancy probability P(s) [Equation (37)] along a 12 kbp long fragment of the budding yeast chromosome 2 as a function of the residual chemical potential ˜μ = μ −E (Chevereau et al. 2009): dark blue corresponds to low probability and red to high probability. The two white occupancy proﬁles are the theoretical proﬁles obtained for ˜μ = −6 kT and −1.3 kT that correspond to a genome nucleosome coverage of 30% and 75% as observed in vitro (Kaplan et al. 2009) and in vivo (Lee et al. 2007) respectively; the corresponding in vitro and in vivo experimental nucleosome occupancy proﬁles are shown in red for comparison. (b) The corresponding energy landscape E(s) computed with the following parameter values: δ = ⟨(E −E)2⟩1/2 = 2 kT and lw = 125 bp (see text). Figure 26. ln (ρ(s1)/ρ(s2)) versus E12 = E(s1) −E(s2), where ρ(s1) (resp. ρ(s2)) is the nucleosome density (computed as explained in the text, from the budding yeast genome) and E(s1) (resp. E(s2)) the nucleosome formation energy at the position s1 (resp. s2). The crosses correspond to two statistical sam- ples in a diluted ( ˜μ = −6 kT, red) and dense ( ˜μ = 0 kT, black) non-uniform ﬂuid. ‘Intrinsic’ nucleosome formation energy landscape is shown the predicted energy landscape when ﬁxing the eﬀective nucleosome wrapping length to lw = 125 bp, a value which is smaller than the typical and well accepted 146 bp nucleosomal DNA length (Luger et al. 1997). This suggests that sequence-speciﬁcity is eﬀectively dominated by the wrapping around the H3–H4 tetramer. Note that the hard-core length l that we consider for the computation of the nucleosome density remains l = 146 bp. Interestingly, Figure 25(a) enlightens the fundamental role of the energy landscape and its topography (amplitude, size and distri- bution of favorable and unfavorable regions) that entirely control the ﬂuctuations in the nucleosome occupancy pro- ﬁle but in a non-trivial (nonlinear and non-local) manner that depends on the chemical potential. Figure 26. ln (ρ(s1)/ρ(s2)) versus E12 = E(s1) −E(s2), where ρ(s1) (resp. ρ(s2)) is the nucleosome density (computed as explained in the text, from the budding yeast genome) and E(s1) (resp. E(s2)) the nucleosome formation energy at the position s1 (resp. s2). The crosses correspond to two statistical sam- ples in a diluted ( ˜μ = −6 kT, red) and dense ( ˜μ = 0 kT, black) non-uniform ﬂuid. As shown in Figure 26, the histone octamer sequence speciﬁcity can be estimated at a low value of the chemi- cal potential ˜μ (= −6 kT) for a diluted system where the ratio of the nucleosome densities at two diﬀerent points s1 and s2 (here separated by 1 kbp) is given by ρ(s1)/ρ(s2) ≃ e−β(E(s1)−E(s2)). For a more concentrated system at a higher value of ˜μ (= 0 kT), the diﬀerence in energy E12 = E(s1) −E(s2) is no longer suﬃcient to specify the den- sity diﬀerence between the two points as the consequence of the interactions between particles. Thus at high nucle- osome density, a same nucleosome formation energy can lead to very diﬀerent nucleosome densities due to diﬀerent energetic environments. When the nucleosome density is increased, ordering progressively appears leading to an overall organization with ‘crystal-like’ phases of regularly positioned nucle- osomes conﬁned near or in between excluding energy barriers, coexisting with ‘ﬂuid-like’ phases where ordering is lost, in agreement with the statistical ordering principles described in the section ‘Statistical positioning’. ‘Intrinsic’ nucleosome formation energy landscape The main feature in the nucleosome occupancy heat map shown in Figure 25(a) is the fact that the highest energy barriers present in the energy landscape (Figure 25b) corre- spond to regions that are robustly depleted in nucleosomes whatever the overall nucleosome density. At low density (low ˜μ values), the conﬁnement is weak and the nucleo- somes distribute everywhere in between the highest energy barriers according to the energy landscape ﬂuctuations. Remark: As illustrated in Figure 25(b) and in Figure 43, see later on, the energy landscape computed in budding yeast presents a disordered topography as the consequence of the ‘disordered’ organization of the underlying genome; similar genome wide behavior is observed for Hemias- comyta yeasts as well as for S. pombe. As we shall see in the concluding section, at the genome scale, this genomic ‘disorder’ is characterized by (almost) gaussian statistics 22 (a) (b) (c) (e) (d) Figure 27. Energy landscape E(s) computed with the following parameter values: δ = ⟨(E −E)2⟩1/2 = 2 kT and lw = 125 bp along 10 kbp fragments of the chromosome 1 of C. elegans (green). Disordered patterns (regions D1, D2 and D3) are alternating with regular patterns, either quasi-ﬂat (regions F1, F2, F3, F4 and F5) or periodic with periodic stretches of barriers/wells (R1 and R2). In red are reported the experimental occupancy data δY(s) of Valouev et al. (2008). (a) (b) (c) (e) (d) Figure 27. Energy landscape E(s) computed with the following parameter values: δ = ⟨(E −E)2⟩1/2 = 2 kT and lw = 125 bp along 10 kbp fragments of the chromosome 1 of C. elegans (green). Disordered patterns (regions D1, D2 and D3) are alternating with regular patterns, either quasi-ﬂat (regions F1, F2, F3, F4 and F5) or periodic with periodic stretches of barriers/wells (R1 and R2). In red are reported the experimental occupancy data δY(s) of Valouev et al. (2008). et al. (2008) (resp. Kaplan et al. 2009) model based on sta- tistical learning. Furthermore, this very satisfactory mean Pearson correlation value really reﬂects the pertinence and consistency of our physical model all along the S. cere- visiae chromosomes (Chevereau et al. 2009). ‘Intrinsic’ nucleosome formation energy landscape As shown in Figure 29, the histogram of correlation values computed in 1 kbp sliding windows along the entire genome is mainly concentrated over a range 0.7 < r < 1, with a well-deﬁned maximum for a value as high as r = 0.85. For the sake of comparison, we have also reported the histogram of correla- tion values obtained between the predictions of our physical model and those of the Field et al. (2008) statistical model. Thishistogramisevenmoreconcentratedatverylarger val- ues with a rather sharp maximum for r = 0.92. This brings the demonstration that our model based on the structural and mechanical properties of the DNA double helix performs as well as rather sophisticated models requiring statistical learning (Field et al. 2008; Kaplan et al. 2009; Tillo and Hughes 2009). This is conﬁrmed in Figure 30 where the experimental in vitro distribution of nucleosome occupancy values and auto-correlation function (Figures 10c and d) are quite well reproduced by our physical model. Impor- tantly as observed in the data (Figure 10d), our physical model for low ˜μ predicts (as dictated by the energy pro- ﬁle) no oscillatory modulation of the (power-law) decay of the auto-correlation function of the nucleosome occupancy and long-range correlations. However, besides this dis- ordered organization, genomes of higher eukaryotes are largely composed by repeated sequences (Richard et al. 2008) that lead to ‘periodic’/regular patterns in the energy landscape as shown in Figure 27 for C. elegans. The shape of the regular patterns depends on the repeated motif (in par- ticular the motif’s size) and interestingly, for short repeats (<10 bp) the E(s) proﬁle can become almost ﬂat (see part F5 in Figure 27b). However, due to their periodic nature, current biochemical mapping methods cannot provide any reliable chromatin pattern in these regions (see parts F1, F5 and R2 in Figure 27). Modelling of in vitro nucleosome occupancy data in S. cerevisiae As shown in Figure 28 for 10 kbp fragments of various S. cerevisiae chromosomes, when adjusting the chemi- cal potential ˜μ = −6 kT to obtain the nucleosome density (30%) observed in vitro by Kaplan et al. (2009) (subsection ‘In vitro nucleosome occupancy proﬁles’), we get nucleo- some occupancy proﬁles that reproduce quite impressively the data (Chevereau et al. 2009). The mean Pearson cor- relation computed along the 12 Mbp of the budding yeast genome is r = 0.74, a result which is as good as the correla- tion value r = 0.74 (resp. r = 0.89) obtained with the Field 23 (a) (b) (c) (d) Figure 28. Comparison between the experimental occupancy proﬁle from the in vitro MNase-seq experiment of Kaplan et al. (2009) (orange), the theoretical low-density occupancy proﬁle (blue) and the energy landscape (green) (subsection “Intrinsic’ nucleosome for- mation energy landscape’) over regions of 10 kbp of several S. cerevisiae chromosomes. The theoretical predictions were obtained with the following parameter values: ˜μ = −6 kT, δ = 2 kT and lw = 125 bp. (a) (b) (c) (d) Figure 28. Comparison between the experimental occupancy proﬁle from the in vitro MNase-seq experiment of Kaplan et al. (2009) (orange), the theoretical low-density occupancy proﬁle (blue) and the energy landscape (green) (subsection “Intrinsic’ nucleosome for- mation energy landscape’) over regions of 10 kbp of several S. cerevisiae chromosomes. The theoretical predictions were obtained with the following parameter values: ˜μ = −6 kT, δ = 2 kT and lw = 125 bp. Figure 29. Histograms of Pearson correlation values r as mea- sured in a 1 kbp sliding window between our physical modelling ( ˜μ = −6 kT, δ = 2 kT and lw = 125 bp) and the Kaplan et al. S. cerevisiae in vitro MNase-seq data (Kaplan et al. 2009) (light blue), Field et al. statistical model (Field et al. 2008) (pink) and a random occupancy landscape (black). thereby conﬁrming the absence of statistical ordering at low nucleosome density. Remark: Let us mention that other attempts to describe nucleosome positioning in S. cerevisiae based on sequence- dependent nucleosome score (N-score) models built from the learning on two training sets of sequences, one corre- sponding to nucleosome sequences and the other one to linker sequences, do not provide competitive predictions (Peckham et al. 2007; Yuan and Liu 2008). For example, the Peckham et al. Modelling of in vitro nucleosome occupancy data in S. cerevisiae model performance for Kaplan et al. in vitro MNase-seq data is rather modest as quantiﬁed by a mean Pearson correlation r = 0.48. Concerning other physical models proposed using either diﬀerent di- or tri-nucleotide coding tables (Miele et al. 2008) rather than the PNuc cod- ing table used here or constructed ab initio (Tolstorukov et al. 2007; Morozov et al. 2009; Tolkunov and Moro- zov 2010), the obtained performances are even poorer, e.g. r = 0.38 for the Miele et al. model, and r = 0.01 for the Tolstorukov et al. model. However, very recent physical modelling based on a new ab initio computation of DNA Figure 29. Histograms of Pearson correlation values r as mea- sured in a 1 kbp sliding window between our physical modelling ( ˜μ = −6 kT, δ = 2 kT and lw = 125 bp) and the Kaplan et al. S. cerevisiae in vitro MNase-seq data (Kaplan et al. 2009) (light blue), Field et al. statistical model (Field et al. 2008) (pink) and a random occupancy landscape (black). 24 (a) (b) Figure 30. Comparing the predictions of our physical modelling ( ˜μ = −6 kT, δ = 2 kT and lw = 125 bp) with the Kaplan et al. S. cere- visiae in vitro MNase-seq data (Kaplan et al. 2009). (a) Histograms of nucleosome occupancy Y(s) values centered at their typical value: model (blue), in vitro data (orange). (b) Corresponding auto-correlation function C(s) = ⟨δY(s)δY(s + s)⟩; the green curve corresponds to the auto-correlation function of the theoretical nucleosome formation energy proﬁle (see the chromosome 7 panel in Figure 28). (b) (a) Figure 30. Comparing the predictions of our physical modelling ( ˜μ = −6 kT, δ = 2 kT and lw = 125 bp) with the Kaplan et al. S. cere- visiae in vitro MNase-seq data (Kaplan et al. 2009). (a) Histograms of nucleosome occupancy Y(s) values centered at their typical value: model (blue), in vitro data (orange). (b) Corresponding auto-correlation function C(s) = ⟨δY(s)δY(s + s)⟩; the green curve corresponds to the auto-correlation function of the theoretical nucleosome formation energy proﬁle (see the chromosome 7 panel in Figure 28). (a) (b) (c) (d) Figure 31. Comparison between the experimental occupancy proﬁle from the in vivo MNase-chip experiment of Lee et al. Modelling of in vitro nucleosome occupancy data in S. cerevisiae (2007) (red), the theoretical high-density occupancy proﬁle (blue) and the energy landscape (green) (subsection “Intrinsic’ nucleosome formation energy landscape’) over regions of 10 kbp of several S. cerevisiae chromosomes. The theoretical predictions were obtained with the following parameter values: ˜μ = −1.3 kT, δ = 2 kT and lw = 125 bp. (a) (b) (c) (d) Figure 31 Comparison between the experimental occupancy proﬁle from the in vivo MNase chip experiment of Lee et al (2007) (red Figure 31. Comparison between the experimental occupancy proﬁle from the in vivo MNase-chip experiment of Lee et al. (2007) (red), the theoretical high-density occupancy proﬁle (blue) and the energy landscape (green) (subsection “Intrinsic’ nucleosome formation energy landscape’) over regions of 10 kbp of several S. cerevisiae chromosomes. The theoretical predictions were obtained with the following parameter values: ˜μ = −1.3 kT, δ = 2 kT and lw = 125 bp. sequence-dependent elasticity provides a very good match at TSS and TTS of yeast genes (Deniz et al. 2011). the in vivo nucleosome density (75%), our physical model predictsnucleosomeoccupancyproﬁlesthatarestillingood agreement with the experimental data (Chevereau et al. 2009). However, as reported in Figure 32(a), the histogram ofPearsoncorrelationvaluesissigniﬁcantlyshiftedtolower values as compared to the one previously obtained at lower (in vitro) nucleosome density in Figure 29, with a mean value r = 0.33 and a rather wide support. The weakest cor- relations observed with our model are also shared by other Modelling of in vivo nucleosome occupancy data in S. cerevisiae As shown in Figure 31 for various budding yeast chro- mosomes (same 10 kbp contigs as in Figure 28), when increasing the chemical potential ˜μ (= −1.3 kT) to reach 25 (a) (b) (c) Figure 32. Histograms of Pearson correlation values r between the Lee et al. S. cerevisiae in vivo MNase-chip data (Lee et al. 2007) and our physical modelling ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) (blue), Yuan and Liu (2008) model (pink) and a random occupancy landscape (black). The Pearson correlation was measured in a 1 kbp sliding window over the 16 yeast chromosomes: (a) no shift d = 0 between the theoretical and experimental signal; (b) for a shift dM that maximizes the correlation; (c) histogram of optimal shift dM values. In (a,b), the green dots correspond to the histogram of Pearson correlation values obtained between the in vivo data and the theoretical nucleosome formation energy proﬁle (actually with the aﬃnity −E(s)) (see the chromosome 7 panel in Figure 31). (a) (c) (b) of Pearson correlation values r between the Lee et al. S. cerevisiae in vivo MNase-chip data (Lee et a ng ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) (blue), Yuan and Liu (2008) model (pink) and a random occ Figure 32. Histograms of Pearson correlation values r between the Lee et al. S. cerevisiae in vivo MNase-chip data (Lee et al. 2007) and our physical modelling ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) (blue), Yuan and Liu (2008) model (pink) and a random occupancy landscape (black). The Pearson correlation was measured in a 1 kbp sliding window over the 16 yeast chromosomes: (a) no shift d = 0 between the theoretical and experimental signal; (b) for a shift dM that maximizes the correlation; (c) histogram of optimal shift dM values. In (a,b), the green dots correspond to the histogram of Pearson correlation values obtained between the in vivo data and the theoretical nucleosome formation energy proﬁle (actually with the aﬃnity −E(s)) (see the chromosome 7 panel in Figure 31). (a) (b) Figure 33. Comparing the predictions of our physical modelling ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) with the Lee et al. (2007) S. cerevisiae in vivo MNase-chip data. Modelling of in vivo nucleosome occupancy data in S. cerevisiae (a) Histograms of nucleosome occupancy Y(s) values centered at their typical value: model (blue), in vivo data (red). (b) Corresponding auto-correlation function C(s) = ⟨δY(s)δY(s + s)⟩. (b) (a) Figure 33. Comparing the predictions of our physical modelling ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) with the Lee et al. (2007) S. cerevisiae in vivo MNase-chip data. (a) Histograms of nucleosome occupancy Y(s) values centered at their typical value: model (blue), in vivo data (red). (b) Corresponding auto-correlation function C(s) = ⟨δY(s)δY(s + s)⟩. Figure 34. Histograms of local NRL: comparison of the pre- dictions of our physical modelling ( ˜μ = −1.3 kT, δ = 2 kT and lw = 125 bp) for budding yeast (blue) (the same histogram is obtained for C. elegans) with the in vivo S. cerevisiae MNase-chip data of Lee et al. (2007) (red) and C. elegans MNase-seq data of Valouev et al. (2008) (green). models based on statistical learning like the Yuan and Liu (2008) model that yields a Pearson correlation histogram very similar to the one obtained with our physical model (Figure 32a), the Field et al. (2008) model (r = 0.39), the Kaplan et al. (2009) model (r = 0.34) and the Peckham et al. (2007) model (r = 0.22). As a careful inspection of the theoretical and in vivo experimental nucleosome occupancy proﬁles in Figure 31 seems to indicate, these weakest correlations result from two main features, namely, (i) experimental NFRs that do not correspond to genomic energy barriers but more likely result from the action of external factors like transcription of other proteic factors (TF, Insulators, PIC…) and (ii) regions (up to 1 kbp) where the experimental nucleosomal pattern is shifted by a few tens bp with respect to the predicted nucleosomal pattern as the possible outcome of ATP consuming remodelling fac- tors. We will come back to this point in more details in the next subsection. Figure 34. Histograms of local NRL: comparison of the pre- dictions of our physical modelling ( ˜μ = −1.3 kT, δ = 2 kT and lw = 125 bp) for budding yeast (blue) (the same histogram is obtained for C. elegans) with the in vivo S. cerevisiae MNase-chip data of Lee et al. (2007) (red) and C. elegans MNase-seq data of Valouev et al. (2008) (green). observed in vivo in the two-point correlation function (Figure 33b). From in vitro to vivo: ‘Intrinsic’ versus ‘extrinsic’ nucleosome positioning From in vitro to vivo: ‘Intrinsic’ versus ‘extrinsic’ nucleosome positioning have introduced and discussed in subsection ‘Homoge- neous energy proﬁle E = Eo’, as characterizing the spatial periodic modulation of the two-point nucleosome distri- bution. The equality holds in the particular case of the semi-conﬁned ﬂuids, i.e. around vertical energy barriers, where both density and pair distribution function indeed coincide. The extreme situation is in regions where the density (occupancy) is almost ﬂat (or unorganized) and where, consequently, the two-point correlation function (or the Fourier spectrum analysis) cannot account for the internal periodic ordering of the nucleosomal array. These results conﬁrm that the stretches of well-ordered nucleo- somes observed in vivo but not in vitro (i.e. at high but not at low nucleosome density) are the direct consequence of the organizing role of eﬀective nucleosome energy barri- ers that condition nucleosome ordering over rather long distances consistent with the statistical physics princi- ples (see section ‘Thermodynamical model of nucleosome assembly’). The small-scale chromatin structure, as deﬁned by the local nucleosome occupancy, conditions the regulation of tran- scriptioninparticularbymodulatingtheaccessibilityofTFs to their cognate regulatory sites (Kornberg and Lorch 1999; Li et al. 2007; Morse 2007; Rando and Ahmad 2007; Segal and Widom 2009b). Actually, as seen in the previous sec- tions, the nucleosome occupancy proﬁle predicted, directly from the DNA sequence, by our physical model using a grand canonical description, accounts remarkably well for the nucleosome occupancy proﬁle observed in vitro (see subsection ‘Modelling of in vitro nucleosome occupancy data in S. cerevisiae’) (Vaillant et al. 2007; Chevereau et al. 2009). However, the comparison with in vivo data reveals that the ‘intrinsic’ nucleosome positioning encoded in the sequence can be inﬂuenced and perturbed by the action of ‘extrinsic’ factors like TFs and ATP-dependent remodellers (Segal and Widom 2009; Radman-Livaja and Rando 2010). Remark: As noticed in the section ‘In vivo and in vitro genome-wide primary structure of chromatin’, the fact that our physical modelling reproduces quite well the distribu- tion of in vivo nucleosome occupancy values obtained in S. cerevisiae by Lee et al. (2007) (MNase-chip), questions the reliability of the in vivo Kaplan et al. (2009) (MNase-seq) data that surprisingly yields the same nucleosome occu- pancy histogram as obtained in vitro (Figures 7b and 10c) whereas the nucleosome density is more than twice bigger in vivo (75%) than in vitro (30%). Modelling of in vivo nucleosome occupancy data in S. cerevisiae As reported in Figure 34, the distribution of the local NRL values obtained from the predicted nucleosome proﬁle is indeed shifted towards larger val- ues (centered around l∗= 172 bp), but present a ﬂuctuation range similar to the experimental distribution. Note that the NRL computed as the periodic modulation of the two-point correlation of the (one-point) nucleosome dis- tribution is not stricto sensus equal to the NRL that we Let us emphasize that at a statistical level, our phys- ical model accounts very well for the in vivo distri- bution of nucleosome occupancy values obtained from the Lee et al. (2007) MNase-chip data (Figure 33a) as well as for the harmonic modulation with a period l∗= 172 bp which is slightly larger than the NRL l∗= 167 bp 26 Transcription factors Even more interesting, TFs reside in the predicted nucleosome depleted linker regions of the intergenic regular nucleosome array suggesting some coop- erativity in TF binding (Adams and Workman 1995; Polach and Widom 1996; Vashee et al. 1998; Miller and Widom 2003) as the result of a collaborative competition against ‘intrinsic’ collective nucleosome ordering and not just of speciﬁc protein–protein interactions. Note that the theoret- ical NFR of the anti-sense gene at the left of Figure 35(b) has been disturbed (mainly shifted) by the binding of TFs that have induced a second NFR nearby possibly catalyzed by remodelling factors. value r = 0.5. Note that a similar histogram is now obtained with a random control where our 1 kbp theoretical proﬁles were compared to randomly chosen 1 kbp experimental pro- ﬁles along the 16 yeast chromosomes. But as shown in Figure 32(c), whereas the distribution of optimal shift dM values is rather ﬂat for the control, it is narrowly peaked around dM = 0 with a width ≃60 bp, conﬁrming that up to some local shift of a few tens bp, this sequence-dependent model accounts remarkably well for in vivo nucleosome positioning data. Let us point out that these distances are typical of distances over which the ATP consuming remod- ellingfactorsareknowntooperateinvivo(Whitehouseetal. 2007; Shivaswamy et al. 2008; Hartley and Madhani 2009). This ‘extrinsic’ nucleosome positioning under the action of remodelling factors is likely to explain the strong phasing of the 5′ NFR with respect to the gene TSS observed in vivo as compared to our physical model predictions as discussed in Chevereau et al. (2009), and Vaillant et al. (2010). But as noticed by Rippe et al. (2007), some remodelling complexes only change the relative nucleosome occupancy without altering nucleosome positions. This suggests that equilibrium positioning can also be relevant even during remodelling activity. Since these chromatin remodellers are found all over the yeast chromosomes, they may contribute to increasing the eﬀective temperature so that thermal equi- librium is attained much faster, in particular along most yeast genes as discussed in the subsection ‘Bistability induced by statistical conﬁning in between two energy bar- riers’. Transcription factors TFs can inﬂuence nucleosome positioning in vivo by com- peting with histones to access to their DNA target sites (Koerber et al. 2009). The outcome of this competition likely depends on the relative aﬃnities of the nucleosomes and TFs to the underlying DNA sequence but also on their relative concentrations (Segal and Widom 2009c). As shown in Figures 35(a) and (b), when comparing the nucle- osome occupancy proﬁle predicted by our physical model (a) (b) (c) (d) Figure 35. Nucleosome occupancy proﬁles observed in vivo (red) and predicted by our physical model for parameter values ˜μ = −1.3 kT, δ = 2 kT and lw = 125 bp (blue) along fragments of S. cerevisiae chromosome 2 (a), 7 (b), 2 (c) and 6 (d). For comparison also represented are the corresponding theoretical energy landscapes (green). The symbols represent the positions of TSS (red dots) and TFS (black triangle). The arrows at TSS indicate the transcription sense. (b) (d) (b) (d) Figure 35. Nucleosome occupancy proﬁles observed in vivo (red) and predicted by our physical model for parameter values ˜μ = −1.3 kT, δ = 2 kT and lw = 125 bp (blue) along fragments of S. cerevisiae chromosome 2 (a), 7 (b), 2 (c) and 6 (d). For comparison also represented are the corresponding theoretical energy landscapes (green). The symbols represent the positions of TSS (red dots) and TFS (black triangle). The arrows at TSS indicate the transcription sense. 27 27 at high nucleosome density with the in vivo yeast data, we observe mainly two kinds of diﬀerences. There are locations where a NFR is observed in vivo but not in vitro as predicted by our physical model (Figure 35a). At many other loca- tions where some nucleosome depletion is predicted by our model, the in vivo nucleosome occupancy proﬁle displays a deeper and pronounced depleted region. At a larger scale, as seen in the intergenic region between the two divergent yeast genes in Figure 35(b), the high concentration of TFs coincides with a signiﬁcant lowering of the in vivo mean nucleosome occupancy. Importantly, this lowering has not disturbed the regular nucleosome ordering predicted by the physical model consistent with the emerging view that nucleosome and TFs compete to occupy DNA in thermo- dynamic equilibrium (Raveh-Sadka et al. 2009; Segal and Widom 2009b, 2009c). Transcription factors In that context, our grand canonical physical model can be used as a theoretical reference for in vitro nucleo- some positioning whose comparison with in vivo data is likely to provide very instructive information for future modelling of both remodeller (Teif and Rippe 2009) and TF (Raveh-Sadka et al. 2009) driven ‘extrinsic’ nucleosome positioning. ATP-dependent chromatin remodelling factors Nucleosome positioning can also be controlled by a family of enzymes that consume the energy from ATP hydroly- sis to move nucleosomes to diﬀerent locations along the DNA or even to disassemble nucleosomes (Tsukiyama and Wu 1997, Längst et al. 1999; Lorch et al. 1999; Travers 1999; Whitehouse et al. 1999, 2007; Peterson and Work- man 2000; Hamiche et al. 2001; Angelov et al. 2003; Boeger et al. 2008; Shivaswamy et al. 2008; Hartley and Madhani 2009; Clapier and Cairns 2009). In vitro these molecular motors were shown to actively drive nucleosomes away from presumed equilibrium positions (Montel et al. 2007; Rippe et al. 2007) and this regardless of the underlying DNA sequence. As illustrated in Figures 35(c) and (d), some in vivo nucleosome occupancy patterns including the 5′ NFR (resp. 3′ NFR) and the ﬂanking ordered nucleo- somes inside the corresponding genes are globally shifted by ∼50–100 bp from their predicted positions by our DNA sequence directed grand canonical modelling. This obser- vation is conﬁrmed statistically in Figures 32(b) and (c) where we have re-computed the Pearson correlation his- togram between our physical theoretical occupancy proﬁles and the in vivo experimental ones of Lee et al. (2007), when allowing a possible shift (−200 < d < 200 bp) between the numerical and experimental proﬁles. When optimizing the shift (dM) for each 1 kbp sliding window, the histogram is signiﬁcantly shifted towards higher values with a mean Remark: Let us note that at high density, our ‘intrinsic’ model could be improved by allowing a short-range attrac- tion between nucleosomes as the result of the folding of the nucleosomal array into higher-order chromatin struc- ture (Chereji et al. 2011; Riposo and Mozziconacci 2012). Works in this direction are under progress. S. kluyveri As shown in Figure 36(a) on a 10 kbp fragment of S. kluyverii chromosome C, when using the same parame- ters ( ˜μ = −1.3 kT, δ = 2 kT and lw = 125 bp) as used in the previous subsection to model the nucleosome occupancy proﬁles observed in vivo in S. cerevisiae, our physical model again predicts theoretical proﬁles that match rather well the experimental data of Tsankov et al. (2010) (Figure 3). When averaging over the eight S. kluyveri chromosomes, we get a mean Pearson correlation r = 0.32 in good agreement with the value previously obtained for S. cerevisiae (r = 0.33). 28 (a) (b) Figure 36. Comparison between our physical model predictions ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) (blue) and in vivo nucleosome occupancy MNase-seq data (Tsankov et al. 2010, 2011) (orange): (a) S. kluyveri; 10 kbp fragment on chromosome C; (b) S. pombe; 10 kbp fragment on chromosome 2. (a) Figure 36. Comparison between our physical model predictions ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) (blue) and in vivo nucleosome occupancy MNase-seq data (Tsankov et al. 2010, 2011) (orange): (a) S. kluyveri; 10 kbp fragment on chromosome C; (b) S. pombe; 10 kbp fragment on chromosome 2. Let us point out that quite similar results are obtained for other Hemiascomycota fungi (data not shown). (Figure4a)andthepredictionsofmodelsbasedonstatistical learning (Field et al. 2008; Kaplan et al. 2009). S. pombe As can be seen by a simple visual inspection of Figure 37, our physical model accounts much better for the C. ele- gans in vivo nucleosome occupancy MNase-seq data of Valouev et al. (2008) (see Figure 5) and this without requir- ing any change in the model parameters estimated from S. cerevisiae data. As shown in Figure 38, the histogram of Pearson correlation values computed in a 10 kbp slid- ing window along the entire C. elegans genome is rather wide with a maximum at r∗= 0.51 and a mean r = 0.43. This mean value is signiﬁcantly larger than the mean values previously obtained for the diﬀerent yeast genomes. Again The comparison of our physical model predictions with the nucleosome occupancy proﬁles observed in vivo in S. pombe (Figure 4) looks much less satisfactory as illus- trated in Figure 36(b) on a 10 kbp fragment of chromo- some 2. Indeed if the model parameters estimated in the previous subsection for S. cerevisiae are still rather opti- mal, the mean Pearson correlation obtained when averaging over the three S. pombe chromosomes r = 0.1 is much weaker than for S. cerevisiae and S. kluyverii. Let us men- tion that a similar observation was reported by Lantermann et al. (2010) when comparing their in vivo MNase-chip data (a) (b) Figure 37. Comparison between our physical model predictions ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) (blue) and the C. elegans in vivo nucleosome occupancy MNase-seq data (Valouev et al. 2008) (red): (a) and (b) correspond to two 15 kbp fragments of chromosome 1. Figure 37. Comparison between our physical model predictions ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) (blue) and the C. elegans in vivo nucleosome occupancy MNase-seq data (Valouev et al. 2008) (red): (a) and (b) correspond to two 15 kbp fragments of chromosome 1. 29 29 Figure 38. Histogram of Pearson correlation values r between our physical model ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) and the Valouev et al. MNase-seq in vivo nucleosome occupancy data (Valouev et al. 2008) (see Figure 37). The Pearson correlation r was measured in a 10 kbp sliding window over the six C. elegans chromosomes. Figure 38. S. pombe the ‘shifting’ distance above which the overlapping between predicted and experimental nucleo- somes is decreasing) is about 35 bp (Figure 40a). The fact that our physical model performs as modestly on well- positioned nucleosomes as the Peckham et al. (2007) and Yuan and Liu (2008) models, is an indication that these Human As previously noticed in the subsection ‘In vivo nucleo- some occupancy proﬁles’ (Figure 6), the sequencing depth in the human in vivo nucleosome occupancy MNase-seq data obtained by Schones et al. (2008) is too weak to allow us to perform a quantitative comparison with the predictions of our physical model. However, as shown in Figure 39 on several 10 kbp fragments of the human chromosome 6, the agreement is good. In particular, when focusing on the nucleosome occupancy proﬁles observed around human gene TSS, the numerical and experimental mean proﬁles are quite consistent, in particular for CpG poor gene pro- moters (data not shown). A more quantitative comparison with the recent in vivo and in vitro nucleosome occupancy MNase-seq data of Valouev et al. (2011) is currently under progress. S. pombe Histogram of Pearson correlation values r between our physical model ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) and the Valouev et al. MNase-seq in vivo nucleosome occupancy data (Valouev et al. 2008) (see Figure 37). The Pearson correlation r was measured in a 10 kbp sliding window over the six C. elegans chromosomes. have concluded that a large set of well-deﬁned nucleosome positions could eﬀectively be related to a ‘genomic nucle- osomal pattern’ based on the 10 bp periodicity in the distri- bution of given dinucleotide steps (e.g. AA/TT) (Satchwell et al. 1986; Ioshikhes et al. 1996; Widom 1996, 2001). Indeed, using probabilistic models that take into account the matching of their patterns to the sequence and the steric hindrance between nucleosomes, they both obtained nucleosome occupancy proﬁles that correlate rather well with the in vivo experimental data, specially in regions of well-positioned nucleosomes. However, later studies by Peckham et al. (2007) and Yuan and Liu (2008) have seri- ously questioned the conclusions of Ioshikhes et al. (2006) and Segal et al. (2006). When using a HMM (Hidden Markov Model) prediction algorithm to compare the per- formances of their models to those of Ioshikhes et al. and Segal et al. models, they both found that no more than 20% of the in vivo nucleosome positioning above what is expected by chance is determined by intrinsic signals in the genomic DNA. As shown in Figure 40(a), when applying a similar HMM approach to the numerical nucle- osome occupancy proﬁles predicted by our physical model and then comparing to the set of well-positioned nucleo- somes obtained by Lee et al. (2007) on their S. cerevisiae in vivo data, we get performances quite similar to those obtained with the Yuan and Liu (2008) model. Indeed our physical model predicts 48.7% of true positive within a distance of 35 bp as compared to 42% by chance. This is actually nothing but the expression, at the level of (HMM derived) well-positioned nucleosomes, of the remodelling shiftingeﬀectspreviouslydiscussedinthesubsection‘ATP- dependent chromatin remodelling factors’. When focusing on well-positioned nucleosomes, the characteristic ‘remod- elling distance’ (i.e. the ‘shifting’ distance above which the overlapping between predicted and experimental nucleo- somes is decreasing) is about 35 bp (Figure 40a). The fact that our physical model performs as modestly on well- positioned nucleosomes as the Peckham et al. Discussion Rôle of the genomic sequence on nucleosome positioning: 10 bp periodicity versus long-range correlations? S. pombe (2007) and Yuan and Liu (2008) models, is an indication that these it compares and competes remarkably well with the mean correlation values obtained with models based on statistical learning, namely the Kaplan et al. (2009) model (r = 0.47), the Field et al. (2008) model (r = 0.46) and the Peckham et al. (2007) model (r = 0.29). As previously observed for S. cerevisiae, the performances of formely proposed physical models are by far less satisfactory, e.g. the much weaker r values obtained with the Miele et al. (2008) model (r = 0.21) and the Tolstorukov et al. (2007) model (r = −0.001). have concluded that a large set of well-deﬁned nucleosome positions could eﬀectively be related to a ‘genomic nucle- osomal pattern’ based on the 10 bp periodicity in the distri- bution of given dinucleotide steps (e.g. AA/TT) (Satchwell et al. 1986; Ioshikhes et al. 1996; Widom 1996, 2001). Indeed, using probabilistic models that take into account the matching of their patterns to the sequence and the steric hindrance between nucleosomes, they both obtained nucleosome occupancy proﬁles that correlate rather well with the in vivo experimental data, specially in regions of well-positioned nucleosomes. However, later studies by Peckham et al. (2007) and Yuan and Liu (2008) have seri- ously questioned the conclusions of Ioshikhes et al. (2006) and Segal et al. (2006). When using a HMM (Hidden Markov Model) prediction algorithm to compare the per- formances of their models to those of Ioshikhes et al. and Segal et al. models, they both found that no more than 20% of the in vivo nucleosome positioning above what is expected by chance is determined by intrinsic signals in the genomic DNA. As shown in Figure 40(a), when applying a similar HMM approach to the numerical nucle- osome occupancy proﬁles predicted by our physical model and then comparing to the set of well-positioned nucleo- somes obtained by Lee et al. (2007) on their S. cerevisiae in vivo data, we get performances quite similar to those obtained with the Yuan and Liu (2008) model. Indeed our physical model predicts 48.7% of true positive within a distance of 35 bp as compared to 42% by chance. This is actually nothing but the expression, at the level of (HMM derived) well-positioned nucleosomes, of the remodelling shiftingeﬀectspreviouslydiscussedinthesubsection‘ATP- dependent chromatin remodelling factors’. When focusing on well-positioned nucleosomes, the characteristic ‘remod- elling distance’ (i.e. 10bp periodicity With the objective of ﬁnding consensus nucleosome posi- tioning sequences, in a pioneering analysis of the original S. cerevisiae in vivo nucleosome occupancy data of Yuan et al. (2005), Ioshikhes et al. (2006) and Segal et al. (2006) 30 9. Comparison between our physical model predictions ( ˜μ = −1.3, δ = 2 kT, lw = 125 bp) (blue) and the in vivo nucleosome y MNase-seq data obtained by Schones et al. (2008) in human CD4+ T cells (red). The four panels correspond to 10 kbp fragments osome 6. Figure 39. Comparison between our physical model predictions ( ˜μ = −1.3, δ = 2 kT, lw = 125 bp) (blue) and the in vivo nucleosome occupancy MNase-seq data obtained by Schones et al. (2008) in human CD4+ T cells (red). The four panels correspond to 10 kbp fragments of chromosome 6. on strongly positioning sequences that present large varia- tions in nucleotide contents (from 0.1 to 0.5, Figure 41a). Among these sequences that have an anomalously large aﬃnity to the histone octamer to form the nucleosome, the 601 sequence (Lowary and Widom 1998; Thåström et al. 2004) was shown to have a gain in the formation energy of the DNA–histone complex of E = −4.9 ± 0.55 kT rel- ative to a reference sequence (the sea urchin 5S rRNA gene sequence (Dong et al. 1990)). This 601 sequence was recently shown to prevent the nucleosome from sliding (Shlyakhtenko et al. 2009) which explains that, for obvi- ous functional reasons, no organism actually possesses this sequence in its genome. As shown in Figure 41(a), the sequences that are known to bind to the histone octamers in eukaryotic organisms have a much weaker nucleotide content variability, typically 5–10% for the chicken and 3–12% for S. cerevisiae. As illustrated in Figure 41(b), this simply means that nucleosomes adapt themselves bet- ter on sequences that display a 10 bp periodicity with AA/TT/AT that oscillate in phase with each other and authors were right to moderate the too hasty statement made by Ioshikhes et al. (2006) and Segal et al. (2006) that 10 bp periodic dinucleotide patterns could explain a majority of well-positioned nucleosomes. However, let us point out that when computing the performance of the HMM predictions of our physical model in a 5 kbp sliding windows over the 16 S. Long-range correlations Actually when performing power-spectrum and correlation analysis on the in vivo nucleosome occupancy data of Lee et al. (2007) (Figure 1), we mainly reveal the existence of a mean period l∗∼167 ± 10 bp that corresponds to regular arrays of well-ordered nucleosomes. This chacteristic NRL manifests as a bump in the power spectrum (Figure 42) at high frequencies (1/167 bp−1) and not as a peak for strict periodicity, as the signature of some ﬂuctuations in the NRL values. As previously reported in subsection ‘ModellingofinvivonucleosomeoccupancydatainS.cere- visiae’ (Figure 33), this statistical nucleosome ordering can also be diagnosed from the periodic modulations observed in the auto-correlation function C(s) = ⟨δY(s)δY(s + s)⟩. But in addition and very importantly, when plot- ted in a logarithmic representation, the power spectrum displays a very convincing power law decay S(k) ∝k−ν, with exponent ν = 2H −1 = 0.74 ± 0.02 (H = 0.87) that is likely to be a direct consequence of the large-scale (low frequency k < 1/200) LRC regime observed in the yeast DNA bending proﬁle in Audit et al. (2001, 2002, 2004). Figure 40. Performances of our physical model ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) and of Yuan and Liu (2008) N-score model in terms of well-positioned nucleosomes as obtained by HMM methods. The comparison is made against the set of well-positioned nucleosomes obtained by Lee et al. (2007) on their S. cerevisiae in vivo experimental data using a similar HMM algorithm. Performance is measured by the proportion of true positive i.e. well-predicted positioned nucleosomes at a given overlapping distance of an experimental nucleosome. (a) Mean performance value versus the overlapping distance for the the- oretical predictions of our physical model for the nucleosome occupancy proﬁle (blue) and the energy landscape (green), and the Yuan and Liu N-score model (magenta). (b) Statistics of the performance values (at 35 bp accuracy) computed in a sliding win- dow of size 5 kbp along the entire S. cerevisiae genome for our theoretical nucleosome occupancy predictions (blue) and for the random control (black). The vertical dashed lines (black and blue) indicate the corresponding mean values. facing the minor groove, and out of phase with GC fac- ing the major groove. As originally pointed out in Audit et al. 10bp periodicity cerevisiae chromosomes, we obtain a histogram of performance values that now strikingly deviates from random expectation (Figure 40b). If performance is rather homogeneously distributed along the genome around the expectation value for the random control (42%), there is a greater heterogeneity for our theoretical predictions with a large proportion of genomic regions where performance is much better than expected by chance, while in some other regions it is worse. Again, this reﬂects at the level of well- positioned nucleosomes, which is what we observed for the Pearson correlation distribution in Figure 32. Actually the rather modest predictive power of methods based on a 10 bp periodicity of some di- or tri-nucleotides is not surprising since this periodicity has been established 31 (a) (b) Figure 40. Performances of our physical model ( ˜μ = −1.3 kT, δ = 2 kT, lw = 125 bp) and of Yuan and Liu (2008) N-score model in terms of well-positioned nucleosomes as obtained by HMM methods. The comparison is made against the set of well-positioned nucleosomes obtained by Lee et al. (2007) on their S. cerevisiae in vivo experimental data using a similar HMM algorithm. Performance is measured by the proportion of true positive i.e. well-predicted positioned nucleosomes at a given overlapping distance of an experimental nucleosome. (a) Mean performance value versus the overlapping distance for the the- oretical predictions of our physical model for the nucleosome occupancy proﬁle (blue) and the energy landscape (green), and the Yuan and Liu N-score model (magenta). (b) Statistics of the performance values (at 35 bp accuracy) computed in a sliding win- dow of size 5 kbp along the entire S. cerevisiae genome for our theoretical nucleosome occupancy predictions (blue) and for the random control (black). The vertical dashed lines (black and blue) indicate the corresponding mean values. (a) (b) peak that emerges in the power spectrum at the frequency 1/10 bp−1. This conﬁrms that if locally the 10 bp periodic- ity sketched in Figure 41(b) can help to phase and position some nucleosomes, at the genome scale this periodicity is clearly not exploited to position the majority of well-deﬁned nucleosomes observed in vivo. Long-range correlations In particular the power-law decay of the power-spectrum is well reproduced withanexponentν = 0.65(H = 0.82)and0.46(H = 0.73) respectively at low and high nucleosome densities which corroborates the existence of LRC in the numerical nucle- osome occupancy proﬁles and further strengthens the rele- vance of these LRC in the experimental in vitro and in vivo data. But the most important result reported in Figure 42 is the fact that these LRC are also observed in the power spectrum of the nucleosome formation energy landscape with a power-law exponent ν = 0.68 (H = 0.84), as likely dictated by the LRC encoded in the DNA sequence (Audit et al. 2001, 2002). Furthermore, as shown in Figure 43, when recomputing the energy landscape after randomly shuﬄing the DNA sequence, the obtained energy proﬁle displays uncorrelated Gaussian ﬂuctuations without any- more tail at large E > 0 values corresponding to the presence of excluding energy barriers in the genuine DNA sequence. This is a strong indication that the sequence sig- nalingwhichprevailsisexcludingenergybarriersthatresult from the presence of LRC in the DNA bending proﬁle. As reported in the subsection ‘Modelling of in vitro nucleo- some occupancy data in S. cerevisiae’, they explain the Figure 42. Power spectrum analysis of nucleosome occupancy proﬁles obtained from the in vivo data of Lee et al. (2007) (red), the in vitro data of Kaplan et al. (2009) (orange), the physical model described in the section ‘A sequence-dependent physical model of nucleosome occupancy’ for δ = 2 kT and low ˜μ = −6 kT (cyan) and high ˜μ = −1.3 kT (dark blue) nucleo- some density. For comparison, the green curve corresponds to the power spectrum of the formation energy landscape. The dashed lines correspond to the power-spectrum scaling expo- nent values ν = 0.65, 0.74, 0.68, 0.74 and 0.46 from top to bottom corresponding to the following Hurst exponent values H = 0.82, 0.87, 0.84, 0.87 and 0.77, respectively. at high frequency (l∗)−1 = (167 bp)−1 as an indication of a signiﬁcant weakening of periodic nucleosome ordering. But what is remarkable is the fact that the power spectrum (as well as the auto-correlation function) still presents a very convincing power-law behavior with exponent ν = 0.74 corresponding to a Hurst exponent value H = 0.85 > 1/2, the hallmark of the presence of LRC. (a) (b) (c) Figure 43. Long-range correlations (2001, 2002, 2004), when performing time–frequency analysis of eukaryotic genomic sequences using either di- nucleotide codings or more elaborated di- or tri-nucleotide experimentaltablescodingforthestructuraland/orbending properties of the DNA double helix, there is no signiﬁcant When reproducing this statistical analysis on the in vitro nucleosome occupancy data of Kaplan et al. (Figure 10), consistently with the disappearance of the periodic mod- ulations in the auto-correlation function (Figure 10d), the power spectrum in Figure 42 no longer displays a bump Figure 41. (a) Fraction in dinucleotides AA/TT/TA (3 bp moving average) at each position of centre aligned yeast, chicken and random chemically synthesized nucleosome-bound DNA sequences showing ∼10 bp periodicity of these dinucleotides. (b) Key dinucleotides inferred from the alignment are shown relative to the three-dimensional structure of one-half of the symmetric nucleosome. Adapted from Segal et al. (2006). Figure 41. (a) Fraction in dinucleotides AA/TT/TA (3 bp moving average) at each position of centre aligned yeast, chicken and random chemically synthesized nucleosome-bound DNA sequences showing ∼10 bp periodicity of these dinucleotides. (b) Key dinucleotides inferred from the alignment are shown relative to the three-dimensional structure of one-half of the symmetric nucleosome. Adapted from Segal et al. (2006). 32 32 Figure 42. Power spectrum analysis of nucleosome occupancy proﬁles obtained from the in vivo data of Lee et al. (2007) (red), the in vitro data of Kaplan et al. (2009) (orange), the physical model described in the section ‘A sequence-dependent physical model of nucleosome occupancy’ for δ = 2 kT and low ˜μ = −6 kT (cyan) and high ˜μ = −1.3 kT (dark blue) nucleo- some density. For comparison, the green curve corresponds to the power spectrum of the formation energy landscape. The dashed lines correspond to the power-spectrum scaling expo- nent values ν = 0.65, 0.74, 0.68, 0.74 and 0.46 from top to bottom corresponding to the following Hurst exponent values H = 0.82, 0.87, 0.84, 0.87 and 0.77, respectively. In Figure 42 are also shown for comparison the power spectra of the nucleosome occupancy proﬁles predicted by our physical model at low ( ˜μ = −6 kT) and high ( ˜μ = −1.3 kT) nucleosome density. As previously observed for the auto-correlation functions in Figures 30(b) and 33(b) respectively, the theoretical power spectra are in good agreement with the experimental ones. Long-range correlations Energy landscape statistics (E(s) = E(s) −¯E) computed with the following parameter values: δ = ⟨(E −¯E)2⟩1/2 = 2 kT and lw = 125 bp. The colors correspond to the LRC genomic DNA sequence (green) and to its randomly shuﬄed uncorrelated version (black). (a) E(s) along a 50 kbp long fragment of budding yeast chromosome 2. (b) Energy pdfs computed for the 16 yeast chromosomes. (c) Energy auto-correlation function C(s)/C(0) vs s. (a) (c) (b) Figure 43. Energy landscape statistics (E(s) = E(s) −¯E) computed with the following parameter values: δ = ⟨(E −¯E)2⟩1/2 = 2 kT and lw = 125 bp. The colors correspond to the LRC genomic DNA sequence (green) and to its randomly shuﬄed uncorrelated version (black). (a) E(s) along a 50 kbp long fragment of budding yeast chromosome 2. (b) Energy pdfs computed for the 16 yeast chromosomes. (c) Energy auto-correlation function C(s)/C(0) vs s. 33 33 with our physical model. Overall, it turns out that in terms of the genomic rule, the (G + C) content is the one that appears to contribute most signiﬁcantly to ‘intrinsic’ nucle- osome occupancy. As a consequence, all the models which are based on an aﬃnity of DNA to the histone octamer that strongly correlates to the (G + C) content are likely to pre- dict nucleosome occupancy proﬁles that match quite well the in vitro experimental data (Tillo and Hughes 2009). This is typically the case of the ‘Pnuc’ structural bending cod- ing table (Goodsell and Dickerson 1994; Gabrielian and Pongor 1996) used in our physical model (see subsection ‘Intrinsic’ nucleosome formation energy landscape) and of other di- or tri-nucleotide coding tables proposed in other studies (Anselmi et al. 2000; Miele et al. 2008; Santis et al. 2010). As previously mentioned, another well-established and major compositional factor is the poly(dA:dT) which areknowntocorrespondtorigidDNAfragmentsthatimpair nucleosome formation (Yuan et al. 2005; Bao et al. 2006; Segal and Widom 2009a; Tillo and Hughes 2009) and favor nucleosome disassembly (Iyer and Struhl 1995; Suter et al. 2000) by increasing the DNA wrapping free energy cost. Overall, as estimated on synthetic 150 bp oligonu- cleotides (Kaplan et al. 2009; Tillo and Hughes 2009), over the range of (G + C) content from 20% to 60%, the DNA/histone aﬃnities can be ﬁve-fold or greater. Long-range correlations This is typically the case of the ‘Pnuc’ structural bending cod- ing table (Goodsell and Dickerson 1994; Gabrielian and Pongor 1996) used in our physical model (see subsection ‘Intrinsic’ nucleosome formation energy landscape) and of other di- or tri-nucleotide coding tables proposed in other studies (Anselmi et al. 2000; Miele et al. 2008; Santis et al. 2010). As previously mentioned, another well-established and major compositional factor is the poly(dA:dT) which areknowntocorrespondtorigidDNAfragmentsthatimpair nucleosome formation (Yuan et al. 2005; Bao et al. 2006; Segal and Widom 2009a; Tillo and Hughes 2009) and favor nucleosome disassembly (Iyer and Struhl 1995; Suter et al. 2000) by increasing the DNA wrapping free energy cost. Overall, as estimated on synthetic 150 bp oligonu- cleotides (Kaplan et al. 2009; Tillo and Hughes 2009), over the range of (G + C) content from 20% to 60%, the DNA/histone aﬃnities can be ﬁve-fold or greater. This means that the diﬀerence in nucleosome formation energy between the genomic low (G + C) unfavorable sequences and high (G + C) more favorable sequences is ∼1.6 kT (maximum value ∼3 kT). As far as poly-A are concerned, the depletion observed with respect to random sequences actually depends on their size, namely 2 (−0.7 kT) to 6 (−1.7 kT) fold for 5 to 15 bp fragments up to 30 fold (−3.4 kT) for 25 bp fragments. Thus according to these in vitro data (Kaplan et al. 2009) which are ‘MNase indepen- dent’ (see below), the range of aﬃnities, or in other words, the energy variability for a genomic sequence to form the nucleosome is rather weak (as compared to strongly posi- tioning artiﬁcial sequences, as the 601 sequence (Lowary and Widom 1998, Thåström et al. 2004)) but non-negligible as implemented in our physical model in the subsection “Intrinsic’ nucleosome formation energy landscape’ by ﬁx- ing the parameter δ = ⟨(E −E)2⟩1/2 = 2 kT that controls the ﬂuctuation range in the energy landscape. (G + C) content drives nucleosome occupancy: experimental bias or reality? As originally pointed out by Miele et al. (2008), the nucle- osome occupancy proﬁle observed in vivo in S. cerevisiae turns out to be signiﬁcantly correlated to the local (G + C) content when estimated in a 125 bp sliding window. As shown in Figure 44, the (G + C) content provides an excellent prediction of the in vitro nucleosome occupancy MNase-seq data obtained by Kaplan et al. (2009) in S. cere- visiae. The mean Pearson correlation is as large as r = 0.78 and is comparable to the performances of our physical model (r = 0.74) and of more sophisticated models based on statistical learning (see the subsection ‘Modelling of in vitro nucleosome occupancy data in S. cerevisiae’). In vivo, this correlation is still strong and comparable with the ones obtained with the nucleosome proﬁles predicted by other models (see the subsection ‘Modelling of in vivo nucleo- some occupancy data in S. cerevisiae’). This is particularly true for the S. cerevisiae in vivo MNase-seq data of Kaplan et al. (2009), r = 0.40. A signiﬁcantly smaller but still sig- niﬁcant value r = 0.25 is obtained with the S. cerevisiae in vivo MNase-chip data of Lee et al. (2007). Note that a similar correlation between (G + C) content and nucleo- some occupancy is observed in the C. elegans in vivo data of Valouev et al. (2008): r = 0.42, as compared to r = 0.43 Long-range correlations This means that the diﬀerence in nucleosome formation energy between the genomic low (G + C) unfavorable sequences and high (G + C) more favorable sequences is ∼1.6 kT (maximum value ∼3 kT). As far as poly-A are concerned, the depletion observed with respect to random sequences actually depends on their size, namely 2 (−0.7 kT) to 6 (−1.7 kT) fold for 5 to 15 bp fragments up to 30 fold (−3.4 kT) for 25 bp fragments. Thus according to these in vitro data (Kaplan et al. 2009) which are ‘MNase indepen- dent’ (see below), the range of aﬃnities, or in other words, the energy variability for a genomic sequence to form the nucleosome is rather weak (as compared to strongly posi- tioning artiﬁcial sequences, as the 601 sequence (Lowary and Widom 1998, Thåström et al. 2004)) but non-negligible as implemented in our physical model in the subsection “Intrinsic’ nucleosome formation energy landscape’ by ﬁx- ing the parameter δ = ⟨(E −E)2⟩1/2 = 2 kT that controls the ﬂuctuation range in the energy landscape NFRs observed in vitro in S. cerevisiae as well as the regular nucleosome ordering observed nearby in vivo (subsection ‘ModellingofinvivonucleosomeoccupancydatainS.cere- visiae’) as the result of statistical conﬁning according to thermal equilibrium principles (Section ‘Statistical posi- tioning’). But as discussed in the subsection ‘From in vitro to in vivo: ‘Intrinsic’ versus ‘extrinsic’ nucleosome posi- tioning’, there are much more (2 or 3 times) NFRs observed in vivo than in vitro and than predicted from the energy barriers encoded in the DNA sequence. These additional NFRs likely result from the action of external factors (TF, chromatin remodellers) and contribute to strengthen the col- lective nucleosomal ordering observed in vivo and specially in S. cerevisiae genes as shown in the subsection ‘Bistabil- ity induced by statistical conﬁning in between two energy barriers’ (Figure 23). with our physical model. Overall, it turns out that in terms of the genomic rule, the (G + C) content is the one that appears to contribute most signiﬁcantly to ‘intrinsic’ nucle- osome occupancy. As a consequence, all the models which are based on an aﬃnity of DNA to the histone octamer that strongly correlates to the (G + C) content are likely to pre- dict nucleosome occupancy proﬁles that match quite well the in vitro experimental data (Tillo and Hughes 2009). Experimental artefact Figure 44. Comparison between the (G + C) content estimated in a 125 bp sliding window (black) and the S. cerevisiae in vitro nucleosome occupancy MNase-seq data of Kaplan et al. (2009) (orange). The horizontal line indicates the genome wide average (G + C) content value of 0.38. The observed important correlation between the (G + C) content and in vitro (and to a less extent in vivo) nucleosome occupancy data raises the issue of a possible experimen- tal artefact. Indeed, it is well known that the MNase presents a sequence speciﬁcity, cutting preferentially at AT steps (Dingwall et al. 1981; Hörz and Altenburger 1981). Recently, Chung et al. (2010) and Fan et al. (2010) have shown that MNase digestion proﬁles obtained on genomic naked DNA (S. cerevisiae) are indeed signiﬁcantly corre- lated to the (G + C) content ﬂuctuations as well as to the Figure 44. Comparison between the (G + C) content estimated in a 125 bp sliding window (black) and the S. cerevisiae in vitro nucleosome occupancy MNase-seq data of Kaplan et al. (2009) (orange). The horizontal line indicates the genome wide average (G + C) content value of 0.38. 34 chromatin digestion proﬁles obtained in vitro and in vivo. Similar observation at the promoters of yeast to ﬂy genes was already reported in Miele et al. (2008). As performed in some experimental studies (Yuan et al. 2005; Lantermann et al. 2010), a way to overcome this possible bias consists in normalizing the chromatin MNase digestion data by the corresponding ones obtained on naked DNA. Hopefully, in an experimental study that just appeared (Allan et al. 2012), equivalent patterns of nucleosome positioning sites were obtained when using two nucleases that have notable diﬀerences in cleavage behavior, namely the MNase and the Caspase-Activated DNase (CAD). These recent results indicate that possible biases (e.g. content) in nucleosome positioning collected using MNase are likely to be insignif- icant. As noticed by Chung et al. (2010) this correlation with (G + C) content might be related to the reconstitu- tion procedure (salt-gradient dialysis) that would naturally favor the sequence speciﬁcity of the tetramer (H3/H4)2 which has been shown to preferentially bind to (G + C) rich sequences at high salt concentration. The physical ori- gin of this sequence speciﬁcity (of the tetramer) and its eﬀectiveness in the in vivo context remain elusive. Experimental artefact (G + C) content dependent DNA/histones interactions Increasing experimental evidence seems to validate the central eﬀect of the mean (G + C) content on the nucleoso- mal organization of the 10 nm chromatin ﬁber as recently reported by Valouev et al. (2011) for diﬀerent types of human cells. This has led to a renewal in the modelling of DNA sequence evolution and specially of its (G + C) content across the eukaryotic kingdom in relationship with the availability of nucleosome occupancy data (Kenigs- berg et al. 2010). Let us note that this correlation of the nucleosome occupancy proﬁle with the (G + C) content is not revealed by most of the energetic models used in molecular dynamics to account for the detailed atomic (G + C) content dependent DNA/histones interactions Increasing experimental evidence seems to validate the central eﬀect of the mean (G + C) content on the nucleoso- mal organization of the 10 nm chromatin ﬁber as recently reported by Valouev et al. (2011) for diﬀerent types of human cells. This has led to a renewal in the modelling of DNA sequence evolution and specially of its (G + C) content across the eukaryotic kingdom in relationship with the availability of nucleosome occupancy data (Kenigs- berg et al. 2010). Let us note that this correlation of the nucleosome occupancy proﬁle with the (G + C) content is not revealed by most of the energetic models used in molecular dynamics to account for the detailed atomic (a) (b) (c) (d) (e) 45. Comparison between the S. kluyveri in vivo nucleosome occupancy proﬁle δY(s) of Tsankov et al. (2010) and the theoretical predicted by our physical model (see sections ‘Statistical positioning’ and ‘A sequence-dependent physical model of nucleosome ncy’). (a) Formation energy landscape E(s) = E(s) −E along the chromosome C computed with the following parameters T, lw = 125 bp; the high (G + C) content (52%) contig corresponds to the ﬁrst 1 Mbp of the chromosome; low (G + C) content = 40%) part corresponds to the last 250 kbp. (b,c) Comparison of the predictions of our physical modelling ( ˜μ = −1.3 kT) (dark yan) with Tsankov et al. (2010) data (Figure 3) (red/orange) along a 10 kbp fragment of the high/low (G + C) content contig ted in (a) by the red/orange segments). (d,e) Corresponding auto-correlation functions C(s) = ⟨δY(s)δY(s + s)⟩. Experimental artefact 2009): the ﬁrst 1 Mbp of the chromosome C has an average (G + C) content of 52.9% which is signiﬁcantly higher than the 40.4% of the rest of the genome. As revealed by the auto-correlation functions of the experimental in vivo data of Tsankov et al. (2010) reported in Figure 45(e), the observed NRL in the high (G + C) domain is actually the same as the one observed in the rest of the genome, i.e. a typical value of l∗= 167 bp (Figure 9). This contrasts with our physical model that clearly predicts a smaller NRL for the high (G + C) region (l∗= 167 bp) than for the rest of the genome (lower (G + C)) (l∗= 172 bp) (Figure 45d). This results from a lower nucleosome formation energy and thus a higher resid- ual chemical potential in high as compared to low (G + C) Acknowledgements We are very grateful to Y. d’Aubenton-Carafa, B. Audit, M. Barbi, C. Lavelle, R. Lavery, V. Miele, C. Thermes, A. Travers and J.M. Victor for very stimulating and fruitful discussions. This research was supported by the Conseil Régional Rhône-Alpes (project ‘Le rôle de la séquence sur la structure et la dynamique de la chroma- tine’, Emergence 2005) and the Agence Nationale de la Recherche ‘Programme Physique et Chimie du Vivant’ (project ‘DNAnucl: Inﬂuence de la séquence ADN sur la structure et la dynamique du nucléosome’, PCV2006). Experimental artefact But this raises the issue of modelling nucleosome occupancy data in higher eukaryotic organisms like in human where the so-called isochore structure mani- fests as large-scale domains (several 100 kbp) with uniform (G + C) content and appreciable scatter of the average (G + C) content when comparing domains (Bernardi 1989, 1995, 2000, 2001; Mouchiroud and Bernardi 1993; Lander et al. 2001; Li et al. 2003; Duret et al. 2006). Is the variabil- ity in the ﬂuctuations in the energy landscape, namely large δ in (G + C) rich isochores and small δ in (G + C) poor isochores, suﬃcient to reproduce the experimental nucle- osome densities and NRL in these regions? Do we need to change accordingly the chemical potential ˜μ as the pos- sible signature of compensatory co-evolutive mechanisms of regulation of the nucleosomal array? If yes, what are these sequence-dependent ((G + C) dependent) chromatin regulation mechanisms (Dekker 2008)? In Figure 45 are shown the results of a preliminary analysis of regions in the S. kluyveri genome that have a signiﬁcantly diﬀerent (G + C) content (Payen et al. 2009): the ﬁrst 1 Mbp of the chromosome C has an average (G + C) content of 52.9% which is signiﬁcantly higher than the 40.4% of the rest of the genome. As revealed by the auto-correlation functions of the experimental in vivo data of Tsankov et al. (2010) reported in Figure 45(e), the observed NRL in the high (G + C) domain is actually the same as the one observed in the rest of the genome, i.e. a typical value of l∗= 167 bp (Figure 9). This contrasts with our physical model that clearly predicts a smaller NRL for the high (G + C) the high (G+C) domain is exactly the one observed in vivo whatever the genomic content. This indicates that either our sequence-dependent model overestimates the inﬂuence of the (G + C) content variations or there is a mechanism that fully compensates/eliminates this ‘intrinsic’ nucleo- some energy diﬀerence. A similar comparative analysis of our physical model predictions with the human in vivo nucleosome occupancy data of both Schones et al. (2008) and Valouev et al. (2011) at the genome scale is currently under progress. It is likely to provide new insight on the isochore stucture of mammalian genomes in relation with their primary nucleosomal chromatin structure. interactions involved in the nucleosome complex (Moro- zov et al. Experimental artefact 2009; Tolkunov and Morozov 2010). Indeed the physical models that perform as well as the one used all along in this paper (Vaillant et al. 2007; Chevereau et al. 2009) are based on energetic coding tables that were established from the local structural and mechanical prop- erties of the DNA double helix, namely the ‘Pnuc’ coding table (Goodsell and Dickerson 1994; Gabrielian and Pon- gor 1996) in our study and the Anselmi et al. (2000) coding table in the Miele et al. (2008) model. In both these coding tables, the dominating contribution is given by the roll angle which deﬁnitely strongly correlates with the (G + C) content. A posteriori, it is therefore not so surprising that our physical model accounts so well for in vitro nucleosome positioning data in organisms like S. cerevisiae (see subsection ‘Modelling of in vivo nucleo- some occupancy data in S. cerevisiae’) where the (G + C) content is rather homogeneous around 39% with a typi- cal variability in the nucleosome formation energy proﬁle δ = 2 kT (see the subsection ‘Rôle of the genomic sequence on nucleosome positioning: 10 bp periodicity versus long- range correlations?’). But this raises the issue of modelling nucleosome occupancy data in higher eukaryotic organisms like in human where the so-called isochore structure mani- fests as large-scale domains (several 100 kbp) with uniform (G + C) content and appreciable scatter of the average (G + C) content when comparing domains (Bernardi 1989, 1995, 2000, 2001; Mouchiroud and Bernardi 1993; Lander et al. 2001; Li et al. 2003; Duret et al. 2006). Is the variabil- ity in the ﬂuctuations in the energy landscape, namely large δ in (G + C) rich isochores and small δ in (G + C) poor isochores, suﬃcient to reproduce the experimental nucle- osome densities and NRL in these regions? Do we need to change accordingly the chemical potential ˜μ as the pos- sible signature of compensatory co-evolutive mechanisms of regulation of the nucleosomal array? If yes, what are these sequence-dependent ((G + C) dependent) chromatin regulation mechanisms (Dekker 2008)? In Figure 45 are shown the results of a preliminary analysis of regions in the S. kluyveri genome that have a signiﬁcantly diﬀerent (G + C) content (Payen et al. Experimental artefact (a) (b) (c) (d) (d) (e) (e) Figure 45. Comparison between the S. kluyveri in vivo nucleosome occupancy proﬁle δY(s) of Tsankov et al. (2010) and the theoretical proﬁle predicted by our physical model (see sections ‘Statistical positioning’ and ‘A sequence-dependent physical model of nucleosome occupancy’). (a) Formation energy landscape E(s) = E(s) −E along the chromosome C computed with the following parameters δ = 2 kT, lw = 125 bp; the high (G + C) content (52%) contig corresponds to the ﬁrst 1 Mbp of the chromosome; low (G + C) content (G + C = 40%) part corresponds to the last 250 kbp. (b,c) Comparison of the predictions of our physical modelling ( ˜μ = −1.3 kT) (dark blue/cyan) with Tsankov et al. (2010) data (Figure 3) (red/orange) along a 10 kbp fragment of the high/low (G + C) content contig (indicated in (a) by the red/orange segments). (d,e) Corresponding auto-correlation functions C(s) = ⟨δY(s)δY(s + s)⟩. 35 interactions involved in the nucleosome complex (Moro- zov et al. 2009; Tolkunov and Morozov 2010). Indeed the physical models that perform as well as the one used all along in this paper (Vaillant et al. 2007; Chevereau et al. 2009) are based on energetic coding tables that were established from the local structural and mechanical prop- erties of the DNA double helix, namely the ‘Pnuc’ coding table (Goodsell and Dickerson 1994; Gabrielian and Pon- gor 1996) in our study and the Anselmi et al. (2000) coding table in the Miele et al. (2008) model. In both these coding tables, the dominating contribution is given by the roll angle which deﬁnitely strongly correlates with the (G + C) content. A posteriori, it is therefore not so surprising that our physical model accounts so well for in vitro nucleosome positioning data in organisms like S. cerevisiae (see subsection ‘Modelling of in vivo nucleo- some occupancy data in S. cerevisiae’) where the (G + C) content is rather homogeneous around 39% with a typi- cal variability in the nucleosome formation energy proﬁle δ = 2 kT (see the subsection ‘Rôle of the genomic sequence on nucleosome positioning: 10 bp periodicity versus long- range correlations?’). References Adams CC and Workman JL. 1995. Binding of disparate tran- scriptional activators to nucleosomal DNA is inherently cooperative. Mol Cell Biol. 15:1405–1421. Albert I Mavrich TN, Tomsho LP Qi J, Zanton SJ, Schuster SC, Pugh BF. 2007. Translational and rotational settings of H2A.Z nucleosomes across the Saccharomyces cerevisiae genome. Nature. 446:572–576. Alberts B, Bray D, Lewis J, RaﬀM, Roberts K, Watson JD. 2002. Molecular biology of the cell. 4th ed. New-York: Garland Publishing. Allan J, Fraser R, Owen-Hughes T, Keszenman-Pereyra D. 2012. Micrococcal nuclease does not substantially bias nucleosome mapping. J Mol Biol. 417:152–164. Angelov D, Molla A, Perche P-Y, Hans F, Cote J, Khochbin S, Bouvet P, Dimitrov S. 2003. The histone variant macroH2A interferes with transcription factor binding and SWI/SNF nucleosome remodeling. Mol Cell. 11:1033–1041. Anselmi C, Bocchinfuso G, Santis PD, Savino M, Scipioni A. 2000. A theoretical model for the prediction of sequence- dependent nucleosome thermodynamic stability. Biophys J. 79:601–613. Arneodo A, Vaillant C, Audit B, Argoul F, d’Aubenton Carafa Y, Thermes C. 2011. Multi-scale coding of genomic informa- tion: From DNA sequence to genome structure and function. Phys Rep. 498:45–188. Audit B, Thermes C, Vaillant C, d’Aubenton Carafa Y, Muzy JF, Arneodo A. 2001. Long-range correlations in genomic DNA: a signature of the nucleosomal structure. Phys Rev Lett. 86:2471–2474. Audit B, Vaillant C, Arneodo A, d’Aubenton Carafa Y, Ther- mes C. 2002. Long-range correlations between DNA bending sites: relation to the structure and dynamics of nucleosomes. J Mol Biol. 316:903–918. Audit B, Vaillant C, Arneodo A, d’Aubenton-Carafa Y, Thermes C. 2004. Wavelet analysis of DNA bending proﬁles reveals 36 Dong F, Hansen JC, van Holde KE. 1990. DNA and protein determinants of nucleosome positioning on sea urchin 5S rRNA gene sequences in vitro. Proc Natl Acad Sci USA. 87:5724–5728. structural constraints on the evolution of genomic sequences. J Biol Phys. 30:33–81. Bao Y, White CL, Luger K. 2006. Nucleosome core particles containing a poly(dA.dT) sequence element exhibit a locally distorted DNA structure. J Mol Biol. 361:617–624. Dorigo B, Schalch T, Kulangara A, Duda S, Schroeder RR, Richmond TJ. 2004. Nucleosome arrays reveal the two- start organization of the chromatin ﬁbers. Science. 306: 1571–1573. distorted DNA structure. J Mol Biol. 361:617–624 Baxter R. 1982. Exactly solved models in statistical mechanics. London: Academic Press. Bednar J, Horowitz RA, Grigoryev SA, Carruthers LM, Hansen JC, Koster AJ, Woodcock CL. 1998. References Calladine CR, Drew HR. 1999. Understanding DNA. San Diego: Academic Press. Chereji RV, Tolkunov D, Locke G, Morozov AV. 2011. Statistical mechanics of nucleosome ordering by chromatin-structure- induced two-body interactions. Phys Rev E: Stat Nonlin Soft Matter Phys. 83:050903. g Giaquinta P. 2008. Entropy and ordering of hard rods in one dimension. Entropy. 10:248–260. Gkikopoulos T, Schoﬁeld P, Singh V, Pinskaya M, Mellor J, Smolle M, Workman JL, Barton GJ, Owen-Hughes T. 2011. A role for Snf2-related nucleosome-spacing enzymes in genome-wide nucleosome organization. Science. 333: 1758–1760. y Chevereau G. 2010. Thermodynamique du positionnement des nucléosomes [Ph.D. thesis]. [Lyon (France)]: Université de Lyon-Ecole Normale Supérieure. Chevereau G, Palmeira L, Thermes C, Arneodo A, Vaillant C. 2009. Thermodynamics of intragenic nucleosome ordering. Phys Rev Lett. 103:188103. Goodsell DS, Dickerson RE. 1994. Bending and curva- ture calculations in B-DNA. Nucleic Acids Res. 22: 5497–5503. Chung H-R, Dunkel I, Heise F, Linke C, Krobitsch S, Ehrenhofer- Murray AE, Sperling SR, Vingron M. 2010. The eﬀect of micrococcal nuclease digestion on nucleosome positioning data. PLoS One. 5:e15754. Hall MA, Shundrovsky A, Bai L, Fulbright RM, Lis JT, Wang MD. 2009. High-resolution dynamic mapping of histone- DNA interactions in a nucleosome. Nat Struct Mol Biol. 16:124–129. Clapier CR, Cairns BR. 2009. The biology of chromatin remod- eling complexes. Annu Rev Biochem. 78:273–304. Hamiche A, Kang JG, Dennis C, Xiao H, Wu C. 2001. His- tone tails modulate nucleosome mobility and regulate ATP- dependent nucleosome sliding by NURF. Proc Natl Acad Sci USA. 98:14316–14321. Davis HT. 1990. Density distribution functions of conﬁned Tonks–Takahashi ﬂuids. J Chem Phys. 93:4339–4344. Dekker J. 2008. Mapping in vivo chromatin interactions in yeast suggests an extended chromatin ﬁber with regional variation in compaction. J Biol Chem. 283:34532–34540. Hansen J, McDonald I. 2006. Theory of simple liquids. London: Academic Press. Deniz O, Flores O, Battistini F, Pérez A, Soler-Lòpez M, Orozco M. 2011. Physical properties of naked DNA inﬂuence nucle- osome positioning and correlate with transcription start and termination sites in yeast. BMC Genomics. 12:489. Hartley PD, Madhani HD. 2009. Mechanisms that specify pro- moter nucleosome location and identity. Cell. 137:445–458. Hörz W, Altenburger W. 1981. Sequence speciﬁc cleavage of DNA by micrococcal nuclease. Nucleic Acids Res. 9: 2643–2658. y Depken M, Schiessel H. 2009. Nucleosome shape dictates chro- matin ﬁber structure. Biophys J. 96:777–784. Ioshikhes I, Bolshoy A, Derenshteyn K, Borodovsky M, Trifonov EN. 1996. References Nucleosomes, linker DNA, and linker histone form a unique structural motif that directsthehigher-orderfoldingandcompactionofchromatin. Proc Natl Acad Sci USA. 95:14173–14178. Duret L, Eyre-Walker A, Galtier N. 2006. A new perspective on isochore evolution. Gene. 385:71–74. Fan X, Moqtaderi Z, Jin Y, Zhang Y, Liu XS, Struhl K. 2010. Nucleosome depletion at yeast terminators is not intrinsic and can occur by a transcriptional mechanism linked to 3’-end formation. Proc Natl Acad Sci USA. 107: 17945–17950. Bernardi G. 1989. The isochore organization of the human genome. Annu Rev Genet. 23:637–661. Bernardi G. 1995. The human genome: Organization and evolu- tionary history. Annu Rev Genet. 29:445–476. Field Y, Kaplan N, Fondufe-Mittendorf Y, Moore IK, Sharon E, Lubling Y, Widom J, Segal E. 2008. Distinct modes of regu- lation by chromatin encoded through nucleosome positioning signals. PLoS Comput Biol. 4:e1000216. Bernardi G. 2000. Isochores and the evolutionary genomics of vertebrates. Gene. 241:3–17. Bernardi G. 2001. Misunderstandings about isochores. Part 1. Gene. 276:3–13. Fire A, Alcazar R, Tan F. 2006. Unusual DNA structures asso- ciated with germline genetic activity in Caenorhabditis elegans. Genetics. 173:1259–1273. Bernstein BE, Liu CL, Humphrey EL, Perlstein EO, Schreiber SL. 2004. Global nucleosome occupancy in yeast. Genome Biol. 5:R62. Floer M, Wang X, Prabhu V, Berrozpe G, Narayan S, Spagna D, Alvarez D, Kendall J, Krasnitz A, Stepansky A, Hicks J, Bryant GO, Ptashne M. 2010. A RSC/nucleosome complex determines chromatin architecture and facilitates activator binding. Cell. 141:407–418. Blank TA, Becker PB. 1995. Electrostatic mechanism of nucleo- some spacing. J Mol Biol. 252:305–313. Boeger H, Griesenbeck J, Kornberg RD. 2008. Nucleosome retention and the stochastic nature of promoter chromatin remodeling for transcription. Cell. 133:716–726. g Fu Y, Sinha M, Peterson CL, Weng Z. 2008. The insulator binding protein CTCF positions 20 nucleosomes around its binding sites across the human genome. PLoS Genet. 4:e1000138. Burcin M, Arnold R, Lutz M, Kaiser B, Runge D, Lottspe- ich F, Filippova GN, Lobanenkov VV, Renkawitz R. 1997. Negative protein 1, which is required for function of the chicken lysozyme gene silencer in conjunction with hormone receptors, is identical to the multivalent zinc ﬁnger repressor CTCF. Mol Cell Biol. 17:1281–1288. Gabrielian A, Pongor S. 1996. Correlation of instrinsic DNA curvature with DNA property periodicity. FEBS Lett. 393: 65–68. Garcia JF, Dumesic PA, Hartley PD, El-Samad H, Mad- hani HD. 2010. Combinatorial, site-speciﬁc requirement for heterochromatic silencing factors in the elimination of nucleosome-free regions. Genes Dev. 24: 1758–1771. References Nucleosome DNA sequence pattern revealed by multiple alignment of experimentally mapped sequences. J Mol Biol. 262:129–139. Dingwall C, LomonossoﬀGP, Laskey RA. 1981. High sequence speciﬁcity of micrococcal nuclease. Nucleic Acids Res. 9:2659–2673. 37 LionnetT,DawidA,BigotS,BarreF-X,SalehOA,HeslotF,Alle- mand J-F, Bensimon D, Croquette V. 2006. DNA mechanics as a tool to probe helicase and translocase activity. Nucleic Acids Res. 34:4232–4244. Ioshikhes IP, Albert I, Zanton SJ, Pugh BF. 2006. Nucleo- some positions predicted through comparative genomics. Nat Genet. 38:1210–1215. Iyer V, Struhl K. 1995. Poly(dA:dT), a ubiquitous promoter element that stimulates transcription via its intrinsic DNA structure. EMBO J. 14:2570–2579. Lorch Y, Zhang M, Kornberg RD. 1999. Histone octamer transfer by a chromatin-remodeling complex. Cell. 96:389–392. Johnson SM, Tan FJ, McCullough HL, Riordan DP, Fire AZ. 2006. Flexibility and constraint in the nucleosome core land- scape of Caenorhabditis elegans chromatin. Genome Res. 16:1505–1516. Lowary PT, Widom J. 1997. Nucleosome packaging and nucleo- some positioning of genomic DNA. Proc Natl Acad Sci USA. 94:1183–1188. Lowary PT, Widom J. 1998. New DNA sequence rules for high aﬃnity binding to histone octamer and sequence-directed nucleosome positioning. J Mol Biol. 276:19–42. Kaplan N, Moore IK, Fondufe-Mittendorf Y, Gossett AJ, Tillo D, Field Y, LeProust EM, Hughes TR, Lieb JD, Widom J, Segal E. 2009. The DNA-encoded nucleosome organization of a eukaryotic genome. Nature. 458:362–366. Luger K, Mäder AW, Richmond RK, Sargent DF, Richmond TJ. 1997. Crystal structure of the nucleosome core particle at 2.8 Å resolution. Nature. 389:251–260. Kenigsberg E, Bar A, Segal E, Tanay A. 2010. Widespread com- pensatory evolution conserves DNA-encoded nucleosome organization in yeast. PLoS Comput Biol. 6:e1001039. Mavrich TN, Ioshikhes IP, Venters BJ, Jiang C, Tomsho LP, Qi J, Schuster SC, Albert I, Pugh BF. 2008a. A barrier nucleosome model for statistical positioning of nucleosomes throughout the yeast genome. Genome Res. 18:1073–1083. Kepper N, Foethke D, Stehr R, Wedemann G, Rippe K. 2008. Nucleosome geometry and internucleosomal interac- tions control the chromatin ﬁber conformation. Biophys J. 95:3692–3705. Mavrich TN, Jiang C, Ioshikhes IP, Li X, Venters BJ, Zanton SJ, Tomsho LP, Qi J, Glaser RL, Schuster SC, Gilmour DS, Albert I, Pugh BF. 2008b. Nucleosome organization in the Drosophila genome. Nature. 453:358–362. Klenova EM, Nicolas RH, Paterson HF, Carne AF, Heath CM, Goodwin GH, Neiman PE, Lobanenkov VV. 1993. CTCF, a conserved nuclear factor required for optimal transcriptional activity of the chicken c-myc gene, is an 11-Zn-ﬁnger pro- tein diﬀerentially expressed in multiple forms. References Mol Cell Biol. 13:7612–7624. Mergell B, Everaers R, Schiessel H. 2004. Nucleosome inter- actions in chromatin: ﬁber stiﬀening and hairpin formation. Phys Rev E: Stat Nonlin Soft Matter Phys. 70: 011915. Miele V, Vaillant C, d’Aubenton Carafa Y, Thermes C, Grange T. 2008. DNA physical properties determine nucleosome occu- pancy from yeast to ﬂy. Nucleic Acids Res. 36:3746–3756. Koerber RT, Rhee HS, Jiang C, Pugh BF. 2009. Interaction of transcriptional regulators with speciﬁc nucleosomes across the Saccharomyces genomes. Mol Cell. 35:889–902. Milani P, Chevereau G, Vaillant C, Audit B, Haftek-Terreau Z, Marilley M, Bouvet P, Argoul F, Arneodo A. 2009. Nucle- osome positioning by genomic excluding-energy barriers. Proc Natl Acad Sci USA. 106:22257–22262. Kornberg RD, Lorch Y. 1999. Twenty-ﬁve years of the nucleo- some, fundamental particle of the eukaryote chromosomes. Cell. 98:285–294. Kornberg RD, Stryer L. 1988. Statistical distributions of nucle- osomes: nonrandom locations by a stochastic mechanism. Nucleic Acids Res. 16:6677–6690. Miller JA, Widom J. 2003. Collaborative competition mechanism for gene activation in vivo. Mol Cell Biol. 23:1623–1632. Mito Y, HenikoﬀJG, HenikoﬀS. 2007. Histone replacement marks the boundaries of cis-regulatory domains. Science. 315:1408–1411. Lam FH, Steger DJ, O’Shea EK. 2008. Chromatin decouples pro- moter threshold from dynamic range. Nature. 453:246–250. Lander ES et al. 2001. Initial sequencing and analysis of the human genomes. Nature. 409:860–921. Möbius W, Gerland U. 2010. Quantitative test of the barrier nucle- osome model for statistical positioning of nucleosomes up- and downstream of transcription start sites. PLoS Comput Biol. 6:e1000891. Längst G, Bonte EJ, Corona DF, Becker PB. 1999. Nucleosome movement by CHRAC and ISWI without disruption or trans- displacement of the histone octamer. Cell. 97:843–852. Montel F, Fontaine E, St-Jean P, Castelnovo M, Faivre- Moskalenko C. 2007. Atomic force microscopy imaging of SWI/SNF action: mapping the nucleosome remodeling and sliding. Biophys J. 93:566–578. Lantermann AB, Straub T, Strålfors A, Yuan G-C, Ekwall K, Korber P. 2010. Schizosaccharomyces pombe genome-wide nucleosome mapping reveals positioning mechanisms dis- tinct from those of Saccharomyces cerevisiae. Nat Struct Mol Biol. 17:251–257. Moreira JM, Holmberg S. 1999. Transcriptional repression of the yeast CHA1 gene requires the chromatin-remodeling complex RSC. EMBO J. 18:2836–2844. Lee C-K, Shibata Y, Rao B, Strahl BD, Lieb JD. 2004. Evi- dence for nucleosome depletion at active regulatory regions genome-wide. Nat Genet. 36:900–905. Morozov AV, Fortney K, Gaykalova DA, Studitsky VM, Widom J, Siggia ED. 2009. References Revisiting polymer statistical physics to account for the presence of long-range-correlated structural disorder in 2D DNA chains. Eur Phys J E: Soft Matter. 34:119. Salsburg ZW, Kirkwood J, Zwanzig R. 1953. Molecular distri- bution functions in a one-dimensional ﬂuid. J Chem Phys. 21:1098. Noll M, Kornberg RD. 1977. Action of micrococcal nuclease on chromatin and the location of histone H1. J Mol Biol. 109:393–404. Santis PD, Morosetti S, Scipioni A. 2010. Prediction of nucle- osome positioning in genomes: limits and perspectives of physical and bioinformatic approaches. J Biomol Struct Dyn. 27:747–764. Ohlsson R, Bartkuhn M, Renkawitz R. 2010. CTCF shapes chro- matin by multiple mechanisms: the impact of 20 years of CTCF research on understanding the workings of chromatin. Chromosoma. 119:351–360. Satchwell SC, Drew HR, Travers AA. 1986. Sequence peri- odicities in chicken nucleosome core DNA. J Mol Biol. 191:659–675. Ozsolak F, Song JS, Liu XS, Fisher DE. 2007. High-throughput mapping of the chromatin structure of human promoters. Nat Biotechnol. 25:244–248. Schones DE, Cui K, Cuddapah S, Roh T-Y, Barski A, Wang Z, Wei G, Zhao K. 2008. Dynamic regulation of nucleosome positioning in the human genome. Cell. 132:887–898. Payen C, Fischer G, Marck C, Proux C, Sherman DJ, Coppée J-Y, Johnston M, Dujon B, Neuvéglise C. 2009. Unusual composition of a yeast chromosome arm is associated with its delayed replication. Genome Res. 19:1710–1721. Segal E, Fondufe-Mittendorf Y, Chen L, Thåström A, Field Y, Moore IK, Wang J-PZ, Widom J. 2006. A genomic code for nucleosome positioning. Nature. 442:772–778. Segal E, Widom J. 2009a. Poly(dA:dT) tracts: major determi- nants of nucleosome organization. Curr Opin Struct Biol. 19: 65–71. Peckham HE, Thurman RE, Fu Y, Stamatoyannopoulos JA, Noble WS, Struhl K, Weng Z. 2007. Nucleosome positioning signals in genomic DNA. Genome Res. 17:1170–1177. Segal E, Widom J. 2009b. From DNA sequence to transcrip- tional behaviour: a quantitative approach. Nat Rev Genet. 10: 443–456. Percus JK. 1976. Equilibrium state of a classical ﬂuid of hard rods in an external ﬁeld. J Stat Phys. 15:505–511. Percus JK. 1982. One-dimensional classical ﬂuid with nearest- neighbor interaction in arbitrary external ﬁeld. J Stat Phys. 28: 67. Segal E, Widom J. 2009c. What control nucleosome positions? Trends Genet. 25:335–343. Peterson CL, Workman JL. 2000. Promoter targeting and chro- matin remodeling by the SWI/SNF complex. Curr Opin Genet Dev. 10:187–192. Shivaswamy S, Bhinge A, Zhao Y, Jones S, Hirst M, Iyer VR. 2008. References Using DNA mechanics to predict in vitro nucleosome positions and formation energies. Nucleic Acids Res. 37:4707–4722. Lee JT. 2003. Molecular links between X-inactivation and auto- somal imprinting: X-inactivation as a driving force for the evolution of imprinting? Curr Biol. 13:R242–R254. Morse RH. 2007. Transcription factor access to promoter ele- ments. J Cell Biochem. 102:560–570. Lee W, Tillo D, Bray N, Morse RH, Davis RW, Hughes TR, Nislow C. 2007. A high-resolution atlas of nucleosome occupancy in yeast. Nat Genet. 39:1235–1244. Moshkin YM, Chalkley GE, Kan TW, Reddy BA, Ozgur Z, van Ijcken WFJ, Dekkers DHW, Demmers JA, Travers AA, Ver- rijzer CP. 2012. Remodelers organize cellular chromatin by counteracting intrinsic histone-DNA sequence preferences in a class-speciﬁc manner. Mol Cell Biol. 32:675–688. Lesne A, Victor J-M. 2006. Chromatin ﬁber functional organi- zation: some plausible models. Eur Phys J E Soft Matter. 19:279–290. Li B, Carey M, Workman JL. 2007. The role of chromatin during transcription. Cell. 128:707–719. Mouchiroud D, Bernardi G. 1993. Compositional properties of coding sequences and mammalian phylogeny. J Mol Evol. 37:109–116. Li W, Bernaola-Galván P, Carpena P, Oliver JL. 2003. Isochores merit the preﬁx iso. Comput Biol Chem. 27:5–10. Moukhtar J, Faivre-Moskalenko C, Milani P, Audit B, Vaillant C, Fontaine E, Mongelard F, Lavorel G, St-Jean P, Bouvet P, Argoul F, Arneodo A. 2010. Eﬀect of genomic long-range Lieb E, Mattis D. 1966. Mathematical physics in one dimension. London: Academic Press. 38 correlations on DNA persistence length: from theory to single molecule experiments. J Phys Chem B. 114:5125–5143. Rippe K, Schrader A, Riede P, Strohner R, Lehmann E, Längst G. 2007.DNAsequence-andconformation-directedpositioning of nucleosomes by chromatin-remodeling complexes. Proc Natl Acad Sci USA. 104:15635–15640. correlations on DNA persistence length: from theory to single molecule experiments. J Phys Chem B. 114:5125–5143. Moukhtar J, Fontaine E, Faivre-Moskalenko C, Arneodo A. 2007. Probing persistence in DNA curvature properties with atomic force microscopy. Phys Rev Lett. 98:178101. Robinson P, Fairall L, Huynh V, Rhodes D. 2006. EM measure- ments deﬁne the dimensions of the “30-nm” chromatin ﬁbre: evidence for a compact, interdigitated structures. Proc Natl Acad Sci USA. 103:6506–6511. Moukhtar J, Vaillant C, Audit B, Arneodo A. 2009. Gen- eralized wormlike chain model for long-range correlated heteropolymers. Europhys Lett. 86:48001. Robledo A, Rowlinson J. 1986. The distribution of hard rods on a line of ﬁnite length. Mol Phys. 58:711–721. Moukhtar J, Vaillant C, Audit B, Arneodo A. 2011. References Dynamic remodeling of individual nucleosomes across a eukaryotic genome in response to transcriptional perturbation. PLoS Biol. 6:e65. Piasecki J, Peliti L. 1993. Harmonic properties of hard-spheres crystals: a one-demensional study. J Phys A: Math Gen. 26:4819–4825. p Shlyakhtenko LS, Lushnikov AY, Lyubchenko YL. 2009. Dynamics of nucleosomes revealed by time-lapse atomic force microscopy. Biochemistry. 48:7842–7848. Polach KJ, Widom J. 1996. A model for the cooperative binding of eukaryotic regulatory proteins to nucleosomal target sites. J Mol Biol. 258:800–812. Solis FJ, Bash R, Yodh J, Lindsay SM, Lohr D. 2004. A sta- tistical thermodynamic model applied to experimental AFM population and location data is able to quantify DNA-histone binding strength and internucleosomal interaction diﬀerences between acetylated and unacetylated nucleosomal arrays. Biophys J. 87:3372–3387. Pusarla R-H, Vinayachandran V, Bhargava P. 2007. Nucleo- some positioning in relation to nucleosome spacing and DNA sequence-speciﬁc binding of a protein. FEBS J. 274:2396– 2410. y Strick TR, Dessinges MN, Charvin G, Dekker NH, Allemand JF, Bensimon D, Croquette V. 2003. Stretching of macro- molecules and proteins. Rep Prog Phys. 66:1–45. Radman-Livaja M, Rando OJ. 2010. Nucleosome positioning: how is it established, and why does it matter? Dev Biol. 339:258–266. Rando OJ, Ahmad K. 2007. Rules and regulation in the pri- mary structure of chromatin. Curr Opin Cell Biol. 19: 250–256. Suter B, Schnappauf G, Thoma F. 2000. Poly(dA.dT) sequences exist as rigid DNA structures in nucleosome-free yeast promoters in vivo. Nucleic Acids Res. 28:4083–4089. Raveh-Sadka T, Levo M, Segal E. 2009. Incorporating nucleo- somes into thermodynamic models of transcription regula- tion. Genome Res. 19:1480–1496. Takahashi H. 1942. A simple method for treating the statistical mechanics of one-dimensional substances. Proc Phys Math Soc Jpn. 24:60. Rayleigh Lord (1891). On the virial of a system of hard colliding bodies. Nature London. 45:80–82. Teif VB, Rippe K. 2009. Predicting nucleosome positions on the DNA: combining intrinsic sequence preferences and remodeler activities. Nucleic Acids Res. 37:5641– 5655. Richard G-F, Kerrest A, Dujon B. 2008. Comparative genomics and molecular dynamics of DNA repeats in eukaryotes. Microbiol Mol Biol Rev. 72:686–727. Thåström A, Bingham LM, Widom J. 2004. Nucleosomal locations of dominant DNA sequence motifs for histone- DNA interactions and nucleosome positioning. J Mol Biol. 338:695–709. Richmond TJ, Davey CA. 2003. The structure of DNA in the nucleosome core. Nature. 423:145–150. Riposo J, Mozziconacci J. 2012. Nucleosome positioning and nucleosome stacking: two faces of the same coin. Mol Biosyst. 8:1172–1178. References Thåström A, Lowary PT, Widlund HR, Cao H, Kubista M, Widom J. 1999. Sequence motifs and free energies of selected natural 39 Vanderlick TK, Scriven LE, Davis HT. 1986. Solution of Percus equation for the density of hard-rods in an external-ﬁeld. Phys Rev A. 34:5130–5131. and non-natural nucleosome positioning DNA sequences. J Mol Biol. 288:213–229. Tillo D, Hughes TR. 2009. G + C content dominates intrinsic nucleosome occupancy. BMC Bioinforma. 10:442. Vashee S, Melcher K, Ding WV, Johnston SA, Kodadek T. 1998. Evidence for two modes of cooperative DNA binding in vivo that do not involve direct protein–protein interactions. Curr Biol. 8:452–458. Tirosh I, Barkai N. 2008. Two strategies for gene regulation by promoter nucleosomes. Genome Res. 18:1084–1091. Tolkunov D, Morozov AV. 2010. Genomic studies and compu- tational predictions of nucleosome positions and formation energies. Adv Protein Chem Struct Biol. 79:1–57. Wang MD, Schnitzer MJ, Yin H, Landick R, Gelles J, Block SM. 1998. Force and velocity measured for single molecules of RNA polymerase. Science. 282:902–907. Tolstorukov MY, Colasanti AV, McCandlish DM, Olson WK, Zhurkin VB. 2007. A novel roll-and-slide mechanism of DNA folding in chromatin: implications for nucleosome positioning. J Mol Biol. 371:725–738. p y Wang X, Bai L, Bryant GO, Ptashne M. 2011a. Nucleosomes and the accessibility problem. Trends Genet. 27:487–492. Wang X, Bryant GO, Floer M, Spagna D, Ptashne M. 2011b. An eﬀect of DNA sequence on nucleosome occupancy and removal. Nat Struct Mol Biol. 18:507–509. p g Tonks L. 1936. The complete equation of state of one, two and three-dimensional gases of hard elastic spheres. Phys Rev. 50:955–963. Weiner A, Hughes A, Yassour M, Rando OJ, Friedman N. 2010. High-resolution nucleosome mapping reveals transcription- dependent promoter packaging. Genome Res. 20:90–100. Travers A. 1999. An engine for nucleosome remodeling. Cell. 96:311–314. Tsankov A, Yanagisawa Y, Rhind N, Regev A, Rando OJ. 2011. Evolutionary divergence of intrinsic and trans-regulated nucleosome positioning sequences reveals plastic rules for chromatin organization. Genome Res. 21:1851–1862. Whitehouse I, Flaus A, Cairns BR, White MF, Workman JL, Owen-Hughes T. 1999. Nucleosome mobilization catalysed by the yeast SWI/SNF complex. Nature. 400:784–787. Whitehouse I, Rando OJ, Delrow J, Tsukiyama T. 2007. Chro- matin remodelling at promoters suppresses antisense tran- scription. Nature. 450:1031–1035. Tsankov AM, Thompson DA, Socha A, Regev A, Rando OJ. 2010. The role of nucleosome positioning in the evolution of gene regulation. PLoS Biol. 8:e1000414. Whitehouse I, Tsukiyama T. 2006. References Antagonistic forces that posi- tion nucleosomes in vivo. Nat Struct Mol Biol. 13:633–640. Tsukiyama T, Wu C. 1997. Chromatin remodeling and transcrip- tion. Curr Opin Genet Dev. 7:182–191. Widom J. 1996. Short-range order in two eukaryotic genomes: relation to chromosome structure. J Mol Biol. 259:579–588. Vaillant C, Audit B, Arneodo A. 2005. Thermodynamics of DNA loops with long-range correlated structural disorder. Phys Rev Lett. 95:068101. Widom J. 2001. Role of DNA sequence in nucleosome stability and dynamics. Q Rev Biophys. 34:269–324. Vaillant C, Audit B, Arneodo A. 2007. Experiments conﬁrm the inﬂuence of genome long-range correlations on nucleosome positioning. Phys Rev Lett. 99:218103. y p y Wolﬀe AP. 1998. Chromatin Structure and Function, 3rd ed. London: Academic Press. Woodcock CL, Skoultchi AI, Fan Y. 2006. Role of linker histone in chromatin structure and function: H1 stoichiometry and nucleosome repeat length. Chromosome Res. 14:17–25. Vaillant C, Audit B, Thermes C, Arnéodo A. 2006. Formation and positioning of nucleosomes: eﬀect of sequence-dependent long-range correlated structural disorder. Eur Phys J E: Soft Matter 19:263–277. Yuan G-C, Liu JS. 2008. Genomic sequence is highly predictive of local nucleosome depletion. PLoS Comput Biol. 4:e13. Yuan G-C, Liu Y-J, Dion MF, Slack MD, Wu LF, Altschuler SJ, Rando OJ. 2005. Genome-scale identiﬁcation of nucleosome positions in S. cerevisiae. Science. 309:626–630. Vaillant C, Palmeira L, Chevereau G, Audit B, d’Aubenton Carafa Y, Thermes C, Arneodo A. 2010. A novel strategy of transcription regulation by intragenic nucleosome ordering. Genome Res. 20:59–67. Zhang Y, Moqtaderi Z, Rattner BP, Euskirchen G, Snyder M, Kadonaga JT, Liu XS, Struhl K. 2009. Intrinsic histone-DNA interactions are not the major determinant of nucleosome positions in vivo. Nat Struct Mol Biol. 16:847–852. Valouev A, Ichikawa J, Tonthat T, Stuart J, Ranade S, Peck- ham H, Zeng K, Malek JA, Costa G, McKernan K, Sidow A, Fire A, Johnson SM. 2008. A high-resolution, nucleo- some position map of C. elegans reveals a lack of universal sequence-dictated positioning. Genome Res. 18:1051–1063. Zhang Z, Pugh BF. 2011. High-resolution genome-wide mapping of the primary structure of chromatin. Cell. 144:175–186. Valouev A, Johnson SM, Boyd SD, Smith CL, Fire AZ, Sidow A. 2011. Determinants of nucleosome organization in primary human cells. Nature. 474:516–520. Zhang Z, Wippo CJ, Wal M, Ward E, Korber P, Pugh BF. 2011. A packing mechanism for nucleosome organization reconstituted across a eukaryotic genome. Science. 332: 977–980. van Holde KE. 1988. Chromatin. New York: Springer-Verlag. References 40 40
https://openalex.org/W4242154068	https://research-information.bris.ac.uk/ws/files/156914583/document.pdf	English	null	What is the impact of regulatory guidance and expiry of drug patents on dementia drug prescriptions in England: a trend analysis in the Clinical Practice Research Datalink	null	2,017	cc-by	8,957	Walker, V., Davies, N., Kehoe, P., & Martin, R. (2018). What is the impact of regulatory guidance and expiry of drug patents on dementia drug prescriptions in England? A trend analysis in the Clinical Practice Research Datalink. Alzheimer's Research and Therapy, 10(1), Article 51. https://doi.org/10.1186/s13195-018-0379-6, https://doi.org/10.1186/s13195-018-0379-6 Publisher's PDF, also known as Version of record License (if available): CC BY Link to published version (if available): 10.1186/s13195-018-0379-6 10.1186/s13195-018-0379-6 This is the final published version of the article (version of record). It first appeared online via BioMed Central at https://alzres.biomedcentral.com/articles/10.1186/s13195-018-0379-6 . Please refer to any applicable terms of use of the publisher. Walker, V., Davies, N., Kehoe, P., & Martin, R. (2018). What is the impact of regulatory guidance and expiry of drug patents on dementia drug prescriptions in England? A trend analysis in the Clinical Practice Research Datalink. Alzheimer's Research and Therapy, 10(1), Article 51. https://doi.org/10.1186/s13195-018-0379-6, https://doi.org/10.1186/s13195-018-0379-6 Walker, V., Davies, N., Kehoe, P., & Martin, R. (2018). What is the impact of regulatory guidance and expiry of drug patents on dementia drug prescriptions in England? A trend analysis in the Clinical Practice Research Datalink. Alzheimer's Research and Therapy, 10(1), Article 51. https://doi.org/10.1186/s13195-018-0379-6, https://doi.org/10.1186/s13195-018-0379-6 RESEARCH Open Access University of Bristol – Bristol Research Portal General rights University of Bristol – Bristol Research Portal General rights This document is made available in accordance with publisher policies. Please cite only the published version using the reference above. Full terms of use are available: http://www.bristol.ac.uk/red/research-policy/pure/user-guides/brp-terms/ Walker et al. Alzheimer's Research & Therapy (2018) 10:51 https://doi.org/10.1186/s13195-018-0379-6 What is the impact of regulatory guidance and expiry of drug patents on dementia drug prescriptions in England? A trend analysis in the Clinical Practice Research Datalink Venexia M. Walker1,2, Neil M. Davies1,2, Patrick G. Kehoe3,4 and Richard M. Martin1,2 Abstract Background: Drugs for dementia have been available in England since 1997. Since their launch, there have been several changes to national guidelines and initiatives that may have influenced prescribing. These include changes in National Institute for Health and Care Excellence (NICE) guidance, several government dementia strategies, the addition of dementia to the Quality and Outcomes Framework (QOF), and the expiry of drug patents. Despite this, there has been little research into the effect of these events on prescribing. This paper examines prescribing trends in England using data from the U.K. Clinical Practice Research Datalink since the launch of drugs for dementia up to 1st January 2016. Methods: We considered the monthly proportion of patients eligible for treatment, with a diagnosis of probable Alzheimer’s disease, receiving their first prescription for each drug class—namely, acetylcholinesterase (AChE) inhibitors (donepezil, rivastigmine, galantamine) and N-methyl-D-aspartate (NMDA) receptor antagonists (memantine). Trend analysis using joinpoint models was then applied to identify up to two trend changes per treatment of interest. Results: The overall trend was for increasing prescriptions in each drug class over the period in which they were studied. This was indicated by the average monthly percentage change, which was 6.0% (95% CI, −6.4 to 19.9; June 1997 to December 2015) for AChE inhibitors and 15.4% (95% CI, −77.1 to 480.9; January 2003 to December 2015) for NMDA receptor antagonists. Prescriptions of AChE inhibitors increased at the end of 2012, probably in response to the patent expiry of these drugs earlier that year. The Prime Minister’s Dementia Challenge launched in May 2012 may also have contributed to the observed increase. However, neither this strategy nor patent expiry appeared to influence prescriptions of NMDA receptor antagonists. Instead trend changes in this drug class were driven by NICE guidance released in 2011 that allowed access to these drugs outside of clinical trials. Conclusions: Dementia drug prescribing does not always respond to factors such as regulatory guidance, recommendations, or patent expiry, and when it does, not necessarily in a predictable way. This suggests that communication with clinicians may need to be improved to use drugs for dementia more cost-effectively. Keywords: Alzheimer disease, Dementia, Donepezil, Rivastigmine, Galantamine, Memantine, Clinical Practice Research Datalink, National Institute for Health and Care Excellence, Quality and Outcomes Framework, England Correspondence: venexia.walker@bristol.ac.uk , 2MRC Integrative Epidemiology Unit, University of Bristol, Bristol, UK Full list of author information is available at the end of the article Background Table 1 Events prior to 1st January 2016 that potentially affected prescription rates Event date Event May 1997 Donepezil first recorded in CPRD September 1998 Rivastigmine first recorded in CPRD January 2001 Galantamine first recorded in CPRD and first NICE guidance released December 2002 Memantine first recorded in CPRD November 2006 NICE recommended restricting drug access September 2007 QOF revised to include dementia February 2009 First National Dementia Strategy launched March 2011 NICE removed recommendation restricting drug access January 2012 Galantamine patent expired February 2012 Donepezil patent expired May 2012 Prime Minister’s Dementia Challenge launched July 2012 Rivastigmine patent expired April 2014 Memantine patent expired February 2015 Prime Minister’s Challenge on Dementia 2020 launched CPRD Clinical Practice Research Datalink, NICE National Institute for Health and Care Excellence, QOF Quality and Outcomes Framework g There are currently four licensed treatments that pro- vide symptomatic relief for patients with Alzheimer’s disease in England—three acetylcholinesterase (AChE) inhibitors (donepezil, rivastigmine, galantamine) and one N-methyl-D-aspartate (NMDA) receptor antagonist (memantine). These drugs are collectively referred to as drugs for dementia in the British National Formulary, despite their licensing for Alzheimer’s disease only [1]. Since the first of these drugs became available in 1997, there have been several changes in national guidelines for the treatment of Alzheimer’s disease, as well as several initiatives to encourage better diagnosis and treatment of the disease. Despite this, there has been lit- tle research into whether such changes to guidelines and initiatives have directly influenced clinical practice [2, 3]. We examined how prescription rates in England have changed since the launch of these drugs up to 1st January 2016, using data from the U.K. Clinical Practice Research Datalink (CPRD). We investigated how prescribing was affected by changes in National Institute for Health and Care Excellence (NICE) guidance (includ- ing the 2006 guidance that was subject to legal challenges), the addition of dementia to the Quality and Outcomes Framework (QOF), the introduction of ambi- tious government dementia strategies, and the expiry of drug patents. The timing of each of these changes, which may have influenced aspects of drug prescribing and clinical practice, is discussed further below and summarized in Table 1. © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Walker et al. Alzheimer's Research & Therapy (2018) 10:51 Page 2 of 11 Background CPRD Clinical Practice Research Datalink, NICE National Institute for Health and Care Excellence, QOF Quality and Outcomes Framework restrict access to these drugs, and was ultimately the subject of a high court challenge by the Alzheimer’s Society and two drug manufacturers, Eisai and Pfizer [8–10]. This led to a further revision being made to the NICE guidance at the end of March 2011, which recom- mended AChE inhibitors for patients with mild to moderate Alzheimer’s disease and memantine for patients with moderate to severe Alzheimer’s disease or who could not tolerate AChE inhibitors [11]. For the duration of our present study, treatment had to be initiated by a specialist and deemed effective as long as there has been ‘an improvement or no deterior- ation in MMSE score, together with evidence of global improvement on the basis of behavioral and/or functional assessment’ [6]. NICE guidance on the prescribing of drugs for dementia In the past NICE guidance has used scores from the Mini Mental State Examination (MMSE), in combin- ation with other measures, to guide whether a patient should be prescribed a drug for dementia. The test, proposed in 1975 by Folstein et al., assesses a patient’s cognition out of a total possible score of 30, where nor- mal cognition is considered as a score of 24 or more [4]. The original NICE guidance, issued in 2001, on the use of drugs to treat Alzheimer’s disease recommended that the three AChE inhibitors should be used for all patients scoring 12 or above on the MMSE until the drugs were deemed no longer effective [5, 6]. In November 2006, NICE revised their guidance so that the use of AChE inhibitors was restricted to patients with moderate Alzheimer’s disease; this was defined as patients scoring between 10 and 20 points on the MMSE. The 2006 guidance was also the first to consider the use of the NMDA receptor antagonist memantine, which was rec- ommended for use only in clinical trials for patients with moderate to severe disease [7]. This revision of the guidance was controversial because of the way in which it assessed cost-effectiveness, which was expected to Inclusion of dementia on the QOF QOF is a voluntary incentive program, introduced in 2004, to improve services in primary care [12]. Dementia first appeared in QOF as an ‘indicator’ in September 2007 [13]. There are currently three indicators for de- mentia included in the framework. The first requires that the practice establish and maintain a register of patients diagnosed with dementia, and the other two in- dicators refer to the ongoing management of the disease [14]. The inclusion of dementia on the QOF could there- fore have encouraged a greater focus on the diagnosis Page 3 of 11 Page 3 of 11 Page 3 of 11 Walker et al. Alzheimer's Research & Therapy (2018) 10:51 and pharmacological management of the disease in par- ticipating practices. and pharmacological management of the disease in par- ticipating practices. for the treatment of interest in each month. We defined patients as eligible for their first prescription if they had the diagnosis of interest with no previous prescription for the treatment of interest. The time period was mea- sured in units of 1 month because this was the smallest clinically meaningful measure we could realistically de- fine. We investigated treatment rates as a proportion of eligible patients because the underlying rate of diagnosis of Alzheimer’s disease, as well as non-Alzheimer’s dis- ease and mixed dementias, has changed over time in the CPRD (Fig. 1). The joinpoint analysis used in this study has been developed for incidence rates, so prevalent drug use, which requires consideration of both incidence and continued drug use, was not studied. Government dementia strategies The first National Dementia Strategy was launched by the Department of Health in February 2009. The aim of that strategy was ‘to ensure that significant improve- ments are made to dementia services across three key areas: improved awareness, earlier diagnosis and inter- vention, and a higher quality of care’ [15]. This strategy was followed in 2012 by the Prime Minister’s Dementia Challenge, which looked to improve care and research by 2015, and more recently by the Prime Minister’s Challenge on Dementia 2020 [16, 17]. The most recent strategy aims to build on the work of its predecessors to make England the best place for both dementia care and research. In general such strategies may help to increase the awareness of dementia for both the public and health services [18, 19]. The four drugs for dementia were separated according to drug class (i.e., AChE inhibitors and NMDA receptor antagonists). Exposure date was taken to be the date on which the first prescription requesting the drug(s) being considered was recorded. This allowed patients who had previously been prescribed AChE inhibitors to be in- cluded in the NMDA receptor antagonist analysis. This is necessary because NMDA receptor antagonists may be prescribed alongside AChE inhibitors and are often given to patients later in the course of their disease, po- tentially following exposure to AChE inhibitors. Methods The aim of the study was to examine prescribing trends in England from the launch of the drugs for dementia up to 1st January 2016, using data from the CPRD. Drug patents h ’ The King’s Fund charity found that the prescription of generic drugs over their patented alternatives has ‘saved the NHS around £7.1 billion and allowed more than 490 million more items to be prescribed to patients’ between 1976 and 2013 [20]. AChE inhibitors for the treatment of Alzheimer’s disease became available generically from 2012, whereas NMDA receptor antagonists became available generically from 2014 (Table 2) [21]. Therefore, in recent years the cost of drugs for dementia has de- creased significantly from previous years. This serves as a potential factor in rates of prescribing, particularly in publicly funded health care services such as the NHS in England. Setting h In this study we used data from the CPRD, an ongoing U.K.-based primary care database established in 1987. The data used in this study were obtained as part of a larger project investigating whether commonly pre- scribed drugs can be repurposed for the prevention or treatment of Alzheimer’s and other neurodegenerative diseases [22]. For this project, we sampled patients older than 40 years of age with at least 12 consecutive months of records classified as ‘acceptable’ by the CPRD from an ‘up to standard’ practice. The data were taken from the March 2016 CPRD GOLD database snapshot, which covered the period from 1st January 1987 to 29th Febru- ary 2016, inclusive. Sample Read and product codes uniquely identify clinical terms and prescriptions, respectively, in the CPRD and are recorded by the general practitioner at the time of the consultation with the date [24]. The val- idity of codes for dementia diagnoses in the CPRD has previously been studied and was found to be in concord- ance with depersonalized written records relating to the diagnosis [25]. The diagnoses and their definitions as used in the present study are provided in Table 3. We used treatment to define diagnosis under the assumption that treatment implies diagnosis. Diagnosis date was taken to be the first date on which a code from any of ensured that all data were complete for the time frame being considered. We also restricted the data to English practices. This is because guidelines and initiatives can differ by nation in the United Kingdom; for example, all nations are subject to patent expiry, but the National Dementia Strategy is applicable only to England, with other nations having their own strategies. Additional file 1 presents a sensitivity analysis investigating the effect of limiting the study to practices in England. The analysis concludes that because the majority of the CPRD data is obtained from English practices and the proportion of people included in the study is similar for England and the CPRD as whole, the representativeness of the CPRD is likely to have been preserved. To be included in the study, a patient had to have a diagnosis of dementia as determined by a read or product code (see reference [23] for code lists). Read and product codes uniquely identify clinical terms and prescriptions, respectively, in the CPRD and are recorded by the general practitioner at the time of the consultation with the date [24]. The val- idity of codes for dementia diagnoses in the CPRD has previously been studied and was found to be in concord- ance with depersonalized written records relating to the diagnosis [25]. The diagnoses and their definitions as used in the present study are provided in Table 3. We used treatment to define diagnosis under the assumption that treatment implies diagnosis. Diagnosis date was taken to be the first date on which a code from any of the lists was recorded. We performed a sensitivity ana- lysis to test the diagnosis definitions, which is presented in full in Additional file 2. Sample The analysis considered the sensitivity and specificity of the diagnoses in the CPRD dataset using linked data from the Office of National Statistics (ONS) death registry and the Hospital Episode Statistics (HES) inpatient dataset. We found there to be high specificity (HES, 62.9–79.1%; ONS, 57.5–75.1%) and variable sensitivity (HES, 37.3–71.6%; ONS, 36.0– 80.4%). The high specificity demonstrated in this analysis reflects our conservative approach when constructing the code lists. Consequently, we expected a lower sensi- tivity, and this is in line with what we observed. Sample For this study, we considered the data available from 1st January 1987 to 31st December 2015, inclusive, from practices with a last data collection date in 2016; this This study was a joinpoint analysis of the proportion of patients eligible for treatment, with a diagnosis of ‘prob- able Alzheimer’s disease’, receiving their first prescription Table 2 Patent information for the drugs used for dementia [21] Generic name Patent name (manufacturer) Drug class Patent expiry Donepezil Aricept (Eisai / Pfizer) AChE inhibitor January 2012 Rivastigmine Exelon (Novartis) AChE inhibitor February 2012 Galantamine Reminyl (Shire) AChE inhibitor July 2012 Memantine Ebixa (Lundbeck) NMDA receptor antagonist April 2014 AChE Acetylcholinesterase, NMDA N-methyl-D-aspartate Walker et al. Alzheimer's Research & Therapy (2018) 10:51 Page 4 of 11 Fig. 1 Bar graph illustrating the number of patients diagnosed with dementia, by diagnosis type. The data presented are restricted to patients who received a diagnosis prior to 1st January 2016 and are from an English practice with a last data collection date in 2016 to reflect the main analysis. Definitions for each of the diagnoses are presented in Table 3. Fig. 1 Bar graph illustrating the number of patients diagnosed with dementia, by diagnosis type. The data presented are restricted to patients who received a diagnosis prior to 1st January 2016 and are from an English practice with a last data collection date in 2016 to reflect the main analysis. Definitions for each of the diagnoses are presented in Table 3. ensured that all data were complete for the time frame being considered. We also restricted the data to English practices. This is because guidelines and initiatives can differ by nation in the United Kingdom; for example, all nations are subject to patent expiry, but the National Dementia Strategy is applicable only to England, with other nations having their own strategies. Additional file 1 presents a sensitivity analysis investigating the effect of limiting the study to practices in England. The analysis concludes that because the majority of the CPRD data is obtained from English practices and the proportion of people included in the study is similar for England and the CPRD as whole, the representativeness of the CPRD is likely to have been preserved. To be included in the study, a patient had to have a diagnosis of dementia as determined by a read or product code (see reference [23] for code lists). Analysis h l The analysis of each treatment of interest started on the first day of the month following the first recorded pre- scription for that treatment. For example, the first pre- scription for NMDA receptor antagonists occurred on 16th December 2002, so the analysis of this drug class started on 1st January 2003. For each patient, we used the month and year of diagnosis (Table 3) and first pre- scription. For each month, we calculated the following: (A) the number of patients receiving their first prescrip- tion in that month and (B) the number of patients with a diagnosis who had not received treatment before the first of the month. Dividing A by B provided the propor- tion of patients with diagnoses who received their first Walker et al. Alzheimer's Research & Therapy (2018) 10:51 Page 5 of 11 Table 3 Diagnosis definitions used in the study, presented with the number of patients Diagnosis Definition Patients Probable AD Patients with one or more codes on the list ‘probable AD’. Patients may also have codes on the lists ‘possible AD’, ‘donepezil’, ‘rivastigmine’, ‘galantamine’ and ‘memantine’. 10,651 Possible AD Patients with one or more codes on the list ‘possible AD’. Patients may also have codes on the lists ‘donepezil’, ‘rivastigmine’, ‘galantamine’ and ‘memantine’. 12,167 Non-AD and mixed dementias Patients with one or more codes on any of the following lists: ‘probable AD’, ‘possible AD’, ‘other dementia’, ‘vascular dementia’, ‘non-specific dementia’, ‘donepezil’, ‘rivastigmine’, ‘galantamine’ and ‘memantine’, who do not meet the above criteria. 17,384 AD Alzheimer’s disease The data presented are restricted to patients who received a diagnosis prior to 1st January 2016 and are from an English practice with a last data collection date in 2016 to reflect the main analysis. The total number of patients with ‘any dementia’ is 40,202 Table 3 Diagnosis definitions used in the study, presented with the number of patients f Trend analysis for AChE inhibitors The proportion of patients with probable Alzheimer’s disease receiving their first prescription for an AChE in- hibitor increased throughout the study period (Fig. 2). This is reflected in the average monthly percent change, which was 6.0 (95% CI, −6.4 to 19.9) for the period from June 1997 to December 2015. For much of the study period, the trend was for an increasing proportion of pa- tients to receive their first prescription for an AChE in- hibitor with a monthly percent change of 5.4 (95% CI, 4.2 to 6.7). In October 2012 (95% CI; September 2011 to April 2013; p = 0.816), the prescription rate surged with a monthly percent change of 67.2 (95% CI, −96.6 to 8179.8). Less than 1 year later, in May 2013 (95% CI; November 2012 to April 2014; p = 0.789), the trend re- versed so that prescription rates were falling. In the months that followed, the monthly percent change had a value of −1.6 (95% CI, −10.4 to 8.1), falling below zero for the first time since the launch of these drugs. Trend analysis for NMDA receptor antagonists Figure 3 presents the equivalent analysis for the NMDA receptor antagonist memantine. Memantine became available in January 2003 and was prescribed much less often than the other drugs, despite similar numbers of eligible patients. This is partly related to the indication of these drugs. Memantine is generally recommended for more advanced disease than the AChE inhibitors and is often added to a prescription of AChE inhibitors following progression of the disease. Despite this, as ob- served for the AChE inhibitors, the proportion of pa- tients with probable Alzheimer’s disease receiving their first prescription for an NMDA receptor antagonist increased on average throughout the study period. The average monthly percent change for the period from January 2003 to December 2015 was 15.4 (95% CI, − 77.1 to 480.9), though the 95% CI around this estimate is large. The initial trend for prescribing of this drug showed a reduced number of prescriptions in the time that followed the launch with a monthly percent change of −5.3 (95% CI, −12.6 to 2.6). This changed around Results prescription for the treatment of interest each month. We also calculated the SE of this proportion [26]. Trend analysis using joinpoint models was then conducted. The optimal number of joinpoints, as determined by the software and up to a maximum number of 2, was used to select the model. We refer to the period between two joinpoints as a ‘segment’ and number them chronologic- ally. Our model assumes that the rate of prescription ‘changes at a constant percentage of the rate of the pre- vious year’ [27] and so is determined by the following equation: lny = xb. This allows us to consider the monthly percent change. The trend over the entire study period is summarized using the average monthly percent change. This is calculated as the average of the monthly percent changes, weighted by segment length [28]. All analysis was conducted using Joinpoint Regression Program (version 4.3.1.0; National Cancer Institute, Bethesda, MD, USA) and Stata (version 14.1; Stata- Corp, College Station, TX, USA) software [29, 30]. The model specifications for the joinpoint analyses are the software’s default with dependent variable type set to ‘proportion’ and the maximum number of join points set to 2. The Stata code used in this analysis is available from GitHub (https://github.com/venexia/ DementiaDrugsCPRD) [31]. News search S l f Several of the national guidelines and initiatives considered in this study may have increased aware- ness of dementia, including Alzheimer’s disease. To investigate this, we downloaded the Google Trends (https://trends.google.com/trends/) data for news searches in England for the disease term ‘Alzheimer’s disease’ from 1st January 2008 up to 1st January 2016 [32]. Unfortunately, data were not recorded prior to 2008, so we cannot comment on the effect media coverage may have had on trend changes iden- tified before this point in time. As with the main analysis, data were processed and plotted using Stata (version 14.1; StataCorp, College Station, TX, USA), and the code is available from GitHub (https:// github.com/venexia/DementiaDrugsCPRD) [29, 30]. Walker et al. Alzheimer's Research & Therapy (2018) 10:51 Page 6 of 11 Fig. 2 Indicative graph of acetylcholinesterase (AChE) inhibitor prescriptions in patients with probable Alzheimer’s disease. This graph shows the proportion of patients with probable Alzheimer’s disease receiving their first prescription for an AChE inhibitor each month from June 1997 to December 2015. The fixed lines indicate events with the potential to affect prescription rates during the study period. The joinpoints, monthly percent change (MPC) for each segment, and the average monthly percent change (AMPC) for the entire study period are also presented Fig. 3 Indicative graph of N-methyl-D-aspartate (NMDA) receptor antagonist prescriptions in patients with probable Alzheimer’s disease. This graph shows the proportion of patients with probable Alzheimer’s disease receiving their first prescription for an NMDA receptor antagonist each month from January 2003 to December 2015. The fixed lines indicate events with the potential to affect prescription rates during the study period. The joinpoints, the monthly percent change (MPC) for each segment, and the average monthly percent change (AMPC) for the study period are also presented Fig. 2 Indicative graph of acetylcholinesterase (AChE) inhibitor prescriptions in patients with probable Alzheimer’s disease. This graph shows the proportion of patients with probable Alzheimer’s disease receiving their first prescription for an AChE inhibitor each month from June 1997 to December 2015. The fixed lines indicate events with the potential to affect prescription rates during the study period. The joinpoints, monthly percent change (MPC) for each segment, and the average monthly percent change (AMPC) for the entire study period are also presented Fig. 2 Indicative graph of acetylcholinesterase (AChE) inhibitor prescriptions in patients with probable Alzheimer’s disease. News search S l f This graph shows the proportion of patients with probable Alzheimer’s disease receiving their first prescription for an AChE inhibitor each month from June 1997 to December 2015. The fixed lines indicate events with the potential to affect prescription rates during the study period. The joinpoints, monthly percent change (MPC) for each segment, and the average monthly percent change (AMPC) for the entire study period are also presented Fig. 3 Indicative graph of N-methyl-D-aspartate (NMDA) receptor antagonist prescriptions in patients with probable Alzheimer’s disease. This graph shows the proportion of patients with probable Alzheimer’s disease receiving their first prescription for an NMDA receptor antagonist each month from January 2003 to December 2015. The fixed lines indicate events with the potential to affect prescription rates during the study period. The joinpoints, the monthly percent change (MPC) for each segment, and the average monthly percent change (AMPC) for the study period are also presented Fig. 3 Indicative graph of N-methyl-D-aspartate (NMDA) receptor antagonist prescriptions in patients with probable Alzheimer’s disease. This graph shows the proportion of patients with probable Alzheimer’s disease receiving their first prescription for an NMDA receptor antagonist each month from January 2003 to December 2015. The fixed lines indicate events with the potential to affect prescription rates during the study period. The joinpoints, the monthly percent change (MPC) for each segment, and the average monthly percent change (AMPC) for the study period are also presented Page 7 of 11 Page 7 of 11 Walker et al. Alzheimer's Research & Therapy (2018) 10:51 March 2011 (95% CI, August 2010 to April 2011, p = 0.892) to a very strong trend for increased prescribing. From the second trend change in June 2011 (95% CI, April 2011 to November 2011, p = 0.896) until the end of the study in December 2015, this trend reduced to a monthly percent change of 20.7 (95% CI, 15.3 to 26.4). This indicates a continuing increase in the prescriptions for NMDA receptor antagonists in recent years, albeit substantially reduced from the rise observed between March and June 2011. The complete output for both this analysis and that relating to AChE inhibitors is provided in Additional file 3. interest for the search term, with values indicating both low and high interest occurring throughout the period studied. Sensitivity analyses y y We repeated the main analysis, which considers the diagnosis ‘probable Alzheimer’s disease’, with relaxed diagnosis definitions to test the sensitivity of our results. We did this in two ways: (1) introducing codes that rep- resented what may be lesser degrees of confidence in the accuracy of Alzheimer’s disease diagnosis (termed any Alzheimer’s disease) and (2) introducing codes capturing other types of dementia (termed any dementia). The re- sults of these further analyses are provided in Additional file 3, and a summary of all results can be found in Table 4. For NMDA receptor antagonists, the joinpoint analysis is consistent regardless of the diagnosis defin- ition used. However, for AChE inhibitors, the joinpoint analysis varies according to the diagnosis definition used, though the two sensitivity analyses are reasonably con- sistent with each other. News search S l f Months with insufficient data, indicating little interest in the search term, became less common over the period studied, with the most recent occurring in August 2015. Interest peaked in September 2012 and was also high in January 2011 (88%), January 2010 (82%) and April 2008 (81%). Discussion The first trend change for the proportion of patients with probable Alzheimer’s disease receiving their first prescription for an AChE inhibitor occurred in October 2012 (95% CI, September 2011 to April 2013, p = 0.816). At this time, a long-term steadily increasing trend be- came a very strong increasing trend. This surge could be related to two factors. Firstly, the patents expired on the three drugs in this class in 2012—galantamine in January 2012, donepezil in February 2012 and rivastigmine in July 2012. Secondly, the Prime Minister’s Dementia Challenge launched in May 2012. It is likely that the re- duction in cost of these drugs, which resulted from their patents expiring, in combination with increased aware- ness of dementia due to the Prime Minister’s Dementia Challenge, led to this substantial change in prescription rates we observed. In addition to these factors, a large amount of literature concerning AChE inhibitors had been published ahead of the revisions to the NICE guid- ance in 2011. Although this is unlikely to have caused the sharp surge that we observed, it could have contrib- uted to the long-term steadily increasing trend observed prior to this change. A systematic review which covers the literature through November 2014 (i.e., after all join points identified in our analysis but 13 months before the end of our study) summarizes the literature available at that time [33]. It shows that several studies published News search Values of zero indicate insufficient data for that month between 2003 and 2008 suggested that patients with mild to moderate Alzheimer’s disease could benefit from AChE inhibitors with estimated ‘improvements on the order of 1.5 MMSE (30-point scale)’. We therefore can- not rule out a potential effect of the literature on pre- scribing, even though the authors of the review questioned whether such an improvement was clinically meaningful when all the evidence was presented to- gether. Further to the support from the literature, the Google Trends data for news searches in England also suggested increased awareness around the time of this trend change. The interest for the search term ‘Alzhei- mer’s disease’ was at its maximum in September 2012 (based on the data available from January 2008 to De- cember 2015, inclusive), which could indicate interest among the public. ‘catch-up prescribing’ was complete, which is suggested by the trend analysis but is not as clear when considering the raw data points. These results differ from the sensitiv- ity analyses that considered relaxed diagnosis definitions, though the ‘any Alzheimer’s disease’ and ‘any dementia’ analyses were in line with each other. This suggests that prescribing for patients with probable Alzheimer’s disease was more consistent, as one might expect, across the study than for other groups. This could indicate that pa- tients with dementias other than probable Alzheimer’s dis- ease (i.e., with unlicensed indications) were receiving these drugs and that their prescriptions were subject to change over the period studied. Further to this, large increases in prescriptions are observed as the diagnosis definition is re- laxed. This could provide further evidence for the possible unlicensed use of this drug class. The literature at that time also reflects ongoing discussion concerning the bene- fit of these drugs for indications other than Alzheimer’s disease. For example, a 2012 review by Rodda and Carter discusses their use in vascular dementia, dementia with Lewy bodies and Parkinson’s disease dementia [34]. Alter- natively, it could be attributed to the fluctuating course of symptoms that some people with dementia experience or increased recognition of mixed diagnoses where there is evidence of Alzheimer’s disease in addition to other forms of dementia, both of which might lead to treatment changes. News search Figure 4 presents the Google Trends data for news searches in England for the disease term ‘Alzheimer’s disease’ each month from January 2008 to December 2015, inclusive. There were no strong trends in the Table 4 Comparison of the sample sizes and joinpoint estimates, presented with 95% confidence intervals, f Probable AD Any AD An Table 4 Comparison of the sample sizes and joinpoint estimates, presented with 95% confidence intervals, for all analyses Probable AD Any AD Any dementia Source Main analysis Additional file 3 Additional file 3 Diagnoses Probable AD Probable AD Possible AD Probable AD Possible AD Non-AD and mixed dementias AChE inhibitors Eligible: 10,456 Treated: 5019 Joinpoint 1: Oct 2012 (Sep 2011–Apr 2013) Joinpoint 2: May 2013 (Nov 2012–Apr 2014) Eligible: 21,342 Treated: 6449 Joinpoint 1: Jun 1999 (Apr 1998–Dec 2000) Joinpoint 2: Jun 2001 (Sep 2000–Mar 2002) Eligible: 38,650 Treated: 9896 Joinpoint 1: Aug 2000 (Jun 1998–Nov 2000) Joinpoint 2: Jan 2001 (Sep 2000–Nov 2001) NMDA receptor antagonists Eligible: 9964 Treated: 1052 Joinpoint 1: Mar 2011 (Aug 2010–Apr 2011) Joinpoint 2: Jun 2011 (Apr 2011–Nov 2011) Eligible: 18,930 Treated: 1309 Joinpoint 1: Sep 2010 (Dec 2009–Apr 2011) Joinpoint 2: Nov 2011 (Apr 2011–Mar 2012) Eligible: 35,625 Treated: 1961 Joinpoint 1: Aug 2010 (Nov 2009–Dec 2010) Joinpoint 2: Nov 2011 (Aug 2011–Mar 2012) AChE Acetylcholinesterase, AD Alzheimer’s disease, NMDA N-methyl-D-aspartate Walker et al. Alzheimer's Research & Therapy (2018) 10:51 Page 8 of 11 Fig. 4 Google Trends data for news searches in England for the disease term ‘Alzheimer’s disease’. This graph shows the interest in the disease term ‘Alzheimer’s disease’ each month from January 2008 to December 2015, inclusive. Interest is given as a percentage scaled against peak popularity, which is represented as a value of 100% and occurred for the downloaded data in September 2012. Values of zero indicate insufficient data for that month Fig. 4 Google Trends data for news searches in England for the disease term ‘Alzheimer’s disease’. This graph shows the interest in the disease term ‘Alzheimer’s disease’ each month from January 2008 to December 2015, inclusive. Interest is given as a percentage scaled against peak popularity, which is represented as a value of 100% and occurred for the downloaded data in September 2012. Strengths and limitations The key strength of this study is the large sample of pri- mary care data with prescribing information, provided by the CPRD. The CPRD is ‘broadly representative of the UK general population’ and was generally compar- able to the last census in 2011 for age, sex and ethni- city despite young people and smaller practices tending to be slightly underrepresented [36, 37]. Our data extract contains 40,202 patients diagnosed with dementia in England up to 1st January 2016 (note that data are re- stricted to practices with a last data collection date in 2016), including 10,651 with probable Alzheimer’s dis- ease and a further 12,167 with possible Alzheimer’s dis- ease. A further strength of our study is the long follow-up of patients that allowed us to consider patients who did not receive immediate treatment. This is im- portant because pharmacological interventions for Alzheimer’s disease have historically considered severity as part of the prescribing decision, so there is likely to be a treatment delay after initial diagnosis for those pre- senting with mild disease. g The main limitation of our study is the likelihood of missed diagnoses. This is demonstrated within our data- set, because there were 1231 patients receiving one of the treatments of interest who did not have any form of recorded dementia diagnosis. Missed diagnoses are likely to be due to (1) outdated or non-specific diagnoses (i.e., type of dementia is not updated once established), (2) diagnoses received outside of primary care (i.e., from a specialist service) and (3) unrecorded diagnoses in pri- mary care (i.e., a diagnosis is given but not added to a record). Missed diagnoses have been explored in sensi- tivity analyses by testing the sensitivity and specificity of our diagnosis definitions (Additional file 2) and by relax- ing the diagnosis definition from ‘probable Alzheimer’s disease’ to include other less certain codes for the disease and other types of dementia (Additional file 3). Neither of these sensitivity analyses provided any cause for concern. A final limitation of this study is the diffi- culty in determining the lag time between an event and a trend change to assess the impact of the event. To allow for this, we have focused on events that are con- sidered to be of greatest impact—for example, changes at a national level—and so we expect any effect associ- ated with them to be evident if present. News search The second trend change in the AChE inhibitor ana- lysis occurred in May 2013 (95% CI, November 2012 to April 2014, p = 0.789), less than 1 year after the initial change for this drug class and with overlapping 95% CIs. This change signals the end of the surge in prescribing and the start of a decreasing trend in prescriptions. This is not unexpected, because patent expiry may have led to a form of ‘catch-up prescribing’ whereby people who were previously denied access to the drug were granted access at this time owing to its newly reduced cost. This would result in the apparent decreasing trend once Page 9 of 11 Walker et al. Alzheimer's Research & Therapy (2018) 10:51 The trend changes in the NMDA receptor antagonist analysis occurred in March 2011 (95% CI, August 2010 to April 2011, p = 0.891) and June 2011 (95% CI, April 2011 to November 2011, p = 0.896). Notably, the 95% CI for the first trend change ends in April 2011, which is when the 95% CI for the second trend change begins. This suggests that the trend changes may be related. The first of these trend changes marks the start of a strong increasing trend that changes to a steadily in- creasing trend following the second trend change. In March 2011 NICE introduced guidelines that recom- mended the prescription of memantine for patients with moderate to severe Alzheimer’s disease or for those people who could not tolerate AChE inhibitors. This replaced existing guidelines that restricted access to memantine to patients participating in clinical trials. It would therefore seem that these trend changes relate to the transition between the existing guidelines and those introduced in March 2011. In addition, we observed the second highest peak in interest (88% of maximum inter- est) for the disease term ‘Alzheimer’s disease’ in the Goo- gle Trends data for news searches in England in January 2011. In this month, the ‘Final Appraisal Determination on Donepezil, galantamine, rivastigmine and memantine for the treatment of Alzheimer’s disease’ was released. NICE defined this document as ‘the appraisal commit- tee’s final draft guidance about using a treatment or group of treatments in the NHS’, which becomes guidance if not appealed [35]. News search The increase in news searches around this time, and its alignment with the re- lease of the final draft guidance, supports the idea of a transition in prescribing practice due to the NICE guid- ance. Finally, the evidence concerning the use of mem- antine is summarized in a technology appraisal conducted by NICE in 2011 to support their guidance [11]. We cannot disentangle the role that this informa- tion from several studies published prior to the trend change might have played in changes to prescribing. indicates a clear pattern in prescribing, suggestive of a distinct change in practice. This provides additional sup- port for our inferences concerning the impact of the 2011 NICE guidance on the NMDA receptor antagonist drug class. Competing interests p g The authors declare that they have no competing interests. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Funding This work was supported by the Perros Trust and the Integrative Epidemiology Unit. The Integrative Epidemiology Unit is supported by the Medical Research Council and the University of Bristol (grant number MC_UU_12013/9). PGK has a professorship supported by the Sigmund Gestetner Foundation. Conclusions Analysis of both drug classes indicates that inclusion of dementia in QOF had no effect on prescribing trends and the other factors had mixed effects. NICE guidance on the prescribing of drugs for dementia aligned with trend changes for NMDA receptor antagonists but not AChE inhibitors. The guidance that had the noticeable effect was released in March 2011 and allowed the NMDA receptor antagonist memantine to be used out- side of clinical trials. All other guidance for both this drug and AChE inhibitors, including that which recom- mended restricting access, did not align with trend changes. Government dementia strategies also appear to have had mixed results, with the Prime Minister’s Dementia Challenge (launched May 2012) being the only strategy to align with a trend change. Although this strategy is likely to have increased awareness of demen- tia around the time of the October 2012 trend change for AChE inhibitors, we believe that the more likely cause of this change is the patent expiry of the drugs in this class. This will have reduced the cost of these drugs and potentially led to a surge in prescribing, such as that observed in our trend analysis. The events considered here highlight the many factors that may have influenced prescribing rates and the challenges in assessing the impact of a given event. Overall it would seem that the proportion of patients receiving prescriptions increased over the period studied, regardless of changing guidelines and other initiatives. Furthermore, given the increase in diagnoses of dementia and, more specifically, Alzheimer’s disease reported in the CPRD (Fig. 1), the absolute number of prescriptions has increased consider- ably over the period studied. Strengths and limitations However, this prevents us from covering all the factors that may influ- ence prescriptions for drugs for dementia during the study period; for example, we cannot comment on all Interestingly, neither of the trend changes in the NMDA receptor antagonist analysis aligns with those observed for the AChE inhibitors. This suggests that the NICE guidelines, which were implemented at the same time for both drug classes, may not have been as effect- ive for AChE inhibitors. This is likely due to the fact that these drugs were available outside of clinical trials prior to the restrictive guidelines recommended in 2006. The sensitivity analyses conducted for the NMDA receptor antagonists were consistent with these results, regardless of the diagnosis definition used. The first of the join- points for all NMDA receptor antagonist analyses oc- curred in the 7-month period between August 2010 and March 2011, and the second occurred in the 6-month period between June 2011 and November 2011. This high level of consistency across diagnosis definitions Page 10 of 11 Page 10 of 11 Page 10 of 11 Walker et al. Alzheimer's Research & Therapy (2018) 10:51 papers concerning these drugs published during this time. treatments and to adjust for them in other forms of analyses, particularly as factors that may modify the rates of disease progression. This study may also help to inform the handling of regulatory guidance and recommendations concerning drugs for dementia in the future. Acknowledgements We acknowledge Dr. Elizabeth Coulthard (University of Bristol) for input during the initial stages of this study. Abbreviations AChE: Acetylcholinesterase; AD: Alzheimer’s disease; AMPC: Average monthly percent change; CPRD: Clinical Practice Research Datalink; HES: Hospital Episode Statistics; MMSE: Mini Mental State Examination; MPC: Monthly percent change; NICE: National Institute for Health and Care Excellence; NMDA: N-methyl-D-aspartate; ONS: Office of National Statistics; QOF: Quality and Outcomes Framework Availability of data and materials This study was based in part on data from the Clinical Practice Research Datalink obtained under licence from the U.K. Medicines and Healthcare products Regulatory Agency. These data are provided by patients and collected by the NHS as part of patient care and support. The interpretation and conclusions contained in this article are those of the authors alone. In this study, we also used linked data from the Hospital Episode Statistics and Office of National Statistics datasets. Copyright © 2018. Re-used with the per- mission of the Health & Social Care Information Centre. All rights reserved. To our knowledge, there are two other studies that have considered prescribing trends, and these were fo- cused mainly on the impact of the National Dementia Strategy [2, 3]. Our study extends the findings of these previous studies because it considers trends since the launch of these drugs and implements a joinpoint model as a hypothesis-free approach for the factors affecting prescribing. We have observed that prescription rates in England do not always respond to factors such as regula- tory guidance, recommendations or patent expiry, and when they do, not necessarily in a predictable way. This suggests that communication with clinicians may need to be improved to use drugs for dementia more cost-effectively. In addition to this, the present study provides insight into the factors that may have influ- enced prescription rates of drugs for dementia in Eng- land since their launch in 1997. This is essential for accurate assessment of the effectiveness of these Authors’ contributions All authors contributed to planning the analysis. VMW conducted the analysis and drafted the manuscript. All other authors edited and revised the manuscript. PGK and RMM were responsible for securing the funding. All authors read and approved the final manuscript. Additional files Additional file 1: Analysis of the representativeness of the CPRD. (PDF 343 kb) Additional file 2: Analysis of the sensitivity and specificity of the diagnoses. (PDF 438 kb) Additional file 3: Analysis using alternative diagnoses. (PDF 1728 kb) Additional file 2: Analysis of the sensitivity and specificity of the diagnoses. (PDF 438 kb) Additional file 3: Analysis using alternative diagnoses. (PDF 1728 kb) Additional file 3: Analysis using alternative diagnoses. (PDF 1728 kb) Ethics approval and consent to participate The protocol for this study was approved by the CPRD’s Independent Scientific Advisory Committee (ISAC protocol 15_246R). This study did not directly involve patients, so further ethical approval was not required. References 21. Boarer C, Solomons K. Acetylcholinesterase inhibitors: maximising benefits to the Surrey healthcare economy from the loss of exclusivity of donepezil, galantamine and rivastigmine in 2012–13. 2012. http://pad.res360.net/ Content/Documents/Loss%20of%20patent%20exclusivity%20of%20 acetylcholinesterase%20inhibitors.pdf. Accessed 16 Dec 2016. 1. British National Formulary (BNF). Section 4.11: drugs for dementia. London: BNF; 2015. https://www.medicinescomplete.com/mc/bnflegacy/current/ PHP3236-dementia.htm. Accessed 11 Jan 2016 2. Donegan K, Fox N, Black N, Livingston G, Banerjee S, Burns A. Trends in diagnosis and treatment for people with dementia in the UK from 2005 to 2015: a longitudinal retrospective cohort study. Lancet Public Health. 2017;2: e149–56. 22. Walker VM, Davies NM, Jones T, Kehoe PG, Martin RM. Can commonly prescribed drugs be repurposed for the prevention or treatment of Alzheimer’s and other neurodegenerative diseases? Protocol for an observational cohort study in the UK Clinical Practice Research Datalink. BMJ Open. 2016;6:e012044. 3. Mukadam N, Livingston G, Rantell K, Rickman S. Diagnostic rates and treatment of dementia before and after launch of a national dementia policy: an observational study using English national databases. BMJ Open. 2014;4:e004119. 23. Walker V, Davies N, Kehoe PG, Martin R. CPRD codes: neurodegenerative diseases and commonly prescribed drugs. 2017. https://doi.org/10.5523/bris. 1plm8il42rmlo2a2fqwslwckm2. Accessed 16 Oct 2017. 4. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189–98. 4. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189–98. 24. Health & Social Care Information Centre. Read codes. http://webarchive. nationalarchives.gov.uk/20160921135209/http://systems.digital.nhs.uk/data/ uktc/readcodes. Accessed 16 Dec 2016 5. National Institute for Health and Care Excellence (NICE). NICE consults on revised first draft guidance on the use of drugs to treat Alzheimer’s disease (including a review of existing guidance no. 19) [press release]. https:// www.nice.org.uk/guidance/TA111/documents/2006001-nice-consults-on- revised-first-draft-guidance-on-the-use-of-drugs-to-treat-alzheimers-disease2. Accessed 24 Nov 2016. 25. Seshadri S, Zornberg GL, Derby LE, Myers MW, Jick H, Drachman DA. Postmenopausal estrogen replacement therapy and the risk of Alzheimer disease. Arch Neurol. 2001;58:435–40. 26. Statistical Methodology and Applications Branch, Surveillance Research Program, National Cancer Institute. Dependent variable—Joinpoint help system 4.5.0.1. https://surveillance.cancer.gov/help/joinpoint/setting- parameters/input-file-tab/dependent-variable. Accessed 20 Nov 2017. 6. National Institute for Health and Care Excellence (NICE). Alzheimer’s disease (mild to moderate) - donepezil, galantamine, rivastigmine and memantine (part review): final scope. London: NICE; 2009. https://www.nice.org.uk/ guidance/TA217/documents/alzheimers-disease-mild-to-moderate- donepezil-galantamine-rivastigmine-and-memantine-part-review-final- scope2. Accessed 6 Feb 2017 27. Statistical Methodology and Applications Branch, Surveillance Research Program, National Cancer Institute. References APC Definition - Joinpoint Help System 4.6.0.0. https://surveillance.cancer.gov/help/joinpoint/tech-help/frequently- asked-questions/apc-definition. Accessed 18 May 2018. 7. National Institute for Health and Care Excellence (NICE). Dementia: supporting people with dementia and their carers in health and social care. London: NICE; 2006. http://www.nice.org.uk/guidance/cg42/resources/ dementia-supporting-people-with-dementia-and-their-carers-in-health-and- social-care-975443665093. Accessed 6 Feb 2017. 28. Kim HJ, Fay MP, Feuer EJ, Midthune DN. Permutation tests for joinpoint regression with applications to cancer rates. Stat Med. 2000;19:335–51. 29. Statistical Methodology and Applications Branch, Surveillance Research Program, National Cancer Institute. Joinpoint regression program. Bethesda, MD: National Institutes of Health; 2016. 8. Dyer C. NICE’s decision on dementia drugs was “irrational,” High Court is told. BMJ. 2007;334:1337. 30. StataCorp. Stata statistical software. College Station, TX: StataCorp LP; 2015. 31. Walker V. DementiaDrugsCPRD. GitHub. 2017. https://github.com/venexia/ D ti D CPRD A d 7 S 2017 31. Walker V. DementiaDrugsCPRD. GitHub. 2017. https://github.com/venexia/ DementiaDrugsCPRD. Accessed 7 Sep 2017. V. DementiaDrugsCPRD. GitHub. 2017. https://github.com/venexia/ 9. Kmietowicz Z. NICE hears appeals over dementia drugs. BMJ. 2006;333:165. DementiaDrugsCPRD. Accessed 7 Sep 2017. 10. Iliffe S. The National Institute for Health and Clinical Excellence (NICE) and drug treatment for Alzheimer’s disease. CNS Drugs. 2007;21:177–84. 32. Google. Interest over time on Google Trends for Alzheimer’s disease - England, 1/1/08–12/1/15. https://g.co/trends/TTrq9. Accessed 4 Apr 2018. 11. National Institute for Health and Care Excellence (NICE). Donepezil, galantamine, rivastigmine and memantine for the treatment of Alzheimer’s disease. NICE technology appraisal guidance 217. London: NICE; 2011. https://www.nice.org.uk/guidance/ta217/resources/donepezil-galantamine- rivastigmine-and-memantine-for-the-treatment-of-alzheimers-disease-pdf- 82600254699973. Accessed 20 Nov 2017. 33. Buckley JS, Salpeter SR. A risk-benefit assessment of dementia medications: systematic review of the evidence. Drugs Aging. 2015;32:453–67. 34. Rodda J, Carter J. Cholinesterase inhibitors and memantine for symptomatic treatment of dementia. BMJ. 2012;344:e2986. 34. Rodda J, Carter J. Cholinesterase inhibitors and memantine for symptomat treatment of dementia. BMJ. 2012;344:e2986. 35. National Institute for Health and Care Excellence (NICE). Glossary. https:// www.nice.org.uk/glossary?letter=f. Accessed 4 Apr 2018. 35. National Institute for Health and Care Excellence (NICE). Glossary. https:// www.nice.org.uk/glossary?letter=f. Accessed 4 Apr 2018. 12. NHS Digital. Quality and Outcome Framework. 2015. http://content.digital. nhs.uk/qof. Accessed 18 Jan 2017. 36. Herrett E, Gallagher AM, Bhaskaran K, Forbes H, Mathur R, van Staa T, et al. Data resource profile: Clinical Practice Research Datalink (CPRD). Int J Epidemiol. 2015;44:827–36. 13. Information Centre, Government Statistical Service. National Quality and Outcomes Framework statistics for England 2006/07. 2007. http://content. digital.nhs.uk/catalogue/PUB05997/qof-eng-06-07-bull-rep.pdf. Accessed 11 Jan 2017. 37. Received: 7 September 2017 Accepted: 1 May 2018 Received: 7 September 2017 Accepted: 1 May 2018 Received: 7 September 2017 Accepted: 1 May 2018 20. Alderwick H, Robertson R, Appleby J, Dunn P, Maguire D. Better value in the NHS: the role of changes in clinical practice. https://www.kingsfund.org.uk/ sites/default/files/field/field_publication_file/better-value-nhs-Kings-Fund- July%202015.pdf. Accessed 13 Dec 2016. Author details 1 1Bristol Medical School: Population Health Sciences, University of Bristol, Bristol, UK. 2MRC Integrative Epidemiology Unit, University of Bristol, Bristol, Page 11 of 11 Page 11 of 11 Walker et al. Alzheimer's Research & Therapy (2018) 10:51 UK. 3Dementia Research Group, University of Bristol, Bristol, UK. 4Bristol Medical School: Translational Health Sciences, University of Bristol, Bristol, UK. 18. Abdi Z, Burns A. Championing of dementia in England. Alzheimers Res Ther. 2012;4:36. 19. Wortmann M. Importance of national plans for Alzheimer’s disease and dementia. Alzheimers Res Ther. 2013;5:40. References Mathur R, Bhaskaran K, Chaturvedi N, Leon DA, vanStaa T, Grundy E, et al. Completeness and usability of ethnicity data in UK-based primary care and hospital databases. J Public Health (Oxf). 2014;36:684–92. 14. NHS Digital. Summary of QOF indicators. https://assets.publishing.service. gov.uk/government/uploads/system/uploads/attachment_data/file/213226/ Summary-of-QOF-indicators.pdf. Accessed 18 Jan 2017. 15. Department of Health. Living well with dementia: a national dementia strategy. 2009. https://assets.publishing.service.gov.uk/government/uploads/ system/uploads/attachment_data/file/168220/dh_094051.pdf. Accessed 14 Sept 2016. 16. Department of Health. Prime minister’s challenge on dementia - delivering major improvements in dementia care and research by 2015. 2012. https:// assets.publishing.service.gov.uk/government/uploads/system/uploads/ attachment_data/file/215101/dh_133176.pdf. Accessed 16 Dec 2016. 17. Department of Health. Prime minister’s challenge on dementia 2020. 2015. https://assets.publishing.service.gov.uk/government/uploads/system/ uploads/attachment_data/file/414344/pm-dementia2020.pdf. Accessed 16 Dec 2016.
W4389279662.txt	https://ejournal.uin-suska.ac.id/index.php/IBF/article/download/23493/9721	en		Pengaruh Pengetahuan Investasi, Risk Tolerance, Dan Self Efficacy Terhadap Intensi Investasi Pasar Modal Pada Generasi-Z	Islamic Business and Finance	2,023	cc-by-sa	6,119	Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 Pengaruh Pengetahuan Investasi, Risk Tolerance, Dan Self Efficacy Terhadap Intensi Investasi Pasar Modal Pada Generasi-Z 1),2),3) Reza Fitriana Rizkia1, Fitria Ardia Pramesti2, Siti Amaroh3 Fakultas Ekonomi dan Bisnis Islam, Institut Agama Islam Negeri Kudus rezafitrianarizkia@gmail.com Abstract Investments in the country's actual economy and its capital market will increase as a result of globalization. Because of this, the capital market requires members of Generation Z to have a high level of self-awareness. The purpose of this study is to examine how generation Z investors' levels of investing expertise, risk aversion, and confidence influence financial market performance. The goal of this research was to identify the factors (investment knowledge, risk tolerance, and self-efficacy) that influence people's decisions to put money into the stock market. Gen Z is the focus of this investigation. Non-probability type purposive sampling was used to choose the 200 respondents who would make up this study's sample of members of Generation Z. The validity and reliability tests were used to evaluate the research instrument, and the classical assumption test and multiple regression analysis were used to analyze the collected data. Capital market investing intentions are positively related to investment knowledge, risk tolerance, and confidence, according to the investigation. The findings of this study have important implications for encouraging younger generations' interest, trust, and self-awareness in the capital market. Keywords: Investment Intensity, Investment Knowledge, Risk Tolerance, Self Efficacy. Abstrak Investasi dalam ekonomi aktual negara dan pasar modalnya akan meningkat sebagai akibat dari globalisasi. Oleh karena itu, pasar modal menuntut anggota Generasi Z memiliki kesadaran diri yang tinggi. Tujuan dari penelitian ini adalah untuk menguji bagaimana tingkat keahlian investasi, penghindaran risiko, dan kepercayaan investor generasi Z mempengaruhi kinerja pasar keuangan. Tujuan dari penelitian ini adalah untuk mengidentifikasi faktor-faktor (pengetahuan investasi, risk tolerance, dan self efficacy) yang mempengaruhi keputusan masyarakat untuk memasukkan uang ke pasar saham. Gen Z adalah fokus dari investigasi ini. Jenis purposive sampling non-probability digunakan untuk memilih 200 responden yang akan menjadi sampel penelitian ini anggota Generasi Z. Uji validitas dan reliabilitas digunakan untuk mengevaluasi instrumen penelitian, dan uji asumsi klasik dan analisis regresi berganda digunakan untuk menganalisis data yang terkumpul. Niat investasi pasar modal berhubungan positif dengan pengetahuan investasi, toleransi risiko, dan kepercayaan, menurut penyelidikan. Temuan penelitian ini memiliki implikasi penting untuk mendorong minat, kepercayaan, dan kesadaran diri generasi muda di pasar modal. Kata Kunci: Intensi Investasi, Pengetahuan Investasi, Toleransi Risiko, Efikasi Diri. PENDAHULUAN Perubahan sistem ekonomi dan teknologi informasi saat ini berdampak besar pada banyak sektor bisnis yang berbeda. Hal ini terlihat dari semakin banyaknya pebisnis yang gencar mengembangkan usahanya dengan memanfaatkan berbagai keuntungan dan kemudahan Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 144 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 yang diberikan oleh kemajuan teknologi serta berlomba-lomba untuk tetap menjalankan usahanya agar dapat berkembang. Perusahaan dapat mendorong partisipasi publik dalam keberhasilannya melalui pasar modal dengan memberikan kemungkinan bagi investor untuk membeli saham di perusahaan tersebut. Pasal 1 ayat 13 Undang-Undang Nomor 8 Tahun 1995 tentang Pasar Modal mendefinisikan Pasar Modal sebagai “kegiatan yang berkaitan dengan penawaran umum dan perdagangan efek, perusahaan publik yang terkait dengan efek yang mereka terbitkan, serta lembaga dan profesi yang terkait dengan efek” (Ortega, 2023). Perdagangan pasar modal merupakan sebagian besar dari semua investasi. Pengenalan Bursa Efek Indonesia telah membuat investasi pasar saham lebih mudah diakses oleh masyarakat awam. Tingkat kemajuan ekonomi dan teknis di Indonesia saat ini sebagian besar dapat dikaitkan dengan meningkatnya kecepatan pencatatan perusahaan baru di Bursa Efek Indonesia (BEI) dan banyaknya investor internasional yang membuka rekening di Indonesia (Wardani, 2020). Banyak orang yang tertarik untuk berinvestasi, namun tidak banyak yang berhasil dalam jangka panjang. Alasan utama mengapa hal ini terjadi karena orang-orang yang terlibat tidak memiliki tujuan khusus untuk uang mereka yang penting bagi mereka saat berinvestasi. Hal ini mengakibatkan dua hal, yaitu pemahaman yang kurang terhadap potensi investasi dan kurangnya motivasi untuk berpartisipasi (Lubis, 2019). Pertumbuhan investasi di pasar modal dapat dilihat dari meningkatnya jumlah emiten serta jumlah investor pasar modal Indonesia. Jumlah investor yang terdaftar pada tahun 2022 mengalami peningkatan, menurut data Otoritas Jasa Keuangan (OJK) pada tahun 2022. Gambar 1 Data Pertumbuhan Investor periode 2019-2022 Sumber: PT. Kustodian Sentral Efek Indonesia (KSEI), 2022 Pasar modal telah mengembalikan lebih dari 50% setiap tahun kepada investor, menurut Kustodian Sentral Efek Indonesia (KSEI) untuk tahun 2022. OJK semester I Juni 2022, pasar modal akan memiliki 9,1 juta investor, di mana 3,9 juta di antaranya akan memiliki saham. Menurut Kementerian Dalam Negeri (Kemendagri, 2020), dari total penduduk Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 145 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 Indonesia sebanyak 275,36 juta jiwa per Juni 2022, sebanyak 190,83 juta (atau 69,3%) merupakan usia kerja (15-64 tahun). Hanya sekitar 3,9 persen orang Indonesia dalam populasi usia kerja yang terlibat di pasar saham, dan hanya 2,2 persen orang Indonesia yang memegang saham di pasar saham, menurut statistik. Perkembangan pasar modal di Indonesia terhambat oleh rendahnya rasio investor terhadap penduduk, yang merupakan fenomena sekaligus hambatan tersendiri. Indeks Harga Saham Gabungan (IHSG) naik nilainya sebesar 58,70% antara tahun 2013 dan 2022, menurut data Bursa Efek Indonesia (BEI). Untuk indeks saham, ini adalah langkah ekspansi yang sangat sehat. Menurut statistik CMBC, milenial bukan lagi kelompok usia yang paling umum di kalangan pelaku pasar saham Indonesia. Pada 14 Oktober 2022, KSEI mencermati sebanyak 58,91% investor dengan total Rp52,77 triliun dan jumlah single investor identity (SID) berusia di bawah 30 tahun. Uriep Budhi Prasetyo, Direktur Utama KSEI, optimis menatap masa depan sejak muda orang-orang saat ini menjadi semakin terdidik tentang nilai investasi, khususnya di pasar saham. KSEI telah mengerjakan proyek untuk membuat akun online bagi investor sejak 2019 sebagai tanggapan atas fakta bahwa mayoritas investor pasar modal adalah milenial dan Gen Z. Kenaikan 31,85% dibanding tahun sebelumnya membuat jumlah investor di pasar modal menjadi 9,87 juta per 14 Oktober 2022. Saat ini ada 9,18 juta orang yang berinvestasi di reksa dana, naik 34,30 persen dari tahun sebelumnya. Berdasarkan data di atas, terlihat dari hasil bahwa meskipun pasar modal Indonesia mengalami pertumbuhan investasi yang relatif tinggi, namun tidak serta akan meningkatkan intensi dalam berinvestasi. Hal ini dikarenakan kurangnya pengetahuan investasi yang dimiliki investor atau calon investor. Selain dengan pengetahuan investasi, intensi investor juga dipengaruhi oleh faktor psikologis. Karena salah satu faktor penentu keberhasilan dalam berinvestasi adalah tingkat pengetahuan seseorang, maka penting untuk mulai menumbuhkan minat berinvestasi sejak dini agar dapat menumbuhkan semangat investasi. Pemahaman menyeluruh tentang instrumen saham diperlukan untuk melakukan analisis mendalam tentang hasil perdagangan (Syaputra, 2022). Saat melakukan investasi pertimbangan harus dilakukan agar harapan, risiko yang terjadi, dan tingkat kepercayaan pada seseorang sudah diketahui. Indikasi pertama potensi investasi datang dari harga yang lebih tinggi dan terus naik di masa depan (Kurniawan, 2021). Niat investasi perilaku masa depan investor sangat dipengaruhi oleh sikap mereka terhadap perilaku, norma subyektif, kontrol perilaku yang dirasakan, dan kecenderungan risiko, seperti yang diungkapkan oleh Alleyne dan Broome (2011) oleh karena itu, pengetahuan investasi sangat penting bagi seseorang yang berniat berinvestasi sehingga mereka dapat Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 146 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 menilai risiko yang ada, menentukan kemungkinan pengembalian, memperkuat motivasi mereka, dan pada akhirnya mencapai tujuan mereka melalui strategi investasi yang dipikirkan dengan matang (Sobaih, 2023). Pada penelitian terdahulu yang dilakukan oleh Firdaus & Ifrochah (2022) mencapai hasil yang substansial antara pengetahuan investasi dan keinginan untuk berinvestasi di pasar modal, karena seseorang tidak ingin berinvestasi tanpa pemahaman tersebut. Sedangkan menurut penelitian Fitriasuari & Simanjuntak (2022) menunjukkan bahwa konsumen lebih tertarik untuk melakukan pembelian daripada investasi pasar modal, sehingga keahlian investasi menjadi tidak relevan. Kemungkinan kerugian finansial karena perubahan harga saham, berurusan dengan penjual yang tidak jujur, berinvestasi pada penerbit palsu, dan bahaya lainnya adalah bagian dari ketidakpastian yang melekat pada pasar saham. Kemampuan seseorang untuk mengambil risiko bervariasi dari orang ke orang. Toleransi risiko investor merupakan faktor utama dalam menentukan apakah akan berinvestasi atau tidak dan, jika demikian, berapa banyak. Kategori pengambilan risiko, penghindaran risiko, dan penghindaran risiko menggambarkan tiga jenis investor (Frans, 2020). Menurut penelitian yang dilakukan oleh Laurency & Arifin (2022), ada korelasi yang menguntungkan antara perasaan seseorang tentang risiko dan kecenderungan mereka untuk berinvestasi. Karena kecenderungan investasi individu meningkat sebanding dengan tingkat persepsi risikonya. Sementara Wardani (2020) menemukan bahwa persepsi investor tentang risiko memengaruhi keputusan mereka untuk berpartisipasi di pasar saham, studi saat ini menemukan bahwa persepsi tersebut tidak berpengaruh pada perilaku investasi aktual. Menurut Ajzen dan Fishbein (1991), sikap terhadap perilaku adalah evaluasi keseluruhan seseorang atas perilakunya sendiri, sikap positif atau negatif, untuk mendukung situasi berbahaya tertentu. Norma subyektif adalah penerimaan seseorang terhadap pendapat orang lain tentang apa yang perlu bagi mereka untuk melaporkan atau tidak melaporkan peristiwa tertentu yang sesuai dengan kesepakatan. Kontrol perilaku yang dirasakan adalah persepsi seseorang tentang kapasitas mereka untuk secara efektif menangani situasi tertentu (Frans, 2020). Dalam penelitian yang dilakukan oleh Florencia & Arifin (2022), menemukan bahwa mayoritas responden berdomisili Jabodetabek melaporkan adanya hubungan yang menguntungkan antara financial self-efficacy dan niat investasi. Menurut temuan Wibowo & Purwohandoko (2019) tingkat efikasi diri individu berpengaruh terhadap kecenderungan Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 147 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 melakukan investasi keuangan. Orang memiliki pemahaman umum tentang nilai investasi, tetapi mereka berhati-hati melakukannya tanpa tingkat keahlian tertentu. Saat melakukan investasi saham, masyarakat Indonesia merasa rentan karena kondisi saham tertentu di pasar modal. Pasar Modal adalah sumber daya berharga yang dapat digunakan penduduk Indonesia untuk berinvestasi dalam segala hal (dalam aset keuangan). Kehadiran pasar modal membuat investor memiliki akses ke berbagai pilihan investasi yang sesuai dengan tujuan mereka. Investasi mengandung masalah penawaran yang belum terselesaikan. Ada investasi yang menawarkan tingkat pengembalian modal yang tinggi tetapi tingkat modal investasi yang rendah. Sebaliknya, ada investasi yang menawarkan tingkat pengembalian modal yang rendah tetapi tingkat modal investasi yang tinggi. Oleh karena itu, dunia investasi beroperasi berdasarkan prinsip "High Risk High Return", yang menyatakan bahwa investasi apa pun akan memberikan potensi imbalan tingkat tinggi bagi investor sekaligus juga memiliki tingkat risiko yang tinggi (Florencia & Arifin, 2022). Ketiga faktor dalam penelitian ini tergolong baru karena bertujuan untuk menentukan apakah faktor tersebut memiliki dampak yang menguntungkan atau tidak terhadap kecenderungan berinvestasi generasi milenial. Dengan menggunakan data dan deskripsi yang diberikan, ditentukan bahwa ada kekurangan studi tentang topik niat investor di pasar modal, serta fenomena kesenjangan. Tujuan dari penelitian ini adalah untuk mengetahui ada tidaknya perbedaan antara Generasi Z dan generasi sebelumnya dalam hal pengetahuan investasi, risk tolerance, dan self efficacy. KERANGKA TEORI Pada tahun 1975, Fishbein dan Ajzen menciptakan Theory of Reasoned Action (TRA), di mana TPB adalah subsetnya. Menurut Ajzen, TPB dirilis secara keseluruhan agar para peneliti dapat menganalisis korelasi antara ciri-ciri kepribadian dan perilaku aktual. Theory of Predictive Behavior (TPB) pertama kali diusulkan oleh Ajzen dan Fishben pada tahun 1988. Menurut TPB, situasi seseorang adalah hasil dari penyebab internal dan eksternal, dan kinerja seseorang adalah produk dari kondisi tersebut. Pandangan tentang perilaku, evaluasi perilaku, standar subjektif, pandangan normatif, dan keinginan untuk mematuhi semua aspek sikap individu terhadap perilaku (Florencia & Arifin, 2022). Menurut Ajzen (1991), norma subyektif adalah kekuatan sosial yang mendorong atau mencegah seseorang untuk melakukan tindakan tertentu. Keinginan seseorang untuk ikut serta atau menahan diri dari aktivitas yang dimaksud akan dipengaruhi oleh cara pandang atau Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 148 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 pandangan seseorang terhadap keyakinan orang lain. Keberhasilan investasi teman dan keluarganya akan mendorongnya untuk mencobanya sendiri. Menurut Ajzen (1991), sejauh mana pengalaman masa lalu, hambatan, dan pengetahuan tentang rintangan ini mempengaruhi persepsi seseorang tentang betapa mudah atau sulitnya melakukan suatu kegiatan. Niat individu untuk berpartisipasi dalam tindakan tertentu cenderung meningkat sebanding dengan daya tarik sikap terkait, standar subyektif, dan tingkat kontrol perilaku yang dirasakan. Ini menunjukkan bahwa keakraban dengan masalah uang mengarah pada peningkatan keinginan untuk berinvestasi. Ketika orang menerima informasi yang mereka butuhkan, mereka lebih cenderung ingin berinvestasi. Intensi Investasi Sikap terhadap aspek penting secara pribadi dari aktivitas dan kesan individu terhadap tekanan sosial untuk berpartisipasi atau tidak terlibat dalam tindakan (juga dikenal sebagai norma subyektif) adalah dua pendorong niat yang mendasar. Singkatnya, menurut TRA (Amhalmad & Irianto, 2019), sikap dan standar subjektif mendorong niat, yang pada gilirannya memengaruhi praktik atau perilaku. Memiliki niat investasi adalah ingin belajar tentang instrumen investasi, memiliki waktu dan sumber daya untuk mempelajari apa dan bagaimana berinvestasi dengan berpartisipasi dalam pelatihan dan seminar terkait investasi, dan kemudian berupaya mempraktikkan pembelajaran tersebut. Pengetahuan Investasi Dengan cara yang sama bahwa pengetahuan adalah bagian dari jiwa, pikiran manusia yang sadar diisi ke insang dengan informasi. Panca indera mengambil informasi tentang dunia dan menggunakannya untuk membuat representasi mental, pengamatan, apersepsi, konsepsi, dan imajinasi (Fitriasuari & Simanjuntak, 2022). Memahami cara menginvestasikan uang dan sumber daya Anda akan membantu Anda mendapatkan keuntungan dalam jangka panjang. Penelitian yang menggunakan literatur kanonik yang telah diinternalisasi oleh sistem ingatan manusia dapat memberikan hasil seperti itu. Tujuan pemberian edukasi investasi kepada peserta adalah untuk meningkatkan motivasi peserta dalam berinvestasi saham sendiri (Maulana & Kaukab, 2020). Keyakinan diri seseorang di pasar saham meningkat seiring dengan keahlian investasinya. Menurut penelitian Darmawan (2019), Minat mahasiswa dalam berinvestasi tidak banyak dipengaruhi oleh keahlian investasinya. Menurut penelitian Hidayat (2019) berbeda dengan penelitian Darmawan & Japar (2019) yang menyatakan bahwa pengetahuan investasi berpengaruh terhadap minat berinvestasi di pasar modal. Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 149 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 Risk Tolerance Ekspektasi investor untuk pengembalian berkorelasi positif dengan jumlah risiko yang bersedia mereka ambil. Seorang investor dengan toleransi risiko tinggi adalah orang yang siap menghadapi berbagai ketidakpastian saat melakukan investasi. Menurut penelitian yang dilakukan oleh Florencia & Arifin (2022) kesediaan seseorang untuk mengambil risiko berkorelasi positif dengan keinginannya untuk berinvestasi di saham. Sementara Wardani (2020) menemukan bahwa persepsi investor terhadap risiko memengaruhi keputusan mereka untuk berpartisipasi di pasar saham, studi saat ini menemukan bahwa persepsi tersebut tidak berpengaruh pada perilaku investasi aktual. Self Efficacy Kecerdasan emosional investor mungkin berdampak signifikan pada hasil mereka di pasar. Salah satunya adalah self efficacy, atau kepercayaan pada kemampuan diri sendiri untuk melakukan tugas tertentu dengan sukses. Self efficacy ditemukan memiliki pengaruh yang menguntungkan terhadap minat mahasiswa FE UNY dalam berinvestasi saham oleh Trisnatio (2017) studi self efficacy yang mendapatkan pendanaan, bagaimanapun, masih jarang. Meskipun Ismail (2017) mengklaim bahwa seseorang dengan self-efficacy yang tinggi dapat membantu mereka mencapai perilaku keuangan yang baik dan mengatasi semua rintangan, terutama dalam hal masalah keuangan, self efficacy lebih sering digunakan sebagai prediktor minat dalam berwirausaha. Menurut Frans (2020) self efficacy tidak secara signifikan meningkatkan semangat seseorang untuk berdagang saham. Penelitian yang dilakukan oleh Kurniawan (2021) efikasi diri mahasiswa akuntansi tingkat magister tidak secara signifikan dan positif mempengaruhi niat investasi mereka. Oleh karena itu, jelas bahwa studi lebih lanjut diperlukan untuk mengetahui dampak self-efficacy terhadap kecenderungan untuk berinvestasi. Pengembangan Hipotesis 1. Hubungan antara Pengetahuan Investasi dengan Intensi Investasi di Pasar Modal Memiliki pemahaman yang kuat tentang dasar-dasar investasi sangat penting. Setiap orang diharapkan dapat memperoleh manfaat dari pendidikan investasi, khususnya yang berkaitan dengan pasar modal. Kemauan seseorang untuk berinvestasi berbanding lurus dengan tingkat keahlian seseorang di bidangnya. Menghadiri seminar atau membaca literatur baru tentang pengetahuan investasi di pasar modal merupakan salah satu cara untuk belajar berinvestasi di pasar guna mendapatkan informasi tentang cara menggunakan dana atau sumber daya yang tersedia untuk mendapatkan keuntungan di masa depan (Nabilah & Hartutik, 2020). Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 150 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 Kemampuan untuk berinvestasi dengan baik membutuhkan pemahaman tentang konsep-konsep penting. Langkah pertama untuk mempelajari lebih lanjut tentang investasi adalah memperoleh informasi dasar. Oleh karena itu, menginvestasikan informasi membantu dalam persiapan tujuan dan manajemen tujuan. Penelitian Firdaus dan Ifrochah (2022) mendukung hal tersebut dengan menunjukkan bahwa variabel keahlian berinvestasi berpengaruh positif terhadap pilihan investasi di pasar saham. Sejauh mana variabel ini memengaruhi pilihan seseorang untuk berinvestasi di pasar saham dapat disimpulkan dari apa yang diketahui tentang keahlian berinvestasi. 2. Hubungan antara Risk Tolerance dengan Intensi Investasi di Pasar Modal Toleransi risiko investor diukur dengan seberapa baik mereka mampu mengidentifikasi, mengevaluasi, dan mengendalikan risiko yang melekat pada investasi tertentu. Toleransi risiko investor bervariasi dari orang ke orang dan dari waktu ke waktu berdasarkan sejumlah variabel, termasuk kepribadian dan keadaan khusus di mana seorang investor berada. Menurut Halim (2005), eksposur risiko investor dibagi menjadi tiga kategori, yaitu: 1) Pencari risiko (Risk Seeker) adalah jenis investor yang secara aktif mencari risiko. Seorang yang percaya pada strategi investasi berisiko tinggi dengan imbalan tinggi, investor semacam ini sering kali agresif dan senang melakukan investasi berdasarkan spekulasi. 2) Investor dengan sikap "netral risiko" adalah mereka yang melihat dunia secara keseluruhan. Jenis investor ini seringkali mudah beradaptasi dan berhati-hati saat mempertimbangkan pilihan investasi. 3) Investor yang menghindari risiko adalah orang yang berhati-hati dan lebih suka bermain aman. Saat dihadapkan pada kebutuhan untuk memilih banyak produk yang identik, investor dalam kategori ini akan memilih produk dengan risiko serendah mungkin. Hal ini didukung dengan penelitian yang dilakukan oleh Utsman (2021) bahwa kesediaan seseorang untuk mengambil risiko memengaruhi apakah mereka memasukkan uang ke pasar saham atau tidak. Informasi sebelumnya menunjukkan bahwa toleransi risiko individu memiliki peran dalam pilihan mereka untuk berpartisipasi di pasar saham. 3. Hubungan Self Efficacy dengan Intensi Investasi di Pasar Modal Kemanjuran diri seseorang mengacu pada keyakinannya pada keterampilannya sendiri untuk mengambil langkah-langkah yang diperlukan menuju tujuannya. Kemampuan seseorang untuk berinvestasi demi keuntungan masa depan dievaluasi dalam skenario ini. Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 151 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 Dalam hal mencapai tujuan, Lioera (2022) berpendapat bahwa kepercayaan diri seseorang terhadap kemampuannya sendiri untuk berhasil dalam situasi tertentu sangatlah penting. Oleh karena itu, diperlukan lebih banyak kepercayaan diri untuk penerapan strategi investasi yang efektif. Bergantung pada bakat, tingkat kinerja, aktivitas sehari-hari, dan kondisi emosional seseorang, setiap orang bereaksi berbeda terhadap berbagai rangsangan. Studi Lioera (2022) menunjukkan bahwa kepercayaan pada kemampuan diri sendiri memiliki efek menguntungkan pada investasi pasar keuangan. Oleh karena itu variabel self-efficacy cenderung mempengaruhi pilihan individu untuk memasukkan uang ke pasar saham. Berikut ini menyajikan model kerangka konseptual antar-variabel, yang didasarkan pada teori-teori penelitian aktual, dasar-dasar teoretis saat ini, dan kesulitan-kesulitan yang diakui sebagai dasar untuk menafsirkan hipotesis. Dalam penelitian ini, kami menguji hipotesis bahwa Pengetahuan Investasi (X1), Toleransi Risiko (X2), dan Self-Efficacy (X3) semuanya berpengaruh signifikan terhadap Niat Berinvestasi (Y) sebagai variabel dependen. Diagram yang menggambarkan dasar-dasar teoretis dari penyelidikan ini ditunjukkan pada Gambar berikut ini: Gambar 2 Kerangka Pemikiran Pengetahuan Investasi (X1) Risk Tolerance (X2) Intensi Investasi (Y) Self Efficacy (X3) Hipotesis yang digunakan dalam penelitian ini yaitu: H1: Pengetahuan investasi berpengaruh positif terhadap intensi investasi pasar modal pada generasi Z H2: Risk Tolerance berpengaruh positif terhadap intensi investasi pasar modal pada generasi Z H3: Self Efficacy berpengaruh positif terhadap intensi investasi pasar modal pada generasi Z Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 152 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 METODE PENELITIAN Untuk kepentingan penelitian ini, penelitian kuantitatif adalah suatu teknik untuk menguji kekokohan suatu teori dengan memfokuskan pada hubungan antara bagian-bagian penyusunnya (Amruddin et al., 2022). Populasi atau sampel yang representatif menjadi fokus penelitian kuantitatif. Sampling purposif non-probabilistik digunakan untuk mengumpulkan data untuk penelitian ini, dan respondennya adalah semua anggota generasi Z. Studi ini mengukur pengetahuan investasi, risk tolerance, dan self efficacy di pasar modal dengan menggunakan kuesioner yang meminta responden untuk menjawab pernyataan yang disiapkan yang dirancang untuk mengetahui faktor-faktor tersebut. Data primer dikumpulkan dari responden menggunakan kuesioner online berbasis Google Form yang dikirim melalui aplikasi TikTok. Ada variabel bebas (independen) dan terikat (dependen) dalam penelitian ini. Pengetahuan investasi, riks tolerance, dan self efficacy untuk membuat keputusan keuangan yang baik adalah tiga faktor independen yang diperiksa di sini. Intensi investasi pasar modal adalah variabel dependen. Setiap indikasi sesuai dengan variabel, yang dievaluasi pada skala Likert empat poin. Dalam penelitian ini, data yang dikumpulkan melalui kuesioner dianalisis dengan menggunakan SPSS. SPSS 26.0 akan digunakan untuk mengolah data menggunakan teknik analisis SPSS. Teknik sampling yang digunakan dalam penelitian ini memiliki margin of error 5% berkat karya Isaac dan Michael. Dua ratus orang yang memenuhi kriteria berusia di bawah 27 tahun (Gen Z) dilibatkan dalam penelitian ini. HASIL DAN PEMBAHASAN Uji Validitas Reliabilitas dan akurasi kuesioner dapat dinilai dengan menggunakan uji validitas. Kuesioner dianggap sah jika berisi pertanyaan yang dapat diandalkan untuk mengungkap konstruk yang dinilai. (Darmawan et al., 2019). Tabel 1 Hasil Uji Validitas Variabel R tabel R hitung Intensi Investasi 0,675 0, 138 Sumber: data diolah SPSS, 2023 Hasil penelitian menunjukkan bahwa semua r hitung (0,138) r tabel (0,675) yang menyatakan bahwa semua pernyataan dinyatakan valid, berdasarkan uji signifikan Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 153 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 membandingkan nilai r hitung dengan r tabel untuk derajat kebebasan (df) = n-2 = (200 - 2 = 198), menghasilkan angka 0,675. Uji Reliabilitas Menurut Suryabrata (2006) Keandalan alat ukur menunjukkan seberapa yakin seseorang terhadap hasil pengukuran tersebut. Temuan pengukuran harus konsisten dan stabil agar dianggap dapat diandalkan. Tabel 2 Hasil Uji Reliabilitas Variabel Cronbach's Alpha Pengetahuan Investasi 0.763 Risk Tolerance 0.658 Self Efficacy 0.772 Intensi Investasi 0.773 Sumber: data diolah SPSS, 2023 Memiliki temuan berulang untuk tindakan yang sama merupakan ciri khas penelitian terkemuka. Jika temuan dari beberapa pengukuran tidak konsisten, maka data tersebut tidak dapat diandalkan. Karena perbedaan individu, variabel toleransi risiko rentan terhadap fluktuasi periodik. Uji Asumsi Klasik a. Uji Normalitas Analisis Kolmogorov-Smirnov (K-S) akan digunakan untuk memeriksa normalitas dalam penelitian ini. Nilai signifikansi digunakan dalam uji KS untuk menentukan apakah hasilnya normal atau tidak. Uji normalitas, seperti yang didefinisikan oleh Sugiyono (2017), memeriksa apakah residual yang dihasilkan mengikuti distribusi normal. Uji Kolmogorov-Smirnov digunakan untuk analisis statistik ini. Residual dianggap berdistribusi normal jika dan hanya jika nilai signifikansi lebih besar dari 0,05. Tabel 3 Hasil Uji Normalitas Asymp.sig. (2-tailed) 0.10 Sumber: data diolah SPSS, 2023 Data dikatakan terdistribusi secara normal sejak Asymp.sig. (2-tailed), yaitu 0,10 > 0,05, menunjukkan demikian. Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 154 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 b. Uji Multikolinearitas Uji multikolinearitas dilakukan untuk menguji potensi korelasi antara variabel independen. Dalam model regresi, multikolinearitas terjadi ketika banyak variabel memiliki hubungan linier yang sempurna satu sama lain. (Ajija, 2011). Tabel 4. Hasil Uji Multikolinearitas Variabel Tolerance Pengetahuan 0.640 VIF Investasi 1.562 Risk Tolerance 0.694 1.441 Self Efficacy 0.652 1.533 Sumber: data diolah SPSS, 2023 Hasil analisis data menunjukkan bahwa tidak terdapat bukti multikolinieritas ketika nilai VIF kurang dari 10 dan nilai toleransi kurang dari 0,10 (1,562, 1,441, 1,533). Akhirnya, uji multikolinieritas dianggap berhasil. c. Uji Heteroskedastisitas Menurut Ghozali (2017) Heteroskedastisitas terjadi ketika model regresi mengandung variabel dengan distribusi yang berbeda. Tabel 5 Hasil Uji Heteroskedastisitas Variabel Sig. Pengetahuan Investasi 0.744 Risk Tolerance 0.087 Self Efficacy 0.378 Sumber: data diolah SPSS, 2023 Uji Glejser menunjukkan bahwa self-efficacy secara signifikan lebih tinggi dari tingkat signifikansi 0,378, sedangkan toleransi risiko secara signifikan lebih tinggi dari 0,087. Karena tidak ada bukti masalah dengan heteroskedastisitas di ketiga variabel, ini konklusif. Analisis Regresi Linier Berganda tujuan investasi berfungsi sebagai variabel dependen, dengan pengetahuan investasi, toleransi risiko, dan kepercayaan diri sebagai faktor independen. Persamaan atau fungsi yang mewakili model regresi linier berganda menggunakan variabel-variabel ini terlihat seperti ini: Y = 𝛼 + 𝛽1 X1 + 𝛽2 X2 + 𝛽3 X3 +𝜀 Dimana: Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 155 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 Y: Intensi Investasi X1: Pengetahuan Investasi X2: risk tolerance X3: self efficacy b1, b2, b3: Koefisien regresi e: error Tabel 6. Uji Regresi Linier Berganda Variabel B Sig. Pengetahuan Investasi 0.220 0.007 Risk Tolerance 0.200 0.002 Self Efficacy 0.391 0.000 Sumber: data diolah SPSS, 2023 Berdasarkan output regresi persamaan diatas, maka model analisis regresi berganda yang digunakan dalam penelitian ini dapat dirumuskan sebagai berikut: Y = 3,193 + 0,220 X1 + 0,200 X2 + 0,391 X3 Persamaan regresi tersebut diatas dapat dijelaskan sebagai berikut: 1. Seseorang dapat menyimpulkan bahwa pengetahuan investasi, risk tolerance, dan selfefficacy semuanya berpengaruh pada variabel dependen jika nilai konstanta (α) positif, seperti dalam kasus ini (3,193). Proporsi orang yang berencana menginvestasikan lebih banyak uang naik sebesar 3,193% jika nilai variabel independen di atas nol. 2. Karena tingkat signifikansi koefisien regresi untuk variabel "pengetahuan investasi" kurang dari 0,05 (yaitu, H1 yang menunjukkan bahwa ada hubungan positif antara pengetahuan investasi dan niat investasi pasar modal) kami menerima hipotesis ini. Ini menyiratkan bahwa semakin cerdas anggota Generasi Z tentang pasar modal, semakin besar kemungkinan mereka untuk benar-benar berinvestasi di dalamnya. Dengan asumsi semua faktor lainnya tetap sama. 3. H2 bahwa terdapat hubungan positif antara toleransi risiko dengan niat investasi pasar modal diterima karena variabel toleransi risiko memiliki koefisien regresi sebesar 0,200 pada taraf signifikan 0,002. Hal ini karena tingkat signifikansinya kurang dari 0,05. Aspirasi investasi Gen Z akan meningkat seiring dengan toleransi risiko mereka di pasar modal. Dengan asumsi semua faktor lainnya tetap sama. 4. H3 bahwa ada hubungan positif antara self-efficacy dan intensi investasi pasar modal diterima karena koefisien regresi untuk variabel self-efficacy adalah 0,391 pada tingkat Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 156 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 signifikansi 0,000 yang lebih kecil dari 0,05. Semakin anggota Gen Z merasa kompeten dalam masalah keuangan sebagai hasil dari studi mereka tentang investasi dan pasar modal, semakin besar pula keinginan mereka untuk melakukan investasi pasar keuangan. Dengan asumsi semua faktor lainnya tetap sama. Uji Hipotesis a. Uji t Tabel.7 Hasil Uji t Variabel T Pengetahuan 2.584 Sig. Investasi 0.007 Risk Tolerance 2.748 0.002 Self Efficacy 5.846 0.000 Sumber: data diolah SPSS, 2023 1. Pertama, uji t menunjukkan adanya hubungan positif antara pengetahuan berinvestasi dengan keinginan berinvestasi (H1 diterima) dengan taraf signifikansi 0,007 0,05 (t hitung > t tabel, atau 2,584 > 1,652). Hal ini menunjukkan bahwa keinginan seseorang untuk berinvestasi dapat dipengaruhi oleh tingkat keahlian keuangannya. 2. Signifikansi statistik hubungan antara toleransi risiko dan niat investasi adalah 0,002 0,05 (H2 diterima), dengan t hitung > t tabel, atau 2,748 > 1,652. Aspirasi investasi Gen Z akan meningkat seiring dengan toleransi risiko mereka di pasar modal. 3. Ketiga, t hitung > t tabel, atau 5,846 > 1,652, menunjukkan bahwa efikasi diri ternyata berdampak pada niat investasi, menghasilkan nilai signifikan 0,000 0,05 (H3 diterima). Hal ini menunjukkan bahwa kemungkinan milenial berinvestasi di pasar saham sebanding dengan keyakinan mereka pada kemampuan mereka sendiri untuk menangani aset keuangan dengan sukses di pasar saham. b. Uji f Tabel.8 Uji F Simultan Metode F Sig Uji F 47.872 0.000 Sumber: data diolah SPSS, 2023 Analisis regresi menunjukkan nilai F sebesar 47.872 pada tingkat signifikansi 0,0001. Namun, F-tabel pada tingkat kepastian tertinggi hanya 2,65. Ini berarti bahwa Fnumber lebih signifikan daripada F-tabel (47.872 vs. 2,65) dalam kasus ini. pengetahuan Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 157 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 investasi, toleransi risiko, dan self-efficacy semua memiliki pengaruh besar pada niat berinvestasi, seperti yang ditunjukkan oleh uji F. Uji Koefisien Determinasi Tabel 9 Uji Koefisien Determinasi Metode R Adjusted R Square Uji Koefesien Determinasi 0,650 0.414 Sumber: data diolah SPSS, 2023 Berdasarkan hasil tersebut, nilai Adjusted R-square yang merupakan ukuran koefisien determinasi adalah 0,414. Dengan kata lain, faktor independen pengetahuan investasi, toleransi risiko, dan self-efficacy menjelaskan 41,4% dari variabel dependen, niat investasi. Faktorfaktor lain di luar cakupan investigasi ini menyumbang 58,6% sisanya. Pembahasan 1. Pengaruh Pengetahuan Investasi terhadap Intensi Investasi Pasar Modal Analisis hubungan antara pengetahuan investasi dan niat investasi pasar modal mengungkapkan hubungan positif yang signifikan secara statistik antara pengetahuan investasi dan keinginan investasi saham pasar modal. Hasil ini ditunjukkan oleh nilai diperoleh nilai signifikansi sebesar 0,007 < 0,05 (H1 diterima), dengan t hitung > t tabel yaitu 2.584 > 1.652. Temuan studi ini konsisten dengan gagasan perilaku terencana, yang menyatakan bahwa individu yang memiliki pengetahuan tentang investasi lebih cenderung memiliki keyakinan diri dalam keputusan keuangan mereka. Mengetahui lebih banyak tentang investasi meningkatkan minat seseorang di pasar saham karena kontrol yang meningkat atas perilaku seseorang di bidang ini. Temuan analisis ini sejalan dengan penelitian yang dilakukan oleh Firdaus & Ifrochah (2022) yang menemukan adanya korelasi yang signifikan antara pengetahuan investasi dengan minat berinvestasi di pasar modal. Masyarakat lebih mementingkan membeli produk daripada berinvestasi di pasar modal, bertentangan dengan temuan Fitriasuari & Simanjuntak (2022) yang menyatakan bahwa pengetahuan investasi memiliki pengaruh yang kecil terhadap pilihan investasi. Saham dan instrumen investasi lainnya membutuhkan pemahaman keuangan tingkat tinggi. Saat berinvestasi di pasar saham, pengetahuan sangat penting untuk menghindari hasil yang tidak menguntungkan (kerugian). 2. Pengaruh Risk Tolerance terhadap Intensi Investasi Pasar Modal Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 158 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 Temuan studi menunjukkan bahwa variabel Toleransi Risiko mempengaruhi rencana generasi Z untuk berinvestasi di pasar modal. Hasil penelitian hipotesis kedua diperoleh nilai signifikansi sebesar 0,002 < 0,05 (H2 diterima), dengan t hitung > t tabel yaitu 2.748 > 1.652, maka secara statistik risk tolerance berpengaruh terhadap intensi investasi. Intensi investasi Gen Z akan meningkat seiring dengan toleransi risiko mereka di pasar modal. Temuan penelitian ini menguatkan penelitian Frans (2020) yang menemukan bahwa risk tolerance seseorang secara signifikan mempengaruhi kecenderungan mereka untuk berinvestasi. Temuan Samsuri et al. (2019) bahwa risk tolerance berpengaruh signifikan terhadap minat investasi mahasiswa didukung oleh temuan tersebut. Oleh karena itu, kesediaan individu untuk mengambil risiko berkorelasi positif dengan keinginan mereka untuk terlibat dalam pasar modal. Ini mungkin karena fakta bahwa generasi Z saat ini memiliki lebih banyak kesempatan untuk belajar dan tumbuh baik di dalam maupun di luar kelas. 3. Pengaruh Self Efficacy terhadap Intensi Investasi Pasar Modal Minat investasi di kalangan generasi Z secara signifikan dipengaruhi oleh komponen self-efficacy. Hasil ini ditunjukkan oleh nilai diperoleh nilai signifikansi sebesar 0,000 < 0,05 (H3 diterima), dengan t hitung > t tabel yaitu 5.846 > 1.652 maka secara statistik self efficacy berpengaruh terhadap intensi investasi. Hal ini menunjukkan bahwa kemungkinan generasi Z berinvestasi di pasar modal sebanding dengan keyakinan mereka pada kemampuan mereka sendiri untuk menangani aset keuangan dengan sukses di pasar modal. intensi investasi di antara Generasi Z dipengaruhi oleh rasa self-efficacy mereka. Untuk menggunakan kerangka TPB, kita dapat mengkategorikan sikap kita terhadap suatu item. Sebuah komponen dari jenis kontrol atas perilaku seseorang dikenal sebagai selfefficacy, dan itu terdiri dari keyakinan seseorang dalam kapasitasnya sendiri untuk melakukan tugas-tugas tertentu. Temuan penelitian ini menguatkan temuan Irmiyani et al. (2022) dan Kurniawan (2021), yang menemukan bahwa self-efficacy secara signifikan memprediksi antusiasme investasi. 4. Pengaruh Pengetahuan Investasi, Risk Tolerance, dan Self Efficacy terhadap Intensi Investasi Pasar Modal Intensi investasi pasar modal terbukti secara positif dan signifikan dipengaruhi oleh Pengetahuan Investasi, Risk Tolerance, dan Self Efficacy Gen Z secara simultan. Pengujian hipotesis keempat menghasilkan nilai F hitung 47.872 dimana nilai tersebut lebih besar dari F tabel yaitu 2,65 nilai probabilitas t-hitung (0,000) < Level of Significant (0,05). Berdasarakan hasil nilai adjusted R square (R2) yang diperoleh bernilai 0.414 (41,4%).. Hal Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 159 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 ini menunjukkan bahwa pergeseran 41,4% dalam Intensi Investasi Pasar Modal dapat dikaitkan dengan pertemuan Pengetahuan Investasi, Risk Tolerance, dan Self Efficacy. Niat investasi pasar modal di kalangan generasi Z akan meningkat seiring dengan pengetahuan investasi, toleransi risiko, dan kepercayaan diri mereka. Keputusan investasi dipengaruhi oleh tingkat pengetahuan investasi seseorang, sejauh mana seseorang bersedia mengambil risiko, dan kualitas pengendalian internal dan eksternal. KESIMPULAN Berdasarkan rumusan masalah, tujuan penelitian, landasan teori, serta hasil dan pembahasan maka kesimpulan hasil penelitian ini adalah: 1) Keakraban seseorang dengan pasar modal syariah (ditunjukkan oleh X1) secara signifikan berkorelasi dengan kecenderungan seseorang untuk terlibat dalam pasar tersebut. 2) Faktor kedua, toleransi risiko (X2), berhubungan signifikan dengan rencana masa depan untuk berinvestasi di pasar modal syariah. 3) Ketiga, ukuran keyakinan pada kemampuan sendiri (X3) memprediksi kemungkinan investasi masa depan yang lebih besar di pasar modal syariah. Diharapkan peneliti selanjutnya dapat mengkaji lebih jauh gagasan penulis dalam penelitian ini dengan melakukan penelitian yang lebih luas dan menguji hipotesis dengan faktor-faktor lain yang terkait dengan aspirasi investasi di pasar modal syariah. DAFTAR PUSTAKA Ajija. (2011). Cara Cerdas Menguasai Eviews. Salemba Empat. Amhalmad, I., & Irianto, A. (2019). Pengaruh Pengetahuan Investasi dan Motivasi Investasi Terhadap Minat Berinvestasi Mahasiswa Pendidikan Ekonomi Fakultas Ekonomi Universitas Negeri Padang. Jurnal EcoGen, 2(4), 734–746. Amruddin, Priyanda, R., Agustina, T. S., Ariantini, N. S., Rusmayani, N. G. A. L., Aslindar, D. P., Ningsih, K. P., Wulandari, S., Putranto, P., Yuniati, I., Untari, I., Mujiani, S., & Wicaksono, D. (2022). Metodologi Penelitian Kuantitatif (ke-1). Pradina Pustaka. Darmawan, A., & Japar, J. (2019). Pengaruh Pengetahuan Investasi, Modal Minimal, Pelatihan Pasar Modal dan Motivasi Terhadap Minat Investasi di Pasar Modal (Studi Pada Mahasiswa Feb Universitas Muhammadiyah Purwokerto). Jurnal Majalah Neraca, 15(1), 1–13. Darmawan, A., Kurnia, K., & Rejeki, S. (2019). Pengetahuan Investasi, Motivasi Investasi, Literasi Keuangan dan Lingkungan Keluarga Pengaruhnya Terhadap Minat Investasi di Pasar Modal. Jurnal Ilmiah Akuntansi Dan Keuangan, 8(2), 44–56. Firdaus, R. A., & Ifrochah, N. (2022). Pengaruh Pengetahuan Investasi Dan Motivasi Investasi Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 160 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 Terhadap Minat Investasi Mahasiswa Politeknik Keuangan Negara Stan Di Pasar Modal. Jurnal Keuangan Negara Dan Kebijakan Publik, 2(1), 16–28. Fitriasuari, & Simanjuntak, R. M. A. (2022). Pengaruh pengetahuan investasi, manfaat motivasi, dan modal minimal investasi terhadap keputusan investasi di Pasar Modal. Riset & Jurnal Akuntansi, 6(4), 3333–3343. Florencia, I., & Arifin, A. Z. (2022). Pengaruh Financial Knowledge, Financial Self-Efficacy, dan Risk Perception Terhadap Investment Intention di Pasar Saham. Jurnal Manajerial Dan Kewirausahaan, 4(2), 378–388. Frans, S. E. H. (2020). Pengaruh Ekspektasi Pengembalian , Toleransi Risiko , Dan Efikasi Diri Terhadap Minat Investasi Saham. Jurnal Manajerial Dan Kewirausahaan, 2(1), 22– 32. Ghozali. (2017). Aplikasi Analisis Multivariate dengan Program SPSS (1st ed.). Badan Penerbit UNDIP. Halim, A. (2005). Analisis Investasi (2nd ed.). PT. Salemba Empat Patria. Hidayat, L., Muktiadji, N., & Supriadi, Y. (2019). Pengaruh Pengetahuan Investasi Terhadap Minat Mahasiswa Berinvestasi di Galeri Investasi Perguruan Tinggi. 3(2), 63–70. Irmiyani, N. W. D., Rusadi, N. W. P., Premayanti, K. A., & Pradana, P. A. (2022). Motivasi, Pengetahuan Investasi, Self-Efficacy dan Minat Investasi selama Pandemi Covid-19. Jurnal Akuntansi, 32(10), 3176–3196. https://doi.org/10.24843/EJA.2022.v32.i10.p20 Ismail, S., Faique, F. A., Bakri, M. H., Teknikal, U., Zain, Z. M., & Melaka, U. (2017). The Role of Financial Self - Efficacy Scale in Predicting Financial Behavior. American Scientific Publishers, 1–5. Kurniawan, P. I. (2021). Effect of Expected Return, Self Efficacy, and Perceived Risk on Investment Intention: An Empirical Study on Accounting Master Degree in Udayana, Bali. Journal of Accounting, Finance and Auditing Studies, 7(1), 40–56. https://doi.org/10.32602/jafas.2021.002 Laurency, J. S., & Arifin, A. Z. (2022). Pengaruh Financial Attitude, Financial Self-Efficacy, Dan Risk Perception Terhadap Investment Intention. Jurnal Manajerial Dan Kewirausahaan, 4(3), 617–626. Lioera, G., Susanto, Y. K., & Supriatna, D. (2022). Faktor-Faktor yang Mempengaruhi Minat Investasi Mahasiswa di Pasar Modal. Jurnal Media Bisnis, 14(2), 179–188. Lubis, P. K. D. (2019). Influence of Knowledge Investment and Investment Motivation Against the Interest of Investing in the Stock Market on Economic Education Status of Students of State University of Medan. Journal of Physics: Conference Series, 1387, 1–5. https://doi.org/10.1088/1742-6596/1387/1/012064 Maulana, A. M., & Kaukab, M. E. (2020). Apa Yang Mempengaruhi Mahasiswa Berminat Investasi Saham? Jurnal Neraca, 16(1), 1–13. Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 161 Islamic Business and Finance (IBF) Vol 4, No.2, Oktober 2023 Nabilah, F., & Hartutik. (2020). Pengaruh Pengetahuan, Religiusitas dan Motivasi Investasi Terhadap Minat Berinvestasi Pasar Modal Syariah Pada Komunitas Investor Saham Pemula. Jurnal Taraadin, 1(1), 55–67. Ortega, S. L. (2023). Pengaruh Literasi Keuangan , Kemajuan Teknologi , Investasi di Pasar Modal. Jurnal Ilmiah Bidang Sosial, Ekonomi, Budaya, Teknologi, Dan Pendidikan, 2(2), 709–726. https://doi.org/https://doi.org/10.54443/sibatik.v2i2.648 Samsuri, A., Ismiyanti, F., & Narsa, I. M. (2019). Effects of Risk Tolerance and Financial Literacy to Investment Intentions. International Journal of Innovation, 10(9), 40–54. Sobaih, A. E. E. and I. A. E. (2023). Risk-Taking , Financial Knowledge , and Risky Investment Intention : Expanding Theory of Planned Behavior Using a ModeratingMediating Model. Journal Mathematics, 11, 1–17. Sugiyono. (2017). Metode Penelitian Kuantitatif, Kualitatif, dan R&D. CV. Alfabeta. Suryabrata, S. (2006). Metodologi Penelitian. PT. Raja Grafindo Persada. Syaputra, A. (2022). Pengaruh Pengetahuan Investasi di Pasar Modal Terhadap Minat Berinvestasi Mahasiswa. JOSR: Journal of Social Research, 1(3), 163–168. Trisnatio, Y. A. (2017). Pengaruh Ekspektasi Return, Persepsi Terhadap Risiko, dan Self Efficacy Terhadap Minat Investasi Saham Mahasiswa Fakultas Ekonomi Universitas Negeri Yogyakarta. Skripsi Akuntansi, 16–203. Utsman, M. A. (2021). Analisis Pengaruh Risk Tolerance , Financial Attitude , Dan Literasi Keuangan Syariah Terhadap Keputusan Investasi Saham Syariah ( Studi Kasus : Investor Millenial ). Jurnal Ilmu Ekonomi, 9(2), 1–15. Wardani, D. K. (2020). Pengaruh Sosialisasi Pasar Modal dan Persepsi atas Risiko terhadap Minat Investasi Mahasiswa di Pasar Modal. Jurnal Akuntansi, 12(1), 13–22. Wibowo, A., & Purwohandoko. (2019). Pengaruh Pengetahuan Investasi, Kebijakan Modal Minimal Investasi, dan Pelatihan Pasar Modal Terhadap Minat Investasi. Jurnal Ilmu Manajemen, 7(1), 192–201. Pengaruh Pengetahuan Investasi ……………………. Reza Fitriana Rizkia, dkk 162
https://openalex.org/W2346775280	https://zenodo.org/record/1517638/files/article.pdf	English	null	Notes on the Morphology of Ophioglossum (Cheiroglossa) palmatum, L	Annals of botany	1,911	public-domain	13,322	Notes on the Morphology of Ophioglossum (Cheiroglossa) palmatum, L. BY F. O. BOWER, F.R.S. Regius Professor of Botany in the University oj Glas BY F. O. BOWER, F.R.S. Regius Professor of Botany in the University oj Glasgow. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Annals of Botany, Vol. XXV. No. XCVm. April, With Plates XXII-XXIV. f^PH \~s f^PHIOGLOSSUM PALMATUM is the most outstanding type \~s among the remarkable family to which it belongs. At the same time it is one of the least fully investigated. Its habitat on rotting tree- trunks suggests a saprophytic tendency. Its swollen stock provides storage capacity beyond that of any other member of its genus. Its leafage, with its irregularly lobed sterile region and its varying number of fertile spikes, readily earns for it a position in a distinct section of the genus Ophioglossum, which has been styled § Cheiroglossa. Up to the present time little is known of its methods of propagation, while its gametophyte has never been observed. It was, then, with special interest that efforts were made to secure material of this peculiar plant on a recent visit to Jamaica. Certain decay- ing tree-trunks in the Blue Mountains on which it was known to have grown were visited. But the plant was only found in small quantity, and one plant ->lone was retained whole as a herbarium specimen. It is repre- sented in PI. XXII, Fig. i, A, kindly drawn for me by Dr. A. A. Lawson, who took the specimen himself in the living state. It shows a medium, well- grown plant, with the tuberous stock characteristic of the species, supported by numerous fleshy roots. The upper part of the stock is covered by a dense mat of pale-coloured hairs, from which emerge the two leaves. One of these is simple in form, and barren. The other is larger, and is divided into two divergent lobes, one of which shows a rudimentary third lobe. Upon the adaxial face are three fertile spikes, of which the lower two are seated as a pair side by side; the third is inserted rather higher up, and in a lateral position from the median line. A second larger specimen had a very broad leaf divided distally into three large lobes, and imperfectly sub- divided again into eight minor lobes, while six fertile spikes were borne on the adaxial face. From external observation the lowest of these is approxi- mately median; those which follow upwards are disposed right and left, Annals of Botany, Vol. XXV. No. XCVm. April, Annals of Botany, Vol. XXV. No. XCVm. With Plates XXII-XXIV. April, U Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Bower.—Notes on the Morphology of 278 but as regards external contours all of them are distinctly intramarginal. Their succession is not alternate, as the specimen represented in ' Studies', ii, Fig. i2, might lead us to expect: the lower spike is slightly to the left, the second and third to the right, the fourth and fifth to the left, and sixth to the right. This is not according to the rule for pinnae, for they usually alternate. The stock of this plant, which was the largest found, was pre- served for anatomical study. It thus appears that the leaves of the two fertile specimens collected in Jamaica fall in generally as regards their external characters with those previously compared.1 There is a reasonable, though not an exact numerical parallel between the lobes of the sterile blade and the fertile spikes. The insertion of the latter is distinctly within the margin of the leaf, as defined by the strongest curvature of the surface, but there is no regularity of alternation. In the example here figured there are two fertile spikes, right and left, seated as a pair on either side of the median line. This is like what is seen in the specimen shown as Fig. 118 in my ' Studies', ii, PI. VIII. It is a much less common condition than that where the lowest spike is solitary and median in position. Numerous other plants of smaller size were found, some of them very minute. By carefully rubbing down the decayed humus in which the plant was growing, Dr. Lawson was able to collect a considerable number of tubers too small to bear leaves. Some of them were attached to roots, and it is plain from a comparison of them that the plant is chiefly multiplied by buds formed on the roots of older plants, just as in other species of Ophio- glossutn. Fig. 1, B, shows such a young plant with tuberous stock, bearing its first leaf, and still attached to the parent root. ' Ann. of Bot., 1910, p. 10. p. 315. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 1 Studies, ii, pp. 27-33. 1 Ann. of Bot., 1904, p. 205. 1 Studies, ii, pp. 27-33. ' Ann. of Bot., 1910, p. 10. 1 Ann. of Bot., 1904, p. 205. * p. 315. ANATOMY OF THE STOCK AND THE LEAF-TRACK. The anatomy of the stock of O, palmatiim has never yet been de- scribed, so far as I am aware. Nevertheless Dr. Chrysler * remarks of it that ' as in Ophioderma, the vascular supply of the leaf arises as several strands', and refers to my paper on 0. simplex8 as his authority for the statement. But in that paper * it is specifically laid down that' the case is still open for § Cheiroglossa, in which I am not aware that the stock has been examined anatomically'. And again in the ' Origin of a Land Flora•', which Dr. Chrysler also quotes in his list of literature cited, it is stated on p. 463, after a description of the way in which the strands of the leaf-trace of Ophioderma are separately inserted upon the vascular system of the stock, that ' it is still uncertain whether or not § Cluiroglossa shares this character'. I am now able to supply some data of the anatomy of the Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Ophioglossum {Clieiroglossd) palmalum, L. 279 stock of this species, and they are in the main such as I had anticipated : their demonstration is, however, given for the first time, so far as my infor- mation goes. The large stem investigated was about \| in. in diameter, and a little over an inch in length, bearing one large matured leaf, while younger leaves constituted a terminal bud covered in by numerous hairs. Following the trace of the large leaf downwards, it was found to consist at the base of the petiole of about twenty-five strands arranged in a ring, with their xylems all directed inwards (Fig. 2, I). This is in accordance with what has already been described for this species.1 Passing downwards, they show frequent and irregular fusions, by which their number is diminished (Fig. 2, II), but they still constitute an unbroken series across the adaxial face. In that region there is evidence of repeated fusions, which bear at first no relation to any definite break in the series. Lower down, however, the strands gather into groups right and left of the median plane of the leaf (Fig. 2, III), and these condense by further fusion into a few, and finally into two rela- tively large strands (Fig. 2, IV), which diverge widely apart. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 ANATOMY OF THE STOCK AND THE LEAF-TRACK. Following these deeper into the axis, they attach themselves to the reticulum of the axial system, their phloem naturally fusing first with it (Fig. 2, V). At this level a vascular commissure traverses the space between the points of attach- ment of the two leaf-trace bundles, while from the centre of this usually springs the vascular supply to a root {rt, Fig. 2, V-IX). In sections again lower down the fusion of the two leaf-trace strands with the axial system is seen to be complete (Fig. 2, VIII, IX), but the commissure has faded out, showing the broad leaf-gap upon the divergent margins of which the two leaf-trace strands are inserted (Fig. 2, IX). The root {rt) which has been seen to be attached to the commissure above the leaf-gap is now free to pursue its course, which is usually downwards and obliquely outwards. It is thus seen that in Cheiroglossa also, as already shown for Ophio- derma, the diagnosis will apply ' petioli fasciculi numerosi separatim in rhizomae fasciculos inserti'. The number in Qieiroglossa is, however, the lowest plurality, viz. two. In this character these two sections of the genus Ophioglossum differ from all other Ophioglossaceae, in which the leaf-trace is gathered at its base into a single strand. In this they illustrate an anatomical condition which, on the basis of general comparison among the Filicales, is held to be derivative, as compared with the section of the genus designated Eu-Ophioglossum. The numerous successive leaves of a large stock are found to behave constantly like that above described. Upon this basis the vascular system of the stock is built up. The condition accordingly seen in the upper region of the stock is as shown in the photograph, Fig. 3. Numerous roots are seen on their oblique course outwards, and the section is strewn with 1 Ann. of Bot., xviii, p. 210. U 2 Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Bower.—Notes on tlie Morphology of 280 vascular strands of the leaf-traces, which at first sight appear indefinite in position. But comparison with the diagrammatized drawing in Fig. 4, in which the roots are omitted, shows that the scheme above described is adhered to. There is a central pith of moderate dimensions, round which are disposed three meristeles with their xylems directed inwards, and sepa- rated by wide foliar gaps. ANATOMY OF THE STOCK AND THE LEAF-TRACK. On either side of one of these gaps the two distinct strands of a leaf-trace (No. 1) are coming away, but not yet detached. In the case of No. 2 one of the strands is detached, but the other is still connected with the meristele of the axis. In leaf-trace No. 3 both strands are clear, while that on the anodic side is already branched. It may be noted that constantly the anodic branch takes precedence in time, and is also the larger of the two. Following on through the succes- sion of leaves—which incidentally it may be remarked show no exact regularity of divergence—the strands become wider and wider apart, owing to parenchymatous swelling, and break up into a number of strands, which behave after the manner already shown in Fig. 3. This being the condition in an old stock, it became a matter of interest to see what is the vascular arrangement in those young buds formed upon the roots, which in default of the seedling plants may be taken as giving evidence of the ontogenetic history. The roots correspond to those of other species of the genus, in growing with a single initial cell, and in showing frequent dichotomy. On these, close to the apex, the adventitious buds arise endogenously. The details, so far as they have been observed, correspond to what has already been demonstrated by Rostowzew * for O. vitlgatum. The bud soon swells into an almost spherical tuber, which shows at first little differentiation, but as it grows a deep indentation becomes apparent, lined with closely grouped hairs. At the base of this is the ptmctum vegetationis which gives origin to the young leaves. The first of these is shown emerging from the terminal tuft of hairs in Fig. 1, B. Sections through such tubers show at first little differentiation of the turgid parenchymatous mass, but as the leaf formation is advanced the vascular system develops, and in a case where several leaves were already initiated the condition was traced from below upwards, with the results shown in Fig. 5, I-VIII. The vascular system of the bud is inserted upon that of the root from which it sprang. From the first the stele of the bud was in this case distinctly medullated, though it does not appear to be always so. Fig. 1 Recherches sur V'Ophiogiosmm vulgatum, L., Copenhagen, 1891, p. 12 Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 ANATOMY OF THE STOCK AND THE LEAF-TRACK. 5, I, shows it at the level of its attachment to the vascular tissue of the root, the tracheides projecting right and left forming the junc- tion with it. Externally the stele is delimited by a definite endodermis ; the xylem is, however, only poorly developed, and the phloem is hardly recognizable as such. This condition is continued upwards (II), but shortly the continuity of the xylem-ring is broken by a gap, which resembles the Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Ophioglossum (Clieiroglossd) palmalum, L. 281 foliar gaps formed later (III), but in this case no strands of a leaf-trace were recognized as passing off. A little higher up a second leaf-gap appears on the opposite side of the stele, the xylem being thus separated into two equal bands, while from the middle of each of them a root-strand is given off (IV). It is at this level that the endodermis becomes indefinite, and it is not recognizable higher up. Nor is there at any point any indication of an internal endodermis. Shortly the first foliar strands are separated from the margins of the second gap (V) as two quite distinct strands, widely apart from one another in their origin, and showing further divergence as they pass outwards (VI, VII). Meanwhile the meristeles of the axis again extend, and close the leaf-gaps, so that the ring is reconstituted (VII). Other leaf- traces follow after the same plan, though not with any exact or constant angle of divergence. The two strands entering a leaf soon divide up according to the plan shown for the older leaves (Fig. a), but naturally the branching is here carried out to a less degree. A condition is found in the lower region of the old stock above described, which may be regarded as a natural consequence of the parenchymatous swelling of it for storage purposes. As shown in Fig. 6, the pith becomes greatly distended, as compared with Fig. 3 ; the meristeles pass further apart, and the stelar condition becomes accordingly disturbed, and less readily intelligible. The large pith is traversed by a number of roots. Four of these are shown in Fig. 6, but as many as six have been observed. The roots are usually triarch, but sometimes diarch. They appear to pass into the pith from outside, traversing the foliar gaps. 1 Land Flora, Fig. 356 (5), after Rostowiew. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Bower.—Notes on tlie Morphology of Bower.—Notes on tlie Morphology of 282 that of 0. palmatum. The attachment of the roots is essentially the same. A difference of some importance is, however, seen in the leaf-trace. Whereas in Helminthostachys, Botrychinm, and Eu-Ophioglossum the trace consists of a single strand springing from the base of the foliar gap, here there are two strands which originate widely apart, and laterally from the edges of the foliar gap. The leaf-trace thus divided is found also in O. pendulum and simplex.1 It has already been pointed out that these species, all of which are of saprophytic tendency, may be regarded as specialized offshoots of the Ophioglossaceous stock. And just as in the Ferns the single strand is characteristic of those types which may be held to be more primitive, while the divided leaf-trace is probably a derivative condition, so also here the condition seen in Eu-Ophioglossum and others of the family is probably the primitive state. This conclusion should have its bearing on the compara- tive study of the fertile spikes in those representatives of the genus. It should dispose us to inquire whether the elaboration of the spikes which is seen among them is not also derivative as compared with the simpler condition of Eu-Ophioglossum. 1 Ann. of Bot., 1904, p. 205. ANATOMY OF THE STOCK AND THE LEAF-TRACK. Though this con- dition is unusual, if not actually unique among the Pteridophytes, it is similar in principle to what is seen in Lycopodium Selago and some other species, where the roots traverse the softer middle cortex. In the present case all the tissues except the vascular skeleton are soft, and entry into the pith is as easy as a course through the cortex, while access to the soil is gained through the equally soft basal region of the tuberous stock. An examination of the roots which traverse the humus in which the plant grows shows the presence of endotrophic mycorrhiza, though not in such a profuse development as the habit of the plant might lead one to expect. Comparing what is seen in the stock of 0. pahnatum with the type for Eu-Ophioglossum, the differences appear such as would be natural in a tuberous development of that type, having a saprophytic tendency. The general development of the vascular tissue is poor, both as regards quantity of xylem and of phloem, especially the latter. The stelar construction is, however, fundamentally on the same plan, with its basket-like disposition of meristeles round a bulky pith. If the normal system of O. indgatum 1 be imagined as abbreviated and distended the condition would not be unlike Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 1 Ann. of Bot., 1904, p. 205. * Compare Land Flora, pp. 447-8, Fig. 246. MORPHOLOGY OF THE FERTILE SPIKES IN O. palmaium. Hitherto the development of the spike of Ophioglossum paltnatum has never been described. Fortunately the terminal bud of the large plant under investigation provided two young leaves which were developing as fertile leaves, and gave the opportunity for observing early stages of the spike as seen in transverse sections of the blade (Figs. 7-10). This has a semilunar outline, with relatively blunt margins which are unequal, one being habitually more massive than the other. The cell cleavages at the margin are of the T-type, and are repeated often with some degree of regularity at the point of greatest convexity. The result is a characteristic cell-net such as is shown in Figs. 11, 12. But this structure is most clearly seen in sections of the lamina above the insertion of the spikes. Lower down, where the margin is more rounded, the T-division still holds, but not being repeated in rapid succession the structure appears less characteristic. It is naturally impossible to give any exact account of the segmenta- tions involved in the origin of the spike from two leaves only, and both of them cut in transverse section. But as the details are fairly well known for O. vulgatum, this species will serve as a basis for comparison of the less perfect data from O. palmatum? The spike arises as an outgrowth in which several cells appear to take part (Fig. 13). As the convexity increases, segmentation is seen which is compatible with the type described for O. vulgatum, having an initial cell with the shape of a four-sided pyramid. A comparison of the drawings, meagre though they are, indicates * Compare Land Flora, pp. 447-8, Fig. 246. Ophioglossum (Cheiroglossa) palmatum, L. 283 that the early stages are substantially similar to those described for the Adder's Tongue (Figs. 13, 14). g ( g ) As regards position upon the leaf, the lowest spike is as a rule median, though this is not without exception, as is seen in mature specimens (Fig. 7 also,' Land Flora1, Fig. 238, D). In the younger of the two fertile leaves observed the lowest spike was perfectly median (Fig. 14). Passing upwards, younger spikes in the first stage of initiation were seen right and left, but they were distinctly intramarginal as regards the external contours (Fig. 13). * See aUo Prantl, System der Ophioglosieen, PI. VII, VIII. 1 Bot Zeit., 1875, p. 269. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 MORPHOLOGY OF THE FERTILE SPIKES IN O. palmaium. It thus appears that there is an acropetal succession in the origin of these intramarginal spikes, and it is possible that the succession here seen might have been continued further had the development of the leaf been more advanced. In the older fertile leaf the number of spikes was greater. Figs. 7-10 represent sections successively at higher levels. The lowest (Fig. 7) shows a single median spike as before. Higher up indications of at least four other spikes are seen (Figs. 8, 9); they are disposed right and left of the median line, but all are distinctly intramarginal so far as the external form is concerned. One of these is represented on a larger scale in Fig. 15, and the spike appears seated at a considerable distance from the margin. It may be added that, following up the region of greatest curvature in this section into the upper region of the leaf, it is found to be continuous with that marginal segmentation already described (Figs. 11, 12). One further point is worthy of note: that in Fig. 8, on the left-hand side, the section, though it is in a transverse plane, traverses two spikes. They are both intramarginal, but one more deeply than the other. In Fig. 16 three such spikes are traversed in the single transverse section. Here the group in question lay near to the middle of the adaxial face. In point of origin it seems impossible to refer such developments, varying as they do in degree of intramarginal insertion, to the margin of the leaf. The plain fact is that, whatever the vascular connexions may be, the spikes of O. palmatum are in their prime origin intramarginal. The later phases of the discussion on the morphological nature of the fertile spikes in Ophioglossum palmatum have turned so largely upon the anatomical facts that, while not neglecting other data, it is necessary at the start to be clear as to the vascular connexions. It is common knowledge that the leaf-trace in HelmintJiostachys, Botrychium, and Eu-Ophioglossum leaves the stele as a single strand. Already Holle1 had shown that while in Botrychium that strand divides into two equal parts, in Eu-Ophioglossum there is a central strand which maintains its identity between the lateral strands which spring from it, even up to the apex of the blade.9 The lateral strands vary in number in the different species. Bower.—Notes on the Morphology of 284 Observations have from time to time been made on the vascular supply running from the petiole to the fertile spike in Ophioglossum. Prantl showed in O. lusitanicum how two strands come off laterally from the marginal strands of the petiole, and unite to form the single strand of the spike.1 A somewhat similar state, though simpler still, was traced in O. Bcrgianum? where branches coming off laterally from the single petiolar strand fuse to form the supply of the spike. In such cases the vascular system of the spike is evidently a secondary derivative of that of the petiole. In larger species of Ophioglossum the plan is the same, only with more profuse branching. But no sufficient series of drawings has been published giving the whole story of origin of the supply to the spike, though it has been suggested for various species by several authors. Accordingly, the changes have been depicted as seen in O. reticulatum (Fig. 32, I-X), from which it is clear that the five strands of the spike seen in X arise by successive branchings and fusions from the marginal strands of the petiole, while these are themselves derived by lateral branching from the original leaf-trace strand. This is set down explicitly to meet the statement of Campbell, 'that the bundles which supply the spike are not secondarily given off from the main bundles of the petiole, but are themselves the adaxial bundles which can be traced from the base of the petiole into the spike.'3 I can only read the origin of the vascular supply of the spike in the species quoted as secondary, indicating that the spike is a subsidiary part in its relation to the whole leaf. Passing from the relatively stable type of Eu-Ophioglossum with its single spike to those of Ophioderma and Clieiroglossa with less stability of the fertile tract, it is to be remembered that it has been shown that the leaf-lrace is in them a divided one, a condition which comparison with the Ferns indicates to be probably later and derivative. This suggests that the instability of the spike seen in them is also derivative, a view which had already been developed before the anatomical facts came to hand. It has led to the seriation of the various forms already known for O.palmatum as showing a progression of complexity from the type of the single spike. MORPHOLOGY OF THE FERTILE SPIKES IN O. palmaium. ed from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 D user Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 , , p , , ' Ann. of Bot, 1904, PI. XII, Flgi. 3-5. 1 American Naturalist, vol. xli, No. 483, p. 157. • See Studies, ii, p. 38, &c, and Land Flora, p. 435, &c. 1 loc. tit., PI. VII, Fig. 1. • loc. cit, p. 154, Fig. 13. , , p , , , p , ' Ann. of Bot, 1904, PI. XII, Flgi. 3-5. • loc. cit, p. 154, Fig. 13. 1 American Naturalist, vol. xli, No. 483, p. 157. 1 loc. tit., PI. VII, Fig. 1. • Studies, ii, p. 68, and Land Flora, p. 463. 1 Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 1 loc. tit., PI. VII, Fig. 1. 1 loc. tit., PI. VII, Fig. 1. • Studies, ii, p. 68, and Land 1 1 loc. tit., PI. VII, Fig. 1. • Studies, ii, p. 68, and Land Flora, p. 463. 1 American Naturalist, vol. xli, No. 483, p. 157. , , , p • See Studies, ii, p. 38, &c, and Land Flora, p. 435, &c. Bower.—Notes on the Morphology of The occasional branchings of the spike seen in O. pettdulum are held to be incipient developments in a similar direction.4 It will now be seen how the facts of vascular supply will relate these aberrant forms with the type of Eu-Ophioglossum. It .has already been shown8 how the strands of the fertile leaf of O. pendulum form at the base a semicircle open on the adaxial side (Fig. 3), which, however, closes higher up (Fig. 4), and then becomes flattened as the leaf expands (Fig. 5). This is in accord with Campbell's drawings.6 But Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Opkioglossum {C/ieiroglossa) palniatum, L. 285 it is necessary to be clear how complete the fusion of the marginal series is on the adaxial side. This is shown in Fig. 21, a, b, from which it appears that all possibility of discriminating between the confluent marginal strands is lost. The supply to the spike is afforded by the median adaxial region of the flattened circle. The same is the case with O. intermedium,1 but here the actual coalescence of the marginal strands appears to be delayed till the base of the spike is reached, while the proportion of the strands of the fertile to those of the sterile region matches the relative size of those parts. It may be concluded from my own drawings of O. simplex that the condition there is in principle the same.2 Hence it is seen that the identity of the margins as marked by the vascular strands is entirely obliterated, while the vascular supply to the spike in Ophioderma comes off from the indeterminate vascular supply on the adaxial side of the petiole. Turning now to Cfieiroglossa, in the leaf-stalks of the weaker and first-formed sterile leaves the margins as defined by the vascular strands are widely apart, which is naturally in accord with their flattened form. But as the plant strengthens and works up to the production of propagative organs the leaf-stalks become more nearly cylindrical, while the marginal strands come nearer together, till in the fertile leaf they constitute a com- plete circle (Fig. 2,I-III). Frequent fusion of strands is seen on the adaxial side, so that the identification of the margins by means of the strands is quite impossible. 1 Campbell, loc cit., p. 153, Fig. 17. ' loc cit, PI. XII, Figs. 14-23. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 ' loc. clt., Figs. 34-9. 1 Campbell, loc cit., p. 153, Fig. 17. ' loc. clt., Figs. 34-9. ' loc cit, PI. XII, Figs. 14-23. Bower.—Notes on the Morphology of This obliteration of the margins takes place close to the leaf-base, and it is far above the point where it occurs that the fertile spike or spikes arise. The source of their vascular supply has been in- dicated for a case where there were three spikes, the lowest of them mediant But fresh drawings have been made from the old sections, giving the details more satisfactorily (Fig. 17,1-IX). From these it is seen that the supply to the lowest spike, which was median, comes off from the adaxial region of the circle of strands, where the marginal characters had been completely obliterated by the repeated fusions during their course through the elongated cylindrical petiole (I, II). By the passage of the median strands outwards the remainder of the strands appear ranged in a semi- circle (III). Fig. 17, III-VII, shows successive steps in the separation of the vascular supply for the two higher spikes, which are borne right and left. The origin of the strands is by segregation from the margins of the semi- circle ; it may, however, be noted that their derivation is not by any uniform scheme, though the number ultimately arrived at is in each case three. The sections show further that at the level of separation of the spikes their position on the leaf was intramarginal (VII-IX). It is worthy of note as having some bearing on their morphological character that the number of Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Boiuer.—Notes on the Morpliology of 286 strands traversing the stalk in all the three spikes is equivalent, the number being three, of which the median strand is the largest. As the knowledge of the vascular supply to the spikes in 0. palntatum is still very limited, being hitherto restricted to the above observations together with a very imperfect representation of similar facts by Bertrand,1 the portion of the large dried specimen described above, bearing the two highest spikes, was cut out (Fig. 18). After swelling with caustic potash it was halved, embedded, and cut into sections. These provided the essen- tial points in the origin of their vascular supply. Fig. 19 shows successive sections from below the spike to the left in Fig. x8. It will be seen that the marginal strand from the leaf passes directly into the stalk of the spike without any complications. Bower.—Notes on the Morphology of The next intramarginal strand does the same, but just before the separation of the stalk from the leaf a vascular com- missure runs across, almost transversely connecting it with the third strand from the margin of the leaf (IV). On complete separation of the stalk, a third strand, but weak and imperfectly developed, is seen in the section of the stalk (V), occupying the place of the lateral strand as seen in Fig. 17 (VIII-IX). Evidently the vascular system of this spike, which is small, is incompletely developed as compared with the larger examples. The still smaller spike on the right-hand side, which was the highest borne on this leaf, showed the vascular connexions seen in Fig. ao (I-VII). Here the small marginal strand of the leaf first fuses with the rather stronger strand next adjoining it (I—III). This then passes out into the base of the spike (TV), and subsequently separates again into two (V) ; then a vascular com- missure, as in the previous example, passes from the next inner strand almost horizontally outwards to join it (VI). The stalk then separates from the leaf (VII), its vascular supply being represented in its further course up the spike by a single band-shaped strand, which shows signs of being com- posed of three strands as in other cases. It may be held to be a reduced example of the same type of structure. A careful comparison of the five examples thus described shows that in no two of them are the vascular connexions at the base of the stalk exactly the same, though the structure of the stalk is essentially alike in them all, allowance being made for their varying size. These facts indicate two things. First, that the spikes are probably all morphologically equivalent parts. Secondly, that their attachment is not according to strict rule, a condition that readily accords with the interpretation put upon them in my ' Studies '.2 They were there held to be referable in origin to interpolation or chorisis of a single spike, in which case their position would not neces- sarily be regular. It would then be natural to expect less regular vascular connexions than in the case of parts produced in the normal sequence. ' Travaux et M^moires de l'Univenlte' de Lille, t. x, Me'm. 39, Fig. 97. * it, pp. 43, 44. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Bower.—Notes on the Morphology of A comparison of the specimens here described with those figured in ' T t M^ i d l'U i lt ' d Lill t M ' 39 Fi 97 * it 43 44 Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Ophioglossum (CJieiroglossd) pahnatum, L. 28 ' Studies ', ii, Figs. 120-129, indicates that greater variations still may occur in the vascular connexions of the spikes than those above noted. • In order to check my former statements and verify the drawings, a re-examination has been made of the specimens at Kew. This confirms the accuracy of the description and drawing already given as Fig. 120,.in point of the insertion of the median branched spike which it shows; its vascular connexion also is clearly intramarginal relatively to that of the lower-lying, but more marginal spikes. The specimen of Miers from S. Brazil, represented in my Fig. 121, was also re-examined. I state specifically that my former representation of it is substantially correct. The spike fourth from above is inserted in an almost median position, and its vascular connexion, which can be quite easily followed since the specimen is young and transparent, runs internally (i. e. more near to the median line), as compared with those of the next two spikes on the right, which are more marginal in insertion. To show that these are not isolated cases, it may be stated that there is a second specimen at Kew, also collected by Miers from S. Brazil, very like that of Fig. 121, bearing two spikes in approximately a median position: these are the fifth and sixth counting from below, and they are inserted between those which are lateral, and they have their vascular connexions visibly joining a strong strand which runs up the median region of the leaf. This strand lies clearly in a more median position than the vascular connexion of the next lower lateral pinna. Another specimen, collected by L'Herminier, but without locality, shows on a leaf bearing ten spikes a median insertion of the spike sixth from the base ; those directly below it are inserted right and left of it, one of them being branched. At almost the same level as this median spike are other lateral spikes, while the vascular connexions, which are readily recognized in the dried condition of this specimen, pass internally to those of the next lower marginal spike. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 1 loc. cit, Flgi. 130-9. 3 Ann. of Bot, 1904, PI. XV. • loc cit, p. 43 ; s«e Uo Land Flora, p. 439. Studies, ii, Figi. 133, 134. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 1 loc. cit, Flgi. 130-9. 3 Ann. of Bot, 1904, PI. XV. * Studies, ii, Figi. 133, 134. Bower.—Notes on the Morphology of This is again a clear case of intra- marginal insertion, with the vascular connexion also intramarginal relatively to that of a lower spike. It is to be noted that these examples of pro- nounced intramarginal insertion are all seen in cases where the spikes are numerous. It is precisely in such cases that the clearest departure from the marginal position would be probable if the origin of the plurality of spikes were by some process of interpolation or chorisis. 1 Dr. Chrysler, in his paper,1 makes certain deprecatory remarks on ' inspection of the external surface'. This is at least better than no examina- tion at all. It is to be borne in mind by those who criticize, that specimens in the Herbaria of Kew and the British Museum cannot be cut up at will. Nevertheless, evidence from examination of them without cutting them into sections, though neither final nor so convincing as this would be, has its value. It cannot be summarily dismissed, however great may be the 1 loc. cit., p. 11. Bower.—Notes on tfu Morphology of 288 preference for evidence by sections. But Dr. Chrysler has not only ignored the evidence afforded by my drawings,1 he has also neglected the theory which was based upon them. This theory is not even mentioned by him. But on the basis of observations on other genera, combined with the citation of my own observations on a certain specimen of 0, palmatum above quoted,2 he concludes that ' according to the reasoning here employed it may be inferred that the lowest spike in this specimen represents two fused lobes of the leaf, while the next two spikes represent single lobes'. He continues thus : ' Bower lays much stress on the observation that the spikes do not generally arise from the margin of the leaf, yet the only transverse sections which he figures (6, p. 463) clearly show that in the case of the three spikes so represented, the origin of the vascular supply at any rate is truly marginal, i. e. derived from the free edges of a curved leaf-trace made up of a number of separate strands. Probably most morphologists would place more reliance on the disposition of the vascular skeleton, than on the super- ficial " flesh " which clothes the skeleton.' preference for evidence by sections. But Dr. 1 See note explaining thU term, p. 396. Bower.—Notes on the Morphology of Chrysler has not only ignored the evidence afforded by my drawings,1 he has also neglected the theory which was based upon them. This theory is not even mentioned by him. This rough and ready way of deciding a rather intricate morphological problem will hardly commend itself to morphologists who take other facts besides those of vascular anatomy into their view, least of all to those who have personally examined a large series of specimens of 0. palmatum. Most morphologists have little or no personal knowledge of the plant in question, and Dr. Chrysler does not bring any evidence that he has studied specimens of it himself. Under these circumstances it would only be reason- able to expect some degree of reticence in amending conclusions which are explicitly based upon examination of a large series of specimens. In explana- tion of the facts derived from such study in 1896, a theory of chorisis or duplication was advanced.3 But as this theory is not criticized, but passed over in silence by Dr. Chrysler, it seems necessary to state it afresh, and to recapitulate the reasoning upon which it is still held to be the correct interpretation of the facts for O. palmatum. It is found in such normal types of Eu-Ophioglossum as O. vulgatum that the spike is susceptible of occasional branching, especially when the conditions are hypertrophic4 In 0. pendulum the branching is more fre- quent, and sometimes involves the lower sterile stalk of the spike (Figs. 130-2). It was suggested that the condition which appears as an occa- sional hypertrophic abnormality in O. vulgatum became frequent in O. pendulum, and has become fixed as an almost constant condition in well- nourished plants of O. palmatum. The comparison was drawn with those Angiospermic flowers in which an increase in number of the parts from the probable primitive type is referable to interpolation or chorisis. Comparison may be made with the androecium of Hypericaceae, Malvaceae, or Rosaceae. 3 Ann. of Bot, 1904, PI. XV. * Studies, ii, Figi. 133, 134. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Ophioglossum (Cheiroglossa) palmatum, L. Ophioglossum (Cheiroglossa) palmatum, L. 289 Some indication of a like process is seen in the Sphenophyllales, while it is not going too far to trace a distant parallel in the spread of the son over the enlarged leaf surface in many phyla of Ferns. ' The following are the facts and considerations which make me think that we may find in duplica- tion in its widest sense, or pleiogeny} as it might better be termed, the correct interpretation of the peculiarities of O. pahnatum. Some indication of a like process is seen in the Sphenophyllales, while it is not going too far to trace a distant parallel in the spread of the son over the enlarged leaf surface in many phyla of Ferns. ' The following are the facts and considerations which make me think that we may find in duplica- tion in its widest sense, or pleiogeny} as it might better be termed, the correct interpretation of the peculiarities of O. pahnatum. I. Frequent branched conditions of the fertile spike occur in O. palma- tum. This is especially common in specimens where the number of spikes is large (' Studies,' ii, Figs. 130-9). The details in some of the spikes near the margin in O. palmatum are closely similar to the branchings observed in 0. pendulum, in cases where the single though branched spike is inserted in a median position (' Studies,' ii, Figs. 131,132). Thus the branching is not a feature of the median spike only, nor is the branching always a simple bifurcation, as Dr. Chrysler's interpretation would demand. II. A parallelism has been traced between the number of the lobes of the sterile leaf and the number of the spikes in O. palmatum, showing that the number of the fertile spikes bears a general relation to the nutritive leaf area. As the plant grows stronger the higher complexity in both regions is attained. There are signs of a like though less exact parallelism in 0. pendulum. These facts readily accord with a theory of pleiogeny. III. Certain specimens have been described, such as those shown in ' Studies', ii, Figs. 120,131, which are difficult to harmonize with any pinna- theory, though they fall in readily enough with a theory of pleiogeny. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Ophioglossum (Cheiroglossa) palmatum, L. They show certain of the higher-seated spikes in an approximately median position, nearer, in fact, to the centre than others which are almost at the same level but more nearly marginal in their insertion. The vascular con- nexions of these have not been examined by sections, but the specimens are often sufficiently transparent to give opportunity for tracing them, and from such evidence it seems highly improbable that they would turn out to be constantly marginal like the rest. III. Certain specimens have been described, such as those shown in ' Studies', ii, Figs. 120,131, which are difficult to harmonize with any pinna- theory, though they fall in readily enough with a theory of pleiogeny. They show certain of the higher-seated spikes in an approximately median position, nearer, in fact, to the centre than others which are almost at the same level but more nearly marginal in their insertion. The vascular con- nexions of these have not been examined by sections, but the specimens are often sufficiently transparent to give opportunity for tracing them, and from such evidence it seems highly improbable that they would turn out to be constantly marginal like the rest. IV. The superficial origin of the spikes upon the adaxial face of the leaf is now demonstrated developmentally, and even the later spikes of the leaf have, so far as observed, an intramarginal insertion (Figs. 8, 9, 13, 15), while numerous cases can be quoted where the spikes show no regular alternate arrangement as pinnae usually do. V. The identity of the margins of the leaf, so far as these are defined by the vascular strands, is entirely merged by the repeated fusions of the strands on the adaxial face of the elongated petiole. It is, therefore, a question how far the marginal strands above the first spike really repre- sent the original (phyletic) margins to which normal pinnae might be referred. VI. Dr. Chrysler, following and extending the observations of Bertrand and Cornaille, has shown that marginal vascular origin is not a necessary Bower.— Notes on the Morphology of 290 criterion of pinna-nature within this family, for the supply both to spikes and to pinnae may be intramarginal. It may then be doubted how far a mere marginal origin of the vascular supply can be used as a criterion of pinna-nature, as Dr. Chrysler does in the present case. Ophioglossum (Cheiroglossa) palmatum, L. It cannot be accepted as outweighing other considerations, such as arise from comparison based on a large series of specimens. VII. It has been shown that the vascular supply to the median spike in the leaf examined in 1904 consists, like that of the succeeding two spikes, of three strands, of which the median one is the largest. All the three spikes are in fact equivalent in this respect (Fig. 17,1-IX). Dr. Chrysler's suggestion on the basis of the vascular connexions alone is ' that the lowest spike in this specimen represents two fused lobes of the leaf, while the next two spikes represent single lobes'. The fact that the vascular supply to all of these is equivalent in all except the source, being in the one case median, in the others marginal, should carry some weight with vascular anatomists, raising a doubt of the validity of the conclusion. The vascular supply to two fused pinnae might be expected to be more complex than that to a single lobe. On the other hand, the structure actually observed is such as might be expected on a theory of pleiogeny. VIII. The species which show the peculiarities under discussion belong to the two sections of the genus generally admitted to be specialized. It is in such forms that one may most readily expect developments which will diverge from the usual type and follow a line of their own. Both species are partially-' prophytk, a condition often associated with unusual develop- ments, reducible with difficulty to the ordinary schemes of construction. In plants biologically so restricted as these half-saprophytic dwellers on decaying trunks and humus a large spore-output is almost a necessary condition of survival, or at least of spread. This utilitarian reason for amplification is easily intelligible, and the ready means appears to have been amplification and irregular lobing of the sterile lamina, and repetition of the fertile spike. Both ot these were probably special and phyletically late occurrences. There seems to be no reason to hold that they should necessarily be retrospective developments. Dr. Chrysler appears to assume that they were. I have suggested, on the other hand, that both the sterile and the fertile regions of the leaf have progressed along new lines of development, and that their parts are not necessarily reducible to terms of pinnae such as are seen in the pinnate members of the family. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 1 Ann. of Bot., xvi, 1901. • Bot. Zeit, 1904, and Flora, 1906. • Bot. Gar., Dec, 1905. 4 Ann. of Bot., 1910, p. fin. • Ann. of Bot., 1909, p. 691. Ophioglossum (Cheiroglossa) palmatum, L. The conclusion arrived at may be summed up as follows. The spike of Eu-Ophioglossum is regarded as the unit upon which further develop- ment has played. The facts, both developmental and anatomical, fall in readily with a theory of amplification of that unit, or pleiogeny, as explain- ing the complex forms observed in O. pendulum, and more pronouncedly in O. palmatum. The sterile lamina enlarged and formed irregular lobes, Ophioglosstun {Clieiroglossd) palmatum, L. 29 which are not held as pinnae in the usual sense of the word. The spikes underwent a parallel pleiogenetic amplification. Sometimes this amplifica- tion took the form of repetition of spikes like the original one, but seated at points apart: this is duplication or interpolation. Sometimes two spikes are seated near together, or upon a common stalk; in the latter case they may merely show distal branching. These are conditions indicative of chorisis. Other variants may also be found. But in all these amplifications the spike is the unit throughout, and the branchings are not to be interpreted in terms of pinnae as normally understood. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 MORPHOLOGY OF THE OPHIOGLOSSACEOUS SPIKE. This has paved the way for similar ed from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 D user Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Bower.—Notes on the Morphology of 292 comparisons in the anatomy of the Ophioglossaceae, and the analogy between the vascular condition of the Ophioglossaceous stock and that of Thamnopteris, for instance,1 is, to say the least, very suggestive. In the next place, the work of Bertrand and Cornaille a has shown that the origin of the vascular supply to the pinnae in certain relatively primitive Ferns was intramarginal on the abaxial side. This line of inquiry has been followed up by Chrysler,3 and applying it to the Ophioglossaceae, he has demonstrated that in Botrychium virginianum the intramarginal origin holds not only for the sterile pinnae, but also for the fertile spike. He concludes from this and other facts that the normal spike is the result of fusion of two basal pinnae. In fact, he upholds the old theory of Roeper, and maintains that the Ophioglossaceae are related to the Ferns, and especially to the Osmundaceae and Marattiaceae, springing from a primitive stock at a re- mote period. But in arriving at this conclusion, Dr. Chrysler does not, in his own observations, travel over ground outside that of the anatomy of the leaf, combined with a comparison of external form of certain abnormalities in B. obliquum. This is a rather hazardous course in such questions. While due weight must be accorded to his anatomical results, his conclusions must stand the test of reference to other sources of information before they can be considered as amounting to more than interesting suggestions.4 g g gg It must be admitted that since 1896 the trend of the evidence has been decidedly such as to strengthen the Filical alliance of the family. This follows in the first place from the comparison of the gametophyte, when due allowance is made for the similarity of underground habit which they share with the Lycopods having led to a probable parallel development. It comes out also in some degree (which recent work on the Osmundaceae and Botryopterideae has distinctly strengthened) in the anatomy of the stock. Lastly, the observations of Dr. Chrysler have contributed in marked degree to strengthen the comparison with Ferns on the basis of the anatomy of the leaf. MORPHOLOGY OF THE OPHIOGLOSSACEOUS SPIKE. The above discussion has been concerned with the more elaborate types of the spike-development in the genus Ophioglossum. It has not touched the question of the morphology of the spike itself, which has been held as the unit in the amplifications seen in Ophioderma and Cheiroglossa In my' Studies. II. Ophioglossaceae (1896)', the evidence up to that date was summarized and weighed with a view to tracing the probable affinity of the family, and incidentally the morphological character of the fertile spike which is its most notable feature. This required a careful balancing of evidence for and against alternative views. Of these the one indicated affinity with the Sphenophyllales and Lycopodiales, the spike being regarded as a result of amplification from the sporangiophore, or ultimately from the sporangium. The alternative indicated an affinity with the Filicales, in which case the morphology of the simple spike would be referable in some form or another to ^ ma or a coalescence of pinnae. The evidence up to 1896 ap- peared to me to favour the former alternative, and that conclusion was stated with ample illustration and discussion of the details on which it was based. Since then considerable advance has been made in various directions which bear upon the question. The researches of Lang] and of Bruchmann * have greatly increased the knowledge of the gametophyte and of the embryology. The discovery by Lyon3 of the suspensor in Botrychium obliquum (followed recently by the description of a like body in Hclminthcstachys by Lang4) at first sight appeared to strengthen the Lycopod affinity, but the demonstra- tion of a like organ in Danaea by Campbell8 put matters back in statu quo. A second line along which recent advances affect the question is the anatomy of some of the earlier types of the Filicales. Kidston and Gwynne- Vaughan, by comparison of fossils which are plainly of Osmundaceous alliance, whether or not they are Osmundaceae in the modern sense, have shown it to be highly probable that the structure seen in the modern Osmundaceous stem is in the main the result of an up-grade development with medullation from a ptotostele. , , , 4 The recognition of affinity of the Ophioglossales with certain early Filicales is no new con- ception. It was indicated in unmistakable terms by Renault in 1875 (Ann. Sci. Nat., Se>. 6, torn, i, PP- a33-4)i wnJle Prantl, in 1884, pointed specifically to the Botryopterideae and the Osmnndaceae as possible lines of relationship with the living Ophioglossaceae (Beitr. 1. Syst. d. Ophiogloueen, p. 345). See also Scott, Studies in Fossil Botany, ii, p. 640. 1 FOBSII Osmnndaceae, Part iii, PI. I. 1 Travaux et Mimoires de l'Unlveralti de Lille, x, M<£m. 39. MORPHOLOGY OF THE OPHIOGLOSSACEOUS SPIKE. All this indicates that a reconsideration is necessary of the Lycopod- Sphenophyll-alliance previously recognized for the Ophioglossaceae. But before coming to a definite conclusion there are three other points to be taken into our view, two of them favourable, the other adverse to the Fern alliance. In the first place the comparison has been made in previous dis- cussions with the Sphenophyllales (including Psilotaceae) in respect of the behaviour of the sporogenous cells. It was pointed out that in the Psilota- ceae a proportion of the spore-mother-cells became abortive before the tetrad Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Ophioglossum (Cheiroglossd) palmatum, L. 293 division, acting as a diffused tapetum. In the Memoir of 1896 ' the results of Rostowzew were accepted, without sufficiently critical examination of my own preparations, as showing that in Ophioglossum also cells scattered through the sporogenous group became disorganized without undergoing tetrad division. It has since been found that this is not so,2 and accordingly the comparison with the Psilotaceae on this feature falls away. Secondly, the comparison with the Psilotaceae-Sphenophyllaceae alli- ance was also based on the method of insertion of the sporangiophore on the adaxial face of the leaf. But it now appears that the similarity does not extend to the mode of origin of the vascular supply to the fertile regions in the plants compared. Dr. Chrysler has laid weight upon the marginal attachment of the vascular supply to the spike. The fact is now quite plain that, with certain possible exceptions in the specialized 0. palma- tum, the supply comes off in the Ophioglossaceae either from the margin of the petiolar system, or on the abaxial side from an intramarginal gap. It may be seen, on the other hand, in any series of sections of the sporophyll of TmesipUris that the marginal portions of the strand of the leaf-stalk pass off right and left into the leaf-lobes, while the central portion enters the synangium. The origin of the supply to the sporangiophores in Cheirostrobus is described as coming off from the middle strand that supplies the sterile region of the leaf, by a branching, so that one of its branches lies above and inside the other.8 That is, the origin is median and adaxial. A simi- lar branching appears to be the rule also for Sphenophyllum. 1 p. 20. » Compare Land Flora, p. 451, Fig. 351. » Scott, Phil. Trans., Series B, vol. clxxxtx, p. 12. * Land Flora, p. 414. X Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 MORPHOLOGY OF THE OPHIOGLOSSACEOUS SPIKE. These facts indicate a real difference between the marginal or slightly abaxial origin of the vascular supply to the spike in the Ophioglossales, and the median adaxial supply to the sporangiophores in the Sphenophyllales, which if it prove to be constant will strengthen the alliance of the former with the Filicales. It is of course a possible view, which might be based on the mere anatomical facts, that in Tmesipteris the synangium represents the whole terminal region of the leaf, which has remained fertile, as in Osmunda regalis. But against this are the facts of development; * moreover, that explanation would not fit for Splienophyllum or Cheirostrobus. For these reasons I prefer to accept the anatomical distinction as a real and a valid one. On the other hand, one of the most impressive features in the Filicales is the extraordinary constancy of the progression by which the sporangium suffers reduction in size and complexity of structure, in the thickness of its gradually elongating stalk, and in the numerical output of its spores as we pass from the more primitive to the more specialized forms, while the pre- cision of the mechanism for scattering them increases. This matter has been treated at length in the ' Land Flora' (pp. 637-46, &c). But if the X Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Bower.—Notes oh tlte Morphology of 294 Ophioglossales were derived from a primitive Fern stock, and the sequence be, as the anatomical comparison indicates, from a type like Botrychium virginianum with relatively numerous smaller sporangia through types such as B. ternatum and Lunaria with fewer and larger sporangia, to Ophioglossum itself, it is plain that we are contemplating a progression which is the direct converse of that illustrated by the Ferns at large. Such a progression is of course possible. The question is whether it is probable. The apparent improbability of it weighed with me strongly in previous discussions on the phyletic position of the Ophioglossales. There is, however, a further circumstance which makes this difficulty less serious than it at first appears. It lies in the fact that in Ophioglossum a single initial cell is found at the apex of its stem and root. The leaf-apex and the structure of its margin, on the other hand, approaches that of the Marattiaceae, and is in accord with the bulky sporangial structure. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 1 See Land Flora, p. 650. MORPHOLOGY OF THE OPHIOGLOSSACEOUS SPIKE. A com- parison may be drawn of the meristematic conditions seen in the various groups of Ferns.1 In the Marattiaceae all the segmenting parts are bulky and complex in their cleavages ; in fact the centre of construction lies deep. In the Osmundaceae a middle position is seen between the state of the Marattiaceae and that of the Leptosporangiate Ferns, and this applies here also to all the segmenting parts. In the Leptosporangiate Ferns themselves a more definite scheme of cleavage is found in all the meristems, which is in accord with the definite segmentations of the attenuated sporangium; in fact in them the centre of construction is more superficial. In all of these the various parts share the character consistently of a definite or a less definite reference to single initials. But the Ophioglossaceae show a dis- crepancy which does not find its match in any of the true Filicales. For the leaf and sporangium resemble the Marattiales in'their more complex cleavages, while the apices of stem and root have each a single initial as in the Leptosporangiate Ferns. This discrepancy might be referred either to a progressively more bulky modification of the leaf and sporangium in a plant derived from a type resembling the Osmundaceae, or to a conser- vatism by which the plant has retained the characteristics of some remote ancestry in its leaf and sporangium, while the axis and root have taken on the characters of later forms. In view of the peculiar biological conditions of the family, and especially of the tendency towards saprophytism in the genus Ophioglossum, I am inclined to the former alternative, and to look in the direction of the Osmundaceae and Botryopterideae for the nearest relatives of the Ophioglossaceae. There are various other comparative points which indicate collectively a relation, however distant, with early types of the Filicales rather than with the Sphenophyllales, such as the hairiness in the apical region, especially as seen in O. paliitatum, the conformation of the leaf-base and the vascular 1 See Land Flora, p. 650. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Ophioglossum (Clieiroglossd) palmatum, L. 295 structure in the family at large. But these must be left over at the moment. It is obvious that there are difficulties in assigning any definite place to the family at present. 1 loc cit, pp. 7-9 and PI. II. * Seward, Fossil Plant, il, p. 43.1. 5 Campbell, American Naturalist, xxxviii, 1904, p. 761. See Land Flora, p. 441. MORPHOLOGY OF THE OPHIOGLOSSACEOUS SPIKE. But sufficient has been said to show how clearly the balance of evidence is setting in the direction of an alliance of the Ophio- glossales with the early Filicales. g y A natural consequence of adopting a Filical alliance for the Ophioglos- saceae would be a recognition of the pinna-nature of the spike, probably in most cases with a coalescence of two pinnae as in the old theory of Roeper. I should, however, attach less weight in arriving at this conclusion to such abnormalities as those quoted by Dr. Chrysler than he appears to do.1 For branchings of very various character may be found, which all deserve equal consideration with those which support a hypothesis of fusion. The analogy with the Marsiliaceae also becomes obvious. But probably the condition there seen is an instance of parallel development, and arises along a different phyletic line. For the two families belong to distinct types of Fern- development. The Ophioglossaceae appear to be naturally referable to the Coenopterid2 type, characterized by less definite specialization of the sporangium. But the Marsiliaceae are pronouncedly Leptosporangiate, and, as Campbell has so convincingly shown, their relations appear to be in the direction of the Schizaeaceae.3 5 Campbell, American Naturalist, xxxviii, 1904, p. 761. * See Land Flora, p. 441 1 loc cit, pp. 7-9 and PI. II. 5 Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 SUMMARY. 1. The tuberous stock of Ophioglossum palmatum is traversed by a stele showing a bulky central pith and transversely widened leaf-gaps. It is of the same type as Eu-Ophioglossum. 1. The tuberous stock of Ophioglossum palmatum is traversed by a stele showing a bulky central pith and transversely widened leaf-gaps. It is of the same type as Eu-Ophioglossum. 1. The leaf-trace originates as two distinct strands inserted widely apart, right and left of the leaf-gap. 3. The spikes originate like that of Eu-Ophioglossum. The lowest is usually median, but not always. The insertion of the rest is usually intra- marginal, and shows no regular alternation. The spikes branch frequently and irregularly. 4. The facts indicate that the spike of Eu-Ophioglossum is the morpho- logical unit, and that the conditions seen in less degree in 0. pendulum, and more clearly in O. palmatum, are due to ' pleiogeny', that is an increase from that unit in various ways, such as by partial or completed branching, or by interpolation of accessory spikes. The numerous spikes of 0. palmatum are not then directly referable to normal pinnae, any more than are the irregular lobes of its sterile lamina. 5. Phyletically O. pendulum, intermedium, and simplex form a derivative series from Eu-Ophioglossum.* O. palmatum represents a parallel, but pro- bably a distinct line, which has carried amplification of the leaf further. 5. Phyletically O. pendulum, intermedium, and simplex form a derivative series from Eu-Ophioglossum.* O. palmatum represents a parallel, but pro- bably a distinct line, which has carried amplification of the leaf further. X 2 Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Bower.—Notes on the Morpliology of 296 The divided leaf-trace which they all show confirms the derivative character of both lines. 6. The balance of evidence acquired since 1896 has distinctly favoured an alliance of the Ophioglossaceae with the Filicales. It is in the direction of the Coenopterideae, and of living Ferns the Osmundaceae, that we may look for their true place. 6. The balance of evidence acquired since 1896 has distinctly favoured an alliance of the Ophioglossaceae with the Filicales. It is in the direction of the Coenopterideae, and of living Ferns the Osmundaceae, that we may look for their true place. 7. SUMMARY. The .normal spike of the Ophioglossaceae will, in relation to this comparison, and especially in accordance with the anatomical facts, be held to be ultimately of pinna-nature ; perhaps in most cases a result of pinna- fusion, according to the theory of Roeper. But this unit is subject to repetition in Ophioglossum palmatum. p p g p 8. The vascular supply to the Ophioglossaceous spike being normally marginal, or from an abaxial pinna-gap, it differs from that to the spore- producing organ in the Psilotaceae and Sphenophyllaceae, where it comes off from the adaxial face of the foliar strand, or is the middle region of it. This difference, if it be found to be constant, may provide a real and valid anatomical distinction. p 8. The vascular supply to the Ophioglossaceous spike being normally marginal, or from an abaxial pinna-gap, it differs from that to the spore- producing organ in the Psilotaceae and Sphenophyllaceae, where it comes off from the adaxial face of the foliar strand, or is the middle region of it. This difference, if it be found to be constant, may provide a real and valid anatomical distinction. 1 English Edition, pp. 538-43. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 NOTE. I know only too well the undesirableness of introducing new terms into the vocabulary of the science. But any one who has read attentively those interesting pages of Professor Goebel's ' Organography' in which he deals with the phenomena of increase and decrease in number of parts in the flower 1 must have felt the want of some expression which, without connoting any detailed view as to the method, shall still convey the conception of increase or decrease in number of the parts from that which is regarded as normal or typical. The old terms 'chorisis', 'de'doublement', 'fission', or 'split- ting ', all have had special applications, and convey meanings relating to the method or manner of the increase. In the use of them there is a danger of conveying more than the simple conception of increase. ' Negative chori- sis ' and ' ablast' suffer under the same disability. The readiest way of avoiding such difficulties is by the introduction of some new term which carries no preconceptions, and while stating the fact does not imply any method. I venture therefore to suggest that the term ' pleiogeny ' should be used to connote those phenomena which involve increase in number of parts beyond the normal or typical. The term ' meiogeny ' would similarly connote a decrease from the normal or typical. It is not only in the case of floral structure that such terms may apply. In the present case it would be a phenomenon of ' pleiogeny' that we have been studying in Ophioglossum palmatum. And similarly the increase of sori so frequently traceable in Ferns would be a phenomenon of the same class. Having recognized this, the next step will be by comparison and 1 English Edition, pp. 538-43. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Ophioglossum (C/ieiroglossa) palmatum, L. 297 analysis to decide the method by which it is brought about. It will then appear that more than one method may be involved in any given example. For instance, where one primordium is already initiated, and the activity of tissue-formation becomes localized at more than one centre in relation to it, thus producing more than one part in place of the original one, the method of the pleiogeny would be chorisis. Illustrating Prof. Bower's paper on Ophioglossum palmatum. Fig. 1. A, Fertile plant of Ophioglossum palmatum represented complete, about three-quarters ot the natural size, from a drawing by Dr. A. A. Lawson. B, A bud inserted upon its parent root, showing the tuberous stock (t), surmounted by a dense tuft of hairs (A), and a single yonng leaf, x »J. Fig. 3, I-IX. Successive transverse sections of a leaf-base, following the leaf-trace strandi downwards to their insertion on the stele. I and II show the very numerous small strands, forming a more or less circular series, with frequent signs of fusion across the adaxial face of the petiole. Ill and IV show stages of segregation of the strands into two lateral groups, which in V and VI have united into single strands. V-X show the successive stages of insertion ot these, right and left, at points widely apart upon the margins of a leaf-gap. In V-VII a commissure of the vascular network i» seen just above the leaf-gap, and from the middle of It arises a root (rf), which takes its course obliquely outwards (VII-1X). The leaf-trace strands do not fuse completely with the meristeles of the axi3 till the level of the gap is reached (VIII, IX). x about 6. Fig. 3- Photograph of a transverse section through the old stock, about half-way down its length, showing centrally the pith surrounded by three meristeles. From the edges of two of these a leaf-trace is coming off, as in Fig. 2, VIII. Externally numerous leaf-traces in various stages of advance are traversing the bulky cortex, and several roots are also seen, x 4. Fig. 3- Photograph of a transverse section through the old stock, about half-way down its length, showing centrally the pith surrounded by three meristeles. From the edges of two of these a leaf-trace is coming off, as in Fig. 2, VIII. Externally numerous leaf-traces in various stages of advance are traversing the bulky cortex, and several roots are also seen, x 4. Fig. 4. A slightly diagrammatized key to the vascular system shown in Fig. 3. The roots are omitted for clearness, while the several foliar traces are connected by dotted lines and numbered. It will be seen that they conform to the stages shown in Fig. 1; also that the divergence between the successive leaves is not a constant one. Tig. 5, I-VII. Successive transverse sections of the stock of a young adventitious bud. NOTE. But if two centres of activity originate apart from one another where usually only one exists, or if an additional one appeared in a position usually untenanted, that would be interpolation. But both would fall under the general expression of pleiogeny. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 DESCRIPTION OF FIGURES IN PLATES XXII-XXIV. Illustrating Prof. Bower's paper on Ophioglossum palmatum. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Boiuer.—Ophioglossum (Cheiroglossa) palmalwn, L. Figs. 7-10. Photographs of successive transverse sections thiongh a young fertile leaf. Fig. 7 is at the level of insertion of the lowest spike, which is seen to be median. Fig. 8 shows four spikes in two groups, right and left of the median plane. Fig. 9 shows a higher section traversing two spikes only, which are clearly into amarginal. Fig. 10 is a transverse section of the leaf above the spikes, showing projecting margins similar to those already seen in Fig. 9. x 48. p g p j g g y g Figs. 11, 13. Detail of segmentation of the marginal region of a fertile leaf, showing the T-type of division, but not with constant results, x 125. Fig. 13. A very young spike, as seen in a transverse section of the parent leaf, from which it arises in an intramarginal position. It results from the outgrowth of a number of cells, x 135. Fig. 13. A very young spike, as seen in a transverse section of the parent leaf, from which i arises in an intramarginal position. It results from the outgrowth of a number of cells, x 135. Fig. 13. A very young spike, as seen in a transverse section of the parent leaf, from which it arises in an intramarginal position. It results from the outgrowth of a number of cells, x 135. g p g Fig. 14. A spike from the same leaf, but lower down, and inserted in a median position. It shows a rather more advanced state, thus indicating that there is an acropetal succession in the formation of the spikes, x 135. g p g Fig. 14. A spike from the same leaf, but lower down, and inserted in a median position. It shows a rather more advanced state, thus indicating that there is an acropetal succession in the formation of the spikes, x 135. g p g Fig. 14. A spike from the same leaf, but lower down, and inserted in a median position. It shows a rather more advanced state, thus indicating that there is an acropetal succession in the formation of the spikes, x 135. p Fig. 15. Transverse section from an older leaf, showing a young spike in a distinctly intra margioal position, x 135. g p Fig. 16. Transverse section of the middle region of a fertile leaf, cutting obliquely through the insertion of three spikes. Fig. 32, I-X. Successive transverse sections through the petiole of 0. reticulatum, showing gin of the vascular supply to the fertile spike. Section I is the lowest. Drawn freehand. Illustrating Prof. Bower's paper on Ophioglossum palmatum. I shows the insertion on the vascular system of the parent root ; II, the vascular ring surrounding the central pith, which is present from the first In III the ring opens, but no leaf-trace was given off from the gap. In IV the ring has opened on the opposite side also, while each vascular meristele is giving off a root-trace. The first leaf-trace is shown coming off in V, a3 two separate strands from the edges of the second gap. In VI the strands are separated, but the first gap has closed again by a vascular commissure. In VII the second gap ha3 also closed, reconstituting the ring, while the strands of the trace are diverging widely. Drawn freehand, x about 50. Fig. 6. Photograph of a transverse section of the old stock near to its base, showing the pith greatly dilated, the meristeles separated, and stretc'ued transversely, while four roots are traversing the pith ; others are seen in the cortex, x 4. 298 Boiuer.—Ophioglossum (Cheiroglossa) palmalwn, L. Boiuer.—Ophioglossum (Cheiroglossa) palmalwn, L. Boiuer.—Ophioglossum (Cheiroglossa) palmalwn, L. This indicates the irregularity of their position while young, x 125. g p Fig. 16. Transverse section of the middle region of a fertile leaf, cutting obliquely through the insertion of three spikes. This indicates the irregularity of their position while young, x 125. Fig. 17, I-IX. A series of fresh drawings from the sections of the fertile leaf described in the ' Annals of Botany ', 1904, p. 310, PI. XV. Figs. I-III show the origin of the vascular supply to the lowest median spike. Figs. IV-IX show the same for the two lateral spikes. It will be noted that in either case the number of the Btranda is ultimately three. y Fig. 18. Portion of the leaf of the large dried specimen from Jamaica, bearing to the left the fifth spike, and to the right the sixth, the latter being the smaller. The transverse sections of their stalks are also shown ; the fifth is traversed by three strands, the weaker sixth only by one. x 3. ; y , y y Fig. 19, I-V. Successive transverse sections showing the origin of the vascular supply to the fifth spike. p Fig. 30, I-VII. Successive transverse sections showing the origin of the vascular supply to the sixth spike. p Fig. 30, I-VII. Successive transverse sections showing the origin of the vascular supply to the sixth spike. Fig. 30, I-VII. Successive transverse sections showing the origin of the vascular supply to the sixth spike. p Fig. 21 a, b. Transverse sections through the petiole of 0. pendulum, a, lower down, b, higher on the same part, showing how completely the margin as marked by vascular characters is obliterated. Drawn freehand. Fig. 21 a, b. Transverse sections through the petiole of 0. pendulum, a, lower down, b, higher on the same part, showing how completely the margin as marked by vascular characters is obliterated. Drawn freehand. obliterated. Drawn freehand. Fig. 32, I-X. Successive transverse sections through the petiole of 0. reticulatum, showing origin of the vascular supply to the fertile spike. Section I is the lowest. Drawn freehand. Fig. 32, I-X. Successive transverse sections through the petiole of 0. reticulatum, showing origin of the vascular supply to the fertile spike. Section I is the lowest. Drawn freehand. cfbrruds . PI mi. cfbrruds / h V \ r // I fi- ••; / ; AJL.L BOWER - OPHIOGLOSSUM. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 BOWER—OPHIOGLOSSUM. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 Boiuer.—Ophioglossum (Cheiroglossa) palmalwn, L. Hutiljtl rtun // i / / i / 2 /2/2 /1 806 •t/l 9 0 Fig. 3. •/ /•» I o Fig. 5. u MX. » ft Q !rt \ •t/l •t/l Fig. 3. 9 0 Fig. 3. •/ /•» I o Fig. 5. u MX. » ft Q !rt.\ •V BOWER—OPHIOGLOSSUM. ded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 AD user arch 2018 9 o Fig. 5. Fig. 3. 0 •/ /•» I Fig. 5. u MX. » ft Q MX. BOWER—OPHIOGLOSSUM. icle-abstract/os-25/2/277/158065 BOWER—OPHIOGLOSSUM. Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 »* . 8. Fig.13. »* . 8. Fig.13. • ••• ffrrt JI LcTTVnO ownloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 y INSEAD user n 07 March 2018 . 8. Fig.13. . 8. • ••• ffrrt JI LcTTVnOTl • Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user on 07 March 2018 ffrrt JI LcTTVnOTl • Downloaded from https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 by INSEAD user of Botany, Fig.17.d-n:) Fig.16 BOWER—OPHIOGLOSSUM. rom https://academic.oup.com/aob/article-abstract/os-25/2/277/158065 er 2018 Fig.17.d-n:) , Ti.XXN Fig. 21. . (T-X; [9 d»©<b ."0 Fig . (T-X; Fig. 21. Fig. 21. . (T-X; . (T-X [9 d»©<b ."0
https://openalex.org/W4220761686	https://zenodo.org/records/6367995/files/BOKS%20SPORT%20TURI99.pdf	Quechua	null	Boks sport turi	Zenodo (CERN European Organization for Nuclear Research)	2,022	cc-by	2,262	ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR Annotasiya: Ushbu maqolada Maktablarda sportning 7 turini rivojlantirish boʻyicha 7 muhim qadam, Boks tarixi: kelib chiqishi, muhim sanalari haqida yoritilgan. Аннотация: В этой статье рассматриваются 7 важных шагов в развитии 7 видов спорта в школах, история бокса: истоки, важные даты. Ключевые слова: Бокс, История, Школы, Спорт, Развитие, Олимпиада, Паралимпийские игры. Annotation: This article discusses 7 important steps in the development of 7 sports in schools, boxing history: origins, important dates. Annotation: This article discusses 7 important steps in the development of 7 sports in schools, boxing history: origins, important dates. Keywords: Boxing, History, Schools, Sport, Development, Olympics, Paralympic Games. Keywords: Boxing, History, Schools, Sport, Development, Olympics, Paralympic Games. Yurtimizning tinchlik va osoyishtaligini ko‘ra olmaydigan ichki va tashqi yovuz kuchlar borligi sir emas. Shu sababdan ham mamlakatda yosh-larni harbiy- vatanparvarlik ruhida tarbiyalash ishlariga jiddiy e’tibor berilib, eng avvalo mahallalar, o‘quv yurtlari, ishlab chiqarish korxonalari, muassa-salarda ta’lim- tarbiyaga ojd ishlarni kuchaytirishga katta e’tibor berilmoqda. KONFERENSIYA \| 1 ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR Harbiy-vatanparvarlik tushunchalari va unga doir amaliy tadbirlar, asosan quyidagi yo‘nalishlarda amalga oshirilmoqda: 1. Barcha turdagi o‘quv yurtlarida o‘smirlar bilan ularni harbiy xizmatga chaqirish gacha bo‘lgan harbiy tayyorgarlik mashg‘ulotlarini paxta o‘lish. 2. Harbiy tashkilotlarga ko‘maklashuvchi “Vatanparvar” tashkilotining faoliyat yuritishi. 2. Harbiy tashkilotlarga ko‘maklashuvchi “Vatanparvar” tashkilotining faoliyat yuritishi. 3. Tuman-shahar harbiy komissariatlari. 4. Mahallalar va oilalar. Yuqorida ta’kidlangan yo‘nalishlarda harbiy-jismoniy tayyorgarlik tadbirlari mazmunida jismoniy mashqlar, sport, sayohat (turizm-chamalab topish, belgilash va h.k.) va xalq milliy o‘yinlaridan maqsadli foydalanilmoqda. Ularning ba’zilariga qisqacha tavsif berish mumkin: Jismoniy tarbiya darslari. Umum ta’lim maktablarining yuqori sinflari va o‘rta maxsus hamda oliy o‘quv yurtlarida o‘tkaziladigan jis-moniy tarbiya darslarining mazmuni, asosan o‘quvchi-yoshlarni jismo-nan barkamol bo‘lish va chiniqishiga qaratilgan. Bunda gimnastika, yen-gil atletika, suvda suzish, kurash va boshqa sport turlarining element-laridan foydalaniladi. Chaqiriqqacha bo‘lgan harbiy mashg‘ulotlar darslarida maxsus texnik va taktik tayyorgarliklar mazmunida jismonan chiniqish ham o‘z o‘rniga ega. Bunda turnikda tortilish, osilib chiqish, aylanish, o‘rta va uzoq masofalarga yugurish, suvda suzish, bilak kuchini sinash, to‘siq-lardan sakrab o‘tish va boshqa mashqlar keng qo‘llaniladi. KONFERENSIYA \| 2 BEST ARTICLE RESPUBLIKA ILMIY – ONLAYN KONFERENSIYASI 2022 ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR Harbiy chaqiriq punktlarida (tuman-shahar harbiy komissiariatlari) jismoniy tayyorgarlikni sinov tariqasida tekshirish va bo‘lajak askarlarni maxsus jismoniy tayyorgarliklarga o‘rgatish mashg‘uloti tashkil etiladi. Harbiy qismlar, bo‘linmalardagi turlar hamda ixtisosliklarga qarab, harbiy- jismoniy tayyorgarlik mashg‘ulotlari o‘tkaziladi. Ofitserlar va katta ofitserlar bilan doimiy ravishda o‘tkazib boriladigan jismoniy tayyorgarlikni ta’minlovchi mashg‘ulotlar. Yuqorida ta’kidlangan yo‘nalishlardagi harbiy-jismoniy tayyorgar-lik mashg‘ulotlarining asosiy maqsadi shundaki, xizmatchi va askarlar-ning harbiy- jismoniy tayyorgarligi Vatan mudofaasi uchun qo‘llanilishi bilan bir qatorda, shaxsiy tarkib va barcha a’zolarning sog‘lom bo‘lishini taminlashga ham qaratiladi. Bu jihatlar o‘z navbatida har bir xizmatchi va askarning ma’suliyatini oshirish, Vatan va xalq oldidagi burchlarini sadoqat bilan bajarishga safarbar etadi. Bu jismoniy tarbiya vositalari-ning (mashqlar o‘yinlar, sport va h.k.) ijtimoiy- tarbiyaviy ahamiyatidan dalolat beradi. Shu sababdan o‘quvchi Yoshlar va talabalar o‘quv jarayonlarida jismoniy tarbiya va sport bilan muntazam shug‘ullanishni odat qilib olishlari zarur. Shuningdek, jsimoniy tarbiya va sportning harbiy-jismoniy tayyorgarlikdagi ahamiyatini o‘rganish, ularni yoshlarga singdirish yo‘llarini topishlari kerak. “O‘zbek kurashi” bo‘yicha jahon chempionatlari (1999 - Toshkent, 2000 - Antaliya, 2001 - – Vengriya, 2002 - Armaniston) va Osiyo birinchiliklari (2002 - Hindiston) o‘tkazilishi O‘zbekistonning buyuk istiqlol yo‘lidan borayotganligidan dalolatdir. Bu esa o‘zbek xalqining milliy madaniyati yuksalishini namoyish etuvchi eng ustuvor tadbirdir. KONFERENSIYA \| 3 BEST ARTICLE RESPUBLIKA ILMIY – ONLAYN KONFERENSIYASI 2022 ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR Milliy sport, sportga xos o‘yinlar hamda harakatli o‘yinlarning mazmu-nida jismoniy sifatlar, insoniy fazilat (guruh yoki jamoa uchun kurashish) mujassamlanganligi bilan ularning ijtimoiy-madaniy va tarbiyaviy xususi-yatlari jamiyatimiz hamda davlat tizimida muhim ahamiyatga ega kasb. Xalq milliy o‘yinlarining jismoniy sifatlarini o‘stirish, Vatan himoyasidagi o‘ziga xos xususiyatlarini e’tiborga olingan holda ularning eng sevimli turlari o‘quv yurtlarining “Jismoniy tarbiya” dasturiga (1999) kiritilgan. Xalq milliy o‘yinlarining yana bir ijtimoiy-madaniy xususiyatlari hamda ularning mohiyatlari xalq orasida juda e’zozlanib kelinadi. Ularni qishloq-lardagi to‘ylarda kurash, ko‘pkari, poyga va boshqa ko‘p o‘yinlarning qo‘l-lanilishi timsolida ko‘rish mumkin. Shunga asosan xalq milliy o‘yinlaridan ba’zi birlariga atroflicha baho berish va ularning ijtimoiy-madaniy turmush sharoitimizda tutgan o‘rnini o‘quvchi-yoshlarga eslatish lozim. Boks tarixi qadimgi davrlarga borib taqaladi. Hatto Misrda, zamonaviy olimlar tomonidan miloddan avvalgi ikki, uch ming yilliklardan ko'proq vaqt davomida belgilanadigan relyef rasmlarida, Shumer g'orlarida. Miloddan avval musht urishlarining tasvirlari topilgan.Bag'dod shahri yaqinida Iroqda olib borilgan arxeologik qazishmalar paytida qadimiy jang san'ati tasvirlari ham topildi. Qadimgi Yunonistonda ham, Rim imperiyasida ham o'sha paytlarda mushtlashuvlar bo'lganligi haqida ko'plab dalillar mavjud. KONFERENSIYA \| 4 668 yilda Qadimgi Yunonistonda bo'lib o'tgan Olimpiya o'yinlariga mushtlashuvchilar kiritildi. Shu vaqtdan boshlab ushbu yakka kurash turi sport turi sifatida tan olingan deb hisoblash mumkin. Faqat erkin yunonlar jangchi bo'lishi mumkin edi. Musht janglari juda mashhur bo'lib, jasorat, kuch, epchillik, tezkorlik namunasi sifatida qaraldi. Ularda shoirlar, yozuvchilar va davlat arboblari ishtirok BEST ARTICLE RESPUBLIKA ILMIY – ONLAYN KONFERENSIYASI 2022 ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR etishdi. Masalan, taniqli Pifagoralar, ularning foydalari ko'plab matematik kashfiyotlar deb hisoblanadi, shuningdek, mukammal kurashchi bo'lgan va ko'pincha kurash musobaqalarida qatnashgan. Jang qoidalari vaqt o'tishi bilan o'zgarib bordi. O'sha kunlarda faqat boshiga urish mumkin edi, qo'llar himoya uchun charm chiziqlar bilan o'ralgan edi, janglar juda shiddatli bo'lib, kurashchilardan biri aniq g'olib chiqqunga qadar va raundlar soni aniqlanmagan edi. Bo'limning bunday janglari og'ir jarohat va o'lim bilan yakunlandi. Qadimgi Yunonistonning afsonaviy boks chempioni - Theagen haqida ma'lumot bor. Boks tarixida u 2000 dan ortiq janglarda qatnashgan va 1800 raqibini o'ldirgan. Asrlar davomida qo'llarni o'rash uchun yumshoq teri parchalari yanada qattiqroq bo'lib rivojlanib bordi va keyinchalik ularda mis va temir qo'shimchalar paydo bo'ldi. Ular Rim imperiyasidagi sportchilar tomonidan ishlatilgan va nafaqat qo'llarini himoya qilish uchun, balki ularni dahshatli qurolga aylantirgan. Gladiatorlik janglari paytida jangchilarning qo'llari shu tarzda o'ralgan edi. Zamonaviy boks tarixi Angliya bilan chambarchas bog'liq. Ushbu mamlakat ushbu sport turining ajdodidir. O'tkazilgan boks musobaqasining birinchi yozma eslatmasi 1681 yilga to'g'ri keladi. O'sha kunlarda aniq qoidalar hech qachon o'rnatilmagan, ular jangdan oldin oldindan kelishilgan, hakam tayinlangan, g'olib jang kassasidan mukofot olgan. Og'irlik va vaqt cheklovlari yo'q edi. Biz qo'llarimiz bilan qo'lqopsiz jang qildik, boshimiz, elkamiz, oyoqlarimiz, tirsaklarimizga urdik. Aslida bu qo'l jangi edi. 1719 yilda Jeyms Figg va Ned Saton duelda uchrashishdi. Figg g'olib bo'ldi. Va u chempion unvoniga sazovor bo'ldi. Ilgari ushbu nom ostida sarlavha yo'q edi. Figg davrida boks yanada ommalashgan. Chempion jamoat matbuotiga maqolalar yozdi va hujum va himoya boks texnikasi haqida gapirdi. U birinchi qoidalarni shakllantira boshladi. Ularda jangchilar dushmanni so'zning to'g'ridan-to'g'ri O S \| etishdi. Masalan, taniqli Pifagoralar, ularning foydalari ko'plab matematik kashfiyotlar deb hisoblanadi, shuningdek, mukammal kurashchi bo'lgan va ko'pincha kurash musobaqalarida qatnashgan. Jang qoidalari vaqt o'tishi bilan o'zgarib bordi. O'sha kunlarda faqat boshiga urish mumkin edi, qo'llar himoya uchun charm chiziqlar bilan o'ralgan edi, janglar juda shiddatli bo'lib, kurashchilardan biri aniq g'olib chiqqunga qadar va raundlar soni aniqlanmagan edi. Bo'limning bunday janglari og'ir jarohat va o'lim bilan yakunlandi. Qadimgi Yunonistonning afsonaviy boks chempioni - Theagen haqida ma'lumot bor. Boks tarixida u 2000 dan ortiq janglarda qatnashgan va 1800 raqibini o'ldirgan. Asrlar davomida qo'llarni o'rash uchun yumshoq teri parchalari yanada qattiqroq bo'lib rivojlanib bordi va keyinchalik ularda mis va temir qo'shimchalar paydo bo'ldi. Ular Rim imperiyasidagi sportchilar tomonidan ishlatilgan va nafaqat qo'llarini himoya qilish uchun, balki ularni dahshatli qurolga aylantirgan. Gladiatorlik janglari paytida jangchilarning qo'llari shu tarzda o'ralgan edi. Zamonaviy boks tarixi Angliya bilan chambarchas bog'liq. ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR Ushbu mamlakat ushbu sport turining ajdodidir. O'tkazilgan boks musobaqasining birinchi yozma eslatmasi 1681 yilga to'g'ri keladi. O'sha kunlarda aniq qoidalar hech qachon o'rnatilmagan, ular jangdan oldin oldindan kelishilgan, hakam tayinlangan, g'olib jang kassasidan mukofot olgan. Og'irlik va vaqt cheklovlari yo'q edi. Biz qo'llarimiz bilan qo'lqopsiz jang qildik, boshimiz, elkamiz, oyoqlarimiz, tirsaklarimizga urdik. Aslida bu qo'l jangi edi. 1719 yilda Jeyms Figg va Ned Saton duelda uchrashishdi. Figg g'olib bo'ldi. Va u chempion unvoniga sazovor bo'ldi. Ilgari ushbu nom ostida sarlavha yo'q edi. Figg davrida boks yanada ommalashgan. Chempion jamoat matbuotiga maqolalar yozdi va hujum va himoya boks texnikasi haqida gapirdi. U birinchi qoidalarni shakllantira boshladi. Ularda jangchilar dushmanni so'zning to'g'ridan-to'g'ri KONFERENSIYA \| 5 BEST ARTICLE RESPUBLIKA ILMIY – ONLAYN KONFERENSIYASI 2022 ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR ma'nosida tugatishlari mumkin, oyoqlari va qo'llarini sindirib, ko'zlariga bosishgan. Urush paytida raqibning oyog'ini teshib o'tishi mumkin bo'lgan jangchilar botinkasining tagiga mixlar tiqilib qolgan edi. Bu chindan ham dahshatli manzaralar edi. Figg 1722 yilda Boks akademiyasini yaratdi, u erda bu kurash turini hamma uchun o'rgatdi. Boks tarixi asrlar davomida rivojlanib, o'zgarishlarga duch keldi. 1867 yilda boks musobaqasini o'tkazishni tubdan o'zgartiradigan yangi qoidalar kiritildi. Ular Queensberry markizasi qoidalarida aniq ko'rsatilgan. Ular jangchilarning harakatlari uchun qat'iy asos o'rnatdilar, harakatlarini chekladilar, tirnoqli botinkalardan foydalanishni taqiqladilar, 3 minutlik muddat bilan majburiy turlarni kiritdilar, tepish, tirsaklar, tizzalar va bo'g'ilish taqiqlandi. Agar Bokschi yiqilsa, Hakam 10 soniyani hisoblaydi. Agar bu vaqt ichida bokschi o'rnidan turmasa, hakam unga mag'lubiyatni o'qishi mumkin. Tizzaning ringga tegishi yoki arqonlarga yopishishi bokschining yiqilishi hisoblanadi.Ushbu qoidalarning aksariyati hanuzgacha zamonaviy boksning asosini tashkil etadi. a tugatishlari mumkin, oyoqlari va qo'llarini sindirib, ko'zlariga bosishgan. ma'nosida tugatishlari mumkin, oyoqlari va qo'llarini sindirib, ko'zlariga bosishgan. Urush paytida raqibning oyog'ini teshib o'tishi mumkin bo'lgan jangchilar botinkasining tagiga mixlar tiqilib qolgan edi. Bu chindan ham dahshatli manzaralar edi. Figg 1722 yilda Boks akademiyasini yaratdi, u erda bu kurash turini hamma uchun o'rgatdi. Boks tarixi asrlar davomida rivojlanib, o'zgarishlarga duch keldi. 1867 yilda boks musobaqasini o'tkazishni tubdan o'zgartiradigan yangi qoidalar kiritildi. Ular Queensberry markizasi qoidalarida aniq ko'rsatilgan. Ular jangchilarning harakatlari uchun qat'iy asos o'rnatdilar, harakatlarini chekladilar, tirnoqli botinkalardan foydalanishni taqiqladilar, 3 minutlik muddat bilan majburiy turlarni kiritdilar, tepish, tirsaklar, tizzalar va bo'g'ilish taqiqlandi. Agar Bokschi yiqilsa, Hakam 10 soniyani hisoblaydi. Agar bu vaqt ichida bokschi o'rnidan turmasa, hakam unga mag'lubiyatni o'qishi mumkin. Tizzaning ringga tegishi yoki arqonlarga yopishishi bokschining yiqilishi hisoblanadi.Ushbu qoidalarning aksariyati hanuzgacha zamonaviy boksning asosini tashkil etadi. 1892 yilda Jeyms Jon Korbet va Jon Lourens Sallivan o'rtasidagi jang zamonaviy professional boksning rasmiy tug'ilgan sanasi hisoblanadi. Shu vaqtdan boshlab AQSh va boshqa mamlakatlarda jamoat boks tashkilotlari paydo bo'la boshladi. Ularning mohiyati o'zgarmagan bo'lsa-da, ular ko'p marta o'zgartirilgan. Endi u Butunjahon boks tashkiloti deb ataladi. Qadimgi Rossiyada ular o'z kuchlarini o'lchashni yaxshi ko'rar edilar, ham mushtlar, ham qo'l jangi bo'lgan. Ko'pgina rus ertaklarida Ilya Muromets, Alyosha Popovich va Dobrynya Nikitich qahramonlari bilan janglar haqida so'z boradi. Ularning ajoyib kuchlari haqida aytilgan. Haqiqiy hayotda janglar ham bo'lib o'tdi, u erda jangchilar bir-birlari bilan raqobatlashdilar, ko'pincha devorlardan KONFERENSIYA \| 6 ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR BEST ARTICLE RESPUBLIKA ILMIY – ONLAYN KONFERENSIYASI 2022 devorlarga qarshi kurashlar bo'lib o'tdi, bir vaqtning o'zida har tomondan bir nechta odam qatnashdi. ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR - kurash; - regbi kabi sport federatsiya va assotsiatsiyalari bilan mustahkam hamkorlik yoʻlga qoʻyiladi. - regbi kabi sport federatsiya va assotsiatsiyalari bilan mustahkam hamkorlik yoʻlga qoʻyiladi. Mazkur hamkorlik memorandumlarini imzolash jarayonida Milliy olimpiya qoʻmitasi raisi Rustam Shoabdurahmonov, Xalq taʼlimi vazirining oʻrinbosari Rustam Karimjonov, MOQ Bosh kotibi Oybek Kasimov, MOQ raisi oʻrinbosari, Oʻzbekiston dzyudo federatsiyasi raisi Azizjon Kamilov, boks, voleybol, stol tennisi, regbi, gimnastika, gandbol, kurash boʻyicha mamlakat federatsiya va paralimpiya assotsiatsiyalarning mutasaddi rahbarlari ishtirok etishdi. Memorandum asosida: Memorandum asosida: Voleybol, gandbol,regbi, kurash, stol tennisi federatsiyalari 205 ta maktablarga malakali trenerlar biriktirilib, sport seksiyalari tashkil etiladi. Mazkur maktablarning jismoniy tarbiya oʻqituvchilari malakasi oshiriladi va qayta tayyorlanadi. Bundan tashqari, boks, badiiy gimnastika federatsiyalari— 50 ta maktablarga malakali trenerlar biriktirilib, sport seksiyalari tashkil etiladi. Voleybol sport turi boʻyicha — 410 dona voleybol toʻpi, 205 dona voleybol setkasi va oʻquv-qoʻllanmalar; gandbol sport turi boʻyicha — 1025 dona gandbol toʻpi, 10 dona gandbol darvozasi, 205 dona gandbol setkasi va oʻquv-qoʻllanmalar; kurash sport turi boʻyicha 205 ta kurash gilami va oʻquv-qoʻllanmalari; regbi sport turi boʻyicha 820 dona regbi toʻpi va oʻquv-qoʻllanmalar ajratiladi. ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR devorlarga qarshi kurashlar bo'lib o'tdi, bir vaqtning o'zida har tomondan bir nechta odam qatnashdi. devorlarga qarshi kurashlar bo'lib o'tdi, bir vaqtning o'zida har tomondan bir nechta odam qatnashdi. Pravoslav cherkovi o'yin-kulgining bunday turini ma'qullamagan va ko'pincha qo'l bilan jang qilish taqiqlangan. Ivan Dahshatli davrda va keyinchalik Buyuk Pyotr davrida boks har qanday holatda ham mamlakatga kirib bordi, Angliya va uning madaniyati bilan o'zaro aloqalar bejizga ketmasligi mumkin edi. 1894 yilda Mixail Kister ingliz boksiga bag'ishlangan kitob nashr etdi. 1895 yil 15-iyulda birinchi rasmiy duel bo'lib o'tdi. Ushbu sana Rossiyada boksning tug'ilgan kuni deb hisoblanadi. Respublikada sportning olimpiya, paralimpiya va milliy turlarini rivojlantirish, ommalashtirish, shuningdek, maktablarda bolalarning boʻsh vaqtlarini mazmunli oʻtkazishni tashkil qilish maqsadida Xalq taʼlimi vazirligi hamda sport federatsiya va assotsiatsiyalari oʻrtasida 7 ta muhim memorandum imzoladi. 2018-2020-yillarda Xalq taʼlimi vazirligi futbol assotsiatsiyasi, dzyudo, basketbol federatsiyalari bilan hamkorlik yoʻlga qoʻyilgandi. - boks; - badiiy gimnastika; - gandbol; - voleybol; - stol tennisi; adabiyotlar: KONFERENSIYA \| 8 1. Salomov R.S. Jismoniy tarbiya nazariyasi va uslubiyati. Darslik. Т.: «ITA- PRESS», 2014. 269 бет. 1. Salomov R.S. Jismoniy tarbiya nazariyasi va uslubiyati. Darslik. Т.: «ITA- PRESS», 2014. 269 бет. 1. Salomov R.S. Jismoniy tarbiya nazariyasi va uslubiyati. Darslik. Т.: «ITA- PRESS», 2014. 269 бет. 1. Salomov R.S. Jismoniy tarbiya nazariyasi va uslubiyati. Darslik. Т.: «ITA- PRESS», 2014. 269 бет. KONFERENSIYA \| 8 ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR BEST ARTICLE RESPUBLIKA ILMIY – ONLAYN KONFERENSIYASI 2022 2. Salomov RS., Sharipov A.K. Jismoniy tarbiya nazariyasi va uslubiyati. Darslik. Т.: «ITA-PRESS», 2015.182 бет. ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR BEST ARTICLE RESPUBLIKA ILMIY – ONLAYN KONFERENSIYASI 2022 2. Salomov RS., Sharipov A.K. Jismoniy tarbiya nazariyasi va uslubiyati. Darslik. Т.: «ITA-PRESS», 2015.182 бет. ILM – FAN TA’LIMDA INNOVATSION YONDASHUVLAR, MUAMMOLAR, TAKLIF VA YECHIMLAR 2. Salomov RS., Sharipov A.K. Jismoniy tarbiya nazariyasi va uslubiyati. Darslik. Т.: «ITA-PRESS», 2015.182 бет. 1.www.tdpu.uz KONFERENSIYA \| 9 9 KONFERENSIYA \|

Subsets and Splits

No community queries yet

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